Professional Documents
Culture Documents
2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 1
Centrifugal Troubleshooting
POSSIBLE CAUSES
SYMPTOMS
(See definitions Table 2)
20. Packed box leaks excessively or packing has short life 8/9/45/54/55/57/68/69/70/71/72/73/74
TABLE 2 - DEFINITIONS
Suction Troubles
SYMPTOMS CAUSES
TABLE 4 - VIBRATION
PUMP CLINIC 2
When the pressure of flowing liquids drops to, or below the liquid’s vapour pressure,
the liquid boils and vapour cavities (bubbles) form locally inside the liquid. If the
pressure within the flow path subsequently increases above the vapour pressure, the
vapour cavities implode, releasing energy. The formation and sudden collapse of
these bubbles is called Cavitation.
The generation of head in a centrifugal pump does not commence until the liquid enters
the vane area and is accelerated towards pump discharge. As the liquid flows between
the pump inlet flange and vanes, several points of head loss occur due to:
b) Acceleration losses as the liquid velocity increases from the suction nozzle to the
impeller eye.
c) Shock losses as the liquid contacts the leading edges of the impeller vanes.
The sum of these losses is known as the entry loss. If the suction head minus the
entry loss reduce the liquid pressure to or below the vapour pressure, then a condition
for cavitation exists. Figure 1 Page 2 illustrates the above.
The net positive suction head is a statement of the minimum suction conditions
required to prevent cavitation. The required NPSH (referred to as NPSHR) is the
minimum value of NPSH required at the pump inlet for satisfactory pump operation and
must be determined by test and is stated by manufacturers (appears on the pump
performance curve as an NPSHR curve). The NPSHR is equivalent to the entry loss
as shown in Figure 1.
Pump Clinic 2 Introduction to Cavitation & Net Positive Suction Head 03/05/06 Page 1 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Pump Clinic 2 Introduction to Cavitation & Net Positive Suction Head 03/05/06 Page 2 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
The available NPSH (referred to as NPSHA) is a function of the suction side pumping
system and is defined as the absolute pressure head on the liquid surface, plus the
static liquid level above the pump centre line (negative for suction lift) minus the friction
loss in the piping system leading to the pump minus the vapour pressure head at the
pumping temperature. The discharge pumping system has no effect on NPSHA.
Figure 2 Page 4 shows four typical suction systems with the NPSHA formulae
applicable to each. Please note that the units of the terms in the formulae are metres
absolute of the liquid being pumped.
To avoid cavitation, NPSHA must always be greater than NPSHR at the design flow.
The presence of cavitation due to inadequate NPSH can be diagnosed during pump
operation by a steady crackling noise in and around the pump suction. This should not
be confused with a random crackling noise with high intensity knocks which indicates
another condition termed suction recirculation (not covered in these notes).
If the problem was one of noise alone, it is likely that most situations would call for no
remedial action. However, continual cavitation causes mechanical and operational
problems as follows:
1) Erosion of impeller, particularly at the leading edges of the impeller vanes. In some
cases, the casing itself will show signs of erosion. The extent of damage
experienced is significantly affected by product-related factors such as corrosion
and abrasion. Apart from the damage to the parts, the erosion can cause loss of
pump efficiency and out-of-balance problems with the impeller.
3) The vapour cavities will impede the flow of liquid through the impeller. In some
cases, the flow may be completely blocked. This will result in reduced capacity
plus reduced and/or unstable developed head.
Pump Clinic 2 Introduction to Cavitation & Net Positive Suction Head 03/05/06 Page 3 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Pump Clinic 2 Introduction to Cavitation & Net Positive Suction Head 03/05/06 Page 4 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
The obvious answer is to ensure proper pump selection at the initial stage. Most users
would agree that the majority of pump vendors are sufficiently competent in giving
customers what was asked for in the specifications. Having said this, it is imperative
that the issue of NPSHA vs NPSHR is properly understood and considered by both
user and supplier.
a) Increase the static liquid level above the pump or reduce the suction lift. This can
be done in the case of a flooded suction by raising the liquid level in the suction
tank, raising the suction tank to a higher level or lowering the pump, e.g. one floor
down.
In the suction lift situation, the liquid level in the sump or suction tank can be raised
or the pump can be lowered, e.g. mounting the pump off the sump side or building a
dry sump beside the existing sump.
b) Reduce the friction losses by increasing pipe sizes and reducing the length of pipe
runs and the number of fittings, e.g. tees, bends, valves. Selection of fittings with
lower friction loss, e.g. long radius elbows and full flow ball valves should also be
considered. In particular, resist the use of suction strainers that can clog.
c) Reduce the vapour pressure by reducing the temperature of the product. This can
be done by reducing the operational temperature of the process (if feasible) or
cooling the temperature in the suction line, e.g. cooling annulus on the suction
pipework. It must be noted the reduction of vapour pressure by reducing the
temperature is rarely possible.
a) Reduce the flow rate by throttling on the pump discharge. This will generally
reduce NPSHR (always check the pump curve) and increase NPSHA (due to
reduced friction losses). Care must be taken to ensure that the flow rate is not
reduced below the minimum flow rate recommended by the manufacturer.
b) Reduce the pump speed as this reduces NPSHR. This will require the user to
accept reduced pump performance
c) Reduce the pump speed and install a larger diameter impeller. This will have a
two-fold effect as lower speed means lower NPSHR and in many cases the larger
impeller diameter has lower NPSHR characteristics.
d) Install a different pump. This would normally mean installation of a larger pump as
they generally have a better NPSHR value for the same flow rate. The selection of
a larger pump is sometimes required with speed reduction.
Pump Clinic 2 Introduction to Cavitation & Net Positive Suction Head 03/05/06 Page 5 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
e) Change the impeller material to one that is more resistant to erosion, e.g. from cast
iron to stainless steel. This does not eliminate cavitation but will reduce the impact
of cavitation.
The presence of any of the conditions detailed below significantly increases the
possibility of low NPSHA values and cavitation. Please consider this in your pump
selections and if in doubt, discuss the matter with the pump supplier.
1. High temperature or boiling liquids: This will increase the vapour pressure
head.
3. Suction tank under vacuum: This will reduce the absolute pressure head on the
liquid surface.
6. A high number of fittings in suction pipework: This will increase friction losses.
Conclusion
Cavitation and NPSH seem to be some of the least understood topics associated with
pump applications. To some engineers, these topics appear mysterious or, at best, are
only partially understood.
Many highly technical research papers have been written on this subject. These notes
are an attempt to give a simple introduction to cavitation and highlight areas where
some critical thought should be applied.
It should also be stated that although these notes are based on application to
centrifugal pumps, the majority of the principles apply equally to all other types of
pumps.
Pump Clinic 2 Introduction to Cavitation & Net Positive Suction Head 03/05/06 Page 6 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 3
There is a small flow from impeller discharge to suction through the wearing rings and
any hydraulic balance device present. This takes place at all capacities but does not
usually contribute to raising the liquid temperature very much unless operation is near
shut-off.
When the capacity has been reduced by throttling (or as a result of an increase in
system head), a secondary flow called recirculation begins. Recirculation is a flow
reversal at the suction and/or at the discharge tips of the impeller vanes. All impellers
have a critical capacity at which recirculation occurs. The capacities at which suction
and discharge recirculation begin can be controlled to some extent by design, but
recirculation cannot be eliminated.
Suction recirculation is the reversal of flow at the impeller eye. A portion of the flow is
directed out of the eye at the eye diameter, as shown in Figure 1 Page 1 and travels
upstream with a rotational velocity approaching the peripheral velocity of the diameter.
A rotating annulus of liquid is produced upstream from the impeller inlet and through
the core of this annulus passes an axial flow corresponding to the output capacity of
the pump. The high shear rate between the rotating annulus and the axial flow through
the core produces vortices that form and collapse, producing noise and cavitation in the
suction of the pump.
Discharge recirculation is the reversal of flow at the discharge tips of the impeller
vanes, as shown in Figure 2 Page 1. The shear rate between the inward and outward
relative velocities produces vortices that cavitate and usually attack the pressure side
of the vanes.
The capacity at which suction recirculation occurs is directly related to the design
suction-specific speed S of the pump. The higher the suction-specific speed, the closer
will be the beginning of recirculation to the capacity at best efficiency. Figure 3 Page 2
shows the relation between the suction-specific speed and suction recirculation for
pumps up to 2500 (1530) specific speed and Figure 4 Page 2 shows the same relation
for pumps up to 10,000 (6123) specific speed.
For water pumps, the minimum operating flows can be as low as 50% of the suction
recirculation values shown for continuous operation and as low as 25% for intermittent
operation. For hydrocarbons, the minimum operating flows can be as low as 60% of
the suction recirculation values shown for continuous operation and as low as 25% for
intermittent operation.
The high turbulence produced by recirculation and separation accounts for most of the
high power consumed at shut-off. This may vary from about 30% of the normal power
for pumps of very low-specific speed to nearly three times the normal power for
propeller pumps. Separation and, possibly, cavitation may take place on the casing
tongue or diffusion vanes at very low capacities. Operation near shut-off causes not
only excessive heating, but also vibration and cavitation, which may cause serious
mechanical damage.
Cause and Effect: Recirculation occurs at reduced flows and is the reversal of a
portion of the flow-back through the impeller. Recirculation at the inlet of the impeller is
known as suction recirculation. Recirculation at the outlet of the impeller is discharge
recirculation. Suction and discharge recirculation can be very damaging to pump
operation and should be avoided for continuous operation.
Diagnosis From Pump Operation: Suction recirculation will produce a loud crackling
noise in and around the suction of the pump. Recirculation noise is of greater intensity
than the noise from low NPSH cavitation and is a random knocking sound. Discharge
recirculation will produce the same characteristic sound as suction recirculation except
that the highest intensity is in the discharge volute or diffuser.
Figure 5: Examining the pressure side of the inlet vanes for Figure 6: Damage to the pressure side
suction recirculation. of the vane from discharge recirculation.
PUMP CLINIC 4
The performance of centrifugal pumps will vary when viscous liquids are pumped. For
medium and high viscosities, the power requirement increases considerably, whilst the
head, and to a lesser extent the flowrate, is reduced. With the aid of Table 4 (Page 3),
the characteristics of centrifugal pumps pumping viscous liquids can be calculated,
providing the characteristics for pumping water are known. (This diagram may also be
used as an aid in the selection of a pump for required duty).
The correction factors established from the diagram are sufficiently accurate for general
application within the limits given below. If more accurate values are required, a test
should be performed with the particular liquid.
When pumping highly-viscous liquids, it is recommended that the running costs are
investigated to establish whether other types of pump (eg rotary positive-displacement-
type) could be more economic due to the steep drop in efficiency of centrifugal pumps
under these conditions..
∗ The diagram should only be used for centrifugal pumps with radial impellers
within the normal Q-H range. The diagram must not be used for pumps with
mixed flow and axial flow impellers or for special pumps for viscous or
heterogeneous liquids. Table 4 is not applicable to side-channel pumps.
∗ The diagram may only be used if sufficient NPSH (NPSHavail ) is available to
prevent cavitation.
∗ The diagram may only be used for homogeneous Newtonian fluids. For
gelatinous liquids, widely-scattering results are obtained in practice, depending
upon the special properties of the liquid.
∗ With multistage pumps, the head per stage must be used in the calculation.
∗ When pumps have double-entry impellers, one half of the flowrate must be used
in the calculations.
Pump Clinic 4 Viscosity-How does it affect centrifugal pump performance 27/02/08 Page 1 of 4
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
To establish the correction factors from the diagram, the following procedure is used:
Starting with the flowrate Q on the horizontal axis, move vertically up to the intersection
with the required head H, then proceed horizontally (to the right-hand side or to the left-
hand side) to the intersection with the viscosity v of the liquid, thence vertically up to the
intersection with the lines of the different correction factors.
To determine the correction factory CH for total head, the curve l.0 x Qopt is to be used.
Pump Clinic 4 Viscosity-How does it affect centrifugal pump performance 27/02/08 Page 2 of 4
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Respectively, the selection should be checked by the more accurate method given in the
following section.
Pump Clinic 4 Viscosity-How does it affect centrifugal pump performance 27/02/08 Page 3 of 4
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Pump Clinic 4 Viscosity-How does it affect centrifugal pump performance 27/02/08 Page 4 of 4
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 5
Firstly, let us take a look at the two classifications in question and define both classes discussing
the merits of individual types of pumps in each class.
Rotodynamic
Rotodynamic pumps are rotary machines in which energy is continuously imparted to the pumped
liquid by a rotating impeller, propeller or rotor.
For this discussion we will not consider the special category as they are rarely used and only
under very specific conditions.
Before going on to review centrifugal pumps which account for probably well over 95% of
Rotodynamic applications, when do we use a peripheral pump (sometimes called side-channel or
regenerative turbine pumps)? They are most definitely not suited to handling solids because for
efficient operation, they depend on close clearances between their impellers and guide plates
which also limits their viscosity-handling capabilities to under 20mm²/sec.
However, they are ideal for low capacities limited to 10 l/sec at quite high heads up to 310 metres
through multi-staging, plus they have a built-in self-priming capability. Finally, many peripheral
flow pumps have the ability to handle quantities of vapour mixed with liquid for substantial periods.
Back to what we would all consider true centrifugal pumps, we can, for the purpose of this
discussion, consider the following classes:
1. Closed impeller
2. Open Impellers
3. Slurry pumps
In considering these classes we must look at how they handle solids and viscous liquids. Closed
impeller pumps below 80mm for example, should not be used for liquids of viscosity greater than,
say, 50mm²/sec because the viscous nature of the liquid creates too many internal losses to
operate efficiently, and likewise ‘lightly muddy’ water is about the worst solids they can handle.
However, the larger the centrifugal pump, the higher the viscosity it can efficiently pump, such that
at over 150mm for example, it can handle up to 800mm²/sec.
Open impellers can handle solids up to 5% provided that the individual solids can fit through the
impeller passage-ways and remembering that high velocities within a centrifugal pump encourage
high abrasive wear, if that is a characteristic of the solids. In these cases, a positive displacement
pump with its lower internal speeds could be a more economic option as appropriate materials
must be selected to accommodate the wear and this can be expensive.
For serious solid volumes, firstly these can be handled in sewage style pumps with single and two-
vane impellers that have passages able to pass solids the size of the suction connection,
generally beginning at 80mm.
Then we come to true slurry centrifugal pumps which can handle high volume solids, say to 50-
60% generally at under 100 metres heads. These are built with large clearances with internal
adjustment for wear, plus wear plates and components which are readily replaceable, including
rubber-lined parts.
Positive Displacement
Positive displacement pumps are rotary or reciprocating machines in which energy is periodically
added by application of force to movable boundaries of enclosed fluid containing volumes,
resulting in a direct increase in pressure.
All positive displacement pumps can handle viscous liquids generally to very high viscosities and
most are capable of handling substantial solids with the exception of vane, gear, multiple-screw
and some forms of lobe pumps. For clean liquids of low viscosity, again many positive-
displacement pumps can handle these liquids. However, many rotary types do not do it
economically because of slippage of thin liquids (low viscosities) through their clearances i.e. gear,
lobe and multiple-screw pumps.
For positive displacement pumps it should be remembered that in sizes up to say 50mm
discharge, the capital costs generally are similar to centrifugal pumps.
However, after that, the positive displacement pump rapidly increases in cost, such that a 150mm
discharge-type pump can cost many times that of a standard centrifugal water pump.
Conclusion
However, there are some important instances which do not follow these basic conclusions:
a) Slurry applications, e.g. 80mm discharge and above and generally below 100m head
should be centrifugal.
b) For clean liquid duties below 3 l/sec and above, 180 metres total head positive
displacement pumps should be considered.
c) For viscous liquid applications with up to 800mm²/sec viscosity and capacities above 70
l/sec a centrifugal pump should be considered.
d) Generally, raw sewage applications should use centrifugal pumps
e) All applications below 0.5 l/sec should be positive displacement pumps
To conclude, every pump application should be individually considered as to the type of pump
most suitable.
PUMP CLINIC 6
Affinity Laws
There are fundamental laws which can be used to predict changes in pump performance with
variations in speed. It is important in pump applications to be able to develop performance curves
corresponding to various speeds from standard performance curves. The mathematical
relationships between flowrate, head, power and speed which enable this are known as the
Affinity Laws.
For variation in speed with constant impeller diameter, the following laws apply:
a) Pump flow rate (Q) varies directly with the speed (N)
i.e. Q1/Q2 = N1/N2
b) Pump head (H) varies with the square of the speed (N)
i.e. H1/H2 = (N1/N2)²
In using the above formulae, it is assumed that efficiency remains constant. In practice, the
efficiency if slightly less at lower speeds since friction and drag constitute a larger proportion of
hydraulic power. It is important to note that these laws do not apply to NPSH.
Example:
If you have a pump performance at a speed of 1300rpm, what is the performance at 880rpm (refer
to following performance curve Page 2).
The first step is to select 5 operating points on the 1300rpm performance and tabulate as shown
below:
* kW may be read from the power curve where efficiencies are not detailed or may be calculated
using the formula.
kW = Flow x Head x SG
k x Efficiency
Using the Affinity Laws, the following factors can be applied to the 1300 rpm performance figures
when operating at 880 rpm:
The new performance for the pump at 880 rpm is tabulated below:
The pump and the associated plant equipment such as pipes, valves and tanks must always be
designed to cover for the maximum pumped volume. The following must be taken into
consideration to determine the maximum capacity of the plant:
Forms of Control
Since pumps are selected for the maximum plant capacity, a form of control must be provided to
regulate the volume of flow for variation in pumping demands. An average pumped quantity Qm
may be only a fraction of the maximum pump capacity Qp. The duration curve in Figure 1.2 below
illustrates for example, how for the most part of over a one-year period, the pump may operate at
reduced capacity.
The pump flow can be controlled by using the following control methods:
Throttle control is, even today, the most commonly used control form in industrial applications. Its
efficiency is, however, very low when compared with speed control, which in many cases gives
more than a 50% saving in energy. Pumps at waterworks and sewage water treatment plants are
normally controlled by means of start-stop control. Its efficiency is often also poor (Figure 1.3) and
besides, stress due to frequent starting and stopping may cause damage to the pipes and other
plant equipment.
Energy Efficient
Throttle control means that the flow of liquid in the pipes is restricted by means of a valve. This
results in a waste of energy because the pump is continuously working against the high pressure
imposed by the valve. The power consumed by the pump can be calculated from the formula:
P=QxHxρ
368 x η
Where :
P = Power (kW)
Q = Pumped quantity (m³/hr)
H = Pump head (m)
ρ = Specific gravity of the liquid
η = Pump efficiency
The above formula shows that the power requirement P is directly dependent on the product of the
pumped quantity Q and the pump head H. Figures 2.1 and 2.2 illustrate the power requirements
which are represented by the hatched areas in both figures.
It can be seen that in this example, the power requirement with speed control is less than half of
that with throttle control.
The saving obtained in energy depends essentially on the average pumped quantity. Figure 3
shows how much energy saving can be at different pumped volumes. When the power saving is
known, the saving in energy can always be calculated by multiplying the power saving by the time
factor. The methods of calculation are described in more detail (Page 1).
Control of flow is often arranged by means of two or more parallel-connected pumps which are of
different sizes. Step by step control is thus achieved by running the pumps in turn. Improved
control with a lesser capital cost is achieved if a single large pump is provided with the control as
shown in Figure 4. Parallel-connected pumps and motors require additional valves and piping
which again, increases costs.
Pressure tanks and upper water tanks are used for keeping a uniform pressure in the pipes in
applications where the pump runs on intermittent duty as, for example, in waterworks.
If the pump is provided with a frequency converter, the tanks can be made smaller or may be
totally dispensed with. In addition to the lower investment costs, a better control result is
achieved, which means a more uniform pressure at the consumer end.
The starting current which a pump provided with a frequency converter takes from the electrical
supply line is but a fraction of the starting current required by direct starting. On account of this,
the electrical distribution equipment can be made smaller and be purchased at a lower price. A
typical objective for saving purposes may be a standby generator for critical pumps. When a
frequency converter is used for the speed control of the pump, the generator size need only be 30
to 50% of that previously required.
Squirrel cage motors need reactive power which somehow needs to be generated. To avoid
loading the distribution network unnecessarily with reactive power, compensation is normally
effected by means of capacitors near to the motor. The frequency converter generates the
reactive power required by the motor and no compensating capacitors are needed. The cost of
investment is reduced and an optimum compensation effect achieved.
An improved control effect is more easily achieved with speed control than with other non-linear
forms of control. A drawback of intermittent duty, for example, is the discontinuity of the control.
The controlled parameters, the flow or pressure, for example, keeps varying. An accurate and
linear control is achieved with the converter.
Figure 8 shows the control graph of a plant with three parallel-connected pumps P1, P2 and P3.
When one of the pumps (P1) is provided with the control, a linear control curve (R) is obtained,
whilst the curve (T) of intermittent control is stepped, which can lead to abrupt variations in the
pumped volume of liquid.
When the control is used, the pump, pipes and valves experience less wear, which means
increased service life and reduced maintenance (particularly with plastic pipes).
· Static stress is reduced because the system need not operate with a high pumping pressure
all the time as with choke control. The pressure is as high or low as required.
· Dynamic stresses are far lower with a smooth control than with an intermittent start-stop
control. Pressure strokes (Figure 9) which wear the pipes and other plant equipment can thus
be avoided, and the service life may even be doubled.
P1 = QM[m³/h] x H1 [m] x ρ kW
368 x η1
P2 = QM[m³/h] x H2 [m] x ρ kW
368 x η2
PS = P 1 – P2
0.9
W S = P S x ta = P1 – P2 x ta
0.9
4. Saving in money
The saving in money per year is obtained when the energy saving is multiplied by the unit
price of energy k.
tt = Kp [cost]
Ks [money saved per annum]
Example
The pump (as per following curve Page 11) is designed for a nominal duty of 2200m³/hr at 67.5
metres head. If a secondary flowrate of 1500m³/h is required for 50% of the time, what are the
comparative power costs for throttling versus speed control.
Power saving:
The saving in money and the cost pay-off time can be calculated by inserting the right
values for the unit price of energy k and the capital cost difference Kp in the following
formulae:
tt = Kp [cost]
Ks [money saved per annum]
PUMP CLINIC 7
Instruction Books
Instruction books are intended to help keep the pumps in an efficient and reliable condition at all times.
It is necessary, therefore, that instruction books be available to all personnel involved in this function.
After a pump has been assembled in the manufacturer’s shop, all flanges and exposed machined
metal surfaces are cleaned of foreign matter and treated with an anti-corrosion compound, such as
grease, Vaseline, or heavy oil. For protection during shipment and erection, all pipe flanges, pipe
openings and nozzles are protected by wooden flange covers or by screwed-in plugs.
Usually the driver is delivered to the pump manufacturer, where it is assembled and aligned with the
pump on a common baseplate. The baseplate is drilled for driver mounting, but the final dowelling is
performed in the field after final alignment. When size and weight permit, the unit is shipped
assembled with pump and driver on the baseplate. If drivers are shipped directly to be mounted in the
field, the baseplate should be drilled at the job site.
If the pumping equipment is received before it can be used, it should be stored in a dry location. The
protective flange covers and coatings should remain on the pumps. The bearings and couplings must
be carefully protected against sand, grit and other foreign matter.
More thorough precautions are required if a pump must be stored for an extended period of time. It
should be carefully dried internally with hot air or by a vacuum-producing device to avoid rusting of
internal parts. Once free of moisture, the pump internals should be coated with protective liquid such
as light oil, kerosene or antifreeze. Preferably, all accessible parts, such as bearings and couplings,
should be dismantled, dried and coated with Vaseline or acid-free heavy oil and then properly tagged
and stored.
If rust preventive has been used on stored parts, it should be removed completely before final
installation and the bearings should be relubricated.
Pump Location
Foundations
Foundations may consist of any structure heavy enough to afford permanent rigid support to the full
area of the baseplate and to absorb any normal strains or shocks. Concrete foundations built up from
solid ground are the most satisfactory. Although most pumping units are mounted directly on
baseplates, very large equipment may be mounted directly on the foundations by using sole plates
under the pump and driver feet. Misalignment is corrected with shims.
The space required by the pumping unit and the location of the foundation bolts is determined from the
drawings supplied by the manufacturer. Each foundation bolt should be surrounded by a pipe sleeve
three or four diameters larger than the bolt. After the concrete foundations are poured, the pipe is held
solidly in place but the bolt may be moved to conform to the corresponding hole in the baseplate.
When a unit is mounted on steelwork or some other structure, it should be placed directly over, or as
near as possible to, the main members, beams or walls and should be supported so that the baseplate
cannot be distorted or the alignment disturbed by any yielding or springing of the structure, or of the
baseplate.
A curve ring or soleplate must be used as a bearing surface for the support flange of a vertical wet-pit
pump. The mounting face must be machined because the curb ring or soleplate will be used in
aligning the pump.
If the discharge pipe is located below the support flange of the pump (belowground discharge), the
curb ring or soleplate must be large enough to pass the discharge elbow during assembly. A
rectangular ring should be used. If the discharge pipe is located above the support flange
(aboveground discharge), a round curb ring or soleplate should be provided with clearance on its inner
diameter to pass all sections of the pump below the support flange. A typical method of arranging a
grouted soleplate for vertical pumps is shown below.
If the discharge is belowground and an expansion joint is used, it is necessary to determine the
movement that may be imposed on the structure. The pump casing should be attached securely to
some rigid structural members with tie rods. If vertical wet-pit pumps are very long, some steadying
device is required irrespective of the location of the discharge or of the type of pipe connection. Tie
rods can be used to connect the unit to a wall, or a small clearance around a flange can be used to
prevent excessive displacement of the pump in the horizontal plane.
Alignment
When a complete unit is assembled at the factory, the baseplate is placed on a flat, even surface. The
pump and driver are mounted on the baseplate and the coupling halves are accurately aligned using
shims under the pump and driver mounting surfaces where necessary. The pump is usually dowelled
to the baseplate at the factory, but the driver is left to be dowelled after installation at the site.
The unit should be supported over the foundation by short strips of steel plate or shim stock close to
the foundation bolts, allowing a space of ¾ to 2 inches (2 to 5 cm) between the bottom of the baseplate
and the top of the foundation for grouting. The shim stock should extend fully across the supporting
edge of the baseplate. The coupling bolts should be removed before the unit is levelled and the
coupling halves are aligned.
Where possible, it is preferable to place the level on some exposed part of the pump shaft, sleeve, or
planed surface of the pump casing. The steel supporting strips or shim stock under the baseplate
should be adjusted until the pump shaft is level, the suction and discharge flanges are vertical or
horizontal as required, and the pump is at the specified height and location. When the baseplate has
been levelled, the nuts on the foundation bolts should be made handtight.
During this levelling operation, accurate alignment of the unbolted coupling halves must be maintained.
A straightedge should be placed across the top and sides of the coupling, and at the same time, the
faces of the coupling halves should be checked with a tapered thickness gauge or with feeler gauges
to see that they are parallel. For all alignment checks, including parallelism of the coupling faces, both
shafts should be pressed over to one side when taking readings.
When the peripheries of the coupling halves are true circles of equal diameter and the faces are flat
and perpendicular to the shaft axes, exact alignment exists when the distance between the faces is the
same at all points and when a straightedge lies squarely across the rims at any point. If the faces are
not parallel, the thickness gauge or feelers will show variation at different points. If one coupling is
higher than the other, the amount may be determined by the straightedge and feeler gauges.
The dial readings will indicate whether the driver must be raised, lowered, or moved to either side.
After any movement, it is necessary to check that the coupling faces remain parallel to one another.
For example, if the dial reading at the starting point is set to zero and the diametrically opposite
reading at the bottom or sides shows + 0.202 in (+ 0.508 mm) the driver must be raised or lowered by
shimming or moved to one side or the other by half of this reading. The same procedure is used to
align gear couplings but the coupling covers must first be moved back out of the way and all
measurements should be made on the coupling hubs.
When an extension coupling connects the pump to its driver, a dial indicator should be used to check
the alignment. The extension piece between the coupling halves should be removed to expose the
coupling hubs. The coupling nut on the end of the shaft should be used to clamp a suitable extension
arm or bracket long enough to extend across the space between the coupling hubs. The dial indicator
is mounted on this arm, and alignment is checked for both concentricity of the hub diameters and
parallelism of the hub faces. Changing the arm from one hub to the other provides an additional
check. The dial extension bracket must be checked for sag, and readings must be corrected
accordingly.
The clearance between the faces of the coupling hubs and the ends of the shafts should be such that
they cannot touch, rub, or exert a pull on either pump or driver. The amount of this clearance may vary
with the size and type of coupling used. Sufficient clearance will allow unhampered endwise
movement of the shaft of the driving element to the limit of its bearing clearance. On motor-driven
units, the magnetic centre of the motor will determine the running position of the motor half-coupling.
This position should be checked by running the motor uncoupled. This will also permit checking the
directon of rotation of the motor. If current is not available at the time of installation, move the motor
shaft in both directions as far as the bearings will permit and adjust the shaft centrally between these
limits. The unit should then be assembled with the correct gap between the coupling halves.
Large horizontal sleeve-bearing motors are not generally equipped with thrust bearings. The motor
rotor is permitted to float, and as it will seek its magnetic centre, an axial force of rather small
magnitude can cause it to move off this centre. Sometimes it will move enough to cause the shaft
collar to contact and possibly damage the bearing. To avoid this, a limited-end float coupling is used
between the pump and the motor on all large units to restrict the motor rotor. The setting of axial
clearances for such units should be given by the manufacturer in the instruction books and elevation
drawings.
When the pump handles a liquid at other than ambient temperature or when it is driven by a steam
turbine, the expansion of the pump or turbine at operating temperature will alter the vertical alignment.
Alignment should be made at ambient temperature with suitable allowances for the changes in pump
and driver centrelines after expansion. The final alignment must be made with the pump and driver at
their normal temperatures and adjusted as required before the pump is placed into permanent service.
For large installations, particularly with steam-turbine-driven pumps, more sophisticated alignment
methods are sometimes employed using proximity probes and optical instruments. Such procedures
permit checking the effect of temperature changes and machine strains caused by piping stresses
while the unit is in operation. When such procedures are recommended, they are included with the
manufacturers’ instructions.
When the unit has been accurately levelled and aligned, the hold-down bolts should be gently and
evenly tightened before grouting. The alignment must be rechecked after the suction and discharge
piping has been bolted to the pump to test the effect of piping strains. This can be done by loosening
the bolts and reading the movement of the pump, if any, with dial indicators.
The pump and driver alignment should be occasionally rechecked, for misalignment may develop from
piping strains after a unit has been operating for some time. This is especially true when the pump
handles hot liquids as there may be a growth or change in the shape of the piping. Pipe flanges at the
pump should be disconnected after a period of operation to check the effect of the expansion of the
piping and the adjustments should be made to compensate for this.
Grouting
Ordinarily, the baseplate is grouted before the piping connections are made and before the alignment
of the coupling halves is finally rechecked. The purpose of grouting is to prevent lateral shifting of the
baseplate to increase its mass to reduce vibration and to fill in irregularities in the foundation.
The exposed surfaces of the grout should be covered with wet burlap to prevent cracking from too-
rapid drying. When the grout is sufficiently set so that the forms can be removed, the exposed
surfaces of the grout and foundations are finished smooth. When the grout is hard (72h or longer), the
hold-down bolts should be finally tightened and the coupling halves rechecked for alignment.
There is considerable controversy over whether the levelling strips or wedges should be removed after
grouting. The best practice is to remove these in all cases for reciprocating machinery because
pounding action or vibration will ultimately loosen the unit from the foundation. The space formerly
occupied by shims or wedges must be re-grouted. There is less danger is not removing the strips or
wedges with rotating machinery, provided care is used in mixing the grout material and there is no
shrinkage or drying. The strips or wedges can also be removed from a rotating unit. Erectors can
follow their own preference in this matter.
The pump and driver alignment must be rechecked thoroughly after the grout has hardened
permanently and at reasonable intervals thereafter.
When the pump handles hot liquids, dowelling of both the pump and its driver should be delayed until
the unit has been operated. A final re-check of alignment with the coupling bolts removed and with the
pump and driver at operating temperature is advisable before dowelling.
Large pumps handling hot liquids are usually dowelled near the coupling end, allowing the pump to
expand from that end out. Sometimes the other end is provided with a key and a keyway in the casing
foot and the baseplate.
PIPING
Suction piping
The suction piping should be as direct and short as possible. If a long suction line is required the pipe
size should be increased to reduce frictional losses. (The exception to this recommendation is in the
case of boiler-feed pumps where difficulties may arise during transient conditions of load change if the
suction piping volume is excessive. This is a special and complex subject and the manufacturer
should be consulted).
Where the pump must lift the liquid from a lower level, the suction piping should be laid out with a
continual rise toward the pump avoiding high spots in the line to prevent the formation of air pockets.
Where a static suction head will exist, the pump suction piping should slope continuously downward to
the pump.
Flow velocity should be less than 2m per second and the pipe diameter should be calculated
accordingly. Generally, the suction line is larger than the pump suction nozzle and eccentric reducers
should be used. If the source of supply is above the pump, the straight side of the reducer should be
at the bottom. Installing eccentric reducers with a change in diameters greater than 4 inches (10 cm)
could disturb the suction flow. If such a change is necessary, it is advisable to use properly vented
concentric reducers.
Elbows and other fittings next to the pump suction should be carefully arranged, or the flow into the
pump impeller will be disturbed. Long-radius elbows are preferred for suction lines because they
create less friction and provide a more uniform flow distribution than standard elbows. Pumps should
not be connected near any fitting. There should be a straight length of pipe minimum 10 x pipe
diameter before the pump inlet.
If a common suction head for two or more pumps is used, branches should be designed so that flow
disturbance before the pump inlets is not caused. T-pieces should not be used - a Y-piece is
recommended.
Discharge Piping
Generally, both a check valve and a gate valve are installed in the discharge line. The check valve is
placed between the pump and the gate valve and protects the pump from reverse flow in the event of
unexpected driver failure, or from reverse flow from another operating pump.
The gate valve is used when priming the pump or when shutting it down for inspection and repairs.
Manually operated valves that are difficult to reach should be fitted with a sprocket rim wheel and
chain. In some cases, discharge gate valves are motorised and can be operated by remote control.
Piping Strains
Cast iron pumps are never provided with raised face flanges. If steel suction or discharge piping is
used, the pipe flanges should be of the flat-face and not the raised-face type. Full-face gaskets must
be used with cast iron pumps.
Piping should not impose excessive forces and movements on the pump to which it is connected,
since these might spring the pump or pull it out of position. Piping flanges must be brought squarely
together before the bolts are tightened. The suction and discharge piping and all associated valves,
strainers, etc should be supported and anchored near to, but independent of, the pump so that strain
will be transmitted to the pump casing.
There are four factors to be considered in determining the effect of nozzle loads; material stress in
pump nozzles resulting from forces and bending movements, distortion of internal moving parts
affecting critical clearances, stresses in pump hold-down bolts, and distortion in pump supports and
baseplates resulting in driver coupling misalignment.
With large pumps, or when major temperature changes are expected, the pump manufacturer
generally indicates to the user the maximum movements and forces that can be imposed on the pump
by the piping.
Expansion Joints
Expansion joints are sometimes used in the discharge and suction piping to avoid transmitting any
piping strains caused by misalignment, or by expansion when hot liquids are handled. On occasion,
expansion joints are formed by looping the pipe. More often, they are of the slip-joint or corrugated-
diaphragm-type. However, they transmit to the pump a force equal to the area of the expansion joint,
times the pressure in the pipe. These forces can be of very significant magnitude, and it is impractical
to design the pump casings, baseplates, etc to withstand them. Consequently, when expansion joints
are used, a suitable pipe anchor must be installed between them and the pump proper. Alternately, tie
rods can be used to prevent the forces from being transmitted to the pump.
Suction Strainers
Except for certain special designs, pumps are not intended to handle liquid containing foreign matter.
If the particles are sufficiently large, such foreign matter can clog the pump, reduce its capacity, or
even render it altogether incapable of pumping. Small particles of foreign matter may cause damage
by lodging between close-running clearances. Therefore, proper suction strainers may be required in
the suction lines of pumps not specially designed to handle foreign matter.
In such an installation, the piping must first be thoroughly cleaned and flushed. The recommended
practice is to flush all piping to waste before connecting it to the pump. Then a temporary strainer of
appropriate size should be installed in the suction line as close to the pump as possible. This
temporary strainer may have a finer mesh than the permanent strainer installed after the piping has
been thoroughly cleaned of all possible mill scale or other foreign matter. The size of the mesh is
generally recommended by the pump manufacturer.
Vent valves are generally installed at one or more high points of the pump casing waterways to provide
a means of escape for air or vapour trapped in the casing. These valves are used during the priming
of the pump or during operation if the pump should become air or vapour-bound. In most cases, these
valves need not be piped up away from the pump because their use is infrequent and the vented air or
vapours can be allowed to escape into the surrounding atmosphere.
On the other hand, vents from pumps handling flammable, toxic or corrosive fluids must be connected
in such a way that they endanger neither the operating personnel, nor the installation. The suction
vents of pumps taking liquids from closed vessels under vacuum must be piped to the source of the
pump suction above the liquid level. All drain and drip connections should be piped to a point where
the leakage can be disposed of or collected for re-use if worth reclaiming.
Warm-Up Piping
When it is necessary for a pump to come up to operating temperature before it is started, or to keep it
ready to start at rated temperature, provision should be made for a warm-up flow to pass through the
pump. There are a number of arrangements used to accomplish this. If the pump is operated under
positive pressure on the suction, the pumped liquid can be permitted to drain out through the pump
casing drain connection to some point at a pressure lower than the suction pressure.
Alternately, some liquid can be made to flow back from the discharge header through a jumper line
around the check valve, into the pump and out into the suction header. An orifice is provided in this
jumper line to regulate the amount of the warm-up flow. Care must be exercised in such an installation
to maintain the suction valve open (unless the warm-up line valve is closed, as when the pump is to be
dismantled), lest the entire pump, suction valve, and suction piping be subjected to full discharge
pressure.
The manufacturer’s recommendations should be sought in all cases as to the best means of providing
an adequate warm-up procedure.
Relief Valves
Positive displacement pumps, such as rotary and reciprocating pumps can develop discharge
pressures much in excess of their maximum design pressures. To protect these pumps against
excessive pressures when the discharge is throttled or shut off, a pressure relief valve must be used.
Some pumps are provided with internal integral relief valves, but unless operation against a closed
discharge is both infrequent and of very short duration, a relief valve with an external return connection
must be used and the liquid from the relief valve must be piped back to the source of supply.
Surge Chambers
Generally, centrifugal pumps do not require surge chambers in their suction or discharge piping.
Reciprocating pumps may have a suction and discharge piping layout that does not require
compensation for variations in the flow velocity in the piping system.
In many cases however, reciprocating pump installations require surge chambers when the suction or
discharge lines are of considerable length, when there is an appreciable static head on the discharge,
when the liquid pumped is hot, or when it is desirable to smooth out variations in the discharge flow.
The type, size and arrangement of the surge chamber should be chosen on the basis of the
manufacturer’s recommendations.
Instrumentation
There are a number of instruments which are essential to maintaining a close check on the
performance and condition of a pump. A compound pressure gauge should be connected to the
suction of the pump and a pressure gauge should be connected to its discharge at the pressure taps
which may be provided in the suction and discharge flanges. The gauges should be mounted in a
convenient location so that they can be easily observed.
In addition, it is advisable to provide a flow-metering device. Depending upon the importance of the
installation, indicating meters may be supplemented by recording attachments.
Whenever pumps incorporate various leak-off arrangements, such as a balancing device or pressure-
reducing labyrinths, a check should be maintained on the quantity of these leak-offs by measuring
orifices and differential gauges installed in the leak-off lines.
PUMP CLINIC 8
OPERATION
Pumps are generally selected for a given capacity and total head when operating at rated speed.
These characteristics are referred to as ‘rated conditions of service’ and, with few exceptions,
represent those conditions at or near which the pump will operate the greatest part of the time.
Positive displacement pumps can not operate at any greater flows than rated, except by increasing
the speed; nor can they operate at lower flows except by reducing their operating speed or
bypassing some of the flow back to the source of supply.
On the other hand, centrifugal pumps can operate over a wide range of capacities, from near zero
flow to well beyond the rated capacity. Because a centrifugal pump will always operate at the
intersection of its head-capacity and system-head curves, the pump operating capacity may be
altered either by throttling the pump discharge (hence altering the system-head curve, or by varying
the pump speed (changing the pump head-capacity curve).
This makes the centrifugal pump very flexible in a wide range of services and applications which
require the pump to operate at capacities and heads differing considerably from the rated conditions.
There are, however, some limitations imposed upon such operation by hydraulic, mechanical, or
thermodynamic considerations.
There are certain minimum operating flows which must be imposed on centrifugal pumps for either
hydraulic or mechanical reasons. Four limiting factors must be considered:
· Radial thrust
· Temperature rise
· Internal recirculation
· Shape of the power curve
For sustained operation, it is imperative to adhere to the minimum flow limits recommended by the
pump manufacturer
The thermodynamic problem that arises when a centrifugal pump is operated at extremely reduced
flows is caused by the heating up of the liquid handled.
The difference between the power consumed and the water power developed represents the power
losses in the pump, except for a small amount lost in the pump bearings. These power losses are
converted to heat and transferred to the liquid passing through the pump.
If the pump were to operate against a completely closed valve, the power losses would be equal to
the shutoff power and since there would be no flow through the pump, all this power would go into
heating the small quantity of liquid contained in the pump casing. The pump casing would heat up,
and a certain amount of heat would be dissipated by radiation and convection to the atmosphere.
However, because the temperature rise in the liquid pumped could be quite rapid, it is generally
safer to ignore the dissipation of heat through radiation and the absorption of heat by the casing.
Calculations for determining the temperature rise in the liquid are available from Kelair.
The maximum permissible temperature rise in a centrifugal pump varies over a wide range,
depending on the type of service and installation. The minimum capacity based on thermodynamic
considerations is then established as that capacity at which the temperature rise is the maximum
permitted.
There are also hydraulic considerations which may affect the minimum flow at which a centrifugal
pump can operate. In recent years, correlation has been developed between operation at low flows
and the appearance of hydraulic pulsations both in the suction and in the discharge of centrifugal
impellers.
It has been proven that these pulsations are caused by the development of an internal recirculation
at the inlet and discharge of an impeller at certain flows below the best efficiency capacity. The
pump manufacturer’s recommendations on minimum flows dictated by these considerations should
always be followed.
The NPSHR curve becomes increasingly unstable at low flows. As a rule of thumb, do not operate
pumps at flowrates lower than that equivalent to the left-hand end of the NPSHR curve. This rule
has to be considered in conjunction with other issues detailed in this section.
Priming
With the exception of self-priming pumps, no centrifugal pump should ever be started until it is fully
primed. That is, until it has been filled with the liquid pumped and all the air contained in the pump
has been expelled.
Reciprocating pumps of the piston or plunger type are, in principle, self-priming. However, if quick
starting is required, priming connections should be piped to a supply above the pump.
Positive displacement pumps of the rotating type, such as rotary or screw pumps, have clearances
that allow the liquid in the pump to drain back to the suction. When pumping low viscosity liquids,
the pump may completely dry out when it is idle. In such cases, a foot valve should be used to help
keep the pump primed. Alternately, a vacuum device may be used to prime the pump.
When handling liquids of higher viscosity, foot valves are usually not required because liquid is
retained in the clearances and acts as a seal when the pump is restarted. However, before the initial
start of a rotating positive displacement pump, some of the liquid to be pumped should be introduced
through the discharge side of the pump to wet the rotating element.
The various methods and arrangements used for priming pumps are available from Kelair.
A few last-minute checks are recommended before a pump is placed into service for its initial start.
· Lubricate the bearings with fresh grease if the pump has been standing for a long time
since delivery. With oil-lubricated bearings, fill the bearing housing with the correct
quantity and quality of oil (as per manufacturer’s recommendations).
· Open any sealing and cooling liquid valves where applicable. A mechanical seal should
not normally show a leak, but a soft-packed gland should drip during the start-up period
and be adjusted as required.
· Check that the cooling/sealing liquid outlet pipes are not blocked, where applicable.
· Fill the pump casing by opening the suction valve, or through a filling pipe. Vent
casings by opening the discharge valve, or ventcocks when fitted.
· Ensure the pump shaft turns freely by hand. If it does not, check for foreign matter in
the pump, an over-tightened gland or pipe stresses causing the impeller to seize in the
casing.
· Check the direction of rotation by running the motor for a few seconds.
It is recommended when checking the direction of rotation, that coupling halves are
disconnected to eliminate the risk of loosening the impeller, which could cause damage to
the pump.
The steps necessary to start a centrifugal pump depend upon its type and upon the service on which
it is installed. For example, standby pumps are generally held ready for immediate starting. The
suction and discharge gate valves are held open and reverse flow through the pump is prevented by
the check valve in the discharge line.
The methods followed in starting are greatly influenced by the shape of the power-capacity curve of
the pump. High and medium head pumps (low and medium specific speeds) have power curves
that rise from zero flow to the normal capacity condition. Such pumps should be started against a
closed discharge valve to reduce the starting load on the driver.
A check valve is equivalent to a closed valve for this purpose, as long as another pump is already on
the line. The check valve will not lift until the pump being started comes up to a speed sufficient to
generate a head high enough the lift the check valve from its seat. If a pump is started with a closed
discharge valve, the recirculation bypass line must be open to prevent overheating. If a bypass line
is not installed, start the pump with the discharge valve cracked open.
Low head pumps (high specific speed) of the mixed flow and propeller type, have power curves that
rise sharply with a reduction in capacity; they should be started with the discharge valve wide open
against a check valve, if required to prevent backflow.
Assuming that the pump in question is motor-driven, that its shutoff power does not exceed the safe
motor power, and that it is to be started against a closed or cracked open gate valve, the starting
procedure is as follows:
1. Prime the pump, opening the suction valve, closing the drains, etc to prepare the pump
for operation.
2. Open the valve in the cooling water supply to the bearings, where applicable.
3. Open the valve in the cooling water supply if the stuffing boxes are water-cooled.
4. Open the valve in the sealing liquid supply if the pump is so fitted.
5. Open the warn-up valve of a pump handling hot liquids if the pump is not normally kept
at operating temperature. When the pump is warmed up, close the valve.
6. Open the valve in the recirculating line if the pump should not be operated against dead
shutoff (if fitted) or slightly crack open the discharge valve.
7. Start the motor.
If the pump is to be started against a closed check valve with the discharge gate valve open, the
steps are the same, except the discharge gate valve is opened prior to the motor being started.
In certain cases, the cooling water to the bearings and the sealing water to the seal cages are
provided by the pump. This, of course, eliminates the need for the steps listed for the cooling and
sealing supply.
Just as in starting a pump, the stopping procedure depends upon the type and service of the pump.
Generally the steps followed to stop a pump which can operate against a closed gate valve are:
In general, the starting and stopping of steam-turbine-driven pumps require the same steps and
sequence prescribed for a motor-driven pump. As a rule, steam turbines have various drains and
seals which must be opened or closed before and after operation. Similarly, many turbines require
warming up before starting.
Finally, some turbines require turning gear operation if they are kept on the line ready to start up.
The operator should therefore follow the steps outlined by the turbine manufacturer in starting and
stopping the turbine.
Most of the steps listed for starting and stopping centrifugal pumps are equally applicable to positive
displacement pumps. There is, however, a notable exception and that is:
Never operate a positive displacement pump against a closed discharge. If the gate valve
on the discharge must be closed, always start the pump with the recirculation bypass valve
open.
Standby pumps are frequently started up from a remote location, and several methods of operation
are available for the auxiliary services, such as the cooling water supply to the bearings or to water-
cooled stuffing boxes:
a) A constant flow may be kept through the bearing jackets, seal plates, oil coolers and
through the stuffing box lantern rings, whether the pump is running or on standby service.
b) The service connections may be opened automatically whenever the pump is started up
eg via solenoid valves.
c) The service connections may be kept closed while the pump is idle, and the operator may
be instructed to open them shortly after the pump has been put on line automatically.
The choice among these methods must be dictated by the specific circumstances surrounding each
case. There are, however, certain cases where sealing liquid supply to the pump stuffing boxes
must be maintained, whether the pump is running or not. This is the case when the pump handles a
liquid which is corrosive to the packing or which may crystallise and deposit on the shaft sleeves. It
is also the case when the sealing supply is used to prevent air infiltration into a pump when it is
operating under a vacuum.
Assuming that power failure will not cause the pump to go into reverse rotation, that is, that a check
valve will protect the pump against reverse flow, there is generally no reason why the pump would
be permitted to restart once current has been re-established. Whether the pump will start again
automatically when power is restored will depend on the type of motor control logic used.
Because pumps operating on a suction lift may lose their prime during the time that power is off, it is
preferable to use starters with low load protection for such installations to prevent an automatic
restart. This does not apply, of course, if the pumps are automatically primed, or if some protection
device is incorporated so that the pump can not run unless it is primed.
PUMP CLINIC 9
INTRODUCTION
Maximum noise levels for pumping machinery are being increasingly specified by customers as part
of the overall performance requirements. Alternatively, queries are being raised by users on the
noise levels that can be expected from their existing pumps.
These requirements are as a direct result of increasing awareness of the damage and nuisance to
health of operators that can result in prolonged exposure to high noise levels which in extreme cases
can lead to permanent loss of hearing. Additionally, there is the necessity to comply with statutory
Occupational Health and Safety requirements.
IMPORTANT: It should be recognised that noise levels emitted from similar sized machines
and drivers can vary between quite wide limits. For this reason, the noise levels that can be
estimated from the use of this data should be used with caution and never form part of a
contractual agreement. If guaranteed figures are required, consult your pump supplier.
There is difficulty sometimes in relating a sound pressure level (SPL) to a real practical situation.
Additionally, the overall SPL is a function of high intensity sound producing vibrations over a wide
range of frequencies audible to the human ear.
The graph below indicates typical noise levels by octave band frequencies for various real situations:
Pump Clinic 9 Noise Estimation (Sound Pressure Levels) for General Pumping Equipment 14/12/06 Page 1 of 6
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
The information that follows may be used to determine the expected sound pressure levels (SPL)
expressed in decibels (dB) for conventional single-stage and multi-stage centrifugal pumps. These
cannot be used for solids-handling, vortex or vertical mixed flow and axial flow pumps.
The values are based on test data and the following conditions:
a) Background noise 10dB (minimum) below all noise levels in each octave band.
b) Values are at a distance of 1 metre horizontally from pump surfaces and 1.5 metres above
the floor.
c) Overall noise level in dBA (“A“ scale) and octave mid band frequencies are basis “C” scale.
Table A below shows overall SPL for single, two-stage and three or more stage pumps based on
best efficiency point (BEP) power consumption unit:
When calculating absorbed kW, use the following specific gravity figures:
The values given in Table A are valid for flow rates that fall within the range of 75 - 125% of BEP
flow. For flow rates outside this range, a correction is required as detailed below:
a) For flows in the ranges 62%-75% and 126%-136% of BEP, add 1 to the dBA values on Table A.
b) For flows in the ranges 50%-61% and 137%- 50% of BEP, add 2 to the dBA values in Table A.
c) For flows in the range 38%-49% of BEP, add 3 to the dBA values in Table A.
d) For flows in the range 25%-37% of BEP, add 4 to the dBA values in Table A.
Pump Clinic 9 Noise Estimation (Sound Pressure Levels) for General Pumping Equipment 14/12/06 Page 2 of 6
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
To estimate the noise levels at mid-band frequencies, subtract the appropriate value as shown in
Table B from the overall dBA level.
Table B
When two or more sources produce the total combined noise level (mid-bands) may be obtained by
a simple addition of dB values according to Graph II:
Pump Clinic 9 Noise Estimation (Sound Pressure Levels) for General Pumping Equipment 14/12/06 Page 3 of 6
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
To obtain noise levels (either dBA or any mid-band frequency) for distances greater than 1 metre
from equipment, reduce the levels using the values shown in Graph III:
Noise at constant level of emittance does not have the same irritation at all frequencies. For a given
noise level, the higher the frequency the more objectionable the noise becomes.
Pump Clinic 9 Noise Estimation (Sound Pressure Levels) for General Pumping Equipment 14/12/06 Page 4 of 6
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
The impact of this can be seen by the A scale weighting factor shown on the example below:
Figure
Conversion of sound pressure levels at octave band mid frequencies to an overall SPL in dB(A)
Octave Band mid
31.5 63 125 250 500 1000 2000 4000 8000 1000
frequency (Hertz)
SPL in dB for frequency 84 82 84 85 84 83 82 79 76 -
A scale weighting factor
-39 -26 -10 -9 -3 0 +1 +1 +1 +6
dB
SPL dB(A) 45 56 74 76 81 83 83 80 77 -
dB correction for
.5 2.0 2.0 1.7 0
addition of SPLs Note 1
SPL dB(A) Note 2 56.5 78 85 84.7 77
dB correction for
0 3.0 0
addition of SPLs Note 1
SPL dB(A) Note 2 78 88 77
dB correction for
0 0
addition of SPLs Note 1
SPL dB(A) Note 2 88 0
dB correction for
0
addition of SPLs Note 1
SPL dB(A) Note 2
Worked Example
Find the maximum expected noise levels of a single stage pump at the following design conditions:
For the above duty, the impeller diameter is 540mm and pump efficiency is 83% Best efficiency
point (BEP) conditions are:
kW = l/sec x metres x S G
102.2 x efficiency
= 244.88 kW
Pump Clinic 9 Noise Estimation (Sound Pressure Levels) for General Pumping Equipment 14/12/06 Page 5 of 6
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Using Table A the expected overall SPL at a BEP power of 244.88kW is 87 dBA. A correction needs
to be applied to this value as the duty flow rate (490 l/s) is 68% of the BEP flow rate. The correction
is +1 dB. This will give an overall SPL at the duty point of 88 dBA.
The chart below summarises the estimates for dB values at mid band frequencies utilising Table B:
Assuming that the pump is driven by an 82 dBA turbine and a 91 dBA gear box, find the combined
noise level of the equipment at 1 metre and also at a distance of 15 metres.
Pump - 88 dBA
Turbine - 82 dBA
Gear - 91 dBA
Using Graph III: For a distance of 15 metres from the equipment - subtract 23dB for:
Pump Clinic 9 Noise Estimation (Sound Pressure Levels) for General Pumping Equipment 14/12/06 Page 6 of 6
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 10
Index
Page
5 Rotary Pumps
6 Lobe-Rotor Pumps
6 Condition Monitoring
Open all vent cocks to release trapped air and fill pump and
Lack of prime (see also above)
suction pipe completely with fluid
Check pump inlet for clogging etc. causing excessive friction
Excessive suction lift
head
Check that valves are open
Excessive discharge head Check piping for obstructions or blockage
Check total head
No discharge Check that pump rev/min is consistent with manufacturer’s
Speed too low
recommendations
Pump clogged Check that impeller is not clogged
Wrong direction of rotation Check that pump is running in the correct direction
Bleed suction pipe to clear air lock
Vapour lock
Check that suction pipe is properly submerged
Relief valve not properly adjusted Check adjustment, check for dirt on valve seat
Air leak Check seals, check line(s) for air leaks
Misalignment As above
Out-of-balance As above
Non-rigid mount As above
Bent shaft As above
Lack of vibration Check quantity and quality of lubricant
Dirt in pump Use filter to remove
Excessive wear
Check that pump materials are compatible with fluid being
Corrosion
handled
Check against manufacturer’s recommendations for fluid
Too high operating speed
viscosity
Operating pressure too high Reduce speed or pressure, eg change in system
Abrasives present in fluid Check product/pump suitability
In many designs of pump intended for severe service conditions, provision is made for easy
inspection, removal and replacement of worn parts without completely stripping down the pump, thus
simplifying maintenance and making if both logical and economic to attend to maintenance at
frequent intervals.
The only reason that really justifies shut-down and complete dismantling of the pump for inspection
and overhaul is a marked loss of performance. If this is suspected, the pump can first be tested to
establish if head, capacity or power input figures have deteriorated and whether remedial action is
needed.
Steam pumps are generally trouble-free in operation and require little attention other than routine
lubrication. Gaskets and packings will, however, require periodic replacement and maintenance will
be required to correct for mechanical wear etc. Disassembly is usually straightforward and a specific
procedure is usually detailed in the manufacturer’s instruction manual.
Power-driven reciprocating pump maintenance is largely confined to checking the valve condition
together with periodic attention to seals or packings. Valve seats may be re-ground, re-faced and
ground, or replaced and ground in, depending on the severity of the wear. Operating time before
such attention is required will vary widely with the type and design of pump and the service
conditions.
It is good practice to check the valves on a new pump after three months’ service unless there is
some indication that earlier attention is needed. After that, maintenance periods can be based on
experience.
Rotary Pumps
Rotary pumps in general have a minimum of moving parts and porting rather than valve systems,
hence maintenance is normally held to a minimum and is largely confined to inspection for wear,
corrosion or material defects, together with keeping the seals in good condition.
Disassembly normally follows a logical sequence when individual parts may be inspected for
condition and wear, and clearance is checked against permitted tolerances. Excessively worn or
defective parts should be replaced. Where corrosion is also present, it is advisable to check the
material specifications against the liquid being handled as a replacement part of a better material
may be available.
There may be specific instructions for disassembly, eg a spindle should be withdrawn in one
direction only to avoid damage to a seal. The manufacturer’s instruction book will, therefore, be the
primary guide for maximum permissible wear on individual components.
The performance of many rotary pumps, particularly of the lobe rotor and multi-screw types,
depends on accurate timing and the maintenance of prescribed clearances between the rotating
elements. Any deviation from such settings normally results in a marked reduction in pump life and
loss of efficiency. The importance of periodic timing and clearance checks is also emphasised.
In vane pumps, the principal wear is on the vanes themselves and the liner or casing on which they
rub. With sliding vanes replacement is usually a straightforward matter. In some designs, it is
possible to reassemble a worn vane the other way round and so double its effective life.
With swinging vanes, the shape of the pivoted portion of the rotor is such that a substantial degree of
wear can take place without interfering with the sealing efficiency. There will, however, be a limit to
the amount of wear which can be taken up purely by geometry and once this is exceeded, it will be
necessary to replace the vane. With both types, vanes may usually be replaced merely by removing
the cover plate or head plate, leaving the rotor in position.
Lobe-Rotor Pumps
Lobe-rotor pumps are precisely made machines; undue force must never be used during servicing.
Be especially careful not to damage the means of alignment between rotor case and gear case.
Ensure that sealing surfaces of mechanical seal rings are not scratched or damaged in any way. If
in doubt, obtain correct fitting length from manufacturer.
For shut-down and maintenance, isolate the pump from electrical and hydraulic supplies. Do not
allow the product to solidify in the pumping chamber or on shaft sealing surfaces. Product wetted
parts can often be cleaned with hot water and detergent.
It is advisable to hold components of the following types in stock. Quantities depend on pump
design, service conditions and stocking policy:
a) O-ring seals - pump head, shaft sleeve, front valves, mechanical seals, port
connections, gear case.
b) Oil seals and joints - gear case.
c) Glands - packing sets, preferably pre-formed rings.
d) Mechanical seals - rotating and stationary sealing rings, springs.
e) Transmission - flexible coupling parts, belts, gearbox parts.
f) Miscellaneous - lubricating oil, relief valve springs, fasteners.
Condition Monitoring
The terms predictive maintenance, or machine diagnostics, are often used to anticipate pump failure
and determine a probable cause. Incorrectly stating the problem is one of the most common errors
in troubleshooting pumps. The primary advantages of predictive maintenance are:
· Reduced maintenance
· Increased machinery availability
· Improved plant safety
Condition monitoring plays an important role in avoiding pump failure. In the case of centrifugal
pumps, three areas of monitoring can be distinguished:
Automatic monitoring has become more widely used, particularly in fields where safety is a major
factor. Comprehensive condition monitoring in the form of documental reports and data should be
kept to help identify recurring problems and determine optimum timing for scheduling preventative
maintenance.
Tracking the mean time between repair (MTBR) for all pumps serviced can improve reliability and
identify troublesome units.
Misalignment (where applicable) Check alignment of driver and pump and drive connection
Internal damage Bent or broken rotor
Unbalance If suspected, check rotor for static and dynamic balance
Air entrainment Re-position suction inlet
Excessive noise Air leaks Check and rectify
Cavitation Check against causes of cavitation
Relief valve set too high, adjust to correct setting consistent
Excessive pressure
with pump rating
Deterioration Check for excessive wear or clearances on components
Pump over-heats
Motor over-heats
Under capacity
No discharge
Seizure
Cause Remedy or Action
Pump over-heats
Motor over-heats
Under capacity
No discharge
Seizure
Cause Remedy or Action
DIAGNOSIS WILL BE GREATLY ASSISTED BY TAKING ON-STREAM PRESSURE READINGS AT THE PUMP’S INLET & OUTLET PORTS.
Where:
Before a pump is taken out of service, as much hydraulic performance data as possible should be
obtained. This can be achieved by accurately measuring flows, temperature, pressure, specific
gravity, viscosity etc around the pump and by consulting historical data and other useful information
about its operating condition.
Simple vibration monitoring and analysis can be an accurate and rapid method for detecting
mechanical problems in a pump.
The objective in monitoring rotating machinery vibrations is to determine when the rate of change in
vibration level begins to change.
Sealless pumps generally require more monitoring than mechanical seal pumps. Temperature
measurement, bearing condition monitoring, low current trips (for cavitation protection) are all useful
for magnetic drive pumps. Canned motor pumps also require liquid level monitoring.
Some of the most common problems with pumps include misalignment, oil contamination,
incomplete priming and the absence of detailed up-to-date spare parts records.
Pump Protection
In many cases, the reasons for pumping system failure are other than the pumping system itself.
These include improper selection of the pump, improper handling of the system and improper
selection of protective devices.
Typically, the selection of appropriate protective devices should take into account the following
points:
Protective devices incorporating these characteristics will almost certainly lead to longer pump life.
The development of sophisticated electronics, computers, chemical analysis, lasers, vibration pick-
ups, sonic measurements, ultra-sonics, and radio graphics have permitted the analysis and forecast
of the life of any pump component so that opening up maintenance can be carried out at a
reasonable period before expected failure, rather than at routine times.
It is probable that the cost of monitoring will amount to less than the costs of routine maintenance
disassembly, particularly if predictive maintenance is employed.
PUMP CLINIC 11
Page
1 Pump Purchasing Sequence
2-4 Engineering of System Requirements
-Fluid Type
-System-Head Curves
-Alternate Modes of Operation
-Margins
-Wear
-Future System Changes
5-7 Selection of Pump and Driver
-Pump Characteristics
-Code Requirements
-Fluid Characteristics
-Pump Materials
-Driver Type
7-10 Pump Specifications
-Specification Types
-Codes and Standards
-Alternates
-Bidding Documents
-Technical Specification
-Commercial Terms
10-11 Special Considerations
-Performance Testing
-Pump Drivers
-Intake
-Drawing & Date Requirement Forms
The sequence involved in obtaining a pumping system, following the initial decision that pumping
equipment is required for a system, and culminating with the purchase of the equipment, can be
divided into the following general steps:
In the process of specifying pumping equipment, the engineer is required to determine system
requirements, select the pump type, write the pump specification, and develop all information and
data necessary to define the required equipment for the supplier.
Having completed this phase of work, the engineer is then ready to take the necessary steps to
purchase the equipment. These steps include issuance of the specifications for bids or negotiations,
evaluation of pump bids, analysis of purchasing conditions, selection of supplier, and release of all
data necessary for purchase order issuance.
It is imperative that a pragmatic approach be taken when specifying requirements. Too often,
purchasers provide the same degree of specification, whether a pumpset is valued at $5000 or
$500,000. It is possible that documentation and testing requirements can be greater than the cost of
the equipment. The message is, ‘resist using a sledgehammer to crack a walnut.’
The first decision the engineer must make is to determine the requirements and conditions under
which the equipment will operate.
Fluid Type
One of the initial steps in the defining of the pumping equipment is the development of physical and
chemical data on the fluid handled, such as viscosity, corrosiveness, lubricating properties, chemical
stability, volatility and amount of suspended particles.
Depending upon the process and the system, some or all of these properties may have an important
influence on pump and system design. For example, the degree of corrosiveness of the fluid will
influence the engineer’s choice of materials of construction. If the fluid contains solids in suspension
suitable types of pump seal designs and abrasion-resistant pump construction materials may have to
be considered.
Fluid physical and chemical data of interest to the engineer should cover the entire expected
operating range of the pumping equipment. The influence of such parameters as temperature,
pressure and time upon the fluid properties, should also be considered.
System-Head Curves
The engineer should have a clear concept of the system in which the pumping equipment is to
operate. A preliminary design of the system should be made and include an equipment layout and a
piping and instrumentation diagram (or other suitable diagram), showing the various flow paths, their
preliminary size and length, elevation of system components and all valves, equipment, piping
specialties etc. which establish the system-head losses.
With this information, the engineer can develop system-head curves which show the graphic
relationship between flow and hydraulic losses in a pipe system. In calculating the hydraulic losses,
the engineer may need to include adequate allowances for corrosion and scale deposits etc. in the
system over the life of the plant.
Since hydraulic losses are a function of flow rate, piping sizes, and layout, each flow path in a
system will have its own characteristic curve. Care must be taken when specifying pump
characteristics to take into account the characteristic curve of each possible flow path served by the
equipment.
In specifying pump equipment, it is convenient to add the effect of static pressure and elevation
differences to the system-head curve to form a combined system-head curve. The resultant curve
shows the total head required of the pumping equipment to overcome system resistance. The pump
head must be at, or above, the combined system curve at all expected operating points, and for all
flow paths the pumping equipment is expected to serve.
These and other questions arising from different modes of operation greatly influence such decisions
as to the number of pumps, their capacities, and whether booster pumps are needed in some flow
paths.
The engineer should also consider the continuity of service expected of the pumping system. This
factor will influence the decision on number, type and capacity of installed spares and the quality
expected of the equipment.
Frequently, reliability considerations will dictate the use of multiple pumps, such as 2 full-size pumps
and 3 half-size pumps, or, where continuity is more important than full capacity, 2 half-size pumps.
Where 2 half-size, 3 third-size etc. pumps are used, loss of 1 pump will cause the others to run-out
on their system-head curves.
This run-out case should be evaluated when engineering the system and specifying the pump
characteristics. This loss of a pump can occur no only by pump malfunction, but by motor failure,
external damage, loss of power supply, loss of control power etc. The likelihood of these causes
should be evaluated as part of the pump selection process.
Margins
Pumps are frequently specified with margins over and above the normal rating. Over any long
period of time, it is possible that a system can operate at transient conditions, such as may be
caused by changes in modes of operation, malfunction in system components, or electric system
disturbances.
It is necessary for the engineer specifying the pumping equipment to examine the probability and
duration of such transients and to specify adequate margins to allow the equipment to undergo such
transients without damaging effects.
Margins (Cont.)
This also involves an evaluation of the combined effects of equipment cost, degree of criticality of
the system, inconvenience due to unavailability of the equipment and other economic and
technological factors.
Some transients often considered in design are pressure and temperature fluctuations, electric
voltage and frequency dips. If the maintenance of continuous flow is important, then adequate
margins must be allowed in the pump rating. For example, margins are added to the pump head
and capacity to allow the pumping equipment to maintain rated flow in case of small electric
frequency dips.
In addition, certain design features may be included to allow the pumps to operate without damage
through such transients as suction pressure dips which can cause cavitation. Pumps should not be
purchased for capacities greatly in excess of requirements. An over-sized pump could operate at
capacities less than those recommended by the manufacturer which could present mechanical and
hydraulic problems.
Wear
Wear is an ever-present factor in equipment and system design. No material that is handling fluids
or used in contacting moving surfaces is free from wear. Thus, operating characteristics of both the
pumping equipment and the system can be expected to change due to wear as time goes by. The
engineer should assess the extent of such wear over the life of the plant and provide adequate
margins in the system parameters so that the pumps can provide the expected flow, even at the end
of equipment life.
Where abrasive or suspended materials are handled, pumps with replaceable liners are frequently
specified. These liners are usually made of either resilient material such as rubber compounds, or
extremely hard alloys of cast iron. In addition, plastic linings (including impellers) are also frequently
chosen for these types of services.
In some applications, especially in power plants, the expected pump life is specified as the same as
plant life. However, the design life of a pump is a decision based on an evaluation of economic
factors. The wear margin to be added is a function of such factors as mode of operation (continuous
or intermittent) and fluid properties (abrasiveness, corrosiveness).
Thus it is important to review the possibilities and effects of such future system changes as well as
provide pumping equipment to satisfy the immediate system requirements.
The engineer should attempt to present future requirements based on projection of available data
and then evaluate the possibility and desirability of designing the equipment to allow for the changes
(such as providing extra flow or head margins or specifying a pump with impeller less than the
maximum for a given casing size) versus the desirability of modifying the whole system, including
the pumps, when the changes are made in the future.
In any event, it must always be kept in mind that the equipment must operate satisfactorily in the
present system and this should be a factor in whatever evaluation is being made.
The selection of the pump class and type for a particular application is influenced by such factors as
system requirements, system layout, fluid characteristics, intended life, energy cost, code
requirements and materials of construction.
The behaviour of the system has a very important bearing on the choice of pump:
Pump Characteristics
Constant-speed reciprocating pumps are suitable for applications where the required capacity is
expected to be constant over a wide range of system head variations. This type of pump is available
in a wide range of design pressures, from low to the highest produced. However, the capacity is
relatively small for the size of the equipment required.
That the output from a reciprocating pump will be pulsating is a factor to be considered. Where this
is objectionable, rotary pumps may be required. However, the application of rotary pumps is limited
to low to medium-pressure ranges.
It should be noted that some reciprocating and rotary pumps may be self-priming, but centrifugal
pumps, unless specifically designed as such, are not. This may be an important consideration in
certain applications.
In some cases, the system layout can influence the decision on the choice of pump type. In general,
centrifugal pumps will require less floor space than reciprocating pumps, and vertical pumps less
floor space than horizontal pumps. However, more head room may be required for handling the
vertical pumps during maintenance and installation.
Where the available NPSH is limited, such as when a saturated liquid is being handled, and the
application calls for a centrifugal pump, the engineer may have to investigate the use of a vertical
canned-suction centrifugal pump to gain adequate NPSH. In other cases, the design may call for
the installation of a pump immersed in the liquid handled, and here a vertical turbine pump may be
advantageous.
Code Requirements
The construction ratings and testing of most pumps normally used in industry are governed by codes
such as the ISO, ANSI, API or the Standards of the Hydraulic Institute. However, other codes of
regulatory bodies may impose additional requirements which can affect both pump rating and
construction. For example, the Boiler and Pressure Vessel Code requires feed pumps to be capable
of feeding the boiler when the highest set safety valves are discharging.
Fluid Characteristics
Fluid characteristics such as viscosity, density, volatility, chemical stability and solids content are
also important factors for consideration. Sometimes, exceptionally severe service may rule out
some classes of pumps at once. For example, the handling of fluids having solids content will
exclude the use of reciprocating piston pumps, or pumps with close clearances.
Rotary pumps are suitable for use with viscous fluids, such as oil or grease, whereas centrifugal
pumps can be used for both clean, clear fluids, and fluids with high solids content. On the other
hand, if it is undesirable for the process liquid to come into contact with the moving parts, diaphragm
pumps may have to be used.
Pump Materials
Materials are affected by both the pumped fluid and the environment. Resistance to corrosion and
wear are two of the more important material properties in this regard, and the engineer should
evaluate materials to determine which are most suitable and economical for the purpose intended.
Often this becomes an evaluation for the desirability of specifying expensive long-life materials
versus specifying cheaper materials which must be frequently replaced.
Operating factors, such as type of service (continuous or intermittent, critical or non-critical), running
speed preferences (high or low) and intended life, will also influence the engineer’s decision. For
example, equipment used in continuous and/or critical service will generally demand heavier duty
design and construction than equipment for intermittent and/or non-critical service. High-speed
operation, if allowed, will permit the use of smaller, usually less expensive equipment.
The life of the equipment cannot be predicted with certainty. For a given life expectancy, the
engineer must evaluate the effects of materials of construction, design, severity of service etc.
before making a choice.
Driver Type
The choice of driver type for the pumping equipment is as important as choosing the pump, for
frequently the driver can cost more than the pump. Depending on the available energy sources,
pumps may be driven by electric motors, steam turbines, steam engines, gas turbines, or internal
combustion engines. Also, pumps may be driven at constant speed or at variable speed. Variable
speed can enable centrifugal pumps to operate along the system-characteristic curve and thus save
on power for part-load operations.
Electric motor drives are usually used in constant-speed service unless a hydraulic coupling or other
speed-varying device is introduced into the system. Internal combustion engine drives are usually
chosen because of location (no electric power available), portability, or redundancy (loss of power
back-up) requirements.
They can operate as either constant-speed for variable-speed drivers. Steam turbines, eddy-current
couplings, adjustable-speed motors, fluid couplings, and gears and belts are frequently used where
variable-speed operation is required.
In large, complex installations where the equipment is to be operated continuously, the decision as
to type of driver and variability of pump speed should be based on a comparison to the total
operating and capital costs for the pump system over the intended plant life for the several
alternatives.
Variable-speed operation would usually result in lower operating costs. However, the total first cost
of the driving equipment to accomplish this would frequently be higher than for constant-speed
equipment. The first cost should include cost of equipment, building space etc and the operating
cost should include such factors as energy costs, maintenance, and replacement costs. This
comparison usually results in the choosing of a pumping unit that provides the lowest cost per gallon
pumped over the useful life of the plant.
Pump Specifications
Specification Types
When selecting a pump, the first decision to be made is whether the procurement will be based upon
a formal specification, or whether some abbreviated form of requirements will be suitable. For
relatively simple or inexpensive pumps, or for replacement pumps where duplication is desired, a
specification is frequently not used.
For inexpensive pumps, the time and cost required to write a specification and obtain and analyse
competitive bids, frequently exceed the potential cost-savings. In this case, and where the pump
supplier is already established (replacement/duplication), a direct quotation is frequently requested
from the supplier. It is important, when requesting this quotation, to have the principal requirements
well defined and known to the supplier so that they can be properly included in the technical, and
priced, offering.
Thus, while a formal specification may not be appropriate, the purchaser should have the
requirements well established.
Attached are four (4) data sheets that can be utilised for enquiry purposes. As a minimum, details
under Operating Conditions and the Quantity required are to be defined by the purchaser.
For example, where either leak-off or mechanical shaft seals may be offered, the performance
specification usually states a preference. The performance specification, however, basically
establishes ‘what’ not ‘how’.
The construction specification establishes in some detail the type of design, construction, and
methods to be employed in designing the pump and certain other features which, if the performance
specification is utilised, are left to the manufacturer’s discretion. From the standpoint of legal
responsibility, if a construction specification is used, manufacturers may respond and advise that
since the purchaser has established certain design features of the pump, the manufacturer cannot
be responsible for the performance.
It is therefore, important that care be taken when writing a construction specification not to relieve
the manufacturer of responsibility for applicability, suitability, and performance and that care also be
exercised by the purchaser to avoid any unnecessary assumption of responsibility for the proper
performance of the pump.
In short, unless there are unusual circumstances, it is far more appropriate to specify the pump on
the basis of performance required, rather than construction, unless the purchaser has a high degree
of assurance that the requirements called out in the specification can be met and that the pump
supplier will not be relieved of responsibility.
An example of this is the ASME Boiler and Pressure Vessel Code, Section III. These codes, in
some cases, are extremely detailed in specifying pump construction and are a rather well-defined
specification in themselves. It is, of course, essential to establish the dimensional standards which
apply, such as SI or English, the codes which may apply to the construction and fittings of the pump
and the industrial codes that apply to the application of the pump for the service intended.
In all cases however, the engineer must review each reference to ensure that it does not introduce
conflict. Some codes and standards include alternate choices of material or inspection methods
requiring selection by the engineer. Others may include cross references to additional codes which
the engineer may wish to exclude.
Alternates
It is extremely difficult for a specification to cover all possible pumps offered by various
manufacturers. Coupled with that is the problem faced by a potential user in remaining up-to-date
with the changing state-of-the-art, and the development work being performed by manufacturers.
It is good practice to allow manufacturers to offer alternatives. This gives them an opportunity to
present their best offer and also gives the buyer the advantage of obtaining potentially attractive,
alternate offerings. However, the choice of whether or not to accept the alternates is fully up to the
purchaser who may choose to reject any, and all bids, including alternates.
Bidding Documents
The bidding documents for pumps normally consist of two major parts:
1. Technical specifications
2. Commercial terms
The technical specification establishes the performance requirements, materials of construction and
major technical features. The commercial terms include the contract language and cover such items
as the location of the work, requirements for guarantees/warranties, shipping method, time of
delivery, method of payment, normal inspection and expediting requirements.
Frequently, the commercial terms and conditions are relegated to second place, especially when
standard inexpensive pumps are being bought, but in many cases, the commercial terms can
assume more significance than many of the performance requirements.
Technical Specification
The technical specification should consist of a series of carefully defined and distinct sections. The
more complete and specific the specification, the more competitive will be the bid prices. A typical
specification might contain the following:
1. Scope of work: Pump, baseplate, driver (if included), interconnecting piping, lubricating oil
pump and piping, spare parts, instrumentation (pump-mounted), erection supervision.
2. Work not included: Foundations, installation labour, anchor bolts, external piping, external
wiring, motor starter
3. Rating and service conditions: Fluid pumped, chemical composition, temperature, flow,
head, speed range, preference, load conditions, overpressure, run-out, off-standard
operating requirements, transients.
4. Design and construction: Care should be taken to provide latitude in this section, as this
borders on dictating construction requirements. Codes, standards, materials, type of casing,
stage arrangement, balancing, nozzle, orientation, special requirements for nozzle forces
and movement (if known), weld-end standards, supports, vents and drains, bearing type,
shaft seals, baseplates, interconnecting piping, resistance temperature detectors,
instruments, insulation.
5. Lubricating oil system (if applicable): System type, components, piping, mode of operation,
interlocks, instrumentation.
6. Driver: Motor voltage standards, power supply and regulation, local panel requirements,
wiring standards, terminal boxes, electric devices. For internal combustion drivers, fuel type
preferred (or required), number of cylinders, cooling system, speed governing, self-starting
or manual, couplings or clutches, exhaust muffler.
7. Cleaning: Cleaning, painting, preparation for shipment, allowable primers and finish coats,
flange and nozzle protection, integral piping protection, storage requirements.
8. Performance testing: Satisfactory for service, smooth-running, free of cavitation and
vibration, shop tests for pump and spare rotating elements, hydrostatic tests, test curves,
field testing.
9. Drawing and data: Drawings and data to be furnished, outline, speed versus torque curves,
WK² data, instruction manuals, completed data sheets, recommended spare parts.
10. Tools: One set of any special tools.
11. Evaluation basis: Power, efficiency, proven design.
Supplementing these may be technical specifications relating to other requirements of the order,
such as specifications for the electric motor, steam turbine, or other type of driver, a specification on
marking for shipment, a specification on painting, and requirements for any supplementary quality
control testing.
In addition, it is important that any unusual requirements be listed in the technical specification so
that the manufacturers are aware of them. Examples of these are special requirements for repair of
defects in pump castings, a sketch of the intake arrangement for wet-pit applications and special
requirements regarding unique testing, for example, metallurgical testings which may be required
during manufacture, apart from performance testing.
It is helpful to the pump supplier to provide system-head curves, sketches of the piping system
(dimensioned if this is significant), listing of piping and accessories required etc.
Pump data sheets are extremely useful in providing a summary of information to the bidder and also
in allowing the ready comparison of bids by various manufacturers. As can be noted, by inspecting
these sheets, some of the items are filled in by the purchaser and the balance by the bidder to
provide a complete summary of the characteristics of the pump, the materials to be furnished,
accessories, weight etc. The data sheets should be included with the technical specification.
Commercial Terms
The commercial terms included with the bidding documents should cover the following information:
1. General: Name of buyer, place to which proposals are to be sent, information on ownership
of documents, time allowed to bid, governing laws and regulations.
2. Location of plant site: This establishes the geographic area in which the equipment is to
perform and in a broad way the scope of the work. It should also state maximum
temperatures, humidity, storage provisions (indoor or outdoor), and altitude (so that the
motor drivers can be selected for the proper cooling).
3. Proposal: This establishes the format of the proposal, number of copies, owner’s right to
accept or reject any bids, status of alternates.
4. Schedule: Including requirements for all drawings and design data submittals,
manufacturing schedule, and equipment delivery.
Thus a bid could either include a freight allowance, that is, be FOB manufacturer’s plant
WFA (with freight allowed) to point of use, or be FOB point of use, in which case freight is
included. In either case, the risk of loss remaining with the seller and that assumed by the
purchaser should be clearly stated.
Special Considerations
Performance Testing
An important part of any specification is the requirement for testing. Normally, small commercial
pumps that are routinely produced by a manufacturer, up to about 6in (150mm) are tested on a
sample selection, quality control basis and from that, standardised curves of pump performance are
available. Thus for pumps of this size, it is not necessary to require certified tests unless the pumps
are to be used in critical service, such as fire protection or boiler feed.
However, for larger pumps or pumps with more critical service requirements, a certified performance
test should be required. This test requires the manufacturer to test the pump at several points on its
performance curve to establish its exact head curve. Since it is necessary to assure the pump driver
is of the proper size, power curve must be furnished with the head curve.
Occasionally, pumps for special services or extremely large pumps cannot be tested in the
manufacturer’s shop. Examples of this are very large low-head pumps for circulating water service,
low-lift irrigation pumps and pumps for liquid-metal service. The actual performance testing in this
case takes place following installation of the pump.
It is important that the purchaser and the supplier agree upon a proper (field) test method in some
detail. This method should include the number of points at which the head curve will be determined,
the applicable code, the specific method of traversing the pump discharge characteristics across the
cross-section of the discharge pipe, and the manner in which the head will be varied.
Care should be taken in establishing this procedure to set forth the characteristics of the fluid and
other variables which can affect the performance test. The specification should establish the
performance testing requirements for the pump and whether or not it is necessary the testing of the
pump be witnessed.
Witnessing and furnishing of certified test data (including the test work done) are frequently priced
separately and if not specified, can be a source of dispute between purchaser and supplier.
Pump Drivers
Pump drivers (motors, turbines, engines etc) can be purchased either with the pumps or separately.
With small pumps, pumps using ‘monobloc’ construction (where the pump is mounted on, and
supported by, the motor) and pumps built to special codes, such as ‘Underwriters’ engine-powered
fire pumps, there is not usually any cost-advantage to buying the driver separately.
If the driver is purchased separately and can be economically and conveniently shipped to the pump
supplier’s plant, the pump supplier should be required to mount the driver half of the coupling, as
well as align and mount the driver (for common baseplate installations). For very large drivers, or
where it is costly or impractical to ship the driver to the pump supplier, it will be necessary to perform
this work at the point of installation. To assure compatibility with the other drivers in the plant, it is
important to specify the driver enclosure type, insulation standards, and special features required,
such as heaters and oversize junction boxes.
For steam turbine drivers, speed range, throttle pressure, steam quality, exhaust pressure and
control method should also be specified.
Intake
Vertical wet-pit pumps are sensitive to the geometry of their suction pit. Factors to consider include
clearance beneath the bottom of the suction bell and the floor of the pit, spacing between pumps or
between the pump and the pit walls (both side and rear), the approach angle of the floor of the pit
(including surging and surcharge), submergence and lack of uniform approach flow.
The standards of the Hydraulic Institute include recommendations on the geometry of the intakes.
These, as well as the recommendations of the pump manufacturer, should be carefully reviewed.
Suction piping, where complex or unusual, should be treated in a similar manner
When specifying vertical wet-pit pumps, a layout of the installation should be furnished to the bidders
for their information and comment. In many cases, if the geometry of the installation is not fixed,
bidders can recommend small changes to improve pump performance. Where the geometry is
fixed, it may be necessary to add anti-vortexing baffles, surge walls, or flow-directing vanes (or
walls) to avoid pump operating problems.
For moderate or large installations where any design question exists, model testing may be
considered. Several pump manufacturers offer this as a service, as do a number of universities and
commercial testing laboratories. Responsibility for proper pump performance will rarely be assumed
by the bidder when the intake pit is of non-optimum size or shape. The use of model testing is
usually resorted to in these cases also.
PUMP CLINIC 12
Radial Thrust
A centrifugal pump consists of a set of rotating vanes, enclosed within a housing or casing and used
to impart energy to a fluid through centrifugal force. Thus, stripped of all refinements, a centrifugal
pump has two main parts:
In a centrifugal pump the liquid is forced by atmospheric or other pressure into a set of rotating
vanes. These vanes constitute an impeller which discharges the liquid at its periphery at a higher
velocity. This velocity is converted to pressure energy by means of a volute (Fig 1.1) or by a set of
stationary diffusion vanes (Fig 1.2) surrounding the impeller periphery. Pumps with volute casings
are generally called volute pumps, while those with diffusion vanes are called diffuser pumps.
Diffuser pumps were once quite commonly called turbine pumps, but this term has recently been
more selectively applied to the vertical deep-well centrifugal diffuser pumps usually referred to as
vertical turbine pumps.
The diffuser is seldom applied to a single-stage radial-flow pump. Except for certain high-pressure
multi-stage pump designs, the major application of diffusion vane pumps is in vertical turbine pumps
and in single-stage low-head propeller pumps.
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 1 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
In a single-volute pump casing design (Fig 1.3) uniform or near-uniform pressures act on the
impeller when the pump is operated at design capacity (which coincides with the best efficiency). At
other capacities, the pressures around the impeller are not uniform and there is a resultant radial
reaction. A graphical representation of the typical change in this force with pump capacity is shown
in (Fig 1.4) - NOTE that the force is greatest at shut-off.
For any percentage of capacity, radial reaction is a function of total head and of the width and
diameter of the impeller. Thus a high-head pump with a large-diameter impeller will have a much
greater radial reaction force at partial capacities than a low-head pump with a small-diameter
impeller.
A zero radial reaction is not often realised; the minimum reaction occurs at design capacity. In a
diffuser-type pump which has the same tendency for over-capacity unbalance as a single-volute
pump, the reaction is limited to a small arc repeated all around the impeller with the individual forces
cancelling each other.
In a centrifugal pump design, shaft diameter and bearing size can be affected by allowable deflection
as determined by shaft span, impeller weight, radial reaction forces and the torque to be transmitted.
Formerly, standard designs compensated for reaction forces if maximum-diameter pump impellers
were used only for operations exceeding 50% of design capacity.
For sustained operations at lower capacities, the pump manufacturer, if properly advised, would
supply a heavier shaft, usually at a much higher cost. Sustained operation at extremely low flows,
without informing the manufacturer at the time of purchase, is a much more common practice today.
The result is broken shafts, especially on high-head units.
Because of the increasing operation of pumps at reduced capacities, it has become desirable to
design standard units to accommodate such conditions. One solution is to use heavier shafts and
bearings. Except for low-head pumps in which only a small additional load is involved, this solution
is not economical. The only practical answer is a casing design that develops a much smaller radial
reaction force at partial capacities. One of these is the double-volute casing design, also called twin-
volute or dual-volute.
The application of the double-volute design principle to neutralise reaction forces at reduced
capacity is illustrated in (Fig 1.5). Basically, this design consists of two 180º volutes; a passage
external to the second, joins the two into a common discharge. Although a pressure unbalance
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 2 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
exists at partial capacity through each 180º arc, forces F1 and F2 are approximately equal and
opposite, thereby producing little, if any, radial force on the shaft and bearings.
The pressures generated by a centrifugal pump exert forces on both its stationary and rotating parts.
The design of these parts balances some of these forces, but separate means may be required to
counter-balance others. Axial hydraulic thrust is the summation of unbalanced impeller forces acting
in the axial direction. As reliable large-capacity thrust bearings are not readily available, axial thrust
in single-stage pumps remains a problem only in larger units.
Theoretically, a double-suction impeller is in hydraulic axial balance with the pressures on one side
equal to, and counter-balancing the pressures on, the other (Fig 2.1). In practice, this balance may
not be achieved for the following reasons:
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 3 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
The suction passages to the two suction eyes may not provide equal or uniform flows to the two
sides.
1. External conditions such as an elbow being too close to the pump suction nozzle may cause
unequal flows to the suction eyes.
2. The two sides of the discharge casing may not be symmetrical, or the impeller may be
located off-centre. These conditions will alter the flow characteristics between the impeller
shrouds and casing, causing unequal pressures on the shrouds.
3. Unequal leakage through the two leakage joints will tend to upset the balance.
Combined, these factors create definite axial unbalance. To compensate for this, all centrifugal
pumps, even those with double-suction impellers, incorporate thrust bearings.
The ordinary single-suction radial-flow impeller with the shaft passing through the impeller eye (Fig
2.1) is subject to axial thrust because a portion of the front wall is exposed to suction pressure, thus
exposing relatively more back wall surface to discharge pressure. If the discharge chamber
pressure were uniform over the entire impeller surface, the axial force acting towards the suction
would be equal to the product of the net pressure generated by the impeller and the unbalanced
annular area.
Actually, pressure on the two single-suction impeller walls is not uniform. The liquid trapped
between the impeller shrouds and casing walls is in rotation and the pressure at the impeller
periphery is appreciably higher than at the impeller hub. Although we need not be concerned with
the theoretical calculations for this pressure variation, (Fig 2.2) describes it qualitatively. Generally
speaking, axial thrust towards the impeller suction is about 20% to 30% less than the product of the
net pressure and the unbalanced area.
To eliminate the axial thrust of a single-suction impeller, a pump can be provided with both front and
back wearing rings. To equalise thrust area, the inner diameter of both rings is made the same (Fig
2.3). Pressure approximately equal to the suction pressure is maintained in a chamber located on
the impeller side of the back wearing ring by drilling so-called balancing holes through the impeller.
Leakage past the back wearing ring is returned into the suction area through these holes.
However, with large single-stage suction pumps, balancing holes are considered undesirable
because leakage back to the impeller suction opposes the main flow, creating disturbances. In such
pumps, a piped connection to the pump suction replaces the balancing holes.
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 4 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Another way to eliminate or reduce axial thrust in single-suction is by use of pump-out vanes on the
back shroud. The effect of these vanes is to reduce the pressure acting on the back shroud of the
impeller (Fig 2.4). This design, however, is generally used only in pumps handling gritty liquids
where it keeps the clearance space between the impeller back shroud and the casing free of foreign
matter.
So far, the discussion of the axial thrust has been limited to single-suction impellers with a shaft
passing through the impeller eye and located in pumps with two stuffing boxes, one on either side of
the impeller. In these pumps, suction pressure magnitude does not affect the resulting axial thrust.
On the other hand, axial forces acting on an overhung impeller with a single stuffing box (Fig 2.5) are
definitely affected by suction pressure.
In addition to the unbalanced force found in a single-suction, two-box design (Fig 2.2) there is an
axial force equivalent to the product of the shaft area through the stuffing box and the difference
between suction and atmospheric pressure. This force acts towards the impeller suction when the
suction pressure is less than the atmospheric, or in the opposite direction, when it is higher than the
atmospheric.
When an overhung impeller pump handles a suction lift, the additional axial force is very low. For
example; if the shaft diameter through the stuffing box is 2” (area = 3.14 sq.in) and if the suction lift
is 20ft of water (absolute pressure – 6.06 psia), the axial force caused by the overhung impeller and
acting towards the suction will be only 27lb.
On the other hand, if the suction pressure is 100 psi, the force will be 314lb and acts in the opposite
direction. Therefore, as the same pump may be applied for many conditions of service over a wide
range of suction pressures, the thrust bearing of pumps with single-suction overhung impellers must
be arranged to take thrust in either direction. They must also be selected with sufficient thrust
capacity to counteract forces set up under the maximum suction pressure established as a limit for
that particular pump.
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 5 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
It might seem that the advantages of balanced axial thrust and greater available suction area in a
double-suction impeller would warrant applying such impellers to multistage pumps. But there are
definite shortcomings to this practice. The average multistage pump has relatively low capacity
when compared to the entire range covered by modern centrifugal pumps.
It is seldom necessary, therefore, to use double-suction impellers just to reduce the net positive
suction head (NPSH) required for a given capacity. Even if a double-suction impeller is desirable for
the first stage of a large capacity multistage pump, it is hardly necessary for the remaining stages.
As to the advantage of the axial balance it provides, it must be considered that a certain amount of
axial thrust is actually present in all centrifugal pumps and the necessity of a thrust bearing is
therefore not eliminated.
Most important, the use of double-suction impellers in a multistage pump adds needless length to
the pump shaft span. Additional space is required for the extra passage leading to the second inlet
of each successive stage. In a pump with four or more stages (Fig 2.6) this increase becomes quite
appreciable and causes additional casting difficulties. If shaft diameter is increased to compensate
for the longer span so as to maintain reasonable shaft deflection, the impeller inlet areas are
correspondingly reduced.
The result is that the advantage of superior suction conditions usually offered by double-suction
impellers is considerably reduced. Finally, as it is impractical to arrange the various double-suction
impellers in any but the ascending order of the stages, the impeller at one end of the casing
becomes the last stage impeller and the pressure acting on the adjacent stuffing box becomes the
discharge pressure on the next-to-last stage.
To reduce this pressure, a pressure-reducing bushing must be interposed between the last-stage
impeller and the stuffing box and this bushing further increases the overall length. The result of all
these considerations is that most multistage pumps are built with single-suction impellers.
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 6 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Two obvious single-suction impeller arrangements for a multistage pump are as follows:
Several single-suction impellers may be mounted on one shaft, each having its suction inlet facing in
the same direction and its stages following one another in ascending order of pressure (Fig 2.7).
The axial thrust is then balanced by a hydraulic balancing device.
An even number of single-suction impellers can be mounted on one shaft, one half of these facing in
an opposite direction to the second half. With this arrangement, axial thrust on the one half is
compensated by the thrust in the opposite direction on the other half (Fig 2.8). This mounting of
single-suction impellers back-to-back is frequently called ‘opposed impellers’.
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 7 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
An uneven number of single-suction impellers may be used with this arrangement, provided the
correct shaft and interstage bushing diameters are used to give the effect of an hydraulic balancing
device that will compensate for the hydraulic thrust on one of the stages. It is important to note that
the opposed impeller arrangement completely balances axial thrust only under the following
conditions:
Except for some special pumps that have an internal and enclosed bearing at one end, and
therefore only one stuffing box, most multistage pumps fulfil the first condition. But because of
structural requirements, the last two conditions are not practical. A slight residual thrust is usually
present in multistage opposed-impeller pumps, unless impeller hubs or wearing rings are located on
different diameters for various stages.
Because such a construction would eliminate axial thrust only at the expense of reduced inter-
changeability and increased manufacturing costs, this residual thrust, being relatively small, is
usually carried on the thrust bearing.
If all the single-suction impellers of a multistage pump face in the same direction, the total theoretical
hydraulic axial thrust acting towards the suction end of the pump will be the sum of the individual
impeller thrusts. The thrust magnitude (in pounds) will be approximately equal to the product of the
net pump pressure (in pounds per square inch) and the annular unbalanced area (in square inches).
Actually, the axial thrust turns out to be about 70% to 80% of this theoretical value.
Some form of hydraulic balancing device must be used to balance this axial thrust and to reduce the
pressure on the stuffing box adjacent to the last-stage impeller. This hydraulic balancing device may
be a balancing drum, a balancing disk or a combination of the two.
Balancing Drums
The balancing drum is illustrated in (Fig 3.1). The balancing chamber at the back of the last stage
impeller is separated from the pump interior by a drum that is either keyed or screwed to the shaft
and rotates with it. The drum is separated by a small radial clearance from the stationary portion of
the balancing device, called the ‘balancing drum head’ which is fixed to the pump casing.
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 8 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
The balancing chamber is connected either to the pump suction or to the vessel from which the
pump takes its suction. Thus the back pressure in the balancing chamber is only slightly higher than
the suction pressure, the difference between the two being equal to the friction losses between this
chamber and the point of return. The leakage between the drum and the drum head is, of course, a
function of the differential pressure across the drum and of the clearance area.
The forces acting on the balancing drum in (Fig 3.1) are the following:
1. Toward the discharge end; the discharge pressure multiplied by the front balancing area
(Area B0) of the drum.
2. Toward the suction end; the back pressure in the balancing chamber multiplied by the back
balancing area (Area C) of the drum.
The first force is greater than the second thereby counter-balancing the axial thrust exerted upon the
single-suction impellers. The drum diameter can be selected to balance axial thrust completely or
within 90% to 95% depending on the desirability of carrying any thrust-bearing loads.
It has been assumed in the preceding simplified description that the pressure acting on the impeller
walls is constant over their entire surface and that the axial thrust is equal to the product of the total
net pressure generated and the unbalanced area. Actually, this pressure varies somewhat in the
radial direction because of the centrifugal force exerted upon the water by the outer impeller shroud
(Fig 2.4).
Furthermore, the pressures at two corresponding points on the opposite impeller faces (D and E Fig
3.1) may not be equal because of variation in clearance between the impeller wall and the casing
section separating successive stages. Finally, pressure distribution over the impeller wall surface
may vary with head and capacity operating conditions.
This pressure distribution and design data can be determined by test quite accurately for any one
fixed operating condition and an effective balancing drum could be designed on the basis of the
forces, resulting from this pressure distribution. Unfortunately, varying head and capacity conditions
change the pressure distribution, and as the area of the balancing drum is necessarily fixed, the
equilibrium of the axial forces can be destroyed.
The objection to this is not primarily the amount of the thrust but rather that the direction of the thrust
cannot be pre-determined because of the uncertainty about internal pressures. Still, it is advisable
to pre-determine normal thrust direction as this can influence external mechanical thrust-bearing
design. Because 100% balance is unattainable in practice, and because the slight but predictable
unbalance can be carried on a thrust bearing, the balancing drum is often designed to balance only
90% to 95% of total impeller thrust.
The balancing drum satisfactorily balances the axial thrust of single-suction impellers and reduces
pressure on the discharge side stuffing box. It lacks, however, the virtue of automatic compensation
for any changes in axial thrust caused by varying impeller reaction characteristics. In effect, if the
axial thrust and balancing drum forces become unequal, the rotating element will tend to move in the
direction of the greater force.
The thrust bearing must then prevent excessive movement of the rotating element. The balancing
drum performs no restoring function until such time as the drum force again equals the axial thrust.
This automatic compensation is the major feature that differentiates the balancing disk from the
balancing drum.
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 9 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Balancing Disks
The operation of the simple balancing disk is illustrated in (Fig 3.2). The disk is fixed to, and rotates
with, the shaft. It is separated by a small axial clearance from the balancing disk head which is fixed
to the casing. The leakage through this clearance flows into the balancing chamber and from there,
either to the pump suction or to the vessel from which the pump takes its suction.
The back of the balancing disk is subject to the balancing chamber back pressure, whereas the disk
face experiences a range of pressures. These vary from discharge pressure at its smallest diameter
to back pressure at its periphery. The inner and outer disk diameters are chosen so that the
difference between the total force acting on the disk face and that acting on its back will balance the
impeller axial thrust.
If the axial thrust of the impellers should exceed the thrust acting on the disk during operation, the
latter is moved towards the disk head, reducing the axial clearance between the disk and the disk
head. The amount of leakage through the clearance is reduced so that the friction losses in the
leakage return line are also reduced lowering the back pressure in the balancing chamber.
This lowering of pressure automatically increases the pressure difference acting on the disk and
moves it away from the disk head, increasing the clearance. Now the pressure builds up in the
balancing chamber and the disk is again moved towards the disk head until an equilibrium is
reached.
To assure proper balancing disk operation, the change in back pressure in the balancing chamber
must be of an appreciable magnitude. Thus, with the balancing disk wide open with respect to the
disk head, the back pressure must be substantially higher than the suction pressure to give a
resultant force that restores the normal disk position. This can be accomplished by introducing a
restricting orifice in the leakage return line that increases back pressure when leakage past disk
increases beyond normal.
The disadvantage of this arrangement is that the pressure on the stuffing box packing is variable -
a condition that is injurious to the life of the packing and therefore to be avoided. The higher
pressure that can occur at the packing is also undesirable.
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 10 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
The rotating portion of this balancing device consists of a long cylindrical body that turns within a
drum portion of the disk head. This rotating part incorporates a disk similar to the one previously
described. In this design, radial clearance remains constant regardless of disk position, whereas the
axial clearance varies with the pump rotor position.
1. Towards the discharge end; the sum of the discharge pressure multiplied by Area A, plus
the average intermediate pressure multiplied by Area B.
2. Towards the suction end; the back pressure multiplied by Area C.
Whereas the position-restoring feature of the simple balancing disk required an undesirably wide
variation of the back pressure, it is now possible to depend upon a variation of the intermediate
pressure to achieve the same effect.
Here is how it works. When the pump rotor moves towards the suction end (to the left in Fig 3.3)
because of increased axial thrust, the axial clearance is reduced and pressure builds up in the
intermediate relief chamber, increasing the average value of the intermediate pressure acting on
Area B. In other words, with reduced leakage, the pressure drop across the radial clearance
decreases, increasing the pressure drop across the axial clearance.
The increase in intermediate pressure forces the balancing disk towards the discharge end until
equilibrium is reached. Movement of the pump rotor towards the discharge end would have the
opposite effect, increasing the axial clearance and the leakage, and decreasing the intermediate
pressure acting on Area B.
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 11 of 12
Kelair Pumps Australia Pty Ltd A Mace Group Company ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Call Direct: 1300 789 466 Facsimile: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
There are now in use numerous hydraulic balancing device modifications. One typical design
separates the drum portion of a combination device into two halves, one preceding and the second
following the disk (Fig 3.4). The virtue of this arrangement is a definite cushioning effect at the
intermediate relief chamber thus avoiding too positing a restoring action which might result in the
contacting and scoring of the disk faces.
Pump Handbook ed.Igor J. Karassik, William C. Krutzsch, Warren H. Fraser, Joseph P. Messina (USA:1986)
Igor J. Karassik, Roy Carter, Centrifugal Pumps (USA:1960)
Pump Clinic 12 Radial & Axial Thrust in Centrifugal Pumps 12/09/06 Page 12 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 13
Diaphragms
Flexible diaphragms are clamped at their outer perimeters, between the inner and outer
chambers. The diaphragms are connected at their movable centers by a rod.
Check Valves
As fluid moves through the pump, check valves open and close. This allows each outer chamber
to alternately fill and discharge. The check valves respond to differential pressures. Ball-type
check valves can pass very small particles.
As the Air Distribution Valve directs pressurised air to the left diaphragm, the diaphragm is pushed
outward. This is a discharge stroke, which forces liquid from the left outer chamber. Discharged
liquid moves from the chamber, through an open discharge check valve, and exits the pump at the
discharge manifold.
The position of the discharge port can be top, bottom or side. As the left diaphragm is pressurised
outward, the connecting rod pulls the right diaphragm inward on a suction stroke, which fills the
left chamber with fluid. Liquid enters the pump at the suction manifold, moves through an open
suction check valve and fills the chamber. At the end of the cycle, the Air Distribution Valve
automatically shifts the air pressure to the opposite diaphragm, initiating another pumping cycle.
PUMP CLINIC 14
In the pump business alignment means that the centerline of the pump is aligned with the
centerline of the driver. Although this alignment was always a consideration with packed pumps, it
is critical with sealed pumps especially if you are using rotating seal designs where the springs or
bellows rotate with the shaft.
A little misalignment at the power end of the pump is a lot of misalignment at the wet end, and
unfortunately that is where the seal is located in most pump applications.
Regardless of the alignment method you select, you must start with a pump and driver in good
repair. A perfectly aligned piece of junk is still a piece of junk. You should also check the
following:
· A straight shaft that has been dynamically balanced.
· Good wear rings with the proper clearance.
· The correct impeller to volute, or backplate clearance.
· The elimination of "soft foot".
· Eliminate all pipe strain.
· Good bearings installed on a shaft with the proper finish and tolerances.
· A good mechanical seal set at the proper face load. The closer the seal is to the pump
bearings the better off you are going to be.
I see lots of pumps that have never been aligned properly. When you talk to the people that
should be concerned, you get the following comments:
· Alignment is not important. I have been working with pumps for years and we never do it at
this facility. And we do not do dynamic balancing of the rotating assembly either!
· There is no time to do an alignment. Production wants the unit back on line, and they will
not allow me the time to do it properly.
· We purchase good couplings. The coupling manufacturer states that their coupling can
take a reasonable amount of misalignment.
It turns out that there are at least three methods of getting a good pump to driver alignment, and a
good coupling is not one of them. The coupling is used to transmit torque to the shaft and
compensate for axial thermal growth, nothing else. You install a good coupling after you have
made the pump to driver alignment, not instead of making the alignment.
The reverse indicator method is an acceptable method, but it does take a great deal of time.
There are plenty of schools that teach this method if you are interested in learning how to do it:
The laser is the latest method. It is also the most popular. There are lots of people that can teach
you to use the equipment, once you have made the purchase.
The "C or D" frame adapter is probably the easiest method of all and available from most quality
pump manufacturers. It solves most of the problems with thermal expansion.
If you do not have a C or D frame adapter you will be involved in the last three steps of the four-
step procedure.
Once you have made all the measurements, put in the recommended compensation for thermal
expansion, and figured out all the calculations for how much to move the driver, and in which
direction; now comes the fun part; moving the driver.
You can hit the motor with a big hammer, but small dimensions are hard to get with this method.
Some people use an adjusting wheel that attaches to shims. This will give you a very precise
movement that is necessary for a proper alignment
Another method is to use an adjusting wheel that slips over the motor hold down bolts. Many
mechanics make there own tools and these units also work very well for precise motor movement.
How concerned should you be about alignment? You do it on your automobile when you notice
uneven tyre wear or the car drifts to one side of the road when you loosen your grip on the wheel,
and have no problem justifying the cost and time involved. It is the same logic you use towards
the added cost and time spent balancing the tyres and wheels of your car.
We do not always apply the same logic to our very expensive rotating equipment in the shop, but
we should. A mechanical seal should run trouble-free until the carbon sacrificial face has worn
away.
When we inspect the seals we remove from leaking pumps we find that in better than 85% of the
cases there is plenty of carbon face left on the seals. The seals are leaking prematurely and the
seal movement caused by pump to motor misalignment is a major contributing factor.
PUMP CLINIC 15
In the following paragraphs we will learn how these failures can be separated into:
· Design problems
· Operation problems
· Maintenance problems
The purpose of this paper is to give you an overview of the subject, and assist you in your troubleshooting
function.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 15 20/06/07 Page 1 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
· The carbon must be dense enough to prevent entrained air pockets from expanding and causing pits
in the carbon face. An "unfilled carbon" with four impregnates is the best
Dual seals:
· Rotating "back to back" designs
· Centrifugal force throws solids into the inner faces
· Inner seal blows open if barrier fluid pressure is lost
· Inner stationary face is not positively retained to prevent movement if the pressure is lost between the
faces
· When the outboard seal fails the inboard will fail also due to the pressure drop between the faces
· The inner seal has to move into the sealing fluid as the face wears. This is a major problem if the fluid
contains solids
· Failure to use "two way" hydraulic balance causes the inner faces to open with a reversal in barrier fluid
pressure
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 15 20/06/07 Page 2 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 15 20/06/07 Page 3 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
OPERATION PROBLEMS
Operations that cause excessive shaft movement that will open or damage the seal faces:
· Opening and closing valves in the suction and/or discharge causing the pump to operate off the B.E.P,
and the shaft to deflect
· Pumping the supply tank dry, causing excessive vibration and heat
· Series or parallel pump operation can cause shaft deflection
· Running at a critical speed will cause the shaft to defect
Cavitation problems:
· Low N.P.S.H
· Air getting into the system through packing
· A stuffing box, suction recirculation line is heating the incoming fluid
· A discharge bypass line is heating the suction fluid
· A discharge recirculation line is aimed at the seal face restricting its movement
· Water hammer is opening or damaging the lapped faces
· The piping system has been altered since the pump was installed
· The pump is being started with the discharge valve shut or severely throttled
· Starting a pump with the discharge valve open is just as bad
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 15 20/06/07 Page 4 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
MAINTENANCE PROBLEMS
· The pump and driver are not aligned & emdash; causing excessive seal movement
· Pipe strain
· Thermal growth
· Bad installation techniques that can injure a seal component
· The wrong lubricant was put on the dynamic elastomer
· The impeller clearance was set after the seal installation
· The face is inserted backwards, only one side is lapped
· The seal is set at the wrong installation length
· The sleeve moved when the impeller was tightened to the shaft
· A lubricant was put on the seal face that froze when the product evaporated across the lapped faces
· The rotating assembly is not dynamically balanced
· The shaft is bent
· The sleeve is not concentric to the shaft
· Impeller clearance is not being maintained, causing vibration problems
· The impeller is positioned too close to the cutwater
· The seal has been set screwed to a hardened shaft
· No seal or gasket between the shaft sleeve and the solid shaft. This is a big problem with double ended
pumps
· The seal environmental control is not being maintained
· Flushing fluid is being restricted or shut off
· Quenching steam is shut off
· The barrier fluid tank level is too low
· The convection tank is running backwards
· The cooling jacket is restricted due to a calcium build up
· You are running both a discharge recirculation line and a cooling jacket
· Out of tolerance shaft dimensions will restrict seal movement
· The impeller clearance was made without re-adjusting the seal face load
· The shaft sleeve was removed to accommodate a smaller diameter seal. The sleeve was providing
corrosion resistance
· A gasket is protruding into the stuffing box restricting the seal movement.“
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 15 20/06/07 Page 5 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 16
“Seal problems are almost always associated with face leakage, but as we will soon learn, there are other
leak paths in addition to the obvious one between the lapped seal faces.
In the following paragraphs, we'll be looking at all these leak paths. Keep in mind that seals are classified into
many categories: stationary, rotary, balanced, unbalanced, inside, outside, metallic, non-metallic, single, dual,
elastomer, metal bellows, rubber bellows, cartridge, split, solid, etc. Try to keep these classifications in mind
as we investigate the cause of seal failure.
I will be presenting the troubleshooting hints in an outline form. You should not find these terms confusing
because I've assumed you have a pretty good knowledge of mechanical seals or otherwise you wouldn't be
attempting to troubleshoot them.
The seal face is not flat. (Flatness should be measured within three helium light bands, (0,000033" or 1
micron)
• Poor packaging. The seal should be able to survive a 39" (1 metre) drop. To ensure this, the seal
must be shipped in a reusable box insulated with plenty of foam or any other adequate insulation.
• The face was distorted by high pressure or surges in pressure. "Water hammer" would be an
example.
• The face was distorted when you tightened it against an uneven surface.
• The clamping is not "equal and opposite" across the stationary hard face. This is a common problem
with "L" shaped and "T" shaped stationary faces.
• The "hard" seal face has been installed backwards. You're running on a non-lapped seal face. It is
common practice to lap only one side of a hard face.
• The face is being distorted by a change in temperature. This happens when you forget to vent a
vertical pump.
• The carbon metal composite was not stress-relieved after the carbon was "pressed in".
• Corrosion increases with a temperature increase. A 10°Centigrade (18°F) rise in temperature will
double the corrosion rate of most corrosives.
• A cleaner or solvent is being flushed through the lines and is attacking the carbon.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 1 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• You are using a poor grade of Carbon. Go to an unfilled grade such as Pure Carbon Company grade
658 RC. This is a common occurrence if the seal is being repaired by someone other than the
original manufacturer.
• All coatings are porous. The chemical is penetrating this porous coating and attacking the bond
between the coating and the base material, or the base material itself.
• High temperature is heat checking (cracking) the plated face. This is a common problem with cobalt
based tungsten carbide. The nickel base version is less likely to crack.
• The product is solidifying between the faces and they're breaking at start up. Most face materials
have high compressive strength, but tend to be weak in tension.
• Excessive vibration is causing the drive pins to crack the face. Low cost seals experience this
problem quite often.
• There is a high temperature differential across the ceramic, 7 to 10 cycles can break even good
ceramics in hot water or hot petroleum products.
• Air is trapped in the carbon face. Heat is causing it to expand and blow out pieces of the carbon face.
The carbon usually blisters prior to blowing out. The solution is to go to a more dense carbon.
• The product is vaporizing and allowing solid material to blow across the lapped face. This is a
common occurrence in boiler feed water applications.
• The seal faces have opened, solids penetrated and imbedded into the soft carbon are causing rapid
wear in the hard face. The same problem occurs if the carbon was re-lapped using lapping powder.
• The elastomer is being chemically attacked and swelling up. This can break the face in those seal
applications where the elastomer is positioned in the seal inside diameter. In some instances the
swelling elastomer will open up the two faces, allowing the solids to penetrate. This can be a
problem with boot mounted faces
• The rotating shaft, or sleeve, is hitting the stationary face. This can happen if the pump is running off
of its B.E.P. which almost always occurs at start up.
• The seal is being mishandled during installation. Good packaging and proper training can solve
many of these problems.
• The crack may have occurred during disassembly. Check to see if there is discoloration deep in the
crack. Discoloration means that it occurred during, or before, operation.
• Petroleum products can "coke" at the face causing pieces of carbon to be pulled out as the face
rotates. You will have to select two hard faces for this application.
• The rotating face is not centered in the stationary face and is running off the edge of the stationary
face. Look for rubbing marks around the O.D. of the rotary unit. A bent shaft or out of balance
rotating assembly is the most common cause.
o You will notice a much wider wear track if you are experiencing this problem.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 2 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
o The seal will appear to "spit" as lubricant is dragged across the face and off the seal outside
diameter.
The movable face is not free to follow whip, wobble or run out.
• The recirculation line from the pump discharge is aimed at the seal faces and interfering with their
free movement.
• Dirt or solids are clogging the movable components. Magnetite is a very big problem in most hot
water applications.
• The product is interfering with the free movement of the components. It is:
o Viscous (molasses)
• The elastomer has been chemically attacked causing it to swell up and interfere with free movement
of the face.
• Temperature growth of the shaft is interfering with the free movement of the movable face.
o Solids have attached themselves to that portion of the shaft where the dynamic elastomer is
located.
• Solids from outside the stuffing box are getting under the faces. This is a common problem with
vertical pumps.
• The elastomer is spring loaded and the interference on the shaft is restricting the face movement.
• A foreign object has passed into the seal chamber and is interfering with the free movement of the
seal.
The product has plated, or formed on the face and a piece of it has broken off.
• This problem occurs with products that are sensitive to temperature and/ or pressure changes.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 3 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• They were not replaced when the seal was rebuilt and as a result are not "digging" into the shaft.
• The impeller has been adjusted towards the wet end of the pump.
• The sleeve moved when the impeller was tightened to the shaft.
• The cartridge seal was pushed on the shaft by pushing on the gland and the seal is now over
compressed.
o In a dual seal application this will over compress the inner seal and open up, or unload the
outer seal.
The product is vaporizing and blowing the faces open. This happens in hot applications if there is water in the
product.
• It can also occur if the pump/seal was hydrostatically tested with a water base fluid.
The inner seal, of a dual seal application was not balanced in both directions and is opening up with reversing
pressure. This is a common problem in unbalanced seals that are subject to both vacuum and pressure or if
the barrier fluid pressure varies.
The single spring, found in some seal designs, was wound in the wrong direction for the shaft rotation.
The Bellows seal has lost cooling and the anti vibration lugs are engaging the shaft. Shaft movement will
cause the faces to open.
• Either the product is too hot or there is too much heat being generated at the seal faces. You must
vent vertical pumps to prevent this problem.
• The shelf life has been exceeded. Buna N (Nitrile) has a shelf life of only twelve months because of
its sensitivity to ozone attack.
• Chemical attack. In most cases the elastomer swells, but cracking and shrinking does occur in
isolated cases.
• Cryogenic (cold) temperatures freeze the elastomer and it will crack when hit.
• The rubber bellows did not stick to the shaft because the wrong lubricant was used. The shaft turned
inside the bellows causing high heat.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 4 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• The seal faces stuck together. The shaft was turning inside the rubber bellows causing excessive
heat.
• Mishandling.
• The elastomer was slid over a rough spot on the shaft or sleeve. Be careful of old set screw marks,
splined shafts, key ways, etc.
• The product is penetrating into the elastomer and blowing out the other side. This problem is a
common occurrence when you are trying to seal ethylene oxide.
• Teflon jacketed o-rings can split in the presence of halogenated fluids. The halogen will cause the
elastomer to swell up, inside of the teflon jacket. Halogens can be recognized because most of them
end in the letters "ine", such as bromine, astintine, chlorine, fluorine, iodine, etc..
• Solids have "built up" or penetrated between the elastomer and the shaft.
• Excessive travel can cause the elastomer to "snake". Most o-rings can roll up to one half of their
diameter.
• Product attack. This is the most common cause and usually occurs within five to ten days
• The wrong lubricant was used at installation. As an example, you should never put petroleum grease
on EPR o-rings.
• Solvents or chemicals used to clean the lines are not compatible with the elastomer.
• Oxidizers can attack the carbon black in o-rings and other elastomers.
• A common problem with unbalanced, dual seal applications. Two-way balanced seals are
recommended for these applications.
• Remember that o-rings are the only elastomers that seal in both directions. Wedges, U cups, and
chevrons do not have this ability.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 5 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• Some seal companies, and most seal repair facilities, glue the carbon in place. The glue may not be
compatible with the product you're sealing.
• "Pressed in" carbons can leak in a high temperature application because of the differential expansion
between the carbon and its metal retainer. Low expansion metal is available for these applications
• Distorted sleeve or shaft. Many packed, double ended pumps have this problem because there's no
gasket between the impeller and the sleeve that's holding it in place.
• This leak path isn't always visible. It often looks like face leakage.
Pipe flange leaking above the seal and dripping into the seal area.
• I found this one after every other troubleshooting avenue was exhausted.
At the weld location if a seal face holder is welded to the cartridge sleeve.
At the pipe connections, ancillary hardware, A.P.I. Gland fittings, and recirculation lines.
A scratch or nick in the o-ring groove. Remember that up to 100 psi (6 bar) o-rings seal on the O.D. and the
I.D. not the sides.
Seal faces will not leak visibly if they are lapped flat and we keep them in total contact. Shaft movement is
the main contributor to the opening of the seal faces and allowing solids to penetrate. Shaft movement is
caused by many factors. In the following paragraphs we'll be looking at most of them.
Cavitation
• Vaporization caused by too high a product temperature, or too low a suction head.
• Air is entering the stuffing box. A common problem with pumps that run in a vacuum or taking a
suction from an evaporator or condenser.
• Internal recirculation. Occurs when the Suction Specific Speed is too high, or when either the
impeller or wear ring clearance becomes excessive.
• The vane passing syndrome form of cavitation occurs if the O.D. of the impeller is too close to the
pump cutwater. This clearance should be at least 4% of the impeller diameter in the smaller size
impellers and at least 6% in the larger diameter impellers (greater than 14 inch or 355 mm.)
• Contamination of the lubricant is the biggest cause. Grease or lip seals have a useful life of only
2000 hours (84 days).
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 6 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• Serious misalignment. The misalignment can be the result of pipe strain, or misalignment between
the pump and its driver.
• Usually occurs during sleeve removal, or if the bearing was installed with an arbor press.
• Improper storage with the long shaft supported only on the ends causing it to sag.
• The impeller diameter was reduced and the impeller was not re-balanced
Worn coupling.
• The shaft is hitting the wear ring, or a stationary wear ring is contacting a rotating wear ring.
The stationary seal face is not perpendicular to the rotating shaft. This causes the spring loaded, rotating
face to move back and forth twice per revolution.
• The stuffing box face is not square to the shaft. The stuffing box face is often a rough casting.
• Tightening the gland bolts through a gasket is cocking the stationary face.
• Pipe strain.
• Temperature growth.
• A convection tank, or some other heavy device is hanging off the gland distorting it.
• Coupling alignment.
• Shaft deflection. The deflection can be caused by operating the pump off its best efficiency point, the
rotating assembly is out of balance, or the shaft is bent.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 7 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Harmonic vibration.
• The seal is vibrating in harmony with some rotating component. The same thing that causes a rear
view mirror to vibrate in an automobile. Most harmonic vibration can be stopped by changing the
speed of the equipment or "damping" the vibrating component.
Slipstick (an alternating slipping and sticking of the seal faces,) caused by:
• Hot water.
• Solvents.
• Some detergents.
• Gases
• You are using a high friction face combination. Often occurs if you use two hard faces.
• Each time the impeller passes the recirculation connection it causes a pulse of fluid at the seal face.
• Happens with products that contain water, and are operated at elevated temperature.
• Can occur at the seal face because of high face load caused by using unbalanced seals.
• Temperature growth.
• The impeller was adjusted, after the seal was installed, to compensate for wear.
• The equipment is equipped with sleeve or babbitted bearings and has excessive end play.
• Shaft thrust.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 8 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
o There is a thrust towards the bearings caused by the combination of the fluid changing
direction in the impeller and acting on the shaft and/or impeller surfaces. This thrust is offset
by a thrust towards the wet end caused by the impeller shape.
o In centrifugal pumps the resulting force can be in either direction, depending upon how close
the pump is operating to its best efficiency point. Above 65% of its best efficiency, the thrust
is towards the wet end. Below 65% of the best efficiency the thrust is towards the power or
bearing end. There is little to no movement at 65% of the pump’s best efficiency. This
means that at start up the shaft moves in both directions accounting for a higher percentage
of seal failure at start up.
• Vertical mixer shafts often lift vertically when solids are mixed with liquid.
THE SHAFT IS NOT CONCENTRIC WITH THE STUFFING BOX, this will cause a wiping action in stationary
seals.
• The shaft is bending as you move away from the pump B.E.P.
o It bends at 240 degrees, from the cutwater, at low flow and high head.
o It bends at 60 degrees, from the cutwater, at high flow and low head.
• Coupling misalignment.
• Pipe strain.
• Temperature growth causes the stuffing box to move relative to the shaft.
Heat is always an indication of wasted energy, but it can also have a disastrous affect on seal life and
performance. Let's take a look at what's causing this heat.
• Installation error.
• No print was used, or the mechanic cannot read the print he was given.
• The mechanic used the wrong marking tool. The mark is too wide.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 9 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• A cartridge seal was installed on the shaft, by pushing on the gland. Interference from the sleeve
elastomer has caused an over compression of the seal. In some dual seal applications the outer seal
will become under compressed.
• Unbalanced seals are supplied by original equipment companies. They generate more heat than
balanced seals.
• The elastomer is located too close to the seal faces. The heat generated at the faces is affecting
both the elastomer and the seal face.
• The face is too wide causing the hydraulic force to generate excessive heat.
• The carbon seal face is too narrow causing excessive heat from the spring pressure.
• A vertical seal installation is not being vented. The faces are running dry in a bubble.
• Speeds above 5000 F.P.M. (25 m/sec) require a special hydraulic balance and less spring load. A
60/40 balance and a face load of 8 psi to 15 psi (0,07 to 0,2 n/mm2) would be normal.
• Spring loaded elastomers cause varying seal face loads. The actual load depends upon shaft
tolerance and installation dimension.
• Some seal faces are glued in. The glue acts as an insulator preventing the face heat from conducting
to the metal holder.
• Many single spring designs are uni-directional requiring both right handed and left handed seals on a
double ended pump.
• Stationary seal designs require clean flushing if solids are present. Centrifugal force does not throw
the solids away from the moveable (spring loaded) components.
• The coefficient of friction varies with face combinations and various sealing products.
• Carbon / metal composite faces conduct heat better than plain carbon / graphite, as long as there is a
true interference fit and they're not glued together to hold them in place.
Problems with the pump operation that causes high heat at the faces.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 10 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• Running dry.
• Gases.
• Dry solids.
• Vacuum applications.
• Pump out rings on the back of the impeller, running too close to the pump back plate.
o The gland.
o A protruding gasket.
o A fitting.
o It is bending, because the pump is operating off its best efficiency point.
o Install a large diameter stuffing box. You should be able to get at least 1" (25 mm) all around
the rotating unit.
o Connect a recirculation line from the bottom of the stuffing box to the suction side of the
pump. You can do this in almost every case except when you're pumping a product at its
vapor point, or if the solids have a specific gravity lower than the fluid.
• There is no carbon restriction bushing in the bottom of the stuffing box and you are using the cooling
jacket. The restriction bushing slows down the heat transfer.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 11 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
o Quenching steam or water has been shut off during pump shut down.
o The cooling jacket has become clogged by the calcium in the hard water. Try condensate
instead.
• If you are using double seals, remember that two seals generate twice as much heat and
conventional cooling may not be sufficient. Contact the manufacturer for the rules when using
convection tanks and dual seals. You may need a "built in" pumping ring.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 16 26/07/07 Page 12 of 12
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 17
“All pump manufacturers would like to recommend the perfect pump for your application. To do this they
would like you to provide an accurate system curve that would describe the capacity and head needed for
your various operating conditions. Once they have your system curve, they can plot pump curves on top of
the system curve and hopefully select something that will come close to your needs. Without this system
curve, neither of you has much chance of coming up with the right pump.
To create a system curve we plot the desired capacities against the required head over the total anticipated
operating range of the pump. The head will be measured in feet or meters and the capacity will be measured
in gallons per minute or cubic meters per hour.
Some of the confusion begins when we realise that there are three different kinds of head:
STATIC HEAD This is the vertical distance measured from the centre line of the pump to the height of the
piping discharge inside the tank. Look at figure "A" and note that the piping discharge is below the maximum
elevation of the piping system. We do not use the maximum elevation in our calculations because the
siphoning action will carry the fluid over this point once the piping is full of liquid. This is the same action that
lets you siphon petrol out of a vehicle into a storage can.
The pump will have to develop enough head to fill the pipe and then the siphoning action will take over. The
pump operating point should move back towards the best efficiency point (B.E.P.) if the pump was selected
correctly.
FIGURE "A"
DYNAMIC OR SYSTEM HEAD As the liquid flows through the piping and fittings, it is subject to the friction
caused by the piping inside finish, restricted passages in the fittings and hardware that has been installed in
the system. The resulting "pressure drop" is described as a "loss of head" in the system, and can be
calculated from graphs and charts provided by the pump and piping manufacturers. These charts are not
included with this article, you can find them in the Hydraulic Institute Manuals. This "head" loss is related to
the condition of the system and makes the calculations difficult when you realise that older systems may have
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 17 30/08/07 Page 1 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
"product build up" on the piping walls, filters, strainers, valves, elbows, heat exchangers, etc., making the
published numbers somewhat inaccurate.
A general "rule of thumb" says that the friction loss in clean piping will vary approximately with 90% of the
square of the change in flow in the piping, and 100% of the square with the change of flow in the fittings and
accessories. You calculate the change in flow by dividing the new flow by the old flow and then square the
number. As an example:
At 45m³/hr the piping resistance, calculated from published charts (not included) is twenty-three metres (23m).
What will it be at 67.5m³/hr?
67.5 = (1.5)² = 2.25 x 23m = 51.75 x 90% of the change = 46.58m of resistance head
45
In other words, when we went from 45m³/hr to 67.5m³/hr the piping resistance increased from 23m to 46.58m
The loss through the fittings and hardware was calculated at 7.6m. What will the new loss be?
67.5 = (1.5) ² = 2.25 x 7.6m = 17.1 x 100% of the change = 17.1 new metres of head
45
In the original application system, loss was a combination of the loss through the piping and the loss through
the fittings for a total of 30.6 metres at 45m³/hr. When we increased the flow to 67.5m³/hr our system head
changed to a total of 63.68m (46.58 + 17.1). This change would have to be added to the static and pressure
heads to calculate the total head required for the new pump.
Please note that the pump is pumping the difference between the suction head and the discharge head, so if
you fail to consider that the suction head will be either added to or subtracted from the discharge head, you
will make an error in your calculations. The suction head will be negative if you are lifting liquid from below
ground or if you are pumping from a vacuum. It will be positive if you are pumping from a tank located above
ground. If the suction head is pressurized, this pressure must be converted to head and subtracted from the
total head required by the pump.
A centrifugal pump will create a head/capacity curve that will generally resemble one of the curves described
in figure "B" The shape of the curve is determined by the Specific Speed number of the impeller.
Centrifugal pumps always pump somewhere on their curve, but should be selected to pump as close to the
best efficiency point (B.E.P.) as possible. The B.E.P. will fall some where between 80% and 85% of the shut
off head (maximum head).
The manufacturer generated these curves at a specific R.P.M.. Unless you are using synchronous motors
(you probably are using induction motors on your pumps) you will have to adjust the curves to match your
actual pump speed. Put a tachometer on the running motor and record the rpm. difference between your
pump and the speed shown on the pump manufacturer's published curve. You can use the pump affinity laws
to approximate the change.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 17 30/08/07 Page 2 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
POSITIVE DISPLACEMENT PUMPS have a different shaped curve. They look something like Figure "C".
In this system, the head remains a constant as the capacity varies. This is a typical application for:
• A boiler feed pump that is supplying a constant pressure boiler with a varying steam demand. This is
a very common application in many process systems or aboard a ship that is frequently changing
speeds (answering bells).
• Filling a tank from the top and varying the amount of liquid being pumped, is the normal routine in
most process plants. The curve will look like this if the majority of the head is either static or pressure
head.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 17 30/08/07 Page 3 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
The second system is the ideal one, Figure "E" describes it:
In this system the entire head is system head so it will vary with the capacity. Look for this type of curve in the
following applications:
• Pumping to a non pressurized tank, a long distance from the source with little to no elevation
involved. Filling tank cars is a typical application.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 17 30/08/07 Page 4 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
System curve "G" is a common one. It is a combination of static, pressure and system heads.
Once the pump manufacturer has a clear idea as to the shape of your system curve, and the head and
capacity numbers needed he can then select the proper centrifugal pump. The shape of his curve will be
pretty much determined by the specific speed number of the impeller.
In addition to specific speed he can select impeller diameter, impeller width, pump rpm., and he also has the
option of series or parallel operation along with the possibility of using a multi-stage pump to satisfy your
needs.
The sad fact is that most pumps are selected poorly because of the desire to offer the customer the lowest
possible price. A robust pump, with a low L3/D4, is still your best protection against seal and bearing
premature failure when the pump is operating off of its best efficiency point. Keep the following in mind as you
select your pump:
• A centrifugal pump will pump where the pump curve intersects the system curve. This may bear no
relationship to the best efficiency point (B.E.P.), or your desire for the pump to perform a specific
task.
• The further off the B.E.P. you go, the more robust the pump you will need. This is especially true if
you have replaced the packing with a mechanical seal and no longer have the packing to act as a
support bearing when the shaft deflects. Shaft deflection is always a major problem at start up.
• When you connect pumps in parallel, you add the capacities together. The capacity of a pump is
determined by the impeller width and r.p.m. The head of a centrifugal pump is determined by the
impeller diameter and rpm. If the heads are different, the stronger pump will throttle the weaker one,
so the impeller diameters and rpm's must be the same if you connect pumps in parallel. Check the
rpm's on these pumps if you are experiencing any difficulties.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 17 30/08/07 Page 5 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• If you connect the pumps in series, the heads will add together, so the capacities must be the same
or one of them will probably cavitate. You could also have a problem operating too far to the right of
the best efficiency point with a possible motor "burn out".
• When you vary the speed of a centrifugal pump, the best efficiency point comes down at an angle.
The affect is almost the same as changing the diameter of the impeller. This means that the variable
speed motor will work best on a system curve that is exponential (Figure "F"). Unfortunately most
process and boiler feed pump system curves are not exponential.
• Pump curves are based on a speed of 1450, 1750, 2900, 3500, rpm. Electric induction motors
seldom run at these speeds because of "slip". You can estimate that a 2% to a 5% slip is normal in
these pumps with the "slip" directly related to the price of the motor.
• You should also keep in mind that if the motor is running at its best efficiency point that does not
mean that the pump is running at its B.E.P.
Since you will be using pumps that were supplied at the lowest cost, you can do the following to resist some
of the shaft displacement:
• Use a solid shaft. Sleeves often raise the L3/D4 number to over 60 (2 in the metric system), and this
is too high a number for reliable seal performance.
• Try to keep the mechanical seal as close to the bearings as possible. It is the mechanical seal that is
the most sensitive to shaft deflection and vibration.
• Once the seal has been moved closer to the bearings, you can install a sleeve bearing in the packing
space to support the shaft when the pump is operated off of its B.E.P. This is especially important at
start up, or any time a pump discharge valve is operated.
• Check that the shaft is not bent or the rotating assembly is not out of dynamic balance.
• Use a "C" or "D" frame adapter to solve pump motor alignment difficulties.
• A centre line design wet end can be used if pipe strain, due to temperature expansion, is causing an
alignment problem.
Do not trust the system prints to make your calculations. The actual system always differs from that shown on
the print, because people tap into the lines, using the pumped fluid for a variety of purposes and after having
done so forget to change or "mark up" the original system print. You are going to have to "walk down" the
system and note the pipe length, the number of fittings, etc., to make an accurate system head calculation.
Do not be surprised to find that the discharge of your pump is hooked up to the discharge of another pump
further down the line. In other words, the pumps are connected in parallel and no body knows it. Pressure
recorders (not gauges) installed at the pump suction and discharge is another technique you can use to get a
better picture of the system or dynamic head. They will show you how the head is varying with changes in
flow.”
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 17 30/08/07 Page 6 of 6
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 18
There are occasions when you might want to permanently change the amount of liquid you are pumping, or
change the discharge head of a centrifugal pump. There are four ways you could do this:
• Change the speed of the pump by changing the motor or using a variable speed drive
Of the four methods the middle two are generally the most sensible ones. In the following paragraphs we'll
learn what happens when we change either the pump speed or impeller diameter, and as you would guess,
we will see what other characteristics of the pump are going to change along with these values.
To determine what is going to happen we begin by taking the new speed or impeller diameter and divide it by
the old speed or impeller diameter. Since changing either one will have approximately the same affect we will
refer to only the speed in this part of the discussion.
As an example:
NEW SPEED = A VALUE, or 1500 RPM = 0.5
OLD SPEED 3000 RPM
The capacity, or amount of fluid you're pumping, varies directly with this number.
• Example: A 9 kW motor was required to drive the pump at 1500 rpm. How much is required if you go
to 3000 rpm?
• Likewise if a 12 kW motor was required at 3000 rpm and you decreased the speed to 1500 the new
kilowatts required would be: 12 x 0.125 (0.53) = 1.5 kW required for the lower rpm.
The following relationships are not exact, but they give you an idea of how speed and impeller
diameter affects other pump functions.
The net positive suction head required by the pump (NPSHR) varies by the square of the number.
• Example: If the NPSHR at 1450rpm is 3m, what would be the NPSHR for the given pump if its speed
was increased to 2900rpm
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 18 25/09/07 Page 1 of 2
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3808 8758 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
The amount of shaft run out (deflection) varies by the square of the number
• As an example you had 0.07 mm run out at 2900 rpm and you slowed that shaft down to 1450 rpm
the run out would decrease to 0.07 mm x 0.25 (0.52) or 0.018 mm.
The amount of friction loss in the piping varies by approximately 90% of the square of the number. Friction
loss through fittings and accessories varies by almost the square of the number.
• As an example: If the system head loss was calculated or measured at 65 metres at 1450 rpm, the
loss at 2900 rpm would be: 65 metres x 4 (22) = 260 x 0.9 = 234 metres
• Example: At 1450 rpm the impeller material is wearing at the rate of 0.5 mm per month. At 2900 rpm
the rate would increase to: 0.5 x 8 (23) or 4.0 mm per month. Likewise a decrease in speed would
decrease the wear rate eight times as much.
We started this discussion by stating that a change in impeller speed or a change in impeller diameter has
approximately the same effect. This is true only if you decrease the impeller diameter to a maximum of 10%
This is true because as you cut down the impeller diameter, the housing is not coming down in size
correspondingly so the affinity laws do not remain accurate below this 10% maximum number.
The affinity laws do however remain accurate for speed changes and this is important to remember when we
convert from gland packing to a balanced mechanical seal. We sometimes experience an increase in motor
speed rather than a drop in amperage during these conversions and the affinity laws will help you to predict
the final outcome of the change.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 18 25/09/07 Page 2 of 2
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 19
In past Pump Clinic Articles we have talked about cavitation which most people who have any involvement
with pumps will come across at some stage. In this article however we want to talk only about NPSH and
what it means as it is a term that is misunderstood by many people.
• Bubbles take up space, causing the pumping capacity to diminish. The head also reduces because
energy has to be expended to increase the velocity of the liquid used to fill up the cavities, as the
bubbles collapse. As the velocity goes up, the head or pressure goes down.
• Excessive vibration can occur when part of the impeller is handling a liquid and another part is
handling a vapour. This vibration can lead to pump failure.
• Air is a poor heat transfer medium, meaning that the fluid we are pumping will get hotter and in
almost no cases is there any advantage in heating up the process fluid.
• A bubble is a hole or cavity in the liquid. It is these cavities that are going to cause a cavitation
problem that will damage both the impeller and volute.
Bubbles or cavities form in a liquid when the fluid temperature gets too high, or the fluid pressure gets too
low. This is called vapourisation, or sometimes boiling although the word boiling tends to imply that the liquid
is hot which need not be the case. We all know that if you throw dry ice into cold water it will bubble and
vapourise, but it is not hot.
For the purpose of this article we will use vapourise and further state that a fluid will vapourise any time the
pressure falls below its vapourisation point.
Since temperature is a variable with different fluids, there are charts that will give you the vapour pressure for
any fluid at its various temperatures.
Take a look at the following chart. For the purpose of this article we will use a chart in imperial units. You will
note that the vapour pressure for 60˚F chlorine is 80 psi (540kPa), and the vapour pressure for 68˚F fresh
water is about 0.3 psi (2kPa). These numbers are required to calculate our NPSH available.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 19 31/10/07 Page 1 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• Put the fluid in a container, and then pull a vacuum on the container. This happens in the hot well of
condensers. This can be referred to as a loss of "pressure head"
• Lift the liquid out of a hole. This will diminish the position of the liquid level in respect to the pump
centre line. This can be referred to as a loss of "static head"
• Accelerate the fluid. As its velocity increases its pressure will decrease. This is referred to as
"velocity head"
• As the fluid moves through piping, fittings, restrictions and valving, some friction losses occur that will
drop the fluid pressure. This is referred to as an increase in friction head, resulting in some loss of
"positive suction head."
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 19 31/10/07 Page 2 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Heating of the incoming fluid is not usually a problem, but it can occur several ways:
• Internal recirculation in the pump because of worn wear rings or failure to make an impeller
adjustment.
• Piping, exposed to the elements, can heat up the liquid on hot and sunny days.
We do not know how much pressure a centrifugal pump will develop, but we do know the head it can
produce. The head is a function of the shaft speed and the impeller diameter. The faster the speed the
higher the head.
The larger the diameter, the bigger the head. To determine the pressure we have to know the weight or
"specific gravity" of the fluid we are pumping, and since any given centrifugal pump can move a lot of different
fluids, with different specific gravities, it is simpler to discuss the pump's head and forget about the pressure.
Here are the formulas you can use to convert from one to the other:
The pump manufacturer has decided how much head the pump needs to prevent cold water from vapourising
at different capacities and these values are published on his pump curve. The values have been obtained by
testing the pump at different capacities, throttling the suction side and waiting for the first signs of cavitation.
The pressure was noted, converted to head, and transferred to the pump curve.
This observed number is called the "Net Positive Suction Head Required (NPSHR).
The attached pump curve shows the numbers. On the chart they are located at the bottom of the dotted lines
and they run from 2 to 16. According to this graph a 13-inch impeller, running at its best efficiency point
(60+%), would need a NPSH required of 9 feet. An 11-inch impeller running at its best efficiency point would
need 7 feet of NPSH required. Remember this requirement is for cold water only.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 19 31/10/07 Page 3 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Be sure to keep in mind that any discussion of NPSH or cavitation is only concerned about the suction side of
the pump. There is almost always plenty of pressure on the discharge side of the pump to prevent the fluid
from vapourising.
• If we go back to our formula and put the 0.3 psi/ 2kPa vapour pressure for 68˚ water into the
numbers, it comes out to 0.7 feet or 0.2 metres of head is required to stop the water from vapourising
and forming cavities. So why does the NPSH required increase as the capacity is increasing? It's
because the velocity of the liquid is increasing, and anytime the velocity of a liquid goes up, the
pressure or head comes down.
Now that we know what head is required, we can calculate the head we have available, and remember we
are only interested in the suction side of the pump. Generally we will be looking at three kinds of head.
• The static head measured from the liquid level to the centre line of the pump. If the liquid level is
above the pump centre line you will have a positive number. If the level is below the centre line you
will have a negative number.
• The pressure head. Here we will be using only absolute numbers. In other words atmospheric
pressure is 101kPa at sea level so you will add that number (converted to metres, using the above
formula) to the static head if you have an open tank. If the fluid is under vacuum we will convert the
absolute pressure reading to head and use that number, instead of atmospheric pressure. The friction
loss in the piping will be a minus number. You get the number from charts showing pipes size vs flow,
and flow through fittings and valves.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 19 31/10/07 Page 4 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• The next thing we have to do is subtract the vapour pressure of our fluid (converted to feet of liquid)
using the first formula I gave you. All of the above, added together is the NPSH available. If this
number is equal to, or more than the NPSH required by the pump manufacturer, the liquid will not
form bubbles or cavities on the suction side and the pump will not cavitate.
NPSHA = Atmospheric pressure + static head + pressure head - the vapor pressure of your product - loss in
the piping, valves and fittings.
NPSHA will always need to be greater than the NPSHR for the pump to operate without cavitation. Most
people involved in pumping system design demand an additional safety factor of 1m.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 19 31/10/07 Page 5 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 20
PUMP SELECTION
‘HOW TO PICK THE CORRECT SIZE AND TYPE OF PUMP
FOR YOUR APPLICATION’
Have you ever wondered why sometimes after giving two or more suppliers the same information they come
back with quite different pump selections or, why sometimes pump suppliers ask a lot of questions? Many
readers of these articles will know what goes into a good pump selection but here we will look at a few of the
basics so anyone who is not as familiar with the pump selection process will understand where all the
questions are coming from.
We need to begin by deciding what operating conditions the pump has to meet. Generally when you
approach a pump supplier you will be armed with this information. To clearly define the capacity and
pressure needs of your system sometimes you may construct a system curve. This system curve will then be
given to the pump suppliers and they will try to match it with a pump curve that satisfies these needs as
closely as possible.
• Decide the capacity you'll need. This means the flow rate usually in cubic meters per hour or litres per
second. You must also consider if this capacity will change with the operation of your process. A
boiler feed pump is an example of an application that needs a constant pressure with varying
capacities to meet a changing steam demand The demand for boiler water is regulated by opening
and closing a control valve on the discharge side of the pump with a discharge re-circulation line
returning the unneeded portion back to a convenient storage place, or the suction side of the pump.
Remember that with a centrifugal pump if you change its capacity you change the head also. A
positive displacement pump is different. It puts out a constant capacity regardless of the pressure.
• For other centrifugal pump applications, you're going to have to calculate how much pressure will be
needed to deliver different capacities to the location where you'll need them. You'll need enough
pressure to :
o Reach the maximum static head or height the fluid will have to attain.
o Overcome any pressure that might be in the vessel where the fluid is discharging, such as
the boiler we just discussed. This is called the pressure head.
o Overcome friction resistance in the lines, fittings and any valves or hardware that might be in
the system. As an example: high-pressure nozzles can be tricky, especially if they clog up.
This resistance is called the friction head.
These heads need to be calculated for both the suction and discharge side of the pump. To get the total head
you'll subtract the suction head from the discharge head. This is the head that the pump must produce to
satisfy the application.
The total head of a pump seldom remains static. There are a number of factors that can change the head of
a pump while it is operating and this is the point where most people start adding in safety factors to
compensate for some of the unknowns such as aging or clogged pipework or fittings with unknown
resistance. These safety factors will almost always guarantee the selection of an oversized pump that will run
off its best efficiency point (BEP) most of the time.
The pump itself requires a certain amount of net positive suction head (NPSHR) to prevent cavitation. This
value is shown on the pump curve. When you look at the curve you'll also note that the net positive suction
head required (NPSHR) increases with any increase in the pump's capacity. Remember that the net positive
suction head required (NPSHR) number shown on the pump curve is for fresh water at 20°C and not the fluid
or combinations of fluids you'll be pumping
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 20 12/12/07 Page 1 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Either you or the pump supplier will be calculating the net positive suction head available (NPSHA) to be sure
that the pump selected will not cavitate. Cavitation is caused by cavities or bubbles in the fluid collapsing on
the impeller, and volute and has been covered in previous Pump Clinic articles so we will not expand on it
here other than to say its presence has harmful effects on most pump styles over time and it should be
avoided wherever possible.
• You may have to install an inducer on the pump, add a booster pump, or go to a double suction or
side channel pump design if you don't have enough net positive suction head available (NPSHA)
• Once the duty information has been established you need to look into materials of construction. Will
you need any special materials for the pump components?
o The pump supplier will try to choose pump metal components that are chemically compatible
with what you're pumping, as well as any cleaners or solvents that might be flushed through
the lines. If the temperature of the pumpage changes, the corrosion rate can change also.
Will the material selection have an impact on your stock levels or values if the materials are
exotic. This can sometimes be a bit of a balancing act.
o If the product you're pumping is explosive, or a fire hazard, you should be looking at non-
sparking materials for the pump components. Do not depend totally upon the pump
manufacturer to make this decision for you. If you're not sure what materials are compatible
with your product, how will the pump supplier know? Also, keep in mind that some of the
fluids you'll be pumping could be proprietary products known only by their trade name.
o Food products require high-density seal and pump materials that are easy to clean and
sterilize. They also sometimes affect the design of the pump internal and external
components
o If there are abrasive solids in the pumpage, you'll need materials with good wearing
capabilities. Hard surfaces and chemically resistant materials are often incompatible. You
may have to go to some type of coating on the pump wetted parts or select an expensive
duplex metal.
• Occasionally you'll find an application where metal is not practical. There are many monomer and
polymer materials available for these applications, but their cost is generally higher than comparable
metal parts. Be aware that if you're using a mechanical seal in a non-metallic pump, the seal can't
have metal parts in contact with the fluid for the same reasons the pump was manufactured from non-
metallic materials. Use a non-metallic seal or perhaps magnetic drive in these applications
When the pump supplier has all of this information in his possession he can then hopefully select the correct
size pump and driver for the job. Since we all want to quote a competitive price we are now going to make
some critical decisions:
• If the capacity were going to be very low we would recommend a positive displacement (PD) pump.
This could be a gear or diaphragm type pump for small capacities.
• Between about 5 m3 /hr and 120 m3/hr, depending on the differential head we would probably select a
single stage end suction centrifugal pump. At higher capacities we may go to a split case or double
suction design with a wide impeller or even use two pumps in parallel.
• You might need a high head, low capacity pump in which case we may look at a positive
displacement pump again or a multi stage centrifugal pump
• Does the pump need to be self-priming? A self priming pump removes air from the impeller eye and
suction side of the pump. Some operating conditions dictate the need for a self-priming design. If you
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 20 12/12/07 Page 2 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
do not have a self-priming pump and you're on intermittent service, will priming become a problem
the next time you start the pump?
o If the pump is going to run twenty-four hours a day, seven days a week and you're not going
to open and close valves; you will not need a heavy-duty pump. It's easy to select a pump
that'll run at its best efficiency point and at the best efficiency point (BEP) there's very little
shaft displacement or vibration.
• How important is efficiency in your application? High efficiency is desirable, but you pay a price for
efficiency in higher maintenance costs and a limited operating window. You should be looking for
performance, reliability, and efficiency in that order. Too often the engineer specifies efficiency and
loses the other two. The following designs solve some operation and maintenance problems, but their
efficiency is lower than conventional centrifugal pumps.
o A magnetic drive or canned pump may be your best option if you can live with the limitations
they impose. The main ones being their intolerance of dry running and solids
o A vortex or slurry pump design may be needed if there is a concentration of solids or "stringy"
material in the pumpage.
o A double volute centrifugal pump can eliminate many of the seal problems experienced when
we operate off the pump's best efficiency point. The problem is trying to find a supplier that
will supply one for your application. Although readily available for impellers larger than 14
inches (355 mm) in diameter they have become very scarce in the smaller diameters.
• The supplier should recommend a centerline design to avoid the problems caused by thermal
expansion of the wet end if you're operating at temperatures over 200°F (100°C)?
• Will you need a volute or circular casing? Volute casings build a higher head; circular casing are used
for low head and high capacity.
• Do you need a pump that meets a standard? ANSI, API, DIN or ISO are some of the current
standards.
• The decision to use either a single or multistage pump will be determined by the head the pump must
produce to meet the capacities you need. Some suppliers like to recommend a high speed small
pump to be competitive, other suppliers might recommend a more expensive low speed large pump
to lessen NPSH and wear problems.
There are additional decisions that have to be made about the type of pump the supplier will recommend:
• Will the pump be supplied with a mechanical seal or packing? If the stuffing box is at negative
pressure (vacuum) a seal will be necessary to prevent air ingestion.
• If fitted with a mechanical seal will it also have an oversized stuffing box and any environmental
controls that might be needed?
• Will the pump have a jacketed stuffing box so that the temperature of the seal fluid can be regulated?
How do you intend to control the stuffing box temperature? Will you be using water, steam or maybe
a combination of both? Electric heating is sometimes an option.
• How will the open or semi-open impeller be adjusted to the volute casing or back plate? Can the
mechanical seal face loading be adjusted at the same time? If not, the seal face load will change and
the seal life will be shortened.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 20 12/12/07 Page 3 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• If the pump is going to be supplied with a closed impeller you should have some means of knowing
when the wear rings have to be replaced. If the wear ring clearance becomes too large the pumps
efficiency will be lowered causing heat and vibration problems. Most manufacturers require that you
disassemble the pump to check the wear ring clearance and replace the rings when this clearance
doubles.
• Is the pump fitted with a metric motor frame adapter, or will the pump to motor alignment have to be
done manually using dual indicators or a laser aligner to get the readings? A close-coupled design
can eliminate the need for an alignment between the pump and driver.
• What type of coupling will you use to connect the pump to its driver? Couplings can compensate for
axial growth of the shaft and transmit torque to the impeller. They cannot compensate for pump to
driver misalignment as much as we would like them to although this is much better than it was a few
years ago. Universal joints and ‘Hardi Spicer’ shafts are a potential problem for those not
experienced with their use because they have to be misaligned to be lubricated properly
• Belt drives are an option that allows a reasonable amount of fine tuning of pump speed to be
achieved as well as changes of speed later in the pumps life. They are often used on slurry
applications.
• The supplier may decide to run two pumps in parallel operation if you need a high capacity, or two
pumps in series operation if you need a high head. Pumps that run in parallel or series require that
they are running at the same speed. This can be a problem for some induction motors or installations
where identical location of the pumps is not possible.
• An inline pump design can solve many pipe strain and thermal growth problems.
• The pump supplier must ensure that the pump will not be operating at a critical speed or passing
through a critical speed at start up. If he has decided to use a variable speed drive or motor this
becomes a possibility.
• We all want pumps with a low net positive suction head required to prevent cavitation problems but
sometimes it's not practical. The manufacturer sometimes has the option of installing an inducer or
altering the pump design/application to lower the net positive suction head required, but if he goes too
far all of the internal clearances will have to be perfect to prevent cavitation problems.
• Shaft speed is an important decision. Speed affects pump component wear and NPSH requirements,
along with the head, capacity, and the pump size. High speed pumps cost less initially, but the
maintenance costs can be much higher depending on the pump duty. Speed is especially critical if
you're going to be specifying a slurry pump.
There are multiple decisions to be made about the impeller selection and not all pump suppliers are qualified
to make them:
• The impeller material must be chosen for both chemical compatibility and wear resistance. You
should consider one of the duplex metals because most corrosion resistant materials are too soft for
the demands of a pump impeller.
• The decision to use a closed impeller, open impeller, semi-open, or vortex design is another decision
to be made.
• Closed impellers require wear rings and these wear rings present another maintenance problem.
• Open and semi-open impellers are less likely to clog, but need manual adjustment to the volute or
back-plate to get the proper impeller setting and prevent internal recirculation.
• Vortex pump impellers are great for solids and "stringy" materials but they are up to 50% less efficient
than conventional designs.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 20 12/12/07 Page 4 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
• Investment cast impellers are usually superior to sand cast versions because you can cast compound
curves with the investment casting process. The compound curve allows the impeller to pump
abrasive fluids with less vane wear.
• If you're going to pump low specific gravity fluids with an open impeller, a non-sparking type metal
may be needed to prevent a fire or explosion. You'll be better off choosing a closed impeller design
with soft wear rings in these applications.
• The affinity laws will predict the affect of changing the impeller speed or diameter.
• Either you or the supplier must select the correct size electric motor, or some other type of driver for
the pump. The decision will be dictated by the specific gravity of the liquid you'll be pumping, along
with the specific gravity of any cleaners or solvents that might be flushed through the lines. The
selection will also be influenced by how far you'll venture off the best efficiency point (BEP) on the
capacity side of the pump curve. If this number is under-estimated there is a danger of burning out
some electric motors.
• How are you going to vary the pump's capacity? Are you going to open and close a valve or maybe
you'll be using a variable speed drive, or maybe a petrol or diesel engine. Will the regulating valve
open and close automatically like a boiler feed valve, or will it be operated manually? A variable
speed motor or a frequency inverter might be an alternative if the major part of the system head is
friction head rather than static or pressure head.
• The viscosity of the fluid is another consideration because it'll affect the head, capacity, efficiency and
power requirement of the pump. You should know about viscosity and how the viscosity of the
pumpage will affect the performance of the pump. There are some viscosity corrections you can
make to the pump curve when you pump viscous fluids.
• After carefully considering all of the above, the pump supplier will select a pump type and size,
present the quote and give you a copy of the pump curve. Hopefully you'll be getting the supplier’s
best pump technology.
If all of the above decisions were made correctly, the pump supplier will place his pump curve on top of your
system curve and the required operating window will fall within the pump's operating window on either side of
the best efficiency point (BEP). Additionally, the motor will not overheat and the pump should not cavitate.
If the decisions were made incorrectly, the pump will operate where the pump and system curves intersect
and that will not be close to, or at the best efficiency point, producing radial impeller loading problems that will
cause shaft deflection, resulting in premature seal and bearing failures. Needless to say the motor or driver
will be adversely affected also.
Information reproduced with permission of author, Bill McNally, the McNally Institute
Pump Clinic 20 12/12/07 Page 5 of 5
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 21
One of the more serious operating issues for some air powered double diaphragm pumps is referred to in
the Industry as "Stalled” or “Stalling”. This occurs when a pump would not restart after it was dead
headed or was simply found to have stalled. The condition under which Stalling occurs almost always
includes where the discharge side is air or vapour bound and where operation is at a low speed or dead
headed. It can also be evident where critical air valve components are worn allowing internal air by-pass.
The results are the same. The pump may fail to restart or will stop without warning. The only means to
restart are to remove and reapply the air supply or with some air powered double diaphragm pumps to
use hammer adjustment by hitting the main air valve housing with a hammer or similar object.
HAMMER ADJUSTMENT
A Stalling condition should not be confused with a Sticking condition; both of which may exist. The
conditions that can cause both are outlined as follows:
A pump can be considered Stalled when both the main air valve and the pilot valve spools are centered
in their respective travel. This is when air pressure is equalised on each side of the air valve or pilot
valve. When the air valve is in this position it will divert or block off the supply air (power source)
preventing the unit from reciprocating. If either valve spool is off centre, the unit will again function.
Diaphragm assemblies can float when the discharge system is compressible (air or vapor bound). In
these instances, if the pump were dead headed by closing a valve in the discharge piping, they would
float back and forth in their travel as they compress and decompress the fluid mixture in their respective
chambers. As the pump continues to cycle, i t can repeatedly pulse the pilot valve, hence building up
balanced air pressures causing the valve spools to centre. This phenomenon is regularly encountered in
unbalanced or bias designed air valves and this problem presents itself as a rapid fire rattle or machine
gun type noise in the pump.
A similar condition can also exist with low flow / low speed situations where the same events will occur.
Two applications where both conditions can occur are Filter Press service and On-Demand Spray
service. Both have low speed / low flow coupled with dead head and can have air / vapour build-up on
the discharge side of the pump. Eliminating the condition of air / vapour build-up fixes the problem but this
fix is not always possible.
A pump can be considered to be Sticking if i t fails to restart or stops and inspection reveals that the
main air valve spool is seized or stuck in the valve sleeve. This can come about through the air valve
being contaminated by a poor quality air supply or pumped product that may have migrated into the air
valve through a failed diaphragm.
New Design Enhancement
This enhancement with Sandpiper pumps shifting mechanism has made a dramatic improvement in the
performance of the pumps in so called trouble applications. This has been proved in many field trials in
very tough and not so tough field applications. The not so tough applications were to prove that
Sandpiper did not overcome one problem only to create another one. The design has been applied
across the complete range of Sandpiper pumps. The enhancement has been referred to as "cross-
drilled shifters" for lack of a better term. This term describes the most apparent part of the change, the
cross-drilled porting. This change has been in the main air valve sleeve and/or the pilot valve body. The
concept remains the same across all Sandpiper pump sizes but the means vary according to the unit
designs.
The enhancement provides for what can be called an Air Detent that positively locks the main air valve
spool on one end of its travel or the other. There is no way for the valve spool to float allowing i t to
centre. Only after the pilot valve has made a complete, positive shift can the main air valve spool shift to
the opposite end of its travel. The Sandpiper air valve design has proven tolerant to poor quality air
supplies through its design and selection of materials. Several components were also specifically
developed for use in mine service where poor quality air is not uncommon because of the application
environments in mines. The newer design compensates for these conditions and 'strokes right through
them' with an audible, positive stroke. The improvement can be heard. The components have proven
themselves in this application by extending service life and reliability.
This can only be taken so far in anyone's design until the air valve assembly becomes so fouled that it
must be removed to be cleaned and or serviced. In these instances, the Sandpiper ESADS Plus
feature will prove its worth through its design for ease of access and serviceability. This provides for
servicing of the main valve and pilot valve assemblies without needing to disconnect suction and
discharge pipework.
All the air is ported to where it needs to be through a series of engineered orifices or cross-
drilled ports. It is more than just connecting one part to another, with all the modifications
undergoing rigorous testing both in-house and in the field.
PUMP CLINIC 22
VISCOSITY
The viscosity of a fluid is that property which tends to resist a shearing force. It can be thought of as the
internal friction resulting when one layer of fluid is made to move in relation to another layer. A detailed
discussion on viscosity is a major undertaking and this article serves to provide a basic understanding
of viscosity and how it impacts on pumping.
Consider the model shown in Fig. 1, which was used by Isaac Newton in first defining viscosity. It
shows two parallel planes of fluid of area A separated by a distance dx and moving in the same
direction at different velocities V1 and V2.
Fig. 1
The velocity distribution will be linear over the distance dx, and experiments show that the velocity
The velocity gradient, describes the shearing experienced by the intermediate layers as they
move with respect to each other. Therefore, it can be called the "rate of shear", S. Also, the force per
unit area can be simplified and called the "shear force" or "shear stress," F. With these simplified
terms, viscosity can be defined as follows'.
Newtonian Liquids
Isaac Newton made the assumption that all materials have, at a given temperature, a viscosity that is
independent of the rate of shear. In other words, a force twice as large would be required to move a
liquid twice as fast. Fluids which behave this way are called Newtonian fluids. There are, of course,
fluids which do not behave this way, in other words their viscosity is dependent on the rate of shear.
These are known as Non-Newtonian fluids.
Fig. 2 shows graphically the relationships between shear Stress (F,) rate of shear (S,) and viscosity
(n) for a Newtonian liquid. The viscosity remains constant as shown in sketch 2, and in absolute
units, the viscosity is the inverse slope of the line in sketch 1. Water and light oils are good examples
of Newtonian liquids.
Non-Newtonian Liquids
Fig. 3 shows graphically the three most common types of Non-Newtonian liquids. These liquids can
present problems to the pump suppliers. Group A shows a decreasing viscosity with an increasing rate
of shear. This is known as a pseudo-plastic material. Examples of this type are grease, molasses, paint,
soap, starch, and most emulsions. They present no serious mechanical pumping problems since they
tend to thin out with the high rates of shear present in a pump. They can present problems in positive
displacement pump selection because slippage through clearances may increase due to the drop in
viscosity and pump speeds may need to be increased to compensate.
Group B shows a dilatant material or one in which the viscosity increases with an increasing rate of
shear. Clay slurries and candy compounds are examples of dilatant liquids. Pumps must be selected
with extreme care since these liquids can become almost solid if the shear rate is high enough. The
normal procedure would be to oversize the pump somewhat and open up the internal clearances in an
effort to reduce the shear rate.
Group C shows a plastic material. The viscosity decreases with increasing rate of shear. However, a
certain force must be applied before any movement is produced. This force is called the yield value of
the material. Tomato sauce is a good example of this type of material. It behaves similar to a pseudo-
plastic material from a pumping standpoint.
The viscosity of some Non-Newtonian liquids is dependent upon time as well as shear rate. In other
words, the viscosity at any particular time depends upon the amount of previous agitation or shearing
of the liquid. A liquid whose viscosity decreases with time at a given shear rate is called a thixotropic
liquid. Examples are asphalts, glues, molasses, paint, soap, starch, and grease. Liquids whose
viscosity increases with time are called rheopectic liquids, but they are seldom encountered in
pumping applications.
Units of Viscosity
There are two basic viscosity parameters: dynamic (or absolute) viscosity and kinematic viscosity.
Dynamic viscosities are given in terms of force required to move a unit area a unit distance. This is
usually expressed in pound-seconds per square foot in the English system which is equal to slugs per
foot-second. The Metric system is more commonly used, however, in which the unit is the dyne-
second per square centimetre called the Poise. This is numerically equal to the gram per centimetre-
second. For convenience, numerical values are normally expressed in centipoise, which are equal to
one-hundredth of a poise.
Most pipe friction charts and pump correction charts list kinematic viscosity. The basic unit of
kinematic viscosity is the stoke which is equal to a square centimetre per second in the Metric
system. The corresponding English unit is square foot per second. The centistoke which is one-
hundredth of a stoke is normally used in the charts. The following formula is used to obtain the
kinematic viscosity when the dynamic or absolute viscosity is known:
There are various units used for viscosity and these are determined by the type of viscometers
utilised for determining liquid viscosities, most of which are designed for specific liquids or viscosity
ranges. The Saybolt viscometers are probably the most widely used in the United States. The
corresponding units are the SSU (Seconds Saybolt Universal).
These units are found on most pipe friction and pump correction charts in addition to centistokes.
Conversion charts for various units of viscosity are attached.
Viscosity and Pumping
1. Centrifugal pumps. Centrifugal pump performance curves are primarily based on the viscosity of
water; namely 1cst. Higher viscosities affect the capacity-head performance and more
significantly the pump efficiency and therefore power requirements. The water performance of
pumps may be adjusted for any viscosity and this is covered in a separate Pump Clinic titled
Viscosity Impact on Centrifugal Pump Performance.
Because of the significant impact of viscosity on power requirements, there are general viscosity
limits for centrifugal pumps. These are simply arbitrary figures. The PIA Handbook defines the
limits based on the dimension in millimetres of the pump discharge connections and these are
< 50 mm maximum 300 cst
>50mm but <150 mm maximum 500 cst
>150 mm maximum 800 cst
Our experience has indicated that these viscosities may be a little high and better limits are;
< 50 mm maximum 100 cst
>50mm but <150 mm maximum 250 cst
>150 mm maximum 400 cst
This only shows the arbitrary nature of these limits. It is important to understand the viscosity
characteristics as the liquid is sheared in particular where the viscosity increases with shear and
the pump manufacturer should be consulted in these instances.
2. Positive Displacement Pumps. The application of positive displacement (PD) pumps is easier as
the majority of PD pump selection procedures and software programs use viscosity as one of the
determining parameters for pump size, speed and motor selection. The other parameters are
flow, pressure and other liquid conditions e.g. solids content.
The change in viscosity as the product is sheared is more important with PD pump selection
irrespective of whether viscosity increases or decreases. With decreasing viscosity, the impact of
liquid slippage through pump clearances from pump discharge to suction may increase
significantly (dependent on differential pressure across the pump) and this needs to be
considered in pump size and speed selection. With increasing viscosities, mechanical issues as
well a speed reduction is a major consideration. Contact the pump supplier with viscosity change
information in these instances.
PUMP CLINIC 23
PUMP CLINIC 24
SEALING IN PUMPS
The proper selection of a seal is critical to the success of every pump application. For maximum
pump reliability, choices must be made between the type of seal and the seal environment. In
addition, a sealless pump is an alternative, which would eliminate the need for a dynamic type seal
entirely.
Sealing Basics
There are two basic kinds of seals: static and dynamic. Static seals are employed where no
movement occurs at the Juncture to be sealed. Gaskets and O-rings are typical static seals.
Dynamic seals are used where surfaces move relative to one another. Dynamic seals are used, for
example, where a rotating shaft transmits power through the wall of a tank (Fig. 1), through the
casing of a pump (Fig. 2), or through the housing of other rotating equipment such as a filter or
screen.
Pump Clinic 24 Sealing in Pumps (Reprinted from ITT Goulds Website) 29/05/08 Page 1 of 10
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
In a centrifugal pump, the liquid enters the suction of the pump at the center (eye) of the rotating
impeller (Figures 3 and 4).
Pump Clinic 24 Sealing in Pumps (Reprinted from ITT Goulds Website) 29/05/08 Page 2 of 10
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Fig. 6 Typical Stuffing Arrangement when Stuffing Box Pressure is Above Atmospheric Pressure
When the stuffing box pressure is below atmospheric pressure, a lantern ring is employed and
lubrication is injected into the stuffing box (Fig. 7). A bypass line from the pump discharge to the
lantern ring connection is normally used providing the pumped liquid is dean.
Pump Clinic 24 Sealing in Pumps (Reprinted from ITT Goulds Website) 29/05/08 Page 3 of 10
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Fig. 7 Typical Stuffing Box Arrangement when Stuffing Box Pressure is Below Atmospheric Pressure
When pumping slurries or abrasive liquids, it is necessary to inject a dean lubricating liquid from an
external source into the lantern ring (Fig. 8). A flow of from .05 to .12 m3/hr is desirable and a valve
and flowmeter should be used for accurate control. The seal water pressure should be from .7 to 1.0
bars above the stuffing box pressure, and anything above this will only add to packing wear. The
lantern ring Is normally located In the center of the stuffing box. However, for extremely thick slurries
like paper stock, it is recommended that the lantern ring be located at the stuffing box throat to
prevent stock from contaminating the packing.
Pump Clinic 24 Sealing in Pumps (Reprinted from ITT Goulds Website) 29/05/08 Page 4 of 10
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
6. The wide variety of designs allows use of mechanical seals in almost all pump applications.
The Basic Mechanical Seal
All mechanical seals are constructed of three basic sets of parts as shown in Fig. 9:
1. A set of primary seal faces: one rotary and one stationary shown in Fig. 9 as seal ring and insert.
2. A set of secondary seals known as shaft packings and insert mountings such as 0-rings, wedges
and V-rings.
3. Mechanical seal hardware including gland rings, collars, compression rings, pins, springs and
bellows.
Pump Clinic 24 Sealing in Pumps (Reprinted from ITT Goulds Website) 29/05/08 Page 5 of 10
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUSHER:
Incorporate secondary seals that move axially along a shaft or sleeve to maintain contact at the seal
faces. This feature compensates for seal face wear and wobble due to misalignment. The pusher seals'
advantage is that it's inexpensive and commercially available in a wide range of sizes and configurations.
Its disadvantage is that ft's prone to secondary seal hang-up and fretting of the shaft or sleeve. Examples
are Dura RO and Crane Type 9T.
Pump Clinic 24 Sealing in Pumps (Reprinted from ITT Goulds Website) 29/05/08 Page 6 of 10
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
UNBALANCED:
They are inexpensive, leak less, and are more stable when subjected to vibration, misalignment, and
cavitation. The disadvantage is their relative low pressure limit. If the closing force exerted on the seal
faces exceeds the pressure limit, the lubricating film between the faces is squeezed out and the highly
loaded dry running seal fails. Examples are the Dura RO and Crane 9T.
CONVENTIONAL:
Examples are the Dura RO and Crane Type 1 which require setting and alignment of the seal (single,
double, tandem) on the shaft or sleeve of the pump. Although setting a mechanical seal is relatively
simple, today's emphasis on reducing maintenance costs has increased preference for cartridge seals.
NON-PUSHER:
The non-pusher or bellows seal does not have to move along the shaft or sleeve to maintain seal face
contact, The main advantages are its ability to handle high and low temperature applications, and does
not require a secondary seal (not prone to secondary seal hang-up). A disadvantage of this style seal is
that its thin bellows cross sections must be upgraded for use in corrosive environments Examples are
Dura CBR and Crane 215, and Sealol 680.
Pump Clinic 24 Sealing in Pumps (Reprinted from ITT Goulds Website) 29/05/08 Page 7 of 10
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
BALANCED:
Balancing a mechanical seal involves a simple design change, which reduces the hydraulic forces acting
to close the seal faces. Balanced seals have higher-pressure limits, lower seal face loading, and
generate less heat. This makes them well suited to handle liquids with poor lubricity and high vapor
pressures such as light hydrocarbons. Examples are Dura CBR and PBR and Crane 98T and 215.
CARTRIDGE:
Examples are Dura P-SO and Crane 1100 which have the mechanical seal premounted on a sleeve
including the gland and fit directly over the Model 3196 shaft or shaft sleeve (available single, double,
tandem). The major benefit, of course is no requirement for the usual seal setting measurements for their
installation. Cartridge seals lower maintenance costs and reduce seal setting errors
Mechanical Seal Arrangements
SINGLE INSIDE:
This is the most common type of mechanical seal. These seals are easily modified to accommodate
seal flush plans and can be balanced to withstand high seal environment pressures. Recommended
for relatively clear non-corrosive and corrosive liquids with satisfactory' lubricating properties where
cost of operation does not exceed that of a double seal. Examples are Dura RO and CBR and Crane
9T and 215. Reference Conventional Seal.
SINGLE OUTSIDE:
If an extremely corrosive liquid has good lubricating properties, an outside seal offers an economical
alternative to the expensive metal required for an inside seal to resist corrosion. The disadvantage is
that it is exposed outside of the pump which makes it vulnerable to damage from impact and
hydraulic pressure works to open the seal faces so they have low pressure limits (balanced or
unbalanced).
Pump Clinic 24 Sealing in Pumps (Reprinted from ITT Goulds Website) 29/05/08 Page 8 of 10
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
Pump Clinic 24 Sealing in Pumps (Reprinted from ITT Goulds Website) 29/05/08 Page 9 of 10
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
TANDEM (DUAL UNPRESSURIZED): Due to health, safety, and environmental considerations, tandem
seals have been used for products such as vinyl chloride, carbon monoxide, light hydrocarbons, and a
wide range of other volatile, toxic, carcinogenic, or hazardous liquids. Tandem seals eliminate icing and
freezing of light hydrocarbons and other liquids which could fall below the atmospheric freezing point of
water in air (32? F or 0? C). {Typical buffer liquids in these applications are ethylene glycol, methanol,
and propanol.) A tandem also increases online reliability. If the primary seal fails, the outboard seal can
take over and function until maintenance of the equipment can be scheduled. Examples are Dura TMB-
73 and tandem PTO.
Pump Clinic 24 Sealing in Pumps (Reprinted from ITT Goulds Website) 29/05/08 Page 10 of 10
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 25
PUMP RELIABILITY
Pump Clinic 25 Pump Reliability (Reprinted from ITT Goulds Website) 23/06/08
1
FALL 2001
Innovation...Technology...Leadership
Pump Reliability -
Correct Hydraulic Selection Minimizes
Unscheduled Maintenance
Allan R. Budris, Director Product Development The Laboratory reliability factors presented here (1) IN THIS ISSUE:
Eugene P. Sabini, Director Technology are based on correlation of the Block and Geitner (2)
Feature:
R. Barry Erickson, Vice President Technology reliability factors with laboratory pump bearing
Pump Reliability - Correct Hydraulic
ITT Industrial Pump Group frame oil temperature, and vane pass vibration tests
Selection Minimizes Unscheduled
on 3 API (end suction) pumps, plus published
Maintenance ........................Page 1
Recently, significant attention has been given to the mechanical seal face and abrasive wear rates.
life cycle cost of owning a pump. Major components Tech Talk:
of the cost of ownership are initial cost, installation The field test reliability factors presented are derived Business Environment
cost, operating cost, and maintenance cost. In from curve fits (trend lines) of Mean Time Between Changes Drive Product Cost
process plants it has been found that under many Repair data, on 71 ANSI and 48 split case pumps, in Reduction Initiative to Gain
circumstances the cost of unscheduled maintenance two process plants. There was much scatter of the Competitive Advantage........Page 4
is the most significant cost of ownership. Although data, due to the fact that the records were not
numerous papers have been presented on the cleansed of failures caused by factors other than New Products:
subject of pump reliability, that literature primarily hydraulic selection, such as human error, difficult to New PumpSmart®
addresses mechanical means of improving reliability. handle liquids, system interactions, or the Model PS100 Offers Cost
The results of this attention to the mechanical issues mechanical design of the pumps. The duty cycles Effective Pump Protection
as been a marked increase in the “Mean Time (operating times) varied between pumps, especially Plus Up to 70% Energy
Between Repair” (MTBR) for process plants. This has where pumps were on standby service. Also, the Savings ..................................Page 7
been achieved largely through improved installation pumps were not always operated at the conditions
of service analyzed. Despite the resulting large Goulds Model 3355 ..............Page 8
practices, and increased attention to operating
procedures. scatter in the data, definite trend lines could be ANSI Combo Units Expand
and were developed, on the strength of the large the Polyshield® Foundation
Efforts such as these will continue to yield number of pumps evaluated. Offering.................................Page 9
improvements in MTBF, but will be limited in
potential unless a holistic approach is used. Such OPERATING SPEED: Material Matters:
an approach would give more attention to the best NACE MR0175 ......................Page 9
hydraulic fit to optimize reliability. There are four Operating Speed affects reliability through rubbing
contact, such as seal faces, reduced bearing life Personnel Moves:
basic hydraulic selection factors which can have a
through increased cycling, lubricant degradation and Manna Named Vice President
significant affect on pump reliability. They are Pump
reduced viscosity due to increased temperature, and Global Marketing ...............Page 11
Speed, Percent of Best Efficiency Flow, Suction Energy
and NPSH Margin Ratio. These last two factors have wetted component wear due to abrasives in the
pumpage. Operating Speed also increases the New Positions Support After Sales
further been combined into an NPSH Margin Services................................Page 11
Reliability Factor (NPSH-RF), which has been shown energy level of the pump, which can lead to
to be reasonably effective in predicting the reliability cavitation damage. Service Solutions:
of High Suction Energy pumps. Think Big!
Figure 1 compares the API-610 pump laboratory The CPC Internalift
reliability predictor test results with the reliability Screw Pump Story...............Page 12
trend line from actual MTBR data on 119 actual
process pumps, as a function of the ratio of the View this issue and
actual to maximum rated pump speed. The previous issues of PumpLines
Reliability factor for the field test data was based on on our website at
zero pump repairs in a 48 month period, which was www.gouldspumps.com
assumed to be equal to a MTBR of 72 months. Both
curves show a marked increase in reliability with
reduced speed.
continued on page 2
Figure 1.
2
Pump Reliability...
continued from page 1
PERCENT BEST EFFICIENCY FLOW RATE Budris and Mayleben (3) have also proposed NPSH MARGIN:
(Flow Ratio): distinct gating values for High and Very High
Suction Energy, for End Suction and Radial NPSH Margin Ratio is defined as the NPSH
The Flow Ratio affects reliability through the Suction (also know as split case or double Available to the pump by the application,
turbulence that is created in the casing and suction) pumps, based on the analysis of divided by the NPSH Required by the pump.
impeller as the pump is operated away from hundreds of pumps from several manufacturers. By Hydraulic Institute definition, the NPSHR of a
the best efficiency flow rate. As a result, pump is the NPSH that will cause the total
hydraulic loads, which are transmitted to the Start of High Suction Energy: head to be reduced by 3%, due to flow
shaft and bearings, increase and become End Suction Pumps: S.E. = 160 x 106 blockage from cavitation vapor in the impeller
unsteady. Also, the severity of these unsteady Split Case/Radial Inlet Pumps: vanes. NPSHR is by no means the point at
loads can reduce mechanical seal life. S.E. = 120 x 106 which cavitation starts. That level is referred to
Operation at reduced flow rates that put the Start of Very High Suction Energy: as incipient cavitation. It can take an NPSHA of
pump into its recirculation mode can also lead End Suction Pumps: S.E. = 240 x 106 from 2 to 20 times NPSHR to fully suppress
to cavitation damage in High Suction Energy Split Case/Radial Inlet Pumps: cavitation within a pump, depending on pump
pumps. Refer to ANSI/HI 9.6.3 (5) for more S.E. = 180 x 106 design and Flow Ratio (percent bep). The
guidance on the allowable operating region higher values are normally associated with high
for centrifugal and vertical pumps. The above definition of Suction Energy suction energy, high specific speed, pumps with
The field data to laboratory reliability (Equation (1)), and “High” and “Very High” large impeller inlet areas, or reduced flow
comparison for the Flow Ratio is presented in gating values are consistent with values operation in the region of suction recirculation.
figure 2. The field data is, however, only presented in ANSI/HI 9.6.1 (4). This means that a high percentage of pumps
based on the 48 split case pumps, since no are operating with some degree of cavitation.
definitive trend line could be established from Pumps with values of suction energy below It is the amount of Energy associated with the
the ANSI plant data. Also, for trend purposes, these values are considered to have low collapse of the cavitation bubbles that
the 1.00 Field Reliability Factor is based on a suction energy. Generally speaking, Low determines the degree of noise, vibration or
MTBR of 52 months. Correlation between the Suction Energy pumps are not prone to noise, damage from cavitation, if any.
field and laboratory data is good in the vibration or damage from cavitation. However,
normal operating range, with the maximum there could be detrimental effects on Figure 4 shows the affect of the NPSH Margin
reliability occurring around 90 percent of the mechanical seals from the air or vapors which Ratio on pump reliability, based on the 77 field
best efficiency flow rate. may be liberated from the liquid during the pumps. Again, the Low suction Energy failures
formation of the cavitation bubbles, under low (below 48 months) were deleted, because it is
SUCTION ENERGY: NPSH Margin conditions (below 1.1 – 1.3 unlikely that these failures were caused by
NPSH Margin Ratio).
Suction Energy is another term for the liquid
momentum in the suction eye of a pump Figure 3 is based strictly on the
impeller, which means that it is a function of field data for 77 ANSI and Split
the mass and velocity of the liquid in the inlet. Case pumps, with the 42 Low
Suction Energy, as originally approximated by suction Energy failures (below
Budris and Mayleben (3), is defined as follows: 48 months) being deleted,
because it is unlikely that these
Suction Energy (S.E.) = De x n x S x s.g. failures were caused by factors
Equation (1) related to Suction Energy,
Where: mainly cavitation. Here also, a
De = Impeller Eye Diameter (inches) 1.00 Reliability Factor equates
N = Pump Speed (RPM) to no failures in 48 months, or
S = Suction Specific Speed a MTBR rate of 72 months.
(RPM x (GPM).5 / (NPSHR).75 The trend is unquestionable, Figure 2.
s.g. = Specific Gravity of Liquid pumped with higher suction energy
pumps requiring the most
Since the suction energy numbers are quite frequent repairs.
large, the last six digits are normally dropped
(S.E. x E6). It should be noted that, if not
readily available, the Impeller Eye Diameter
can be approximated as follows:
Figure 3.
3
Pump Reliability...
continued from page 2
factors related to Suction Energy, mainly the more important it is to suppress the CONCLUSIONS:
cavitation. Based on this data, the NPSH residual cavitation that exists above the NPSHR,
Margin Ratio does have a definite influence on to prevent damage. This reliability factor is only The speed,flow ratio,suction energy and
pump reliability, especially for High and Very applicable within the allowable operating flow NPSH margin reliability propositions and
High Suction Energy pumps, due to the fact region, above the start of suction recirculation methodologies were confirmed by field
that some cavitation usually exists below a (see ref. 5). Much higher NPSH Margin values experience.
Ratio of 4.0. are required in the region of suction
recirculation, for High and Very High Suction The “Mean Time Between Repair” (MTBR) and
NPSH MARGIN RELIABILITY FACTOR: Energy pump applications. Life Cycle Cost of most centrifugal pumps can
The NPSH Margin Reliability Factor (Fig. 5) be improved if slower pump speeds are used,
was developed to quantify the relationship The diagonal lines (in figure 5) are lines of and pumps are selected to operate in their
between NPSH Margin and Suction Energy on constant relative Suction Energy (x 106). preferred operating range (70% - 120% of bep
pump reliability. The NPSH Margin Reliability Therefore, (for example) the line marked flow rate – ref. 5).
Factors are based on the fact that, above the “180/240” (Double Suction Suction Energy level
gating suction energy values (start of High / End Suction Suction Enegy level) represents Further, the Mean Time Between Repair of
Suction Energy), the greater the suction energy the start of Very High Suction Energy. Pumps of High and Very High Suction Energy pumps
this suction energy level require a minimum can be increased by keeping the NPSH Margin
NPSH Margin Ratio of 2.5 for Ratio above the values recommended in
maximum reliability. figure 5, and/or by reducing the Suction
Energy Level. The easiest way to lower the
To validate the NPSH Margin Suction Energy and increase the NPSH Margin
Reliability Factors in figure 5, of a pump application is by lowering the
NPSH R.F. values were plotted speed of the pump. ■
against the field reliability of
the 77 ANSI and Split Case REFERENCES:
pumps (without the 42 Low
Suction Energy failures / below 1. Erickson, R. B., Sabini E. P. and Stavale, A. E., October
2000, “Hydraulic Selection to Minimize the Unscheduled
48 months), as shown in figure Maintenance Portion of Life Cycle Cost,” Pump Users
6. Although not perfect, the International Forum 2000, Karlsruhe, Germany.
agreement is quite good. It
must be remembered that the 2. Bloch, H.P. and Geitner, F. K., 1994, “An Introduction to
Figure 4.
NPSH R.F. only applies to “High Machinery Reliability Assessment,” Gulf Publishing
Company, Houston, TX.
Suction Energy” and “Very High
Suction Energy” pumps. 3. Budris, A. R. and Mayleben, P. A., 1998, “Effects of
Entrained Air, NPSH Margin, and Suction Piping on Cavitation
in Centrifugal Pumps,” International Pump Users
Symposium proceedings, Texas A&M University,
Houston, TX.
Figure 5.
Figure 6.
Kelair Pumps Australia Pty Ltd ABN 28 001 308 381 215 Walters Road Arndell Park NSW 2148
Ph: 1300 789 466 Fax: 02 9678 9455 Email: kelair@kelairpumps.com.au www.kelairpumps.com.au
QLD Fax: 07 3805 7309 VIC Fax: 03 9569 7866 TAS Fax: 03 6331 9102 WA Fax: 08 9248 2255
PUMP CLINIC 26
INTRODUCTION
The origins of seal/less magnetically driven pumps date back to 1933 when the first known patent
was granted in the UK.
Early commercial development of the magnetic drive pump was pioneered by Geoffrey Howard of
HMD Pumps Limited, UK in the late 1940's and a few years later by Franz Klaus in West Germany.
This development was in response to a need for 100 percent containment of diphyl heating fluids.
At that time, development of mechanical seals had barely started, and all dynamic seals were prone
to leakage, especially at elevated temperatures. Two companies pioneered the use of magnet-drive
pumps; Imperial Chemical Industries in the UK and Bayer in West Germany.
For the first 30 years, their application was limited essentially to pumping life threatening or
extremely hazardous fluids. Because of the higher cost of the equipment, and possibly the stigma
of some unsuccessful or unreliable products that came on the market, seal less pumps tended to
be considered categorically by many as the solution of the last resort.
However, by the 1970's enough experience had been gained in the chemical processing industries
to bring some engineers to the conclusion that the magnet drive pump had been developed to the
point that it had become the most economical solution in many process systems.
2) The impeller shaft rides in bushings housed within the pump enclosure rather than be
bearings mounted externally.
In Figure 1, the drive motor is coupled directly to the outer magnet ring (OMR) by the motor
coupling. The overhung load of the OMR is carried by bearings in the bearing housing. Figure 1
also shows that the pump impeller is mounted on the same shaft as is the inner magnet ring (IMR).
(NOTE: In models with a non-synchronous drive, this is called a torque ring rather than an IMR
because it does not contain magnets as we shall see later in this discussion). The impeller drive
shaft is carried by two bushings which are within the pumping enclosure. You will note that the
pump enclosure is formed by the pump casing and the containment shell.
The driving torque of the electric motor is transmitted to the pump impeller by the magnetic coupling
of the OMR and the IMR (or torque ring) without breaching the pumping enclosure. It is this
magnetic coupling which replaces the mechanical seals of conventional centrifugal pumps.
MAGNETIC COUPLING
Before the theory of magnetic couplings as applied to sealless pumps can be examined, review of
the fundamentals of magnets and electromagnetism is needed. Recall these basic principles:
1. Magnets have a north pole and a south pole. When two unlike poles are near each
other, they are attracted. When two like poles are near each other, they are repelled.
2. When a magnetic field is moved past an electrical conductor which is in a closed loop, an
electric current will flow in that loop.
3. When an electric current flows in a closed loop, an electromagnet is created with a north
pole at one end of the loop(s) and a south pole at the other end of the loop(s).
4. An electromagnet is formed even if the "core" material is air. The electromagnet can be
made stronger by inserting grain-oriented silicon steel into the "core" area.
5. Electromagnets behave just like permanent magnets with respect to the laws of attraction
and repulsion.
Figure 2 is a schematic representation of the torque ring coupling. The outer magnet assembly is
driven by a separately mounted motor. The OMR consists of a number of permanent magnets
securely attached to a cylindrical frame, evenly distributed to provide a uniform magnetic field. The
torque ring is made of a mild steel core with an outer facing of stainless steel or other metal
compatible with the pumpage. Beneath this layer, a conductive metal (copper bars) is placed to
provide an electromagnetic coupling circuit.
The containment shell is an extension of the pump casing (pressure casing) and therefore is a
pressure containing component which completes the sealing off of the pumpage from the external
environment. Thus the torque ring operates in the process pumpage while the OMR operates in
the ambient atmosphere surrounding the pump. When the outer magnet ring rotates, the magnetic
field passes through the containment shell, through the copper of the torque ring, through the mild
steel beneath the torque ring and then returns to the OMR to complete the circle. The rotating
magnetic field produces eddy currents in the copper and these eddy currents create
electromagnets which tend to follow the rotating magnetic field which created them.
In the torque ring shown in Figure 2 a series of parallel copper strips are laid parallel to the pump
shaft. In practice, these strips are separated and tied together at the ends, much like the "squirrel
cage" of an a.c. motor. This is called a rodded torque ring. In any case, the copper conducting
path(s) in the torque ring are firmly connected to the mild steel cylinder which, in turn, is solidly
attached to the pump shaft.
Incidentally, the greater the "slip" in torque ring speed, the greater the eddy current flowing and the
greater the torque. If a pumpage has high viscosity when it is cold, the eddy-current drive will
provide high starting torques and will also provide greater heating of the pumpage (copper losses
are higher at higher "slip" levels). This heating of the pumpage will be an advantage in liquids with
high viscosity at cooler temperatures, getting the pumping operation under way more quickly than
would be the case for a synchronous drive.
SYNCHRONOUS DRIVE
Figure 3 shows a schematic representation of the synchronous drive coupling. Just as with the
eddy-current coupling, the outer magnet assembly is driven by a separately mounted motor. The
differences between torque ring and synchronous coupling occur within the inner ring. In the
synchronous coupling, the IMR contains the same number of magnets as are mounted in the OMR.
The number of magnets is determined by the torque which must be transmitted. Thus, when the
outer magnet ring rotates, the inner magnet ring rotates in synchronism with the outer magnet ring.
The absence of "slip" means that the magnet to magnet coupling drive has a higher speed than the
torque ring coupling drive.
The inner magnet ring is mounted on the same shaft as the pump's impeller. The containment shell
is an extension of the pump casing and thus the IMR operated in the process pumpage while the
OMR operates in the ambient atmosphere surrounding the pump. The OMR is enclosed by a
coupling housing to protect it from dirt and to shield operating personnel from the high speed OMR.
MAGNETIC PERMANENCE
There are three conditions which can alter the magnetic strength of permanent magnets. They are undue
physical abuse, excessive temperatures and powerful extraneous magnetic fields.
PHYSICAL ABUSE
Mechanical stress or shock has long been known to demagnetise steel bars. Modern high coercive force
permanent magnets such as is used in magnet drive pumps, however, are generally insensitive to these
mechanical degradations. Generally, a mechanical stress large enough to demagnetise modern
permanent magnets would have to be so great that it would physically damage the magnet. For
practically all applications, mechanical stresses can be ignored as contributing to instability.
HIGH TEMPERATURES
In early magnetic materials, changes in magnetic structure could occur at room temperatures. Today's
magnetic materials vary in sensitivity to very high temperatures. AINiCo magnets such as used in torque
ring drives have Curie Temperatures (level above which they are useless as magnets) from 800°C to
900°C. Samarium Cobalt magnets begin permanently losing strength at around 350°C (depending on
grade). Neodymium-Iron-Boron magnets begin permanently losing strength at around 120°C (again,
depending on grade). In general, sintered "rare earth" magnetic materials' flux density is inversely related
to their ability to withstand temperature.
HIGH MAGNETIC FIELDS
To effect a magnet, a magnetic field must be stronger than the field used in the initial formation of the
magnet. Since the fields used to create the magnet drive pump magnets are stronger than any fields
found in plant environments, this mode of demagnetisation is all but eliminated.
In other words, the magnets used today in magnet drive pumps are PERMANENT magnets.
Engineers responsible for preparing specifications for sealless pumps sometimes critically review the
design and manufacture of the containment shell to ensure a level of integrity in the pressure vessel
boundary. Specifications generally dictate;
EDDY CURRENTS
As the inner and outer components rotate, eddy currents are created by the rotating magnetic field
cutting through the stationary containment shell, resulting in losses and reducing the overall drive
efficiency. These eddy currents are a function of;
• Speed²
• Length of magnets in rotor
• Diameter² of containment shell
• Thickness of containment shell
• Field strength²
• Resistivity of shell material
• Number of magnets
The resistivity of containment shell materials commonly used is 80 x 10-3 ohm cm for 316 Stainless Steel
and 130-3 ohm cm for Hastelloy C. This variation in resistivity between materials will result in a 62.5%
increase in losses for the stainless steel assembly over the Hastelloy C design. Temperature losses
similarly increase.
CONSTRUCTION MATERIALS
Several materials in addition to stainless steel and Hastelloy C are readily available for containment shell
design to match the pump application. They include: Hastelloy B, Alloy 20, Inconel 718 and various
space age alloys such as Nimonic 75 and Nimonic 90 for extraordinarily severe pressures.
Recent advances in the manufacture of ceramic and plastic materials for pressure vessel containment
have introduced these materials into magnetic drive pump design. Use of these materials effectively
eliminates eddy current loss. However, the thickness required for the containment shell to contain design
pressure, along with manufacturing tolerances, can result in drive size increases up to 50%.
Corresponding increases in viscous frictional losses and material performance limitations have limited
their application in magnet drive pumps.
Three welds are required to manufacture the containment shell. All welds are performed by qualified
welders and thoroughly inspected with qualified procedures from an independent body.
Extensive fixturing and further non destructive testing (NDT) procedures are carried out to ensure the
integrity of the containment shell. However, one disadvantage of using welds in the manufacturing
process is that it may be considered a weak point in the component and a source for corrosion
susceptibility.
because the material work hardens during the process, it is necessary to heat treat the component prior
to final sizing.
The final fabrication is completed using only one weld. Potential problem areas in this design include
uniformity of thickness and accommodation of the increased length due to rounded end shape.
Metallic Non-Metallic
Typical Materials of Ceramic, ETFE, Polypropylene,
Hast-C / 316SS
Construction PTFE
Eddy Current Losses Yes No
MAGNET RETENTION
Some manufacturers with rare earth OMR assemblies have the magnets mechanically retained in the
body and totally enclosed. The magnets cannot be seen, and it is not possible to damage them. Should
the magnet become damaged while the OMR is being assembled, all flakes and chips will be fully
enclosed. There are no components that can degrade, and any mechanical damage to the assembly can
easily be seen by the assembler.
Other manufacturers use epoxy or other compounds as means of retention. A potential issue with this
process is that epoxy degradation and loss of strength may cause magnets to dislodge form the OMR.
Possibly one of the reasons that magnet drive pumps have not been more fully utilised is the concern of
some engineers over having bearings (bushings) in the process fluid. Recognising that the application of
these pumps is presently on relatively clean liquids of low viscosity, a category into which the majority of
pumping applications fall, it may be reasonable to make comparisons between "bushing and shaft life" in
a magnet drive pump and "shaft, mechanical seal and antifriction bearing life" in conventional sealed
pumps in order to quantify this concern. This would seem to be functionally correct way to make one
comparison between sealed and sealless pumps.
The minimum rated life of ANSI B73.1 pump bearings is 17,500 hours at maximum load. Mechancial seal
life varies widely, but two to three years of life would generally be considered excellent. Feedback from
many maintenance engineers has painted a picture of the combination of bearings and seals seldom
approaching two to three years of life in the real world.
Compare this experience with the operating experience of magnet drive pumps. Spare parts order
records of one manufacturer and field reports support the conclusion that average bushing life is three to
five years in typical magnet drive pump service. Ten years operation with original bushings has been
achieved in a number of instances. Perhaps some of the concern over internal bushing life comes from
over 25 years of experience with canned motor pumps where inherent close clearances cannot provide
the longer term wearing capability which is common in magnet drive pumps.
By far the most common bushing material in use in magnet drive pumps today is special plain carbon.
Bushings are sometimes pressed but most often are an interference fit to handle higher temperatures,
pinned for the highest temperature services. Carbon is consistent in performance with a fluid and has
good lubricating qualities that will normally enable a pump to survive a period of dry running, provided it
is short enough so that the pump itself is not damaged. (Dry running is a fact of life with most pumps at
one time or another). With essentially no binder to be attacked, carbon is suitable for all but a few
services throughout the range of magnet drive pumps. Bushings are provided with spiral grooves of
about 1/8 in diameter on the ID to permit solids to pass through the bushings. They are combined with
hardened shaft journals, except where bushings are lightly loaded and the hardening is not required to
obtain satisfactory shaft life.
Acids such as sulphuric acid and nitric acid do not attack plain carbon, but electrolytic damage makes
many carbons unsuitable. Filled PTFE versus 316 SS has proven to be satisfactory in these services
(although not as consistent as carbon), has good lubricating properties and will survive short duration dry
running. Shorter maintenance intervals must be planned when this material is initiated because of the
inconsistency cited. PTFE bushings may be carbon filled, glass filled or mica filled, each having different
chemical compatibility. The temperature range of PTFE as bushing bearing is limited to 120°C whereas
carbon is suitable for the full range of magnet drive service, -40 to +450°C.
For abrasive service, unusually high bearing loads, and corrosive services where PTFE is inadequate,
silicon carbide bushings and journals are recommended. They have the advantages of extremely high
load capacity compared with carbon or PTFE, and they are more tolerant of solids and abrasive
materials. The negative aspects of their use are:
1. Higher cost
2. Complexity of detail design, because of the need to provide for different coefficients of thermal
expansion with this very brittle material
3. The inability to tolerate even a short period of dry running, which makes it much less abuse resistant
than other materials described. Fragments of a failed silicon carbide bushing can cause extensive
damage in a pump.
However, it has generally been accepted that silicon carbide offers the best long term bearing material
for applications other than high temperature and has become the most widely used bearing material in
HMD pumps.
INTRODUCTION
The recirculation flow for any magnet drive pump has three basic functions. These are:
1. Removal of the heat generated from magnetic losses
2. Lubrication of the internal radial and thrust bearings
3. Thrust balancing of the free floating rotating assembly
If the liquid being pumped is "dirty", recirculation flow must perform a fourth function.
To successfully meet these requirements, the recirculated pumpage must remain in the liquid phase at all
points within the magnetic coupling area.
The determining factors for preventing phase change are the mass flow rate, heat capacity, and localised
pressure at any point within the coupling. The type of recirculation flow design will have impact on the
variables of mass flow rate and localised pressure.
AVAILABLE DESIGNS
There are two standard recirculation flow designs commercially available in magnet drive pumps (more
specialised recirculation flow paths are available as an option);
a) Discharge-to-suction
Fluid flow enters the magnetic coupling area at a high pressure discharge point and returns to the bulk
flow at the suction eye of the impeller.
b) Discharge-to-discharge
Fluid flow enters the magnetic coupling area at a high pressure discharge point and returns to the bulk
flow at a point behind the rear shroud of the impeller.
Each design has its advantages and must be considered on an individual basis.
DISCHARGE-TO-SUCTION RECIRCULATION
The discharge-to-suction design (Fig. 5) involves pulling a slip stream from the high pressure point of the
casing and returning it to the bulk flow at the suction eye of the impeller. The flow is routed to the suction
through either the thrust balance holes in the impeller or, in certain designs, through a hole gun drilled
along the axis of the pump shaft.
The recirculated fluid is driven by the differential pressure between the recirculation inlet and return
locations. Additional pumping action is provided by the rotation of the internal magnetic coupling
components.
The total dynamic head (TDH) generated by the pump provides the differential pressure that drives the
recirculation flow. As the differential pressure increases, internal flow rate increases but at a decreasing
rate. The internal flow will reach a maximum beyond which any additional increase in differential
pressure will have negligible impact. This occurs when the friction losses begin to become the dominant
factor effecting flow.
The observed internal pumping effects are primarily caused by the action of the inner magnet ring and
thrust washers. These components operate at high speeds within tight clearances and behave as
rudimentary impellers. Discharge-to-suction recirculation yields a flow pattern that is characterised by
high fluid velocities. This high velocity profile provides advantages that will be discussed later in this
analysis.
DISCHARGE-TO-DISCHARGE RECIRCULATION
The discharge-to-discharge design (Fig. 6) involves pulling a slip stream from the high pressure point of
the casing (discharge pressure) and returning this recirculation flow to the bulk flow at a point behind the
rear impeller shroud adjacent to the impeller boss.
The recirculated fluid is driven by the differential pressure between the recirculation inlet and return
locations, however, this pressure differential is not as great as that in the discharge-to-suction system. To
insure proper cooling of the magnetic coupling, larger flow passages are provided. The rotation of the
internal magnetic coupling components also provides additional pumping action.
Discharge-to-discharge recirculation yields a flow pattern that is characterised by high localised pressure
and little interference with suction flow. These characteristics also provide advantages that will be
discussed in the following section.
COMPARATIVE ANALYSIS
In comparing the benefits of discharge-to-suction and discharge-to-discharge recirculation, both systems
remove the heat generated from the magnetic coupling and lubricate the internal radial and thrust
bearings equally well. While discharge-to-suction recirculation creates a higher recirculation velocity due
to its high "driving" differential pressure, its mass flow is comparable to that of discharge-to-discharge
recirculation due to the latter's large flow path. Internal pressures in both systems are such that "flashing"
at the magnetic coupling interface or internal bearings are avoided with most liquids.
Discharge-to-suction recirculation tends to have better impeller thrust balancing characteristics than the
discharge-to-discharge system due to its routing of the flow through the impeller eye balance holes. This
advantage is, however, minimal, and with the growing use of silicon carbide thrust bearings it is not a
significant issue.
Discharge-to-suction recirculation also tends to flush solids better due to its higher velocities. In general,
however, the handling of solids can be increased through the use of silicon carbide bearings, since these
bearings are less affected by the abrasive nature of most commonly encountered solids.
A pump equipped with discharge-to-discharge recirculation typically requires less NPSH. Lower NPSH
characteristics are achieved by routing the recirculation return flow to the rear of the impeller so that fluid
flow through the suction eye is not interfered with.
Discharge-to-discharge recirculation also eliminates the chance of "flashing" at the impeller eye. Routing
the recirculation return flow which has been heated by magnetic coupling losses and internal bearing
friction to the low pressure, suction eye of the pump may cause flashing with certain fluids. By routing
return flow to the higher pressure, behind the impeller, the potential for flashing is eliminated.
FUNCTIONS OF RECIRCULATION
The advent of silicon carbide bearings which will grind up particles of a lesser hardness has resulted in
wild claims for the solids handling capabilities of sealless pumps. It is true that the presence of solids can
be less problematic with silicon carbide bearings than with other bearing options such as carbon and
PTFE. However, the upper percentage and size limits have changed little since these limits are dictated
by non bearing-related parameters.
SOLIDS HANDLING
Many years ago, solids limit for the magnet drive pumps was uncertain and generally set at 1.5 percent
up to 150 microns, based on two important considerations:
1. The appreciation that although solids and seal less do not mix, there is always some solids in the
pumpage. The 1.5 percent figure was felt to be low enough to suggest caution but high enough to
allow the presence of pipe scale and occasional pull over of filtrate or crystallate.
2. The result of some lab testing and tests carried out by customers on various pumps, indicate that
wear to carbon bearings increases rapidly above 2 percent of solids.
In instances where solids larger than 150 microns were present in small quantities, a rough in line filter
would be fitted on the feed (from pump discharge) to the magnetic drive.
These programs take into account the nature of solids; (ie: sticky or fibrous solids may block flow
channels). As much information as possible on the type of solids should be obtained from the customers
before pump selection is made.
Other considerations are that very abrasive solids will eventually wear metallic parts, that wear-resistant
materials may require selection to give maximum pump life, and that wear in the impeller will be at a
minimum when the pump operates near to its best efficiency (design) point. At low flows recirculation will
accelerate wear, and at higher flow rates solids will increase wear because of velocity and incorrect
incidence angles to impeller blades.
GENERAL GUIDELINES
The following guidelines are generally recognized but specific pump limits must always be confirmed with
the manufacturer:
2. Up to 30% solids
Up to 30 percent solids may be pumped up to 750 microns (wear ring clearances) if a clean flush is
provided to the magnetic drive or if a closed-loop system is fitted (available only from a small number of
manufacturers) to separate the magnetic drive from the pump head. For a clean flush, 10 to 25 l/m may
be required for cooling. With a closed-loop system, 7 to 15 1/hr is required from a pressurised supply.
Pumping solids of 30 percent concentration would dictate the use of hardened steel or iron if useful pump
life is to be obtained.
3. Magnetic solids
For highly magnetic solids, the torque ring drive which does not have magnets immersed in the liquid is
recommended. Modern magnets are extremely powerful and will attract ferrite particles which will build
up between the inner magnet and containment shell. If the torque ring drive is not available for the
particular model selected, then a magnetic filter may be fitted between the pump discharge and magnetic
coupling.
However, there is considerable variety in size, hardness, chemistry, solubility, abrasiveness and
flocculation of solid particles. Each solids handling application requires evaluation based in testing on
experience.
Conventional pumps have generally been developed for specific applications on the basis of experience
(ie: coal slurry pumps, paper stock etc). The same is true for the sealless market with customer/supplier
cooperation to mutually resolve pumping problems in this difficult area.
PUMP CLINIC 27
This article has been developed from a variety of sources including manufacturers, industry trade
organisations, internet articles and common PD industry knowledge.
By definition, positive-displacement (PD) pumps displace a known quantity of liquid with each
revolution of the pumping elements. This is done by trapping liquid between the pumping elements
and a stationary casing. Pumping element designs include gears, lobes, rotary pistons, vanes,
screws and hoses.
PD pumps are found in a wide range of applications -- chemical-processing; liquid delivery; marine;
biotechnology; pharmaceutical; as well as food, dairy, and beverage processing. Their versatility and
popularity is due in part to their relatively compact design, high-viscosity performance, continuous flow
regardless of differential pressure, and ability to handle high differential pressure.
Positive displacement (PD) pumps are divided into two broad classifications, reciprocating and rotary
(Figure 1). This article covers rotary pumping principles.
Figure 1
By definition, PD pumps displace a known quantity of liquid with each revolution of the pumping
elements (i.e., gears, rotors, screws, vanes). PD pumps displace liquid by creating a space between
the pumping elements and trapping liquid in the space. The rotation of the pumping elements then
reduces the size of the space and moves the liquid out of the pump. The broad category of PD
pumps is able to handle fluids of all viscosities up to 1,320,000 cSt / 6,000,000 SSU, capacities up to
1,150 M3/Hr (delete) and pressures up to 700 BAR (delete). Rotary pumps are self-priming and
deliver a constant, (delete) flowrate, regardless of pressure variations.
Selection of a positive displacement (PD) rotary pump is not always an easy choice. There are many
types of PD pumps available. In this article, we cover the more common ones: internal gear, external
gear, timed lobe, vane, screw and peristaltic. Most PD pumps can be adapted to handle a wide range
of applications, but some types are better suited than others for a given set of circumstances.
The first consideration in any application is pumping conditions. Usually the need for a PD pump is
already determined, such as a requirement for a given amount of flow regardless of differential
pressure, viscosity too high for a centrifugal pump, need for high differential pressure, or other factors.
Inlet conditions, required flow rate, differential pressure, temperature, particle size in the liquid,
abrasive characteristics, and corrosiveness of the liquid must be determined before a pump selection
is made.
A pump needs proper suction conditions to work well. PD pumps are often self-priming, and it is often
assumed that suction conditions are not important. But they are. Each PD pump has a minimum inlet
pressure requirement to fill individual pump cavities. If these cavities are not completely filled, total
pump flow is diminished. Pump manufacturers supply information on minimum inlet conditions
required. If high lift or high vacuum inlet conditions exist, special attention must be paid to the suction
side of the pump.
In addition to their wide viscosity range, internal gear pumps have a wide
temperature range as well, handling liquids up to (delete) 400°C. This is due to the
single point of end clearance (the distance between the ends of the rotor gear teeth
and the head of the pump). This clearance is adjustable to accommodate high temperature,
maximize efficiency for handling high viscosity liquids, and to accommodate for wear.
The internal gear pump is non-pulsing, has some self-priming capability, and can run dry for short
periods. They're also usually bi-rotational, meaning that the same pump can be used to load and
unload vessels. Because internal gear pumps have only two moving parts, they are reliable, simple to
operate, and easy to maintain.
3. The pump head is now nearly flooded, just prior to forcing the liquid out of the discharge port.
Intermeshing gears of the idler and rotor form locked pockets for the liquid which assures volume
control.
4. Rotor and idler teeth mesh completely to form a seal equidistant from the discharge and suction
ports. This seal forces the liquid out of the discharge port.
The crescent internal gear pump has an outer or rotor gear that is generally used to drive the inner or
idler gear (Figure 1).
The idler gear, which is smaller than the rotor gear, rotates
on a stationary pin and operates inside the rotor gear. The
gears create voids as they come out of mesh and liquid
flows into the pump. As the gears come back into mesh,
volumes are reduced and liquid is forced out the discharge
port. Liquid can enter the expanding cavities through the
rotor teeth or recessed areas on the head, alongside the
teeth. The crescent is integral with the pump head and
prevents liquids from flowing to the suction port from the
discharge port.
The speed of internal gear pumps is considered relatively slow compared to centrifugal types. Speeds
up to 1450 rpm are considered common, although some small designs operate up to 3,450 rpm.
Because of their ability to operate at low speeds, internal gear pumps are well suited for high-viscosity
applications and where suction conditions call for a pump with minimal inlet pressure requirements.
For each revolution of an internal gear pump, the gears have a fairly long time to come out of mesh
allowing the spaces between gear teeth to completely fill and not cavitate. Internal gear pumps have
successfully pumped liquids with viscosities above 1,320,000 cSt / 6,000,000 SSU and very low
viscosity liquids, such as liquid propane and ammonia.
Internal gear pumps are made to close tolerances and are damaged when pumping large solids.
These pumps can handle small suspended particulate in abrasive applications, but gradually wear
and lose performance. Some performance loss is restored by adjusting the pump end clearance. End
clearance is the closeness of the rotor gear to the head of the pump
Advantages Disadvantages
• Only two moving parts • Usually requires moderate speeds
• Only one stuffing box • Medium pressure limitations
• Non-pulsating discharge • One bearing runs in the product pumped
• Excellent for high-viscosity liquids • Overhung load on shaft bearing
• Constant and even discharge regardless
of pressure conditions
• Operates well in either direction
• Can be made to operate with one direction
of flow with either rotation
• Low NPSH required
• Single adjustable end clearance
• Easy to maintain
• Flexible design offers application
customization
Applications
Common internal gear pump applications include, but are not limited to:
• Glycol
• Internals (rotor, idler) - Cast iron, ductile iron, steel, stainless steel, Alloy 20, and higher alloys.
• Bushing - Carbon graphite, bronze, silicon carbide, tungsten carbide, ceramic, colmonoy, and other
specials materials as needed.
• Shaft Seal - Lip seals, component mechanical seals, industry-standard cartridge mechanical seals,
gas barrier seals, magnetically-driven pumps.
Tighter internal clearances provide for a more reliable measure of liquid passing through a pump and
for greater flow control. Because of this, external gear pumps are popular for precise transfer and
metering applications involving polymers, fuels, and chemical additives.
1. As the gears come out of mesh, they create expanding volume on the inlet side of the pump.
Liquid flows into the cavity and is trapped by the gear teeth as they rotate.
2. Liquid travels around the interior of the casing in the pockets between the teeth and the casing -- it
does not pass between the gears.
3. Finally, the meshing of the gears forces liquid through the outlet port under pressure.
Because the gears are supported on both sides, external gear pumps are quiet-running and are
routinely used for high-pressure duties such as hydraulic applications. With no overhung bearing
loads, the rotor shaft can not deflect and cause premature wear. Usually, small external gear pumps
operate at speeds up to 3000 rpm and larger versions operate at
speeds up to 640 rpm.
1. As the lobes come out of mesh, they create expanding volume on the inlet side of the pump.
Liquid flows into the cavity and is trapped by the lobes as they rotate.
2. Liquid travels around the interior of the casing in the pockets between the lobes and the casing -- it
does not pass between the lobes.
3. Finally, the meshing of the lobes forces liquid through the outlet port under pressure.
Lobe pumps (Figure 3) are similar to external gear pumps in operation, except the pumping elements
or lobes do not make contact.
Pump shaft support bearings are located in the timing gear case.
Since the bearings are out of the pumped liquid, pressure is
limited by bearing location and shaft deflection. There is not metal-
to-metal contact and wear in abrasive applications is minimal. Use
of multiple mechanical seals makes seal construction important.
Lobe pumps may also be used in other industrial applications as detailed below however specific care
needs to be taken in these applications and manufacturers recommendations should be sought.
• Polymers
• Paper coatings
• Pharmaceuticals
• Internals (rotors, shaft) - Typically 316 or 316L stainless steel, non-galling stainless steel
2. The housing and cam force fluid into the pumping chamber through holes in the cam (small red
arrow on the bottom of the pump). Fluid enters the pockets created by the vanes, rotor, cam, and
sideplate.
3. As the rotor continues around, the vanes sweep the fluid to the opposite side of the crescent where
it is squeezed through discharge holes of the cam as the vane approaches the point of the crescent
(small red arrow on the side of the pump). Fluid then exits the discharge port.
Vane pumps (Figure 4) operate quite differently from gear and lobe types.
Vanes are the main sealing element between the suction and
discharge ports and are usually made of a nonmetallic composite
material. Rotor bushings run in the pumped liquid or are isolated
by seals.
Vane pumps usually operate at 1,000 rpm, but also run at 1,450
rpm. The pumps work well with low-viscosity liquids that easily fill Figure 4. Vane pumps have better
the cavities and provide good suction characteristics. Speeds dry priming capability than other
must be reduced dramatically for high-viscosity applications to positive displacement pumps.
load the area underneath the vanes. These applications require
stronger-than-normal vane material.
Because there is no metal-to-metal contact, these pumps are frequently used with low-viscosity non-
lubricating liquids such as propane or solvent. This type of pump has better dry priming capability
than other PD pumps. Vane pumps can run dry, but are subject to vane wear.
Vane pumps are not well suited to handling abrasive applications. Vane pumps have fixed end
clearances on both sides of the rotor and vanes similar to external gear pumps. Once wear occurs,
this clearance cannot be adjusted, but some manufacturers supply replaceable or reversible end
plates.
Advantages Disadvantages
• Handles thin liquids at relatively higher • Can have two stuffing boxes
pressures • Complex housing and many parts
• Compensates for wear through vane • Not suitable for high pressures
extension • Not suitable for high viscosity
• Sometimes preferred for solvents, LPG • Not good with abrasives
• Can run dry for short periods
• Can have one seal or stuffing box
• Develops good vacuum
• Packing - Available from some vendors, but not usually recommended for thin liquid service
Within the rotary pumps family, single, twin and three screw pumps have earned reputations for some
specific and significant applications.. Single-screw pumps are also known as progressive cavity or
helical rotor pumps have some limitations and require some care in application. Multiple screw pumps
can handle high pressure, temperature, speed and power combined. While other pump types can
handle these variables well individually, their combined force is a challenge.
The pumping element is made from the rotor and stator elements. Normally, the rotor is made of steel
or other metal and has the shape of a single helix (external shape). The stator is normally made from
an elastomer and has the shape of a double helix (internal shape). The rotor is manufactured slightly
larger than the stator so an interference fit exists when the rotor is inserted into the stator.
Some design enhancements could include features to make maintenance simpler and reliability
better. A close-coupled, or so-called "block" design, results in a smaller pump package, less upfront
cost and no drive alignment issues. Sealed pivot style universal joints (as in the above illustration)
keep the joints lubricated.
Easy access to the mechanical seal is important to simplify seal service and reduce downtime with
quick changeover. However, packings are still more accepted for sealing fluids in typical PC
applications (as they rarely pump tough enough chemicals to require mechanical seals). When
equipped with augers, PC pumps produce better NPSHR values with higher volumetric efficiencies
and higher percent solids capabilities. Oversized open hopper inlets handle thicker liquids and
eliminate bridging (for example, filter cake up to 55 percent solids can be handled by the auger-
augmented PC pumps).
Improvements in the pumping element go beyond more traditional 1:2 geometry (single rotor lobe) to
multi-lobe configurations, such as 2:3 (rotor/stator lobes) geometry. A multi-lobe design can increase
flow per revolution and reduce initial pump cost. Equal wall stator doubles pressure capability per
stage as compared to standard designs with constant stator outside wall thickness.
Tie rod construction makes the entire assembly much easier to service, as compared to more
conventional threaded stator designs. Hollow cast rotors reduce inertia and result in lower vibrations.
Coated chromed rotors resist wear and last longer.
A temperature probe, installed at the stator wall, can prevent rapid temperature rise and failure during
dry running. In practice, few installations take advantage of this feature, since many maintenance
departments tend to prefer simpler designs with fewer "gadgets."
Progressive cavity pumps have smooth output flow and good self-priming ability. Capable of pumping
both thick and thin fluids, they are successful pumping liquids with high solids and abrasives content.
These capabilities have made progressive cavity pumps a choice for many tough applications. They
work well in the wastewater treatment industry, but they perform equally well at the "opposite end of
the spectrum" (the food industry) due to their minimal impact on shear sensitive fluids, such as
sauces, cream products and similar fluids.
Through the years, advances in pump design, electronic monitoring and materials of construction
have improved PC pump energy efficiency, decreased maintenance requirements and allowed them
to handle more severe application conditions.
However, some designs actually have a clearance between these, referred to as a "single undersize
rotor," "double-undersized" and even "triple-undersized." The designs with clearance between the
rotating elements are rare (interference fits are probably common in 99 percent of applications). When
designed with a clearance, it is limited to low differential pressures and relatively thick (viscous)
product-otherwise the "slip" (loss of flow) would be substantial. As the rotor turns inside the stator, a
cavity is formed between the two shapes and "progresses" (hence the name) axially from one end of
the element to the other.
PC pumps, like other pump types, have limitations. Typically, they are not useable in high temperature
applications because of the limitations of elastomeric stators. They require significant floor space and
should not be run at speeds much higher than 300- to 400-rpm due to the unbalanced nature of the
rotating element. However, when used correctly, the PC pump's benefits can be substantial. A PC
pump will provide long-lasting service to any plant with proactive maintenance practices including
vibration trending programs, root cause analysis and similar equipment-caring techniques.
Advantages Disadvantages
• Handles thin and highly • Serious
viscous liquids damage if run
• Low shear pumping action dry
• Self priming capability (must • Temperature
have limitations due
Liquid in the pump chamber) to elastomeric
• Can handle abrasive liquids stator
• Have one seal or stuffing • Long floor
box space
• Non pulsing flow requirement
This section compares the twin and triple screw pumps. Unlike gear and lobe pumps, screw pumps
are axial flow rather than radial flow machines; flow moves along the axis of rotation rather than
perpendicular to it. This axial flow allows multiple screw pumps to operate at relatively high direct
drive speeds while still maintaining low fluid inlet velocities and low NPSH requirements. Figure 1
illustrates the flow path within these pumps.
Each wrap of screw thread forms a cavity that moves axially from suction to discharge. The wrap, or
cavity, acts as a pressure stage. Low pressure pumps have only one or two wraps (stages), while
high pressure pumps may have 12 or more wraps. The staging effect allows each stage to handle a
moderate pressure rise, resulting in low stress levels within the pump even at high pressure operation.
Triple screw pumps have one driven screw and two idler screws. There is contact between the driven
and idler screws. In twin screw pumps, external timing gears and bearings keep the screws from
contacting each other or their casing bores and do not rely on pumped fluid characteristics.
The majority of twin screw pumps are double suction designs, which effectively puts two single
suction pumps in parallel in one casing. The double suction designs, both three-screw and two-screw,
are inherently in hydraulic balance in the axial direction due to their symmetry.
In a radial direction, twin screw pumps are not hydraulically balanced and require radial bearings at
each end of each shaft.
Triple screw pumps are fitted with one seal whereas twin screw pumps require four seals; one ate ach
end of the two shafts.
Both twin screw and three screw pumps share many applications including hydraulic services,
machinery lubrication, compressor and expander gas sealing and some refinery (heavy fuel, asphalt,
vacuum tower bottoms, bitumen) and chemical processing (synthetic fibers, explosives, polyol,
isocyanate) applications.
Examples of critical applications include modified twin screw pumps used to handle multiphase flow,
i.e., oil well head flows ranging from nearly 100 percent gas to 100 percent liquid including liquid
slugs. Three screw pumps find use aboard combat ships for hydraulic services where extremely quiet
operation is necessary to avoid acoustic detection. Both twin and three screw pumps are used in
medium and heavy crude oil pipelines operating at efficiencies far above centrifugal pumps.
• Bushing - Carbon graphite, bronze, silicon carbide, tungsten carbide, ceramic, colmonoy,
and other specials materials as needed.
The shoe design hose pump uses two or more fixed shoes to compress the hose twice per revolution
by grinding against the hose.
They are excellent devices for pumping slurries, due to their ability to handle very abrasive slurries.
Hose pumps are also very good at dosing chemicals, since they are a positive placement device that
can very accurately control the flow rate desired.
Hose pumps can run dry for long periods of time without damage to the pump. Typically the only
wearing part is the rubber hose, which is also the only part in contact with the pumped medium.
NPSHA is not a concern for hose pumps because they create their own suction on the inlet side.
Finally, hose pumps will never cavitate.
The biggest challenge is manufacturing the hose itself, which is the main element and only repair
part. As such, the hose is the source of the greatest MTBF.
The slower a hose pump is run, the better, because it places fewer revolutions on the hose. One
school of thought suggests that the abrasiveness of the slurry is what destroys the hose in a hose
pump. This is not the reality, because the number one factor in determining hose life in a hose pump
is how many compressions are placed on the rubber hose. The number two factor that contributes to
hose wear is the amount of stress being placed on the hose during a compression and how much
heat is generated from that compression force.
In other words, the best way to maximize hose life and eliminate pump downtime is to reduce the
number of compressions on the hose and compress the hose in the less damaging manor.
There are basically two types of pumps that can be considered a hose pump, but there is a clear
distinction between hose pumps and tube pumps. The difference between these two types of pumps
are that tube pumps typically do not have a glycerin bath, pump at very low pressures and also are
very small and low flow rate devices.
A hose pump is more of an industrial piece of equipment than a piece of lab equipment. Hose pumps
typically range is size from 12 mm to 150 mm and typically have a maximum pressure capability of
~15 bars, depending on manufacturer. This discussion is focused on industrial hose pumps, rather
than tube pumps.
There are many designs of hose pumps, but there are essentially only three means employed by all
of these designs to compress the hose. The first is the shoe design, where two or more fixed shoes
compress the hose twice per revolution by grinding against the hose. This type of design damages
the hose the most because it generates a lot of heat and creates a lot of stress/damage to the hose
on each revolution.
(delete)
Shoe designed pumps have a significant limitation regarding the speed at which the pump can be
operated. Because of the high drag/friction across the rubber hose, the pump heats up significantly.
Due to this heat and friction, these types of pumps cannot run at very high speeds.
For example, a 80 mm pump may be capable of running at only 40-rpm continuously, which, in turn,
limits the amount of flow that can be produced continuously. Manufacturers of this type of pump tend
to push the user to the next larger size pump so the rpm is kept lower. Though this strategy is
correct, the user of a rolling design peristaltic pump can typically work with a unit that is one size
smaller.
Also, a limiting factor on the shoe type of pump involves running a very low rpm. The high drag
created from a very low rpm may frequently trip the variable frequency drive (VFD).
Advantages Disadvantages
• No pumping moving parts • Requires moderate speeds
• Seal/less configuration • Pulsing flow
• Non-pulsating discharge • Low hose life
• Good for high-viscosity liquids • Large space requirements
• Can handle abrasive • Application may be limited by hose
• Self priming capability material availability
• Low shear pumping action
• Easy to maintain
CORROSION SOLUTIONS
The annual cost of corrosion and of protection against corrosion in In dealing with pump systems, some other factors
the world is staggering. A plant may spend considerable amounts need to be considered:
of money each year in painting steel to prevent rusting. Corrosion
in radiators, exhaust systems and water heaters, increases this
cost further. a. Suction and operation conditions
The world’s economy would be entirely different if it were not for b. Continuous/intermittent service
corrosion. Even though corrosion is here to stay, its cost can be c. Are there several pumps
considerably reduced in industry through proper selection of involved, in series, or parallel?
materials, correct design of products, and preventative d. Type of seal
maintenance. Corrosion contributes to the depletion of our natural
e. Flushing fluid
resources and the recent concern over this is becoming
increasingly influential in inducing people to be ”corrosion-cost“ f. Temperature change
conscious.
In an effort to apply corrosion principles (from fluid flow to The above factors influence the corrosion rates of
underground soil and atmospheric) to contained chemical systems, any given material, and each is dealt with in detail
and in particular to centrifugal pumps, the following schematic when considering a pump’s hydraulics, but the
diagrams have been used (Page 2). These indicate the different choice of a proper material depends on how
forms of corrosion that may take place in a pump. The diagrams accurately these factors have been calculated.
depict an impeller in a casing. The seals, bearing brackets, etc. are Additionally, any changes subsequently made in
to be imagined. operation or processing are critical and may make
To view corrosion engineering in its proper perspective, it is re-evaluation of the sizing and choice of material
necessary to remember that the choice of material depends on necessary.
several factors:
a. Cost
b. Corrosion resistance
c. Availability
d. Strength
e. Fabrication
f. Appearance
1
EIGHT FORMS OF CORROSION AND
TWO FORMS OF CRACKING PHENOMENA
FIG. 1A MECHANICS OF CORROSION
;;;;;
;;;
General Corrosion Erosion FLOW
;
BUBBLE
STATIC DYNAMIC
STRESS STRESS
CASE HISTORY
ILLUSTRATING
GENERAL
ATTACK
Environment:
30% Hydrochloric Acid with
minor impurities.
Temperature 66 o C.
CASE HISTORY
ILLUSTRATING SEVERE
General corrosion leads to relatively uniform thinning. For round bars and wires, GENERAL ATTACK
corrosion proceeds radially inward at an essentially uniform rate around the entire
circumference. Castings suffer corrosion starting at the wall exposed to the fluid (for Alloy: 316 S.S.
example, the impeller face of the casting) and proceeding gradually and uniformly to the Environment:
outer wall. Methods of reducing or eliminating general corrosion are the use of coatings, 70% Sulphuric Acid.
Temperature 60 oC.
the selection of a more corrosion-resistant material (a general rule is to select an alloy
with a higher chrome and/or nickel content), the use of inhibitors, or cathodic protection. Description: A 316 S.S.
General corrosion proceeds by many different means: impeller was substituted for
an R-55 impeller.
A. The corrosion reaction product may be protective; it may form a passivating barrier
Remedy: Do not
that stifles further corrosion. In this case, the material is not inert, but continues to interchange castings
corrode at a low rate, and to continually repair the passive film. Most corrosion-resistant without checking
suitability of
austenitic materials, such as stainless steels, show this type of passivating behaviour application.
with the help of a surface film of oxides. This type of protection is very sensitive
to solids in the pumped fluids. These solids may continually scour away the
protective film which would otherwise form, thus leading to erosion corrosion.
B. The corrosion product may be soluble in the pumped fluid at a rate
determined by the electrode potential of the metal. This is illustrated
very well when steel is used in oxygenated water.
C. A special case of an artificially controlled uniform dissolution
process may be attempted by controlling the pH or current density of a
given solution. This principle is utilised in chemical machining and
electropolishing of stainless steels to improve either the corrosion
resistance or frictional characteristics.
In pumps, the recognition of general corrosion is compounded by velocity and
pressure variations. The surface casing may show whorls and pockets where
velocity variations have influenced the rate of corrosion. These variations may appear to
be caused by solids or erosive products in the fluid. However, close examination will
always reveal the fact that corrosion has left an etched appearance on the surface.
3
EROSION - CORROSION
FLOW
4
B. Purely mechanical or particle erosion
In the case of mechanical erosion, the action is limited to the outer periphery of the casing. It is
here that the velocity and pressure of the liquid are the greatest. The central portion around the
shaft opening is generally untouched.
Methods of preventing or reducing Erosion-Corrosion can be accomplished by use of one or
more of the following methods. Use materials with improved corrosion resistance to provide a
stronger protective oxide film. Improve design of system to reduce turbulence. alteration of the
environment such as filtering to remove solids or reducing the temperature. Use coatings, such
as hard facing, if the coating has the required corrosion resistance, as well, as hardness.
Cathodic Protection has been found to reduce Erosion-Corrosion in some applications.
(B)
CAVITATION
A . Formation of bubbles: At the eye of the impeller the pressure on the liquid is sufficiently
reduced to cause the liquid to vapourise or form bubbles.
B. Collapse of bubbles: As the liquid is now pumped to the outer periphery of the impeller, the
pressure is increased, causing the bubble to implode. Repetition of this process at high speed
causes the bubbles to form and collapse rapidly.
5
CASE HISTORY
ILLUSTRATING
CAVITATION TYPE
DAMAGE
Description: The
imploding of bubbles
during re-absorption on the
pressure side of the blade
causes damage in the form
of closely spaced pits.
CASE HISTORY
ILLUSTRATING
CAVITATION TYPE These rapidly imploding bubbles may produce shock
DAMAGE waves with pressure as high as 4400 bar.
This is well beyond the yield strength of a
Alloy: Sterling R55.
number of materials. Such forces cause
Environment: Water with plastic deformation in metals, which is
50 ppm chloride at 44 o C.
indicated by the presence of slip lines on the
Description: The impeller and casing.
imploding of bubbles
during re-absorption on the An imploding bubble causes the metal to be
pressure side of the blade
causes damage in the form roughened. This roughened area in turn acts as a
of closely spaced pits. nucleating site for a new bubble to form. The
Remedy: a) Check the collapsing bubbles appear to cause closely-spaced
conditions under which the pitted areas and considerable roughening of the
pump is operating.
b) Use a more cavitation- surface. Some measures which can be taken to
resistant material. alleviate this problem are:
A. Change of the design to minimise the hydro-
dynamic pressure differences in the process fluid.
B. Use of a more corrosion-resistant material.
CASE HISTORY C. Smoothing of the finish on the impeller to reduce
ILLUSTRATING nucleating sites for the bubbles.
CAVITATION TYPE -
GAS D. Use of a rubber or plastic coating that inherently
CONCENTRATION possesses a strong metal-coating interface.
CORROSION
E. Cathodic protection can reduce cavitation by
Alloy: 316 S.S.
forming bubbles on the metal surface, thereby
Environment: Unknown. cushioning the shock waves produced by the
collapsing bubbles and thus preventing damage to
Description: A restricted
suction condition caused an the metal surface.
insufficient amount of liquid
to fill the discharge throat, Cavitation damage is located anywhere between the
yet a large amount of gas inlet eye of the impeller and the tip of the blades.
was present. This caused a
vacuum on the low velocity The closely spaced pits are usually seen on the
side as shown.
lagging side of the blades. In certain violent
Remedy: a) Review design instances, damage is noticed on the leading side of
of suction. the blades. The extent and location of the damage is
b) Determine if gas in liquid
can be reduced. dependent on the fluid being handled, the
temperature, partial pressures and the degree of re-
circulation flow inherent in the design.
6
PITTING
CASE HISTORY
ILLUSTRATING PITTING
CORROSION
Environment:
Hydrochloric and nitric acid
mixtures. Temperature and
concentration unknown.
Description: Several
damage marks (caused
during material handling)
served as nucleating sites
for an autocatalytic reaction
to occur. This resulted
in the pits shown.
This form of attack is extremely localised. It usually results in a cavity that has
approximately the same dimensions in breadth and in depth. As the breadth
increases, the depth increases, causing a hole through the wall of the casing.
Pits have the following characteristics:
A. They are difficult to detect because they are often covered with
corrosion products.
B. Pits usually grow in the direction of gravity. This is substantiated by the
fact that they require a dense concentrated solution for continuing activity.
C. Pitting usually requires an extended initiation period before visible pits
appear. This period ranges from months to years, depending on both the
specific metal and the corrosive liquid.
D. Pitting is autocatalytic. That is, the corrosion processes within a pit
produce conditions which are both stimulating and necessary for the
continuing activity of the pit.
E. Pitting is usually associated with stagnant conditions. For example, a
type 304 stainless steel pump would give good service handling sea
water if the pump ran continuously, but would pit if shut down for
extended periods of time.
F. Most pitting is associated with halide ions, such as chlorides, bromides,
and hypochlorites. Fluorides and iodides have comparatively lesser pitting
tendencies. Oxidizing metal ions, such as cupric, ferric, and mercuric in
combination with chlorides are considered to be the most aggressive.
Non-oxidizing metal ions such as sodium chlorides and calcium chlorides
are much less aggressive. This type of corrosion differs from crevice
corrosion in that it creates its own crevice. Materials that are susceptible
to crevice corrosion do not necessarily become susceptible to pitting
corrosion, whereas the reverse may be considered to be true.
7
CREVICE/DEPOSIT CORROSION
This type of corrosion occurs in restricted areas, either metal to metal, (threaded drain plug),
or metal to non-metal, (gasketed joints), where free access to the pumped fluid is restricted. It
is aided by the presence of deposits such as sand, dirt, and carbonaceous material that shield
and create a stagnant condition. In certain cases corrosion products will deposit and form a
crevice.
As with pitting corrosion, an autocatalytic reaction fosters the growth of crevice corrosion.
Thus, the initial driving force is often an oxygen or metal ion concentration cell, but continued
growth by accumulation of acidic hydrolyzed salts within the crevice. The external surfaces
CASE HISTORY
ILLUSTRATING are protected cathodically.
CREVICE CORROSION
This kind of attack occurs in many media, however, it is very common in chloride-containing
Alloy: Sterling R55. environments. It is slow to start, but grows at an ever-increasing rate.There are a number of
actions that can be taken to prevent crevice corrosion:
Environment: Dilute
sulphuric acid with small A. Use of welded joints instead of threaded joints.
amounts of hydrochloric
acid and sodium chloride. B. Weld on both sides of a flange to pipe joint, thus avoiding penetration from either side.
Temperature and C. Ensuring that the pump is completely drained.
concentration unknown.
D. Use of gaskets that are non-absorbent, such as teflon, wherever possible.
Description: The E. Use of flushing in seal areas to avoid stagnant conditions in the bore of the stuffing box cover.
gasketed shielded area
has limited diffusion of
oxidizing ions, thus
creating an imbalance and
the initiation of corrosion.
Note the pitting that is
usually associated with
this type of corrosion.
8
GALVANIC
ALUMINIUM STERLING R53
BASE NOBLE
TABLE 1 - GALVANIC SERIES A potential difference exists between two dissimilar metals when
IN PURE SEA WATER they are immersed in a corrosive and conductive solution. If these
Corroded End (anodic, or least noble)
Magnesium
metals are now connected electrically conductively on the outside,
Magnesium Alloys an electron flow is produced. One of the two metals will corrode
↓ faster than the other. The metal which is corroding at a faster rate
Galvanised Steel or becomes anodic, while the other metal is cathodic.
Galvanised Wrought Iron
↓ The most commonly used series, based on electrical potential
Aluminium measurements and galvanic corrosion tests in unpolluted sea
(5052, 3004, 3003, 1100, 6053 in this order)
water is shown alongside. (See Table 1).
↓
Cadmium A similar series is needed for all of the various situations. The
↓ number of tests required would be almost infinite. Thus the series
Aluminium
should be used only for predicting galvanic relationships. The
(2117, 2017, 2024 in this order)
↓ separation between the two metals or alloys in the series is an
Mild Steel indication of the probable magnitude of corrosive effects. Effects
Wrought Iron such as polarisation (potential shifts as the alloys tend to approach
Cast Iron
each other), area, distance and geometry play a definite role in
↓
NI-RESIST galvanic corrosion.
↓
There are several ways in which to combat galvanic corrosion:
Type 410, Stainless Steel (active)
↓ A. Material selection is extremely important. Substitution of
50-50 Lead Tin Solder impellers of different alloys in an existing system must be done
↓
carefully. Care should be taken to avoid wide separation in the
Type 304, Stainless Steel (active)
Type 316, Stainless Steel (active) relevant galvanic series.
↓
B. The pumped fluid may be controlled by the use of
Lead
Tin
a corrosion inhibitor.
↓ C. Barrier coatings and electrical isolation by means of insulators
Muntz Metal
to break the electrical continuity are sometimes employed.
Manganese Bronze
Naval Brass D. Cathodic protection by way of using sacrificial metals
↓
may be introduced.
Nickel 200 (active)
INCONEL alloy 600 (active) E. Design changes involving the avoidance of the unfavourable area
↓ ratios, using bolts and other fasteners of a more noble metal than
Yellow Brass
the material to be fastened, avoiding dissimilar metal crevices,
Admiralty Brass
Aluminium Bronze (as at threaded connections), and using replaceable sections with
Red Brass large corrosion allowances of the more active member.
Copper
Silicon Bronze
70-30 Copper Nickel
Comp. G-Bronze
Comp. M-Bronze
↓
Nickel 200 (passive)
INCONEL alloy 600 (passive)
↓
MONEL alloy 400
↓
Type 304, Stainless Steel (passive)
Type 316, Stainless Steel (passive)
INCOLOY alloy 825
↓
INCONEL alloy 625
HASTELLOY alloy C
Titanium
Protected End (cathodic, or most noble)
9
HEAT AFFECTED
ZONE WELD
INTERGRANULAR CORROSION
WELD
HEAT AFFECTED
ZONE
CARBIDE PARTICLE
% CARBON AVERAGE CONTENT 0.08%
Sensitisation may be referred to as carbide precipitation in the grain
boundaries. The structure of low carbon austenitic stainless steels SATURATION
VALUE 0.02%
consists of three crystallographic phases: Ferrite, austenite and
carbide under equilibrium condition. Rapid cooling of these steels
will ensure the retention of austenite (a high temperature phase),
18%
but if they are heated to around 800oC for any appreciable length of % CHROMIUM
CARBIDE PARTICLE
time, the carbide will precipitate in the grain boundaries. The effect LEVEL FOR
RESISTANCE12%
of sensitisation on the chromium and carbon concentration is
shown in Figure 3. The figure depicts a transient state only. Figure 2
shows the variation in carbon content in passing from one grain,
GRAIN BOUNDARY
through the grain boundary, to another grain. The chromium
content varies in the manner shown in the lower portion of Figure Fig 3 - EFFECT OF SENSITISATION ON CARBON
3. There is a narrow region at the grain boundary which contains AND CHROMIUM CONCENTRATIONS
less than 12% chromium. This is below the 12% chromium
minimum required for corrosion-resistance. If there was no carbon
present in the alloy, this condition would not be possible.
The sensitisation of an austenitic alloy permits corrosive attack to
start at the grain boundary, (lowest energy level), where there is a
deficiency of free chromium. Since the grains (high energy level),
are more resistant than the boundaries (low energy level),
corrosion follows the boundaries, which is typical of intergranular
corrosion.
In translating this phenomenon to the macro-scale, the attack is first
recognised as ditching along the grain boundaries, on the surface
of the casting. As the attack progresses, it permeates the complete
casting wall and results in leakage and possible grain dropping. It
should be noted that cracking, however, does occur in austenitic
alloys, and is mainly due to stress corrosion.
10
CASE HISTORY
ILLUSTRATING
INTERGRANULAR
CORROSION (a)
Environment:
25% Hydrochloric acid with
100 ppm Chlorine.
Temperature: Ambient.
a) The impeller on the left
was solution annealed and
then put in service.
b) The impeller on the right
is in the sensitised
condition. This impeller has
undergone severe
corrosion with grain
dropping occurring at the
tip of the blades.
CASE HISTORY
ILLUSTRATING
INTERGRANULAR
CORROSION (b)
Environment:
25% Hydrochloric acid with
100 ppm Chlorine.
Temperature: Ambient.
Close-up of the impeller on
the right in (a) above.
In practice at the foundry level, it must be mentioned that sensitisation will occur during slow
cooling in the mould. Immediate cooling of castings from the mould has been tried. However,
the hostility of this production step has limited its applicability. Quench cracks have been
noted in stress concentration areas also. Sensitisation is prevented by providing a solution
anneal above the sensitising range temperature.
11
STRESS CORROSION CRACKING
STATIC
STRESS
CASE HISTORY
ILLUSTRATING STRESS
CORROSION
CRACKING
Environment:
50% Caustic (Sodium
Hydroxide) with trace
amounts of Sodium
Chlorides. Two characteristics are necessary for environment cracking to occur: a tensile stress and a
Temperature 66OC. corrosion reaction. There are several forms of environmental cracking such as stress corrosion
Description: A drastic cracking, hydrogen induced cracking, liquid metal cracking, and corrosion fatigue. Generally,
brittle fracture in a ductile cracks produced by this method are unexpected and sometimes are dangerous. They are often
material in a period wrongly interpreted. For example, intergranular corrosion does not require a tensile stress,
of four months.
however, the morphology of the cracks may be very similar to stress corrosion cracking.
Welding sometimes produces hot-short cracks which may be identified as stress corrosion
Remedy: Use another
alloy type. cracking.
The stresses that exist in a given situation are usually very complex. The surface net stress in
contact with the pumped fluid will be the controlling parameter. The cracks produced are
perpendicular to the stress vector. These cracks may be single (as in corrosion fatigue) or
multiple (as in stress corrosion cracking). They may be intergranular or transgranular.
Several of the parameters and the controlling methods used are discussed below:
A. Stress. For a given alloy/fluid system, a threshold stress for cracking exists. In such instances,
stresses below the threshold will not cause cracking, but as the stress is increased above the
threshold, cracking is immediately evident. Lowering the residual and thermal stresses by heat-
treatment and shot-peening is carried out to decrease the stress levels below the threshold. The
latter generates compressive stresses in the material which often offset the tensile stress
necessary for cracking to occur. Compatibility of various materials in contact with respect to
polarisation of potentials along with geometrics that increase salt ion concentrations (like
crevices) should be considered in detail.
B. Metallurgical: The list of specific environments that aid in stress corrosion cracking is different
for each major alloy classification. For example, caustics being handled by a carbon steel,
chlorides pumped by stainless steel pumps, and copper alloys in an ammonia environment. The
most common method is to utilise another alloy that is not susceptible to this attack.
12
The importance of microstructure may be illustrated by placing:
1. a sensitised 316 material in a nitric acid solution. Stress corrosion cracking is noticed.
2. a solution-annealed 316 material in the same solution. This will not induce cracks.
C. The corrosive liquid in certain cases can be made less effective in causing stress corrosion
cracking by the use of an inhibitor such as chromates in a caustic solution. Elimination of
the critical chemicals from the liquid is probably the most desirable. A review of the entire
system is usually necessary if this attack has been identified.
D. An increase in temperature generally has a detrimental effect, that is, it tends to induce
stress corrosion cracking. If, however, the temperature is high enough to remove the critical
chemicals, then the tendency reduces.
E. Coatings and electro-chemical techniques are also used. The coatings normally act as a
barrier between the metal and pumped fluid. Electro-chemical techniques are generally used
to polarise an alloy to an oxidizing potential out of the range that will cause stress
corrosion cracking.
CORROSION FATIGUE
DYNAMIC
STRESS
CASE HISTORY
ILLUSTRATING
FATIGUE CORROSION
Description: A crack
propagated at an inclusion
present in the material after
several months of operation.
13
CORROSION BY ACIDS
14
at around 60-80% concentration range at Alloy R52 on the other hand, is a nickel-
approximately 66oC. This will maintain the molybdenum alloy. This alloy is also known to
corrosion rate within 20 mm per year. If the possess good corrosion resistance in the
pump, for example, is to be used intermediate and strong concentration range
intermittently, then the temperature limitations of sulphuric acid. It is better suited to reducing
may be increased to 82oC. Ferric sulphate and conditions, and is particularly susceptible to
copper sulphate in the acid act as inhibitors oxidizing contaminants such as nitric acid,
and decrease attack. Ferric chloride and cupric chlorine, cupric and ferric chlorides, ferric
chloride in appreciable concentrations are sulphates, and even aeration.
known to cause pitting.
NITRIC ACID
Sterling Alloy R55: One of the most important ingredients for
R55 is a nickel-chromium-molybdenum-copper resistance to nitric acid is chromium. As the
alloy that shows outstanding resistance to chromium content increases, the corrosion
sulphuric acid and many other media. It will rate decreases. The minimum amount of
withstand the corrosion of both oxidizing and chromium generally accepted is 18%. This
reducing agents to moderately high makes the austenitic stainless steels very well
o
temperatures, (80 C). It is not recommended suited for practically all concentrations and
for halogen acids or halogen salt solutions in temperatures. The addition of molybdenum to
contact with the material, but provides stainless steels, as in type 316, as opposed to
resistance over a wide range of oxidizing and 304, does not improve corrosion resistance to
reducing conditions. The 4% copper in the nitric acid.
alloy is kept in solid solution which is essential
for sulphuric acid service. The R55 alloy has Types 304 and 316 Stainless Steels:
numerous advantages over K26 and is the Type 304 stainless steel exhibits excellent
most widely used Sterling alloy for sulphuric resistance to nitric acid at room temperatures
acid and most sulphur compounds, such as up to 30oC, and also to boiling acids up to
sulphur dioxide and hydrogen sulphide gases. 50% strength. The corrosion resistance
The copper in these alloys does not discolour decreases as the concentration and
the product. temperature are increased beyond 50% and
30oC. Type 304 does, however, show excellent
Sterling Alloys R52 and R53: resistance to red and white fuming nitric acids
Alloy R53 is a nickel-chromium-molybdenum at room temperature.
alloy that shows a great deal of thermal
stability at high temperatures. It is useful over Because of the susceptibility of sensitised Type
the entire concentration range and oxidizing 304 (when exposed in the 430oC to 870oC
conditions. The chromium content in the alloy range) to intergranular attack in nitric acid,
provides excellent resistance to oxidizing boiling 65% nitric acid (Huey test) is often
conditions. This alloy is very suitable for used to detect the existence of this condition
chlorides, up to 220 ppm, at a maximum prior to fabrication. This is only an indicative
o test and is not a prediction of definite behaviour.
temperature of 70 C and over the entire
concentration range.
15
CORROSION BY ACIDS
:- Continued
16
Sterling Alloy R55: up to 50oC. Due to its high chromium content,
This is usefully resistant to all concentrations it provides better resistance to oxidizing
o
of phosphoric at temperatures up to 90 C. environments. It must be kept in mind,
however, that dissolved oxygen is not strong
Sterling Alloy R52: enough to passivate the material.
This is excellent in hot concentrated pure
phosphoric acid. However, copper ions (an Sterling Alloy R52:
impurity) behave somewhat differently in R52 is widely used to handle hydrochloric acid
solution. Copper ions at first decrease the at all concentrations and temperatures up to
corrosion rate, but beyond about 10 ppm, they the boiling point. Due to the absence of
tend to increase the corrosion rate. If this alloy chromium in this alloy, its resistance to
must be used, the copper content must be aeration and oxidizing impurities such as nitric
controlled to extremely low values. acid or ferric chloride (when present even in
small quantities) is often destructive.
Sterling Alloy R48:
This alloy works very well at the same Types 316SS and Sterling Alloy K26:
temperatures and concentrations as K26, is The austenitic stainless steels, including K26
less expensive and has the added advantage are to be used only at very low concentrations
of handling abrasives better than K26. at room temperature. Increasing the
Wrought materials used in conjunction with temperature decreases the critical
R48 castings, such as shafts and piping, are concentrations at which the stainless steels
now available. start to corrode. Rapid corrosion occurs at pH
4 or 5, or below. Pickling solutions which are
HYDROCHLORIC ACID sometimes handled by these materials require
This is the most difficult of acids to handle inhibitors if the pump is to be handling the
from a standpoint of corrosion. Hydrochloric is liquid on a continuous basis.
corrosive to most common metals and alloys.
Oxidizing agents and minor impurities such as Nickel and Nickel Irons:
ferric chloride (or cupric chloride) and nitric Aeration affects these alloys to a great extent.
acid present a very rugged corrosive They are generally not considered to be
condition. suitable for hydrochloric acid service because
Great care and good judgement is required to they are susceptible to influences other than
obtain a balance between service life and cost the acid itself and must be used with caution,
of the equipment. only when specific conditions are
definitely known.
Sterling Alloy R53:
This alloy shows good resistance to all Titanium:
concentrations of hydrochloric acid at room This alloy is good up to 10% at room
temperature and has been used successfully temperature. The presence of ferric and cupric
chlorides actually decreases the corrosion rate
of titanium.
17
CORROSION BY ACIDS
:- Continued
sulphur trioxide increase the corrosion rate. R52 and R53 both provide excellent resistance
Formic acid also increases the corrosion rate. to acetic acid at all concentrations and
This increased corrosivity can normally be temperatures. These alloys are more
tolerated by type 316 stainless steel. expensive than type 316 stainless steel and
2. The presence of aldehydes, ketones and K26, thus a service life-cost justification should
esters in the process stream have been known be done before Sterling types R52 and R53
18
CORROSION BY ALKALIS
Of all the available alkaline materials, caustic Types 304 and 316 Stainless Steels:
soda (sodium hydroxide) is the most widely The cast versions of austenitic stainless steels
used. It is produced along with chlorine by the such as types 304 and 316 are used up to 10%
electrolysis of sodium chloride. The type of concentration and up to the boiling point of
electrolytic cell used for production caustic soda. However, at concentrations
determines the purity that is obtainable. above 10% the critical temperature decreases.
Mercuric cells produce 50% grade caustic, Chlorides in the process stream have been
whereas the diaphragm cells produce 9% to known to contribute to stress corrosion
15% grade caustic, which is further purified cracking of these alloys and consideration
before sale. must be given to the stress and temperature
limitations of these alloys.
The major users of caustic soda are the
chemical, pulp and paper, and aluminium
Sterling Alloy K26:
industries.
Pumps made of this material have been used
Iron and steel are widely used at low for handling caustic soda up to 70% and
temperatures (if iron contamination is not 120oC. Galvanic effects must be considered if
detrimental), whereas nickel and nickel alloys are this alloy is to be used with nickel and nickel-
used at higher temperatures. based alloys.
In concentrations above 75%, and including
molten caustic soda, cast nickel does an Sterling Alloy R55:
excellent job. When temperatures above This alloy is used in similar situations to K26. It
o provides better stress corrosion cracking
320 C are to be considered, the cast nickel
pump castings should be solution-annealed to resistance than K26.
minimise the possibility of graphite
precipitation at grain boundaries and a Sterling Alloy R52 and R53:
resultant loss in ductility. The data available for these alloys is not
sufficient to make any indicative statements.
Velocity and aeration have little effect, except
They have been used up to 50% at the
at high temperatures such as above 540oC.
boiling point.
The thermal decomposition at 260oC of
impurities such as chlorates (present in caustic Ni-Resist Type 2 Cast Iron:
soda produced by the diaphragm cell method) Ni-Resist Type 2 cast iron and spheroidal nickel
increases the corrosion rate of cast nickel. In iron are both used where minimum
such instances, caustic soda produced by contamination of the product by copper is
other methods should be utilised, or reducing desired. These alloys may be used up to 70%
agents such as sucrose or dextrim may be caustic soda concentrations. Stress relief of
added to minimise corrosion and product these alloys may help minimise stress corrosion
contamination. cracking of these alloys.
19
CORROSION BY LIQUID METALS
Heat-transfer characteristics of low boiling Finally, impurities such as dissolved gases can
point metals make them particularly attractive change the solubility limits, the wetting
for use in the power plant industries. The tendencies, and the activity of the solid
efficiency of a power plant is increased by metal ions.
operating at higher temperatures. Water and
The surface area to volume ratio is of utmost
steam require high pressure equipment which,
importance. The greater the ratio, the lower is
besides being hazardous, is also expensive.
the corrosion rate. This is because the greater
The liquid metals and fused salts employed
the liquid volume, the greater is the amount of
are usually high thermal conductors. Due to
solid metal that can be held in solution.
their low melting points, they save a great deal
in initial heat-up during the start-up of a power
Types 304 and 316 Stainless Steels:
plant. In addition, they require lower pumping
These alloys may be used to pump sodium
power due to their lower density. For example,
and sodium-potassium mixtures. They have a
mercury requires very high pumping power,
temperature limitation of 540oC. If used
whereas sodium requires low pumping power.
intermittently, care should be taken to prevent
Liquid metals cause different types of carburisation by carbonaceous material. Both
corrosive attack. The most salient feature is of these alloys can handle lithium, thallium,
the lack of electrochemical reactions. In the mercury, bismuth and bismuth-lead alloys up
simplest type of attack, the solid metal to various temperatures.
dissolves in the liquid metal, resulting in
Grey cast iron: is also good for some of
uniform thinning or preferential leaching of a
these liquid metals, such as cadmium and
selective constituent from the solid metal. This
Bi-Pb-Sn alloys.
dissolution may result in the formation of
brittle alloy phases. Cast nickel: possesses the greatest resistance
to stress cracking in lead, bismuth, tin, and
This kind of uniform thinning also occurs when
their alloys. They do not undergo as many
solid metal may dissolve at a hot zone of a
rupture failures as do the nickel chromium
pump and precipitate on the walls of a cool
steels.
zone, where its solubility is less.
20
STERLING FLUID SYSTEMS PUMP MATERIALS OF CONSTRUCTION
CARBON STEEL BS1504 Gr161-430 C 0.25% max A216 WCA/B/C Carbon Steel A420 ARM FEG42 1.0446
AUSTENITIC (304L) BS1504 Gr304c12 C .03% max, Cr 18%, Ni 8% A351/A743/A744 CF-3 Type 304L Z2 CN 18-10M GX2 CrNi 19 10 SCS 19 / SUS 304L 1.4306 X 2 CrNi 19 11
AUSTENITIC (304) BS1504 Gr304c15 C .08% max, Cr 18%, Ni 8% A351/A743/A744 CF-8 Type 304 Z6 CN 18-10M X5 CrNi 18 10 SCS 13 / SUS 304 1.4308 G-X6CrNi 18 91
AUSTENITIC (304 + Nb) BS1504 Gr347c17 C .08% max, Cr 18%, Ni 8%, Nb 8xC A351/A743/A744 CF-8 Type 347 Z6 CNNb 18-10M X6 CrNi N6 18 11 SCS 21 / SUS 347 1.4552 G-X 5 CrNiNb 18 9
AUSTENITIC (316L) BS1504 Gr316C12 C .03% max, Cr 18%, Ni 10% Mo 2% A351/A743/A744 CF-3M Type 316L Z3 CND 18-12-02 GX2 CrNiMo 19 11 SCS 16 / SUS 316L 1.4404 G-X 2 CrNiMo 18 10
AUSTENITIC (316) BS1504 Gr316C16 C .08% max, Cr 18%, Ni 10% Mo 2% A351/A743/A744 CF-8M Type 316 Z7 CND 17-12-02 X5 CrNi Mo17 12 SCS 14 / SUS 316 1.4401 X 5 CrNiMo 17 12 2
AUSTENITIC (316 + Nb) BS1504 Gr318C17 C .08% max, Cr 18%, Ni 10% Mo 2% Nb 8xC A351/A743/A744 CF-8M Type 316 Z4 CNDNb 18-12M GX6 CrNiMoNb 20 11 SCS 22 1.4581 G-X5 CrNiMoNb 18 10
21
AUSTENITIC (ALLOY 20 TYPE) STERLING ALLOY K26 Ni 26%, Cr 23%, Mo 3%, Cu2%, Nb 12xC A351/A743/A744 CN-7M Type 904 Z6 NCDU 25-20-04M GX5 NiCrCuMo 29 21 1.4536 G-X 2 NiCrMoCu 25 20
DUPLEX STAINLESS STERLING ALLOY R48 Cr 26%, Ni 5%, Cu 3%, Mo 2% A351/A743/A744 CD4-MCU Z3 CNDU 26-05-M 1.4515 G-X 3 CrNiMoCuN 26 6 3
into electronic or photographic storage without permission of copyright owner is expressly forbidden.
Ni-Mo (Hastelloy B type) STERLING ALLOY R52 Ni 62%, Mo 28%, V 0.4% A494 N-12MV 2.4810
Ni-Cr-Mo (Hastelloy C type) STERLING ALLOY R53 Ni 54%, Cr 17%, Mo17%, W 4%, V .4% A494 CW-12MV 2.4819
Ni-Cr-Mo-Cu-Si ALLOY STERLING ALLOY R55 Ni 58%, Cr 22%, Cu 4%, Mo 4%, Si 4%, W 1% NONE
Neither Sterling Fluid Systems, nor any of its officers, directors or employees accept any responsibility for the use of the methods and materials discussed herein.
The information is advisory only and the use of the materials and methods is solely at the risk of the user. Reproduction of the contents in whole or part or transfer
PumpClinic …Issue 29
The factors which determine minimum allowable rate of flow include the following:
• Temperature rise of the liquid as it passes through the pump
• Radial hydraulic thrust on impellers -- This is most serious with single volute pumps and, even
at flow rates as high as 50% of BEP could cause reduced bearing life, excessive shaft
deflection, seal failures, impeller rubbing and shaft breakage. This has been covered in
Pump Clinic 12.
• Flow re-circulation in the pump impeller -- This includes suction and discharge recirculation
when operating at flows other than the best efficiency point. This has been covered in Pump
Clinic 3.
• Total head characteristic curve -- Some pump curves droop toward shut off, and some curves
show a dip in the curve. Operation in such regions should be avoided.
This article considers minimum allowable flows based on temperature rise considerations only. To
avoid thermal problems during low flow operation and to prevent a potential hazardous or
mechanically damaging temperature rise within the pump, the temperature rise at shut off (i.e. fully
closed discharge) and the minimum flow required for thermal protection should be calculated and the
required flow be bypassed to dissipate heat generated due to pump inefficiency.
In the majority of cases, considerations other than thermal issues will dictate minimum allowable flow.
Thermal considerations are important where liquids are at, or close to, the boiling point e.g. boiler feed
pumps.
The rate of temperature rise in the pump at shut off can be calculated by:
TR = P x 14.4
Q x SH x SG
where:
TR= temperature rise per minute, in degrees Centigrade
P = power at shut off in kilowatts
Q = volume of liquid in the pump in litres
SH= specific heat of the liquid in calories/gm. C
SG= specific gravity of the liquid
This calculation disregards any allowance for heat dissipated by radiation from pump casings.
TR = H
432.4 x SH
where:
TR = temperature rise, in degrees Centigrade
H = total head in metres
SH= specific heat in calories/gm. C
E = pump efficiency in % at the flow involved.
Values of total head at various flows can be read from the pump performance curve and the
temperature rise at the various flows can be calculated. The figure below gives a graphical
representation of this formula and allows determination of the minimum allowable operating flow once
the maximum allowable temperature rise has been selected.
As shown, the temperature rise increases very rapidly with a reduction in flow. This is caused by the
fact that the losses at low deliveries are greater when the flow of liquid that must absorb the heat
developed in the pump is low.
If the pump is fitted with a balancing device for axial thrust e.g. balance drum or balance disc, a
certain portion of the suction capacity known as leakoff is returned either to the pump suction or to the
suction supply vessel. In this case, the discharge capacity does not represent the true flow through
the pump.
The formula for the temperature rise can still be used, provided a correction is made to take care of
the increase in pump flow representing the balancing device leakoff. Balance device leak off
information is provided by pump manufacturers.
TR = H X Qd
432.4 x SH Q d + Qb
where:
TR = temperature rise, in degrees Centigrade
H = total head in metres
SH= specific heat in calories/gm. C
E = pump efficiency in % at the flow involved.
Qd = flow through pump discharge in litres/sec
Qb = leak off from balance device in litres/sec
Pump Clinic 30 Designing a Trouble-free Installation - Diaphragm Metering Pumps Page 1 of 5 28/01/09
PumpClinic …Issue 31
Providing sufficient net positive suction head (NPSH) to a centrifugal pump may be costly if it involves
increasing the height of a vessel or maintaining higher than normal inventory levels.
Inducers are a low cost alternative that reduce the NPSH required by pumps, but their application is
not without peril. In this issue we'll discuss inducers as well as the advantages and disadvantages of
using them.
If you rapidly sweep your open palm through the air you'll notice a
higher pressure on the side of your hand that is pushing against
the air than you do on your hand's trailing side. A spinning impeller
blade undergoes the same effect as it sweeps through liquid. The
back side of the blade (the part that you can't see when looking
into an impeller eye) pushes against the liquid, trying to accelerate
it within the impeller passage, as the front side of the blade
experiences a localized reduction in pressure. This localized low
pressure area is where cavitation develops.
PC31 Inducers for Centrifugal Pumps (Reprinted with kind permission of Lawrence Pumps Inc USA) Page 1 of 3 04/02/09
PumpClinic …Issue 31
An inducer invariably has a higher suction specific speed (S) than the adjacent
impeller. S is a dimensionless term that describes the inlet characteristics of a
pump. For a constant RPM and flow, a lower NPSH requirement means a
higher suction specific speed. Inducers commonly have suction specific
speeds of between 15,000 and 25,000. A pump equipped with an inducer
may operate at 1/2 to 1/3 the NPSHR levels of a non-inducer version of the
same pump.
So what's the downside? There are a few that the user should be aware of.
Inducers add mass, cantilevered away from the bearings. This will increase shaft deflection
and reduce the 1st critical speed of the unit. This is not a problem if the pump is designed for
the added mass, but it may be a problem otherwise.
2. The NPSH requirements of the inducer need to be compatible with the entire operating
range of the pump
Many inducers have a steeply rising NPSHR characteristic on either side of the design flow
rate. Often the NPSHR will exceed that of a non-inducer pump when operating off design.
Under these conditions, it is possible to make problems worse by applying an inducer.
3. High suction Specific speed (S) limits the allowable operating range of the pump.
High suction specific speed pumps become unstable when operated off design. Below is a
general chart showing operating range vs. S. The width of the operating range also varies
with the horsepower. Lower power pumps generally have a broader operating range; higher
power pumps have a narrower operating range. Specific pump geometry will also affect the
operating range. But it's good to remember that while a 10 hp pump can use up to 10 hp to
create damage, an 800 hp pump can use 800 hp. It's unreasonable to treat them the same.
PC31 Inducers for Centrifugal Pumps (Reprinted with kind permission of Lawrence Pumps Inc USA) Page 2 of 3 04/02/09
PumpClinic …Issue 31
With a compatible pump design and operating conditions, an inducer can effectively reduce the initial
capital cost related to system construction and can increase equipment reliability through improved
inlet conditions.
PC31 Inducers for Centrifugal Pumps (Reprinted with kind permission of Lawrence Pumps Inc USA) Page 3 of 3 04/02/09
PumpClinic …Issue 32
Pump applications may involve handling liquid and gas mixtures either as part of the process or as a
result of unwanted application conditions or upsets. Although symptoms may be similar to cavitation,
the process of is a little different.
Cavitation is the formation and subsequent collapse of vapour bubbles that are formed when pressure
within the pump drops below the vapour pressure of the pumped product. As pressure increases
within the pump, the vapour bubbles re-condense (implode) resulting in shock waves that cause
damage to pump components.
The pumping of entrained gas is different in that the gas is non condensable. Centrifugal action tends
to cause a separation between the liquid and gas due to differences in density. This centrifugal action
moves the liquid to the outside and concentrates the gas in the eye of the impeller. This restricts the
flow area for the liquid causing significant pressure drop and hence cavitation even though NPSH
calculations will indicate that cavitation should not occur.
Some industries introduce air into products as part of the process. An example is the Pulp and Paper
industry where air in the range of 4-10% is introduced into pulp slurry as part of the ink removal
process in paper recycling plants. Some processes also involve pumping a two-phase flow. Excess
agitation in suction vessels or vortex formation due to inadequate submergence or incorrect sump
designs may result in introduction of undesirable gas entrainment.
The proper selection of a centrifugal pump for liquid and gas (two-phase) mixtures is highly dependent
on the amount of gas and the characteristics of the liquid. The presence of entrained gases will
reduce hydraulic performance of centrifugal pumps and can potentially cause loss of prime. Standard
centrifugal pumps’ pump designs can be used for entrained gases up to 4% by volume.
For gas entrainment values above 4%, specifically modified impellers can be used effectively. Pump
performance corrections are required in all cases with gas content from around 2%. Gas
concentrations above 10% can be handled, however pumps designed with specific gas handling
characteristics are required e.g. vortex impeller pumps, side channel pumps and regenerative turbine
pumps.
Most centrifugal pump can handle low concentrations of entrained gas. As detailed earlier, the gas will
accumulate in the eye of the impeller restricting flow and head generation. Continued gas
accumulation may cause the pump to vapour lock and lose prime. Fig. 1 below shows performance
variation as gas concentrations vary form 0% to 10%. At concentrations up to 2%, the impact is
relatively insignificant. Impact on performance is still reasonably acceptable up to 4%.
PC32 Pumping Liquids with Entrained Gas Page 1 of 2 March 25, 2009
PumpClinic …Issue 32
As the percentage of gas increases further, the performance begins to quickly deteriorate (Fig. 1) until
the pump becomes unstable and loses prime. Increasing the running clearances between the impeller
and pump casing (.004 to .008 mm) allows for additional leakage and this may assist in preventing loss
of prime at the higher gas concentrations.
Much testing has been done by various manufacturers, however the many variables impacting on the
effect of gas entrainment does not allow for presenting specific selection information.
Manufacturers should be contacted for detailed selections on applications where gas entrainment is an
issue.
PC32 Pumping Liquids with Entrained Gas Page 2 of 2 March 25, 2009
PumpClinic …Issue 33
The design of a piping system can have an important effect on the successful operation of a
centrifugal pump.
Selection of the discharge pipe size is primarily a matter of economics. The cost of the various pipe
sizes must be compared to the pump size and power cost required to overcome the resulting friction
head.
The suction system design is far more important. Many centrifugal pump troubles are caused by poor
suction conditions. Such items as sump design, suction piping design, suction and discharge pipe
size, and pipe supports must all be carefully considered.
The information detailed below is to be taken as guidelines only and all readers should have specific
designs undertaken for individual applications.
Suction Piping Design
The function of suction piping is to supply an evenly distributed flow of liquid to the pump suction, with
sufficient pressure to the pump to avoid excessive turbulence in the pump impeller.
The suction pipe should never be smaller than the suction connection of the pump and in most cases
should be at least one size larger. Suction pipes should be as short and as straight as possible.
Suction pipe velocities should be in 1.5 to 2.5 metres per second range unless suction conditions are
unusually good.
Higher velocities will increase the friction loss and can result in troublesome air or vapour separation.
This is further complicated when elbows or tees are located adjacent to the pump suction nozzle, in
that uneven flow patterns or vapour separation keeps the liquid from evenly filling the impeller. This
upsets hydraulic balance leading to vibration, possible cavitation and excessive shaft deflection,
especially on high and very high suction energy pumps. Shaft breakage or premature bearing failure
may result.
Ideally, a straight length of pipe of an equivalent length of five times the pump inlet size (5D) should be
installed before any fitting or valve. Please refer to individual pump instruction books for individual
manufacturer’s recommendations.
On pump installations involving suction lift, air pockets in the suction line can be a source of trouble.
The suction pipe should be exactly horizontal, or with a uniform slope upward from the sump to the
pump as shown in Fig. 1. There should be no high spots where air can collect and cause the pump to
lose its prime. If high spots are unavoidable, automatic vent valves should be installed at the high
points on the piping.
Eccentric rather than concentric reducers should always be used, on horizontal installations, with the
flat side located on top.
If an elbow is required at the suction of a double suction pump, it should be in a vertical position if at all
possible. Where it is necessary for some reason to use a horizontal elbow, it should be a long radius
elbow and there should be a minimum of five diameters of straight pipe between the elbow and the
pump as shown in Fig 2.
Fig 3 shows the effect of an elbow directly on the suction. The liquid will flow toward the outside of the
elbow and result in an uneven flow distribution into the two inlets of the double suction impeller. Noise
and excessive axial thrust will result.
There are several important considerations in the design of a suction supply tank or sump. These are:
Turbulence
It is imperative that the amount of turbulence and entrained air be kept to a minimum. Entrained air will
cause reduced capacity and efficiency as well as vibration, noise, shaft breakage, loss of prime, and/or
accelerated corrosion.
The free discharge of liquid above the surface of the supply tank at or near the pump suction can cause
entrained air to enter the pump. All lines should be submerged in the tank, and baffles should be used
in extreme cases as shown in Fig. 4.
Submergence
Improper submergence of the pump suction line can cause a vortex, which is a swirling funnel of air from
the surface directly into the pump suction pipe. In addition to submergence, the location of the pipe in
the sump and the actual dimensions of the sump are also important in preventing vortexing and/or
excess turbulence.
The amount of submergence required depends upon the size and capacity of the individual pumps as
well as on the sump design. Past experience is the best guide for determining the submergence. The
pump manufacturer should be consulted for recommendations in the absence of other reliable data.
1. Pump Flowrates below 315 lites/sec
For horizontal pumps, Fig. 5 can be used as a guide for minimum submergence and sump dimensions
for flows up to approximately 5000 US gallons/min (315 litres/sec). For larger flowrates , refer to item 2
below.
Baffles can be used to help prevent vortexing in cases where it is impractical or impossible to maintain
the required submergence. Fig. 6 below shows three such baffling arrangements.
On horizontal pumps, a bell should be used on the end of the suction pipe to limit the entrance velocity
to 1 - 2.5 metres per second. Also, a reducer at the pump suction flange to smoothly accelerate and
stabilize the flow into the pump is desirable.
The ideal approach is a straight channel coming directly to the pump or suction pipe. Turns and
obstructions are detrimental, since they may cause eddy currents and tend to initiate deep-cored
vortices.
The amount of submergence available is only one factor affecting vortex-free operation. It is possible to
have adequate submergence and still have submerged vortices that may have an adverse effect on
pump operation. Successful, vortex-free operation will depend greatly on the approach upstream of the
sump.
Complete analysis of intake structures can only be accurately accomplished by scale model tests.
Model testing is especially recommended for larger pumping units.
Subject to the qualifications of the foregoing statements, Figures 7 through 10 have been constructed for
single and multiple intake arrangements to provide guidelines for basic sump dimensions.
t = 240 x V /Q
where V= volume in litres
Therefore for a maximum allowable start frequency of 10 starts/hour (i.e. t =360 secs), V =1.5 x the
pumping rate (Q litres/sec) or 6 minutes
Reprinted with kind permission of Viking Pump Inc, a unit of IDEX Corporation
Ten Misconceptions
on Rotary PD Pumps
John Petersen and C.H. Tan, Viking Pump, Inc.
Rotary positive displacement (PD) pumps are not well understood by many
specifying engineers and users and are sometimes misapplied, incorrectly
specified or simply not used where they could or should be. This article
explores the areas of commonality between centrifugal and PD pumps and
ten common PD pump misconceptions and the reality behind each.
T
here is considerable resident knowledge about cen- • All try to displace the same amount of liquid with each
trifugal pumps, which comprise the majority of all rotation of the shaft.
pump installations. However, rotary positive dis- • Flow is directly proportional to speed.
placement (PD) pumps are not well understood by many • All can pull liquid from below the pump or self prime.
specifying engineers and users. • Most have close fitting internal parts.
Rotary PD pumps account for 15 percent of all installa- • Most have one pumping element driving another (gears,
tions, with sales divided into many different pumping tech- rotor/vanes, etc.).
nologies. As a result, PD pumps are sometimes misapplied, • All have a small amount of liquid that goes from dis-
incorrectly specified or simply not used where they could or charge back to suction. This is called slip, and it varies
should be. Though PD pumps move liquids through many with liquid viscosity and differential pressure.
different approaches, much commonality is shared in design • All require some form of overpressure protection.
and operation.
Centrifugal pumps move liquid in a much different fash- Conventional wisdom says that a PD pump must be
ion than do PD pumps, and their resulting performance dif- used over a centrifugal pump when one or more of the fol-
fers as well. Centrifugal, or kinetic, pumps impart rotational lowing application conditions exist:
energy to the liquid and convert it to potential energy (pres- • Liquid viscosity is too high (generally anything over
sure) through the design of the volute. PD pumps, on the 150-cps requires a PD pump).
other hand, move liquids by transferring confined amounts • Constant flow is needed over varying differential pressure.
of liquid (defined by pumping element geometry) from the • Suction lift or self priming ability is needed.
inlet to the outlet of the pump.
The key here is that centrifugal pumps generate pres- Although these criteria are all sound, PD pumps may
sure and flow results, while PD pumps generate flow and still not be used, or there may be rationale to use a PD pump
pressure results. In other words, a PD pump generates just when none of these conditions exist. At this juncture, many
enough pressure to overcome system resistance created by the of the common misconceptions on PD pumps come into
flow of liquid through it. Because of this, flow output from play. Let’s explore ten of them.
a centrifugal pump varies with differential pressure, whereas
flow from a PD pump is essentially constant with varying 1. PD Pumps Are Not Well Suited for
differential pressure. Thin Liquids
All of the many different types of PD pumps use geom- Ask almost any pump user what type of pumps to use for a
etry of the parts to expand and contract volumes of liquid. thin liquid application and the response is generally a cen-
Volume expansion draws liquid into the pump and volume trifugal pump. This is often the right answer, but sometimes
contraction moves liquid out of the pump. Though a number not. Some believe that PD pumps cannot be used on thin
of geometries are involved, most rotary PD pumps share liquids at all due to their design characteristics. For exam-
common design and operating characteristics: ple, how can a gear pump be used on a thin non-lubricating
liquid when one gear drives the other? can be used very successfully. One particular customer would
The fact is most PD pumps – including gear pumps – can not allow use of gear type PD pumps because of perceived liquid
be used on thin liquids. Water is the most common thin liquid, damage. Actual testing showed this perception to be wrong.
and the internal gear pump was actually invented to handle it. Most important, PD pump manufacturers typically reduce
Liquefied petroleum gas, refrigerants, solvents, fuel oils, gaso- speed and slightly increase internal clearances to minimize the
line and even liquid carbon dioxide are some of the other thin effects of shear. The takeaway is PD pumps can handle shear
liquids handled very successfully with PD pumps. Selecting sensitive liquids without damage if properly applied.
proper pump materials is important when moving a thin liquid,
and most manufacturers offer many choices to handle the low 4. PD Pumps Are Not Suitable for
lubricity and viscosity typically associated with thin liquids. Abrasives or Solids
True, some PD pump principles do not handle solids, but most
2. PD Pumps Do Not Need Overpressure can handle some form of abrasives. Lobe pumps and progressing
Protection cavity pumps do a good job on solids and can handle abrasives
A centrifugal pump that is dead headed, either accidentally or too. Gear pumps handle abrasives quite well, with a few mate-
on purpose, can develop a predictable maximum shut off pres- rial changes to retard wear, but they will not tolerate solids.
sure. It will be above the normal operating pressure, but gener- A large majority of shingle manufacturing relies on gear
ally not much more. pumps to handle a mixture of asphalt and up to 60 percent to
A PD pump that is dead headed tries to displace the same 70 percent finely ground stone. This mixture is quite abrasive,
amount of liquid for each revolution of the shaft. Because of but reasonable life can be attained by reducing pump speed and
this, pressure continues to increase until something breaks in using hard parts in key wear areas.
the system, the pump is damaged or the driver runs out of Abrasive wear is caused by forces acting on the relative
power. None of these are safe or desirable conditions. To pre- motion surfaces within the pump in the presence of an abrasive
vent this, having some form of overpressure protection either media. Pressure and speed of the pump create these forces, so
on the pump or in the system is important. minimizing them can extend pump life. Reducing pump speed
Many manufacturers offer pump mounted relief valves, is the logical way to do this. Most manufacturers recommend
but other ways can accomplish the same purpose. System this in varying degrees, depending upon the abrasive nature of
relief valves, rupture disks, torque limiting couplings and the product being pumped. Pressure is a function of the system,
motor power load monitors can all be used to limit maximum so proper design helps extend pump life by reducing the load
attainable pressure in the pump or system. Though a system the pump is required to handle.
is designed to be continually open, inadvertent valve closing,
plugged filters or other system upsets may cause enough block- 5. PD Pumps Cannot Handle Non-
age to significantly increase pressure. Overpressure protection is Lubricating Liquids
a must with all PD pumps – and this is often overlooked. Non-lubricating liquids range from thin to thick viscosity.
Many thin liquids have poor lubricating properties, but so do
3. PD Pumps Damage Shear Sensitive some thick liquids. For example, number 6 fuel oil (sometimes
Liquids used for heating or diesel fuel) can be as thick as 15,000-cps,
This may well be the most common misconception on PD but it is not very lubricating.
pumps, particularly with gear type pumps. With close fitting Manufacturers tend to limit the load from pumping ele-
internal parts, the liquid is often thought to be simply sheared ment contact in PD pumps by limiting maximum pressure
or damaged by close running components. While it is true with non-lubricating liquids. Doing this minimizes pumping
that some of the liquid is sheared within the close internal element wear on most non-lubricating liquid applications.
running clearances, only a small amount of liquid is actually The critical area in PD pumps is pumping element sup-
being sheared within the pump. The vast majority of liquid port, which can be either a journal bearing operating in the
going through the pump comes through in “chunks” and is not liquid pumped or an external antifriction bearing. PD pump
sheared at all. manufacturers have a wide array of journal bearing and shaft
Numerous applications and actual testing have busted this materials to handle low lubrication situations. The key is choos-
myth. Most notable is wastewater polymers, where excessive ing the right materials for the particular liquid characteristics to
shearing changes the viscosity. Gear pumps are frequently used get the best pump life.
to transfer these polymers at actual manufacturing facilities and External antifriction bearing support takes away the mate-
do no damage to the liquid. Many installations of gear pumps rial problem, but in most designs this moves the support far-
at wastewater treatment facilities operate without affecting ther away from internal pump loads, resulting in higher bearing
liquid properties. loads that may also limit maximum pressure on a particular
Field testing of other sensitive liquids proves these pumps design.
7. PD Pumps Are Expensive tions, but theoretical output remains constant. Traditional lobe
to Own pumps with three lobed rotors do have theoretical flow pulsa-
The initial cost of most PD pumps is more than centrifugal tion, but again this is minimal.
pumps, but one must consider the total cost of ownership.
Using a simplistic approach to cost of ownership, the main con- 10. PD Pumps Cannot Run Dry
tributors are first cost, repair costs and energy to operate. PD pumps self prime, meaning they are capable of pulling a
Assuming an average pump life of seven years, a couple of liquid from a level below the pump into the pump port. This
repairs during its life span and rather modest pressure require- means the pump must run without liquid for the time it takes
ment energy to operate the pump equals one half of the total to get liquid up the elevation. In other situations, PD pumps
cost of ownership. First cost is the lowest, with repair parts are asked to empty tanks, which many times results in periods
coming in second. Figure 1 shows the detail. PD pumps can of running with little, if any, liquid inside the pump.
be quite efficient. A small increase here can save considerable Most PD pumps can run dry for short periods of time
money over the life of the pump. without damage. In many cases there is a small amount of liquid
in the pump, which keeps the parts wetted to the point damage
8. PD Pump Repairs Are Expensive does not occur. Obviously, extended periods of run dry are not
Repair parts costs vary between manufacturers. Since these recommended and some designs are more tolerant than others.
costs differ, understanding them before the pump purchase is In any case, run dry situations are more common than anyone
always a good idea. Misconception 7 above assumed an aver- likes to admit and PD pumps can usually handle them.
age pump life of seven years with two lower level repairs and a
major overhaul (replacing all critical wear parts) during the life Summary
of the pump. Cost for the parts required for a major overhaul is Rotary PD pumps are (for the most part) simple devices, but
only two-thirds the price of a new pump. PD pumps are usually more complex in application. Understanding the basic operat-
easy to work on too, so labor costs are not high. ing characteristics of PD pumps and system requirements is a
great place to start in correctly applying a pump.
9. PD Pumps Have Pulsing Flow Recognizing these common misconceptions opens up a
This is true for reciprocating PD pumps, but not for most whole new application area for these pumps to improve perfor-
rotary PD pumps. This may seem counterintuitive since these mance, extend service life or operate more efficiently. Consider
pumps move liquid by delivering confined volumes of liquid all the alternatives to a pumping problem before making a deci-
from the suction to the discharge port. Although it would seem sion – you’ll be surprised by how many choices you have.
that confined volumes or buckets of liquid would result in flow
pulsations, they actually don’t. P&S
In theory, gear, vane, some lobe pumps and all types of John Petersen is vice president, technical customer service and
screw pumps deliver a continuum of liquid resulting in very CH Tan is regional director for the Asia Pacific Region of
little, if any, flow pulsations. There may be slight differences Viking Pump, Inc., a unit of IDEX Corporation, 406 State
in slip within the pump, depending on rotational position of Street, P.O. Box 8, Cedar Falls, IA 50613-0008, 319-
the pumping elements that result in minimal pressure pulsa- 266-1741, Fax: 319-273-8157, www.vikingpump.com.
Parallel operation may be required to meet variable demands, such as flood control, or to satisfy a
temporary condition that occurs such as when changing over pumps in an uninterruptible process. Figure
2 illustrates the characteristic of single vs. parallel operation.
Two identical pumps operating in parallel are capable of producing twice the flow of a single pump at any
given Total Dynamic Head (TDH). However, the actual flow rate realised in the system is dictated by the
intersection of the system curve with the pump curve. Unless the system curve is variable, the flow
increase may not be that significant. For example, assume that there is a set of fixed spray nozzles,
where the system resistance is purely frictional, and varies only as a result of flow change. When a
second pump is introduced, resistance in the system increases as the flow increases. The flow will
increase only to where the system curve intersects the two-pump curve, as shown in figure 3. The
amount of flow increase is dictated not only by the system curve, but also by the steepness of the pump
curves. Pumps with flat curves will have less TDH separation than pumps with steep curves and therefore
will have less flow rate change.
PC35 Single vs Parallel Operation of Centrifugal Pu8mps June 29, 2009 Page 1 of 4
PumpClinic …Issue 35
A fire pump installation would be an example of a system that has a variable system curve. Figure 4
illustrates this. Each time an additional fire nozzle is activated, the system resistance is decreased. This
causes the system curve to move to a higher flow on the pump curve, increasing the kW load on the pump
and decreasing the amount of TDH available. Eventually, another fire pump may need to be activated to
maintain the system pressure as more nozzles come on line.
In figure 5, the system and TDH curves intersect at a TDH that is greater than the shut-off TDH value and
at a lower TDH than the max TDH of the curve. When the second pump starts, the flowrate will increase
only to the first point where the TDH matches the pressure in the system. Because the first pump’s
PC35 Single vs Parallel Operation of Centrifugal Pu8mps June 29, 2009 Page 2 of 4
PumpClinic …Issue 35
operating TDH point is mirrored by another at a lower flow, the second pump’s flowrate will not increase
beyond the low flow point. At best, the pumps will share the load unevenly, with the second pump
operating at a lower efficiency. Worse, the second pump might operate at less than its minimum allowable
design flow, resulting in damage to the equipment and possible injury to personnel.
A similar condition may exist when one pump is worn. As a pump wears, the amount of TDH produced at
any given flow rate diminishes. If the new pump’s intersection with the system curve is above the shut-off
TDH of the worn pump, it will force the worn pump into a shut-off condition.
Load sharing problems, between pumps operating in parallel, may increase wear, reduce seal and bearing
life, lower operating efficiencies and limit process operations. In the absence of any flow measurement
capability, an uneven performance distribution, between pumps operating in parallel, is easier to avoid
than to detect. Proper pump selection for parallel operation and pump performance monitoring are the
best tools in avoiding load sharing problems and maintaining a well operating parallel pump installation.
Parallel Operation of Unmatched Pumps
When pumps operate in parallel, the flow rate at any given TDH point is additive. In the case of pumps
that have identical operating characteristics, the flow would double. For example, two pumps that each
had a capacity of 100 M3/hr at 50 M TDH would have a combined capacity of 200 M3/hr at 50 M TDH.
Again, the system curve does not change, so the actual change in flow that occurs with bringing a second
pump on line, in parallel, is determined by the characteristic curve intersection with the system curve. This
is shown graphically in Fig. 3.
When a pump is operated in parallel with another pump that has a different operating characteristic, the
same rule applies as for identical pumps: for any TDH, common to both pumps, the flow characteristic will
be additive.
If one pump exhibits a lower shut-off TDH characteristic, it will operate at shut-off until the dominant pump
moves far enough out on its curve so that its TDH falls below the shut-off TDH of the pump with lower
head (Fig 6). The danger here is in the system-pump interaction. If the system curve intersection with the
characteristic curve is at a higher TDH than the shut-off flow of the weak pump, the weak pump will be
forced to run at shut-off and a serious failure could occur.
In Fig. 6, a zero-flow condition for the weak pump will exist when the system curve intersection is to the
left of the vertical dashed line.
PC35 Single vs Parallel Operation of Centrifugal Pu8mps June 29, 2009 Page 3 of 4
PumpClinic …Issue 35
As a general rule it is a good idea to have flow measurement installed for any pumps designed to operate
in parallel. Without flow measurement, it is very difficult to determine what the load sharing is between
two pumps.
Motor power is often a questionable indicator of flow, as many power curves are quite flat and show small
changes in load over relatively large changes in flow. Also, when wear does occur, the power draw may
remain relatively constant even though performance is falling off. This is due to a decrease in pump
efficiency which is not visible to the pump operator.
PC35 Single vs Parallel Operation of Centrifugal Pu8mps June 29, 2009 Page 4 of 4
PumpClinic …Issue 36
METERING PUMPS
The metering pump is a positive displacement chemical dosing device with the ability to vary capacity
manually or automatically as process conditions require. It features a high level of repetitive accuracy
and is capable of pumping a wide range of chemicals including acids, bases, corrosive or viscous liquids
and slurries.
The pumping action is developed by a reciprocating piston, plunger or diaphragm which is either in
direct contact with the process fluid, or is shielded from the fluid by a diaphragm. Diaphragms may be
activated by direct mechanical link or by hydraulic fluid.
Metering pumps are generally used in applications where one or more of the following conditions exist:
• Low flow rates are required
• High system pressure exists
• High accuracy feed rate is demanded
• Dosing is controlled by computer, microprocessor, DCS, PLC, or flow proportioning
• Corrosive, hazardous, or high temperature fluids are handled
• Viscous fluids or slurries need to be pumped
Figure 1
2. Unlike centrifugal pumps, flow rate is not greatly affected by changes in discharge pressure.
Figure 2
3. The metering pump flow vs. stroke characteristic curve is linear. It is not however, necessarily
proportional in that 50% stroke setting may not equal 50% flow. This is due to the fact that the calibration
line may not pass through 0 on both axes simultaneously. By measuring flow at 2 stroke settings, plotting
both points and drawing a straight line through them, other flow rates vs. stroke can be accurately
predicted.
The steady state accuracy of a correctly installed industrial grade metering pump is generally +/- 1.0% or
better. Although a metering pump can generally be adjusted to pump at any flow rate between 0 and its
maximum capacity, its accuracy is measured over a range determined by the pump's turndown ratio.
Most metering pumps have a turndown ratio of 10:1 which simply means that the pump is within its
accuracy rating anywhere between 10% and 100% of capacity. Some newer designs of metering pumps
feature higher accuracy, and a greater turndown ratio of 100:1. Therefore, this design will accurately dose
anywhere between 1% and 100% of capacity.
Figure 3
LIQUID END
The liquid end design and materials of construction are determined by the service conditions, and
the nature of the fluid to be handled. Temperature, flow rate, fluid viscosity, corrosiveness and
other factors are considered.
DRIVE MECHANISM
The drive mechanism translates the rotary motion of the driver into reciprocating movement.
FLOW ADJUSTMENT
Pump flow rate is adjustable by varying stroke length, effective stroke length or stroking speed. Most
metering pumps are supplied with a micrometer screw adjustment similar to the one shown here.
The micrometer can also be replaced by an electronic or pneumatic actuator to adjust pump flow rate in
response to a process signal
DRIVERS
The pump is usually driven by an AC constant speed motor. Electromagnetic drive is available in small
flow pumps.
Figure 4
DRIVE MECHANISMS
Metering pumps can be powered by a variety of drivers however, the almost universal driver is an
electric motor. The motor speed is normally reduced to pump design speed by the use of gearing built
into the pump power end. This rotary power is converted to a linear motion through one of three
methods:
- a crank mechanism with either fixed (Figure 5) or variable (Figure 6) stroke length
- an eccentric or cam arrangement (Figure 7)
Depending on the type of adjustable output flow mechanism used, the power can be utilised on both the
forward thrust of the crank and the back thrust of the crank. The eccentric or cam arrangement,
however, can provide power in only one direction.
Metering pumps with solenoid power ends (Figure 8) are another type of drive and create linear,
reciprocating motion using electromagnets. The solenoid type could be considered the ideal power end
as it does not require any type of transmission to convert motion from rotary to linear. Another
advantage is that it has in built overload protection as the pump simply stops at excessive load. A
disadvantage is that availability is limited to very small powers.
Figure 5
Figure 6
Figure 7
ADJUSTMENT SCREW
Figure 8
LIQUID END
Metering pumps generally fall into five basic types:
Both the piston packed pump and the plunger packed pump allow some degree of leakage past their
dynamic seals. In some cases this is not an objectionable shortcoming; in other cases, it can be very
objectionable and, in most instances, costly as well. Diaphragm and tube pumps are seal/less and
therefore overcome this potential problem.
It is often possible to multiplex pumps i.e. coupling or ganging of two or more liquid ends and use only one
driver. The main purposes of multiplexing are:
- provide greater flow or pressure without significantly increasing the driver size
- provide exact proportioning or synchronization between heads
- provide greater turndown ratio i.e. flow control range
- reduce pulsations
Rings or packing located on the piston move back and forth with the piston to effect a dynamic seal with
the inside diameter of the cylinder and a static seal with the outside diameter of the piston. Packing on
the connecting rod provides at static seal so that it becomes double acting and displacement may occur
on both directions. This provides double the flow of single acting pumps.
The forward travel of the piston reduces the internal volume of the liquid chamber, displacing the metered
liquid out the discharge check valve. The pressure required to move the liquid through the discharge
check valve is also applied to the suction check valve, forcing it into a closed position, ensuring correct
flow direction
The reverse travel of the piston decreases the pressure within the liquid chamber by enlarging the internal
volume of the chamber. This change of pressure results in a rapid closing of the discharge check valve
caused by the external pressure acting on the valve and allows the suction check valve to open because
of an external pressure under the check valve that can be either above or below atmospheric pressure.
The accuracy of the reciprocating metering pump is achieved by the previously described predetermined
controlled piston travel of the pump, the control of the strokes per minute, and the precise opening and
closing of the check valves. The inaccuracy, on the other hand, is caused by leakage past the piston
packing and the check valves.
There is a strong trend to leak free (diaphragm and tube) metering pumps, however plunger pumps
continue to be used. The main reasons are that plunger pumps are cheaper, can develop higher
discharge pressures and are more suitable for extreme temperature conditions.
The mechanical diaphragm pump's principle of positive displacement output is similar to that of the piston
plunger pump except that the diaphragm becomes the displacement measuring element, as it moves back
and forth in the fluid chamber (Figure 8).
pumped fluid, and the actuation of the diaphragm is by hydraulic power instead of mechanical power
(Figure6).
The measuring piston or plunger reciprocates within a precisely sized cylinder at an established stroke
length, displacing a volume of hydraulic liquid, not the product liquid. The hydraulic liquid is stable and
has excellent lubricating qualities. The piston uses the hydraulic oil to move the diaphragm forward and
backward, causing a displacement that expels the product liquid through the discharge check valve and,
on the suction stroke, takes in an equal amount through the suction check valve. The diaphragm isolates
the liquid product being contained within the liquid chamber and check valves. These are the only parts
that must be made of chemically compatible material.
The diaphragm's only job is to separate two liquids. It normally does no work, carries no load, and pumps
no liquid; rather it serves as a moving barrier between liquids during periods of pressure imbalance. It is
simply a moving partition with pressure hydraulically balanced on both its sides; on one side is the liquid
product and on the other side is the hydraulic oil. At full deflection, the diaphragm undergoes total
combined stresses well within the endurance limit of the diaphragm material. Contoured support plates
are provided on either side of the diaphragm to ensure that stresses are kept within limits. When properly
installed and working within the recommended temperature range and not affected by corrosion or
abrasion, the diaphragm has an unlimited life.
As previously stated, the piston or plunger handles only hydraulic oil. Conventional seals are used on the
piston or plunger, which does not require power flushing and complicated drain systems as are found on
conventional piston or plunger pumps handling corrosive or hazardous liquids.
Even the slightest leakage past the piston is replaced on the suction stroke through the automatic
functioning of a compensation system, which draws in replacement oil from the oil reservoir (Figure 11).
Any excess pressure within the hydraulic system or the liquid product chamber is relieved through the
automatic action of a pressure relief valve. This valve blows off oil, under excess pressure ahead of the
piston, back into the oil reservoir. This valve blows off oil, under excess pressure ahead of the piston, back
into the oil reservoir (Figure 12).
The vacuum and pressure compensator systems actually perform three important functions that the other
described types of metering pumps cannot do unless auxiliary equipment is added to their piping systems.
As described previously, they compensate for any leakage occurring within the hydraulic system of the
pump, ensuring a balanced diaphragm movement. In addition, they serve to protect the process system
from an over-pressure condition produced by the pump. For instance, the positive displacement pump,
because of its design, must over pressure the system to the point of damaging the pump, bursting pipes,
or damaging other downstream equipment should an operator mistakenly close a shut-off valve
downstream from the pump.
The hydraulic diaphragm pump will, however, relieve any pump-produced pressure beyond the set
pressure of the pressure relief valve, thus avoiding the dangerous build up of pressure. The
compensation system also serves to protect the pump from a closed suction line or a partially clogged
strainer in the suction line. Should this occur, the backward movement of the diaphragm is prevented and
the vacuum relief system would automatically open to relieve the starved suction condition within the
pump.
In doing so, however, a surplus of hydraulic oil enters into the system between the diaphragm and piston.
As the piston starts forward on its discharge stroke, the diaphragm is displaced forward and will come into
contact with the contoured dish support plate in the process liquid chamber, because of the surplus oil
drawn into the hydraulic chamber.
At the moment of diaphragm contact with its support plate, an over-pressure condition starts to develop
within the hydraulic system. The pressure relief valve now opens to relieve the surplus oil back into the
hydraulic reservoir, preventing a dangerous build up of pressure. The interaction of the two compensation
systems continue stroke after stroke to activate a fluid-clutch-type action to prevent overloading of the
pump's power end until the condition plugging the suction or discharge lines is found and corrected.
FLOW ADJUSTMENT
Flow control can be achieved in a number of ways, however the most common are as follows.
1. Changing the stroke rate within the pump by changing internal gearing in the drive mechanism. This
is done at time of pump manufacture and further changes cannot be made after pump installation.
2. Changing the stroke rate by changing the driver speed. When the driver is an electric motor, using a
frequency inverter will achieve this result. With solenoid operated pumps, this can be done by setting
the switching on the solenoid.
3. Changing the stroke length. This can be achieved in a number of ways and is dependent on the
pump construction. Details are given below.
4. Adjustable cranks and hydraulic bypass have also been used in the past but are rarely seen these
days.
Displacement per stroke is controlled through limiting the rearward travel of the piston. Adjusting the
stroke length mechanism extends and / or retracts the internal adjustment screw. When a stroke length of
less than 100% is desired the internal adjustment screw is rotated and the rearward piston travel is limited
based on the stroke length setting. As a result there is no contact with the cam for a portion of the cam
rotation and the piston stops moving until the cam rotates to a position in which contact is re-established
with the cam follower.
Figure 13
Figure 14
Figure 15
The flow characteristics produced from a lost motion style of pump are shown below. As indicated, the
flow at 100% stroke length can be represented as a sine wave. When the stroke length is decreased the
maximum amplitude of each stroke is maintained while the full potential volume per stroke is decreased.
- - (
360
270
SUCTION
Pumps with larger flow requirements are typically handled by full motion pumps. Full motion pumps rely on
internal linkages for the adjustment of stroke length. The pump motor turns the worm shaft which, in turn,
rotates the eccentric gear within the pump gearbox. The eccentric gear transmits motion to a connecting
rod which is attached to an oscillating housing. The oscillating housing is stationary at its top — the
resulting motion is similar to that of a pendulum. Within the oscillating housing exists a housing block
which is, in turn, connected to a connecting rod.
The connecting rod is attached to the piston. It can be seen from Figure 16 that when the housing block is
at its full bottom position (100% stroke length) the piston will maximize its horizontal movement. As a
result the pump will produce its greatest potential displacement per stroke. When the housing block is
adjusted to its full top position (0% stroke length) the piston will be stationary and, as a result, the pump
will produce no flow. Adjustment of stroke length is actually an adjustment of the housing block position
within the oscillating housing — this adjustment will determine how far back and forth the piston can travel
and, as a result, the volume per stroke that the pump can produce.
Figure 16
Figure 17
Figure 18
With the use of variable speed motors and drives many metering pumps are controlled through variable
speed in lieu of stroke length. The flow characteristics of full motion and lost motion pumps are identical if
stroke length is maintained at 100% and motor speed is used to adjust flow. Instead of changing volume
(amplitude) per stroke the adjustable motor speed will modify strokes per minute (frequency). The
resulting output will be identical regardless of stroke length type.
Finally, a discharge pulsation dampener is a typical recommendation for all styles of metering pumps. The
discharge pulsation dampener transforms a diaphragm metering pump's reciprocating flow to laminar flow.
As a result, the flow characteristics downstream of a metering pump, regardless if it is lost motion or full
motion, will be identical when a discharge pulsation dampener is installed.
Acknowledgements:
Metering Pump Handbook (Pulsafeeder Inc)
www.pulsa.com
www.miltonroy.com
Enclosed impellers
An enclosed impeller incorporates a full front and
back shroud. Fluid flows through the internal
impeller passages without hydraulic interaction with
the stationary casing walls. In a well designed
enclosed impeller, the relative velocity between the
impeller and the fluid at any given radius is quite
small. This results in less wear than other impeller
styles.
Enclosed Impeller
A portion of the fluid exiting an enclosed impeller leaks back to the pump suction by traveling through the
gap between the front impeller shroud and the casing. An enclosed impeller typically has wear rings or
radial pump-out vanes to control this leakage. A centrifugal pump with an enclosed impeller is usually not
dependent on tight axial clearances to manage leakage. Therefore an enclosed impeller pump can tolerate
moderate wear with little adverse effect on overall performance and efficiency.
Pump-out vanes offer a better alternative for handling abrasive solids. Pump-out vanes control both
leakage and axial thrust through a pumping action. Local flow velocities with pump-out vanes are much
lower and spread over a greater area. The lower local velocity results in a much lower wear rate. It is not
uncommon for pump-out vane life to equal or exceed the life of the main impeller blades.
PC37 Solids Handling with Centrifugal Pumps October 21, 2009 Page 1 of 5
PumpClinic …Issue 37
An impeller with wear rings loses efficiency rapidly as the rings wear. If ring wear is severe, the high
velocity zone can shift from the wear rings to the impeller thrust balance holes resulting in an expensive
and premature repair or replacement of the impeller. It is not uncommon to have several outages to
replace wear rings over the life of a single impeller when wear rings are used in an aggressive solids
application.
A disadvantage of enclosed impellers is that the front and back shrouds, rotating in close proximity to the
casing walls, generate disc friction that lowers the efficiency of the pump relative to that found in open
impeller designs. Another disadvantage is that the enclosed impeller is more easily plugged. Large solids
that might otherwise be broken up by the grinding action generated by a rotating open impeller and the
stationary casing wall can easily become lodged in the eye of an enclosed impeller. This may create a
mechanical or hydraulic imbalance that has the potential to damage the pump, or at the least causes a pre-
mature outage to remove the blockage.
Open Impellers
An open impeller is characterized by impeller blades
that are supported almost entirely by the impeller hub.
This is the simplest impeller style and it is primarily
applied to clean, non-abrasive, low horsepower
applications. An open impeller is lighter in weight
than its shrouded counterpart. Less impeller weight
reduces shaft deflection and enables the use of a
smaller diameter shaft, at a lower cost, than an
equivalent shrouded impeller.
Open Impeller
An open impeller typically operates at a higher efficiency than a shrouded impeller of the same specific
speed. The largest contributor to efficiency loss in an enclosed radial impeller is disc friction caused by the
front and back impeller shrouds turning in close proximity to the stationary casing walls. Removing the
shrouds eliminates the disc friction.
One drawback of the open impeller is that it is more susceptible to abrasive wear than a shrouded impeller.
High velocity fluid on the impeller blades in close proximity to the casing walls establishes rotating vortices
that accelerate wear when abrasives are present.
A tight clearance between the impeller and the front and back casing walls is necessary to maximise
PC37 Solids Handling with Centrifugal Pumps October 21, 2009 Page 2 of 5
PumpClinic …Issue 37
efficiency. As the impeller wears, these clearances open and efficiency drops rapidly. The tight operating
clearances required on both sides of an open impeller for efficient operation precludes adjustment of the
impeller axial position to compensate for wear.
Semi-Open Impellers
A semi-open impeller is a compromise between an
open and an enclosed impeller. It incorporates a
single shroud, usually located on the back of the
impeller. A semi-open impeller has a solids passing
capability similar to that found in an open impeller.
With only a single shroud a semi-open impeller is
easy to manufacture and completely accessible for
applying surface hardening treatments. For
moderately abrasive slurries, especially if plugging is
a concern, a semi-open impeller is a good choice.
semi-open impeller
A semi-open impeller operates more efficiently than an enclosed impeller because of lower disc friction and
tighter axial clearances. It has an advantage over an open impeller in that it can be adjusted axially to
compensate for casing wear.
High axial thrust is the primary drawback of a semi-open impeller design. Axial thrust balance is
manageable through design for both open and enclosed impellers. On a semi-open impeller, the entire
backside surface of the shroud is subject to the full impeller discharge pressure. The front side of the
shroud is at suction pressure at the eye of the impeller and increases along the impeller radius due to
centrifugal action. The differential between the pressure profiles along the two sides of the shroud creates
the axial thrust imbalance. This can be managed somewhat through the use of pump-out vanes on the
back side of the shroud, but the vanes will start to lose effectiveness if the impeller is moved forward in the
casing to compensate for wear. Some manufacturers have integrated an adjustable wear-plate into the
casing design so that clearance adjustments can be made. Combined with hard materials or surface
hardening treatments, this option provides a good design in lightly to moderately abrasive applications.
An obvious question is why use a semi-open impeller in a solids application if an open impeller with an
adjustable wear plate could be used instead? It might seem logical that an open impeller of hard metal
construction, used in conjunction with an adjustable wear liner, would combine good solids handling
characteristics, with low thrust imbalance, light weight, and adjustability for wear. Unfortunately, true open
impellers lack the structural support to prevent blade collapse or deformation under the demands of most
industrial applications. A semi-open impeller is well suited for handling solids in applications where the
blades might encounter high impact loads from rocks and the like, or in higher power applications. In both
situations the shroud provides additional structural support and reinforcement to protect against blade
collapse or deformation.
PC37 Solids Handling with Centrifugal Pumps October 21, 2009 Page 3 of 5
PumpClinic …Issue 37
Recessed Impellers
Recessed impeller pumps are characterized by impeller vanes that either do not
extend into the pump casing or extend only partially into the casing, essentially
leaving the casing as an open flow passage. Recessed impeller pumps are well
suited for handling large or stringy solids. The maximum solid size is usually limited
by the pump suction opening such that any solid that enters the pump will pass
through.
In operation, some of the fluid is drawn into the rotating impeller and discharged
back into the casing through centrifugal action. Through fluid dynamics, the partial
flow through the impeller imparts a centrifugal rotating motion to the entire fluid
body within the casing. Large solids entering the casing are transported by the
rotating fluid body from inlet to outlet without necessarily making contact with the
impeller.
In addition to handling large solids, a recessed impeller pump will handle a higher concentration of
entrained gas than a traditional pump with the impeller centered in the casing. In a traditional centrifugal
pump, gas accumulating at the impeller eye prevents fluid from reaching any down stream part of the
impeller. This results in a breakdown of the pumping action. In a recessed impeller pump, gas entering
the pump does not have to pass through the impeller to exit the pump. Additionally, gas present in one
area of the impeller does not prevent other parts of the impeller from pumping. Gas handling capabilities
in excess of 30% by volume have been reported with this style of impeller. However, the actual
concentration of entrained gas that can be handled for any specific application is dependent on the phase
characteristics of the fluid and should be determined by test.
The efficiency of a recessed impeller pump will be less than the efficiency of a
traditional centrifugal pump. Efficiency losses result from flow recirculation
around the impeller passages, and from the inefficiency of a flow pattern
where fluid rotates around the casing numerous times prior to exiting the
discharge. Efficiencies in the 40%-50% range are common for recessed
impeller pumps.
Most screw centrifugal impellers have a single helical vane wrapping around an
expanding hub from inlet to outlet. The single passage allows for an easy
transition of fluid and solids from the pump inlet onto the impeller. Its inducer-
like design exhibits good NPSHR characteristics. Liquid entering the impeller is
accelerated more gradually along the smoothly expanding hub to the pump
outlet than with a traditional impeller design.
The screw centrifugal pump is a popular choice for handling delicate products
such as food and crystals. Its low shear characteristic reduces emulsification
when pumping mixtures. The pump's ability to pass long fibrous materials such
as rope without clogging makes it a frequent choice for municipal waste water applications.
PC37 Solids Handling with Centrifugal Pumps October 21, 2009 Page 4 of 5
PumpClinic …Issue 37
A screw centrifugal pump typically has an operating efficiency of 70% to 85%. It has a relatively steeply
rising head/capacity curve shape giving it good flow control capability over its allowable operating range.
The relatively large size of the screw impeller is a primary disadvantage of this style pump. The heavy
impeller mandates a large shaft and power frame to limit shaft deflection; which increases the unit cost. In
addition, the single vane impeller is prone to high side thrust when operating off-design. A 1x rpm vane
passing frequency vibration, that sets up as the single vane
outlet passes the casing cutwater during each rotation, is not
uncommon. Multi-vane screw impellers that provide smoother
operation are available. The primary trade-off is solid size
capability.
Disc Impellers
Disc impellers incorporate two or more parallel discs and do not
have traditional impeller vanes. Instead this design relies upon
fluid friction and viscosity to generate a pumping action. As liquid
enters the disc impeller, friction between the fluid boundary layer
and the disc's surface accelerates the boundary layer to about
the same speed as the impeller. Resistance to sheer (or
viscosity) between the boundary layer and the adjacent fluid
layer creates motion in the adjacent layer as well. Each layer in
turn is set in motion by the viscous drag from the adjacent layer.
Slip, or the difference in speed between the disc and each layer,
increases with distance from the impeller. The effectiveness of a
disc impeller is related to the spacing of the discs and the
viscosity of the fluid. Close disc spacing and higher viscosity produce better performance than low
viscosities with wider disc spacing.
A disc impeller is well suited for gentle handling of delicate materials. It also performs well in abrasive
services as there is little relative motion between the fluid contacting the impeller and the impeller itself.
Disc impellers have entrained gas capabilities superior to standard impeller pumps, as gas can enter the
impeller and move through the boundary layers without impediment.
The efficiency of a disc impeller is less than that of a standard centrifugal pump. The efficiency of a disc
impeller pump is commonly in the 35%-50% range. Because it relies on close disc spacing for effective
performance, a disc pump is typically not well suited for large diameter solids.
Non-Clog Impellers
In reality, all impellers are able to be clogged in some way so the term “non-clogging” is, in reality, a
misnomer. Non-clog impellers are designed to accommodate large, soft solids and at the same time
limiting clogging the pump. The impeller generally has a small number of vanes (maximum of three) and
large vane width. The impellers can be either of the closed or semi open type. Leading edges of vanes
tend to be rounded so that solids are not caught up and get tangled.
As the vane width is large to allow passing of solids, these pumps tend to be applied on applications
requiring relatively high flowrates.
PC37 Solids Handling with Centrifugal Pumps October 21, 2009 Page 5 of 5
PumpClinic …Issue 38
Production facilities often require increased output and consider means of improving the performance of
installed equipment. For a pump, this is generally the need for increased flow rate.
The applicable concept is the fact that all centrifugal pumps operate at a flow rate that corresponds to the
intersection of the pump curve and the system curve.
To increase the flow rate from an existing pump, either the pump operating characteristics, or the system
resistance characteristics need to change. If other modifications to the pumping system are being
considered such as new piping, heat exchangers, etc, a new system curve needs to be developed as a
baseline reference before considering any of the following modifications
Options for increased flow from an existing pump:
Increase the impeller diameter
In systems where the primary resistance comes from a pressure vessel, any increase in flow will be
primarily dependent on the shape of the pump characteristic curve. Flatter curves will experience a higher
percentage of flow increase than steep curves. The pump operating point relative to the pump best
efficiency point may change. The only way to predict the performance is to plot the intersection of the
pump and system curves.
Summary
Although several methods of increasing flow rate have been discussed, it is our opinion that the
modification of impellers and installation of suction splitter vanes are not really viable. The reasons are the
uncertainty of the improvement and the repeatability of the process particularly in the case of blade
modification. The better means of achieving increased flow are
- increase speed
Other considerations
There are a number of associated mechanical and hydraulic issues that should also be evaluated when
investigating modifications to increase a pump’s flow rate. It is a good idea to consult the OEM’s service
department to assist in evaluating the pump requirements. Some of the items are as follows:
NPSHR: It is likely that the NPSHR will increase if either flow or pump RPM is increased.
Seal flush requirements: The pressure at the seal chamber will increase in proportion to the pump
discharge pressure. Make sure that there is adequate seal flush pressure to give the required flow rate
through the seal.
Power: Increasing flow rate will result in increased power draw unless it is accompanied by an adequate
decrease in system head requirements. When changing speed or impeller diameter, power will vary
approximately as the cube of the ratio of the change. Couplings, base plate dimensions, and electrical
components will all require re-evaluation.
Mechanical: Planned changes to the operating speed should be evaluated with respect to rotor critical
speed, pressure containment, and mechanical seal limitations.
Operating Limits: Changes that move the pump further away from the best efficiency point may decrease
pump reliability due to hydraulic instability.
PERISTALTIC PUMPS
Principles of Operation
The term peristaltic pump is applied to various ranges known as hose or tube pumps. Although there are
specific differences between hoses and tubing, the terms hose and tube are interchangeable in this article.
Generally hose pumps have a large wall thickness, polyamide reinforcing layers and are capable of handling
differential pressure to 15 bars. Tubing is smaller wall thickness and tube pumps are generally capable of
differential pressures to 4 bars.
Peristaltic pumps are self-priming rotary positive displacement pumps that operate on the peristaltic principle.
The pump consists of three major parts: hose or tubing, housing and rotor. The hose is placed in the tubing
bed between the rotor and the housing. The rotor has a number of "rollers” or "shoes" attached to the
external circumference. These move across the hose where it is occluded (squeezed) pushing the fluid.
The hose behind the shoe or roller recovers its shape, creating a vacuum and drawing fluid in behind it.
Liquid is trapped between the rollers specific to the ID of the hose and the geometry of the rotor. Flow rate is
determined by multiplying speed (rpm) by the volume of the trapped liquid. The volume moved is consistent,
even under a wide range of viscosities or density. The flow rate is directly proportional to the gearbox speed
(rpm).
Dry-Running Design
This design incorporates a unique tube bed that always ensures one roller is
occluding the hose. This is termed dry-running because rollers that do not
operate in a lubricated bath occlude the hose. Dry running pumps generally have
lower flow and pressure capabilities than lubricated pumps.
Dry-running pump
design
Hose Pumps
Higher pressure peristaltic hose pumps which can operate at pressures up to 15 bars, typically use shoes
and have casings filled with lubricant to prevent abrasion of the exterior of the pump tube and to aid in the
dissipation of heat, and use reinforced tubes, often called "hoses". This class of pump is often called a "hose
pump".
The hoses in a hose pump are typically reinforced, with a very thick wall. For a given ID the hoses have
much bigger OD than tubing for the roller pump. This thicker wall, combined with a stiffer material typically
used in the hoses make the forces necessary to occlude the hose much greater than for the tubing. This
results in a bigger pump and motor for a given flow rate with the hose pump than the roller pump resulting in
more energy to run.
Tube pumps
Lower pressure peristaltic pumps, typically have dry casings and use rollers and use non-reinforced,
extruded tubing. This class of pump is sometimes called a "tube pump" or "tubing pump".
These pumps employ rollers to squeeze the tube. These pumps have a minimum of 2 rollers 180 degrees
apart, and may have as many as 8, or even 12 rollers. Increasing the number of rollers increases the
frequency of the pumped fluid at the outlet, thereby decreasing the amplitude of pulsing. The downside to
increasing number of rollers it that it proportionately increases the number of squeezes, or occlusions, on the
tubing for a given cumulative flow through that tube, thereby reducing the tubing life.
1. Fixed occlusion - the rollers have a fixed locus as it turns, keeping the occlusion constant as it squeezes
the tube. This is a simple, yet effective design. The only downside to this design is that the occlusion as a
percent on the tube varies with the variation of the tube wall thickness. Typically, the wall thickness of the
extruded tubes vary enough that the % occlusion can vary with the wall thickness (see above). Therefore,
tubing at the high end of the wall thickness, but within the accepted tolerance, will have higher % occlusion,
increasing the wear on it, thereby decreasing the tube life. The tube wall thickness tolerances today are kept
pretty tight so that this is not much of a practical issue.
2. Spring loaded rollers - As the name indicates, the rollers are mounted on a spring. This design is a bit
more elaborate than the fixed occlusion, but helps overcome the variations in the tube wall thickness over a
broader range. Irrespective of the variations, the roller imparts the same amount of stress on the tubing that
is proportional to the spring constant, making this a constant stress operation. The spring is selected to
overcome not only the hoop strength of the tubing, but also the pressure of the pumped fluid.
The operating pressure of these pumps is determined by the tubing, and the motor's ability to overcome the
hoop strength of the tubing and the pressure.
The key design parameter in hose pumps is hose life. Surveys have shown that 95% of spares value used
on peristaltic pumps is on replacement hoses.
Occlusion/Shimming Philosophy
The minimum gap between the roller and the housing determines the maximum squeeze applied on the
tubing. The amount of squeeze applied to the tubing affects pumping performance and the tube life - more
squeezing decreases the tubing life dramatically, while less squeezing decreases the pumping efficiency,
especially in high pressure pumping. Therefore, this amount of squeeze becomes an important design
parameter.
The occlusion is typically 10 to 20%, with a higher occlusion for a softer tube material and a lower occlusion
for a harder tube material. Rubber composition therefore becomes important so that efficiency is maintained
without “over squeezing” the hose and affecting hose life.
Thus for a given pump, the most critical issues are rubber composition, the amount of shimming and wall
thickness. An interesting point here is that the inside diameter of the tubing is not an important design
parameter for the suitability of the tubing for the pump. Therefore, it is common for more than one ID be
used with a pump, as long as the wall thickness remains the same.
The hose needs to be elastomeric to maintain the circular cross section after millions of cycles of squeezing
in the pump. This requirement eliminates a variety of non-elastomeric polymers such as PTFE, PVDF etc.
from consideration as material for pump tubing.
Different hose materials have different mechanical flex life. The most common hose material is natural
rubber. Natural rubber has double the mechanical life of EPDM and three times the life of Buna/Nitrile and
Hypalon.
Fluoroelastomers have been used but have very poor mechanical life that tends to make this material
impractical. Some manufacturers are currently testing various compositions with the aim of improving fatigue
life.
Chemical Compatibility
The pumped fluid contacts only the inside surface of the tubing. There are no other valves, O-rings, seals or
packings to worry about in a peristaltic pump. Therefore, the only compatibility to worry about in a peristaltic
pump is the hoses for the fluid being pumped.
The most popular hose materials (as distinct from tube) are:
• natural rubber
• EPDM (ethylene propylene diene monomer)
• Buna also known as Nitrile
• Hypalon
• Viton flouroelatomer
Hoses are made in different ways by manufacturers. Some use full width in the particular material. Others
use a common outside layer of natural rubber with an inside layer of other material selected for superior
chemical compatibility. Some problems have been encountered with dual material hoses with the integrity
being compromised due to pin holes in the inside layer and delamination between the two layers.
Typical elastomers for pump tubing (as distinct from hose) are silicone, PVC, EPDM+ polypropylene (as in
santoprene), polyurethane, neoprene and a number of proprietary materials. Extruded fluoroplymer tubes
such as FKM (viton, fluorel, etc.) have good compatibility with acids, hydrocarbons, and petroleum fuels.
However, the material has poor fatigue resistance to get meaningful tube life that can be practical.
There are a couple of newer pump tubing developments that offer a broad chemical compatibility - a lined
tubing approach and the use of fluoroelastomer approach.
There are many online sites for checking the chemical compatibility of the tubing material with the pumped
fluid. The manufacturers of these tubing may also have compatibility charts specific to their tubing.
While these charts cover a list of commonly encountered fluids, they may not have all the fluids. If there is a
fluid whose compatibility is not listed anywhere, then a common test of compatibility is the immersion testing.
A 1 to 2 inch sample of the tubing is immersed in the fluid to be pumped for anywhere from 24 to 48 hours
and the amount of weight change from before and after the immersion is measured. If the weight change is
greater than 10% of the initial weight, then that tube is not compatible with the fluid, and should not be used
in that application. This test is still a one way test, in the sense that there is still a remote chance that the
tubing that passes this test can still be incompatible for the application since the combination of borderline
compatibility and mechanical flexing can push the tube over the edge, resulting in premature tube failure.
NOTE. The natural starting point for selection of hose materials is chemical compatibility. This happens
particularly with pump selection software linked to chemical compatibility charts. This can lead to serious
errors as mechanical life is just as important If not more important than chemical compatibility. The correct
selection philosophy is to consider both chemical and mechanical implications.
There have been many examples of where natural rubber, because if its superior mechanical life, has
provided the best hose life notwithstanding chemical compatibility charts giving natural rubber a “poor” rating.
The best selection process is to start with natural rubber and then eliminate it if deemed to be totally unfit
chemically. Natural rubber is also the cheapest material so it is often worth trialling and then eliminating it.
Liquid Temperature
The implications of temperature must be considered. The normal rated temperature for hose pumps is 40°C.
Normal maximum temperature is 80°C subject to adjustments to either maximum pump speed or maximum
discharge pressure. Manufacturer’s selection charts would normally show derating requirements at higher
temperatures. If not, contact the manufacturer.
Discharge Pressure
As discharge pressure increases, additional shimming is required to maintain pump efficiency. This
decreases hose life.
As discharge pressure increases, the maximum allowable pump speed and therefore flow rate decreases to
compensate for the reduction in hose life due to additional “squeeze”.
Care should be taken when moving pumps between applications to ensure that those design parameters are
not exceeded.
Various rupture detection devices are available from manufacturers and these include:
• float type magnetic reed switch which detects a change in liquid level in the hose housing
• conductivity probes however these are limited to use with products that are conductive
• pressure transmitters fitted to the hose housing which detect a change in liquid level.
The best method of mitigating these losses is by preventative maintenance. A hose replacement regime
needs to be established so that hoses are changed prior to rupture occurring. This can be instigated after
hose life is determined. This does not mean that rupture detection systems can be forgotten.
Over Shimming
• This failure occurs in the cheek of the hose as the hose starts to re-vulcanise
• Rubber is built up under the shoe, leading to severe internal friction and a heat build up, resulting in
re-vulcanisation
• The failure will occur over the rotor shoe’s full contact path
Adapter (or Support Head): Connects and aligns the power (bearing) end of an ANSI pump to the wet
end.
A.N.S.I. B73.1 Standard: American National Standards Institute. A set of specifications (envelope
dimensions) for centrifugal pumps.
Affinity laws: They are used to determine changes in capacity, head and power due to changes in shaft
speed. The same laws can be used to estimate changes in capacity, head and power due to changes in
impeller diameter.
Air ingestion: Air is coming into the stuffing box because of a negative suction pressure.
Alignment: The centreline of the pump is perfectly aligned with the centreline of the driver (usually an
electric motor).
Ambient temperature or pressure: The environmental temperature or pressure in the area where the
equipment is located.
API Standards: Standards produced by the American Petroleum Institute for various pieces of
equipment applied in the oil and gas industries. Various API codes are applicable for pumps dependent
on the type of pump. Examples are API610, API685 etc.
Axial Thrust: The resultant of all axial forces (i.e. in direction of the pump shaft) acting in the pump rotor.
Axial Thrust Balancing: Methods by which the axial thrust is balanced to minimize bearing loads.
B
Back to Back Impellers: In multistage pumps, some impellers are located on the shaft facing opposite
directions. This is one method of achieving axial thrust balancing.
Back plate: Used in some centrifugal pumps to position the stuffing box and provide an impeller wear
surface.
Back Vane: A radial narrow vane located on the back shroud of an impeller and is designed to balance
axial thrust.
Back pull out pump: A design that allows the wet end of the pump to be left on the piping yet allows the
power end and adapter to be removed. A.N.S.I. pumps are designed this way.
Back to back double seal: Two mechanical seals located in a pump with the rotating seal faces in
opposite directions.
Balanced seal: A design in which the seal face closing area is reduced to lower the closing force, and
reduce the heat generation between the faces.
Ball bearing: Consists of an inner race, an outer race, and a series of balls between them. Often called
a precision or anti friction bearing.
Barrier fluid: The high-pressure fluid that is circulated between the two seals in a double mechanical
seal. The fluid should enter the bottom and leave the top to prevent air pockets.
Base plate: The pump and motor are mounted on this unit.
Bearing: Supports the rotating shaft and allows it to turn with a minimum amount of friction. Could be
either sleeve or anti-friction type.
Bellows: A component of a mechanical seal that can be manufactured from metal or non-metallic materials
to eliminate flexing, rolling or sliding.
Belt Drive: A combination of belts and pulleys that transfers torque from the driver to the pump.
B.E.P (Best Efficiency Point): The best efficiency point on a pump curve for a specific impeller
diameter. It is the point where the power coming out of the pump (water power) is the closest to the
power used by the pump (absorbed power) from the driver. This is also the point where there is no radial
deflection of the shaft cause by unequal hydraulic forces acting on the impeller.
Buffer fluid: The low pressure fluid that is circulated between the two mechanical seal in a tandem seal.
Buna N: Sometimes called Nitrite. A common elastomer used in the sealing of oil or water.
Bushing: A close fitting support device used to restrict flow between two liquids, thermally isolate a hot
liquid, support the rotating shaft or break down pressure.
Bypass line: Used to either re-circulate fluid from the pump discharge to the stuffing box, the stuffing box
to the pump suction, or the pump discharge to a lower pressure point in the system. This can also be
used as a manual means of flow control.
C
Canned pump: A seal/less pump with the shaft, bearings and rotor contained in a can to prevent product
leakage. These are generally limited to pumping clean lubricating liquids.
Cartridge seal: A self-contained assembly containing the seal, gland, sleeve, and both stationary and
rotating seal faces. Usually needs no installation measurement. Must be used in a pump with impeller
clearance adjustments are made.
Cavitation: Cavities or bubbles form in the fluid low-pressure area and collapse in a higher-pressure
area of the pump, causing noise, damage and a loss of capacity.
Centreline design: The pump is mounted to the base plate by feet attached to the sides of the volute
instead of the bottom. These are used in higher temperature pumping applications e.g. API applications
and allows thermal expansion to occur in without the necessity of re alignment.
Centistoke: One unit for the kinematic viscosity of a liquid. Dynamic viscosity divided by the liquid
specific gravity at the same temperature gives kinematic viscosity.
Centrifugal pump: A pump that imparts energy to a liquid with centrifugal force.
Change of state: This defines a change in phase for any material e.g. liquid to vapour, liquid to solid,
solid to vapour.
Concentricity: When the parts share the same centerline they are concentric to each other.
Condensate: This defines the change on phase of a vapour to liquid e.g. steam to water.
Constant Level Oiler: This is used to maintain the oil level in a bearing housing to the correct level as oil
is used.
Cooling jacket: Cooling jackets can be located on the pump casings and /or the stuffing box of the pump
to control the temperature of the fluid. Cooling jackets can be part of the component casting or separate
bolt on items.
Corrosion: This is a chemical or electrochemical reaction on material surfaces that changes the profile
and/or composition of the surface.
Corrosion resistant: An arbitrary term that indicates a corrosion rate of less than 0.05 mm per year.
Coupling: This is used to connect the pump to the driver. It transmits torque between the driver and
pump. It can be a flexible coupling that allows some axial and radial misalignment or rigid which does not
allow for any misalignment. Allowable misalignment is determined by manufacturers of the coupling.
Critical speed: Any object made of an elastic material has a natural period of vibration. When a pump
rotor or shaft rotates at any speed corresponding to its natural frequency, minor unbalances will be
magnified. These speeds are called the critical speeds.
Cryogenic Pumping: This generally refers to pumping liquid gases at very cold temperatures.
Cutwater: A part of the pump casing that directs the pumped liquid to the pump discharge.
D
D.I.N. standard: This is a German standard that defines various industrial products.
Density: Measured in kg/m3 or lb/in 2. This is the ratio of the mass of a liquid in a given volume to the
magnitude of that volume. Refer to Specific Gravity.
Design Duty Point: This is generally defined for a pump as a capacity at a head or pressure of the liquid
being pumped, ideally the design duty point on a centrifugal pump as at BEP.
Dilatant liquid: A liquid whose viscosity increases with increased shear rate e.g. agitation.
Differential Pressure: The difference between the outlet pressure and the inlet pressure. Differential
pressure is sometimes called the Pump Total Differential Pressure.
Discharge Head: The outlet pressure of the pump converted to head of liquid.
Double seal: An outdated term describing two seals in a pump. The latest terminology is "dual seals",
back-to-back double seals or tandem seal. In the past the term was used to describe a higher-pressure
barrier fluid between dual seals.
Double suction pump: A pump with an impeller where liquid enters the impeller on both sides. The rotor
is generally (but not always) suspended between two bearings. These pumps are generally of higher
capacities.
Dry running: Operating a pump without liquid in the pump and therefore in the seal area.
Dual Seal: Two seals running in various configurations: back to back, tandem, face to face, or
concentric.
Ductility: The property of a metal that allows a great deal of mechanical deformation without cracking.
E
Efficiency: In centrifugal pumps, this is the useful power in the liquid to the expended power. That is
power out of the pump divided by power into the pump.
Elastomer: A rubber-like material that, when compressed and then released will basically return to its
original shape in less than five seconds.
Electrolysis: A process involving chemical change caused by the passage of an electric current through
a liquid.
E.P.D.M or E.P.R: Ethylene propylene rubber. This is a common elastomer used in the sealing of water
based and higher pH materials. Cannot be used in petroleum products.
Erosion: This is wear caused by mechanical action of the liquid on the surface of the materials. It is
obviously more prevalent if the liquid contains solids.
Eye of the impeller: The center of the impeller where the fluid enters.
F
Face combination: The materials chosen for the two lapped seal faces. An example is carbon graphite
running on silicon carbide.
Face-to-face seals: Two seals running against a common seal face. The barrier fluid pressure is always
lower than stuffing box pressure.
Face lubrication: The fluid or vapor that exists between lapped mechanical seal faces.
Face pressure: The sum of all the loads on the seal face including the spring load, hydraulic load and
shaft axial thrust, divided by the area of the seal face. This face load is reduced by friction between the
sliding elastomer and the shaft or sleeve.
Flooded suction: This refers to a situation where the liquid level on the suction side of a pump is higher
than the pump centerline and the liquid flows to the pump by gravity.
Fluorocarbon: This is a genetic term for an elastomer of which Viton (a Dupont product) is a typical
example. The material has good compatibility with hydrocarbons, has high temperature capability but
poor mechanical life.
Flush: Putting an outside liquid into the stuffing box of the pump at a pressure higher than stuffing box
pressure. All of this liquid mixes with and dilutes the pumped fluid.
Foot Valve: A type of check valve with a built-in strainer. Used at the point of the liquid intake to retain
liquid in the system, preventing the loss of prime when the liquid level is below the pump centreline.
Friction Head: This is the head loss due to friction as liquid flows in pipes and fittings.
Gland: The part that holds one half of the mechanical seal and attaches to the stuffing box.
H
Hard face: A seal face either rotating or stationary. The most common materials are silicone carbide,
ceramic, tungsten carbide, Stellite, Ni-resist. The hard face must be the wider seal face.
Hastelloy "C” (also known as Alloy C): A nickel-rich, corrosion-resistant and very hard alloy.
Head: The equivalent height of the liquid that will produce a particular pressure. Can be calculated from
H (metres)= pressure in kPa/(9.8 x specific gravity).
Horizontally Split Pump: This is a pump where the casing is split into two sections in the axial plane.
This means there is a top and bottom-half casing. Connections are normally on the bottom-half casing
to allow removal of the top casing for pump inspection without needing to disconnect pipework.
Hydraulic balance: A method of reducing mechanical seal face loading by reducing the seal face
closing area.
I
I.D.: Inside diameter.
I.S.O.: International Standards Organization. Sets pump and seal standards for the metric community.
Impeller: The rotating component of a centrifugal pump that imparts energy to the fluid being pumped.
Available in open, semi-open and closed design.
Impeller eye: The centre of the impeller or the point where fluid enters the impeller.
Impeller shroud: The plates located on one or both sides of the impeller vanes. Prevents solids from
penetrating behind the vanes.
Impeller vane: Located between the eye and the discharge side of the impeller. Directs the flow of the
liquid to the outside diameter of the impeller.
Inducer: A small axial flow vane that attaches to the impeller of a centrifugal pump that reduces the
N.P.S.H. required by a pump. This improvement occurs across a narrow capacity range and the impact
can be detrimental outside of this range.
Induction motor: The most common type used in industry. Has a slippage of 2 to 5 percent compared
to synchronous motors.
Inline pump: Mounted in the piping generally between two flanges. No base plate or alignment required.
J
Jacket: Usually refers to the heating/cooling jacket surrounding the stuffing box on some pumps.
K
Kalrez®: An "elastomer-like" material manufactured by E.I. Dupont that is used to seal most solvents and
other aggressive fluids. It is available in several different grades.
Kilowatt: One thousand watts. The normal unit for power in the metric system.
L
Labyrinth seal: A non-contacting seal utilising a tortured path for the escape of the fluid. Utilises a series
of pressure drops to reduce the leakage.
Lantern Ring: A device used to supply lubricant to packing. Usually located in the middle of the packing
ring set.
Line bearings: These position the rotor or shaft radially and are normally of the sleeve type. Generally
used in vertical pumps.
Liquefied Natural Gas LNG: This is liquefied gas from natural sources
Liquefied Petroleum Gas LPG: This is liquefied petroleum gas which is a by-product of the refining of
crude petroleum oil.
M
Magnetic drive: A type of seal less pump that utilises permanent magnet technology to provide the
rotation of the impeller.
Mating ring: Another name for the hard face in a mechanical seal. It can be either rotating or stationary.
Mechanical seal: A positive sealing device used to seal all fluids (liquids and gases). The primary seal
is a set of lapped seal faces that are installed perpendicular to the shaft.
Metal bellows: Used in mechanical seal designs to eliminate the need for a dynamic elastomer and
springs.
Metal fatigue: A breakage of the metal caused by the bending and flexing of a metal part beyond its
endurance limit.
Minimum flow: The minimum capacity of a pump to prevent thermal and/or mechanical damage.
Moment of inertia: This represents a magnitude of the inertia in respect of the rotation around the axis
of the pump and drive rotor.
Multistage Pump: This defines a pump that has more than one impeller on the shaft.
N
Negative pressure: A pressure below atmospheric pressure.
Newtonian Fluid: A Newtonian liquid is one whose viscosity does not change with increasing shear rate
e.g. when agitated.
Non Overloading Power: This refers to the maximum power absorbed by a pump with a specific
impeller diameter and liquid. Motors are generally sized at the next size above this power.
N.P.S.H.A: The net positive suction head available to prevent cavitation of the pump. It refers to the
suction side of a pump installation and is defined as the head acting on top of the liquid + static head -
vapor pressure head - friction head loss in the suction piping.
N.P.S.H.R.: Net positive suction head required to prevent cavitation of a pump and is dependent on
impeller and pump design. The pump manufacturer determines the NPSHR by testing. In all cases, it is
imperative that NPSHA >NPSHR to prevent cavitation
O
O.D.: Outside diameter.
Orifice Plate: A plate with a hole smaller than the pipe diameter in which it is located
Over hung impeller: Not supported with bearings on either side of the impeller.
P
Packing: The soft rings that a mechanical seal replaces to stop leakage. Packing must have a small
leak because it works on the theory of a series of pressure drops to reduce the stuffing box pressure to
the point where the leakage is acceptable. Generally, a minimum of five rings of packing is required to
do this.
Parallel operation: This refers to two or more pumps that are discharging to a common header. It is
important that the impeller speed and outside diameters are the same or one of the pumps may cause
other pumps to operate at shut off.
pH: A measure of the acidity or the alkalinity of a fluid. The scale ranges from 0 (acid) to 14 (alkali) with
7 considered neutral.
Pipe strain: The strain on the pump volute caused by the piping. It will cause excessive mechanical seal
movement and can cause contact between rotating and stationary pump and seal components. It can
also cause serious misalignment with resultant damage to bearings and couplings.
Pitting: Surface voids caused by corrosion, erosion or cavitation. It is possible for the three to occur at
the same time.
Positive Displacement Pumps: This is a collective definition of all pumps that operate according to the
positive displacement principle. That is, the liquid being pumped is displaced by a body which
periodically increases and decreases the working volume.
Power end: The end of the pump that attaches to the power source and is not wetted by the liquid. The
bearings are in this part.
Priming: This refers to the filling of a pump with liquid prior to operation.
Q
Quench: The introduction of a fluid outside the seal to cool the product, dilute any leakage across the
seal faces or isolate seal faces from atmosphere.
R
Radial Bearing: This bearing handles most of the radial loads put on the impeller. In an end-suction
centrifugal pump it is the bearing located closest to the stuffing box.
Radial Thrust: This is the thrust produced in the radial direction i.e. at 90 degrees to the centerline of the
shaft, by forces acting on the impeller when operating at points other than BEP.
Radially Split Casing: A pump casing with the casing joint at 90 degrees to the shaft axis.
Rated Operation: This is the basis of selection of both pump and driver. When rated operation is
specified, it generally exceeds the requirement of the design operation.
Ring Section Pumps: These are multistage pumps with several identical stage casings arranged in
tandem behind each other. The stage casings are radially split.
Run out: Twice the distance that the centre of the shaft is displaced from the axis of rotation.
S
Saturation Pressure: Same as vapour pressure.
Seal faces: The lapped faces that provide the primary seal in a mechanical seal.
Self Priming Pump: These pumps are capable of pumping liquids on a suction lift by being able to
remove air from the suction line of its own accord. Most self-priming pumps need to be filled with liquid
(primed) before the initial start, however their design means that subsequent starts do not require
priming. Pumps that can self-prime from dry are limited to diaphragm and peristaltic type pumps only
(both are positive displacement pumps).
Series operation: Two or more pumps connected with the discharge of the first pump discharging to the
suction of the other etc. Rarely used these days.
Shaft packing: The soft packing located in the stuffing box to provide a shaft seal for pumps.
Shut off head: This is the maximum head that the pump can generate with a given impeller outside
diameter. It is normally at zero capacity.
Sleeve bearing: A non-precision or anti-friction bearing. It is usually manufactured from carbon, teflon,
brass, white metal, other synthetic bearing materials.
Solubility: This defines the ability of a liquid to dissolve with another liquid. For example, ethanol will
fully dissolve in water whereas oil will float on water.
Specific Gravity (SG): This is the ratio of the mass of a liquid for a known volume to the weight of water
for the same volume. The reference is water at 4 degrees C with an SG of 1.0. If the liquid you are
questioning will float on water the specific gravity is less than one. If it sinks, it is higher than one. Note
that this is based on the liquid not being soluble in water.
Specific Speed: Specific speed of a pump is determined by the geometry of a pump impeller. The
higher the specific speed the less N.P.S.H. required.
Stainless steel: Alloy steels containing a high percentage of chromium and/or nickel.
Static head: The height of a liquid above a reference point e.g. pump centerline.
Stationary face: The seal face that does not rotate with the shaft.
Stuffing box: The portion of the pump that held the packing and now holds the mechanical seal.
Stuffing box pressure: The pressure in the stuffing box and generally between suction and discharge
pressure but closer to suction pressure.
Submersible pump: A pump/motor pumpset that operates only when totally submersed in the fluid
which is being pumped.
Suction lift: Pumping application where the liquid level on the suction side of the pump is below the
pump centerline.
System resistance Curve: A graphed representation of how total dynamic head varies with capacity. A
pump will operate where the system resistance curve intersects the pump performance curve.
System head: The head caused by friction in the piping valves and fittings.
Thermal conductivity: A measure of the material's ability to conduct heat. This is a very important
factor in the selection of mechanical seal faces.
Thixotropic fluid: The viscosity of the fluid decreases with agitation. Non-drip paint is an example of
such a fluid.
Throttling: This means closing of a valve on the discharge side of a pump to increase friction loss. This
steepens the system resistance curve with a resultant decrease in flowrate. Opening the valve results in
an increase in flowrate.
Thrust bearing: This locates the rotor or shaft axially and is designed to handle any excess axial thrust
load. In an end-suction pump, it is normally located close to the coupling.
Thrust: In a centrifugal pump it refers to the axial movement of the shaft. The thrust can be towards the
wet or power end of the pump and at start up it thrusts in both directions.
Total Discharge Head: This is equal to the pressure at the pump discharge connection converted to
head of liquid.
Total Dynamic Head: Total dynamic head is equal to total discharge head minus total suction head
Total Suction Head: This is equal to the pressure at the pump suction connection converted to head of
liquid.
Tungsten carbide: A common hard face seal material available in several grades depending upon
hardness and corrosion resistance. Cobalt and nickel are the two most common types.
Turbulence: This refers to disturbance of fluid as it enters the suction connection and /or the impeller.
This can cause cavitation problems in a centrifugal pump. This is often caused by an insufficient length
of straight pipe before the pump suction inlet.
U
Unbalanced seal: A mechanical seal not designed to balance the closing force between seals. Refer to
Balanced Seal.
V
Variable speed motor: This is used to control flow in a system by varying the frequency of the motor. A
better system than throttling as it reduces power consumption significantly.
Vapor pressure: Below this pressure, the liquid at this temperature will vaporise.
Vaporisation: The fluid passes from a liquid to a gaseous state. If this happens at the seal faces the
seal faces will be blown open.
Velocity: A measurement of the speed of the liquid in the system. This is measured in metres per
second.
Velocity head: This is part of the total head calculation. This is calculated from the formula H = v2/2g.
Vent: This removes air or gas from the system. It is important to vent the stuffing box in vertical pumps to
prevent the seal faces from running dry.
Viton®: An E.I. Dupont Dow manufactured fluorocarbon elastomer widely used in the sealing industry.
Refer to Fluorocarbon.
Volute casing: This derives its name from a spiral-shaped casing surrounding the pump impeller. It
converts velocity energy to pressure energy.
Vortex Pump: A type of pump used to handle liquids with entrained solids, particularly stringy solids. The
impeller is recessed into the volute. A very low efficiency design, but practical in many applications.
Vortexing liquid: Creating a "whirlpool affect" that can draw air into the suction of the pump. Vortecies
can form both from the surface of the liquid and in vertical pumps, from the floor of a pit or channel in
which they are located.
W
Water hammer: This occurs in a closed piping system as a result of the pressure being rapidly increased
when the liquid velocity is suddenly changed. This damaging effect is usually the result of sudden
starting, stopping, change in pump speed, or the sudden opening or closing of a valve. Water hammer
can usually be controlled by regulating the valve closure time, surge chambers, relief valves or other
means.
Wear ring: This is used with closed impeller pumps to seal leakage from the high-pressure side of the
pump to the low-pressure side. This may need to be replaced as it wears when the recommended
clearance is doubled or when reduction in pump performance can no longer be tolerated.
Welded metal bellows: A seal design used to eliminate the use of elastomers. Excellent for cryogenic
and hot applications. Not as effective for hot petroleum applications because of "coking" problems.
Wet end: The part of the pump that gets wet from the pumping fluid. Includes the volute, stuffing box,
impeller wear rings, and shaft or sleeve.
80
Efficiency - Head
Metres of Head
Another consideration is NPSHR. In a centrifugal pump the NPSHR varies as a function of flow and flow is
determined by the system resistance (total dynamic head). If the total head varies for any reason, the flow
will change and NPSHR will also change as a consequence. In a PD pump NPSHR varies as a function of
flow which is determined by speed. At a fixed speed the flow is constant irrespective of pressure and there-
fore NPSHR is constant.
……..2
When comparing the two types of pumps, it is important to understand that a centrifugal pump has an
ideal operating flowrate i.e. the flowrate at the best efficiency point. At flowrates other than best effi-
ciency flowrate, other factors need to be considered. Radial loads (the load that applies a bending
moment to the shaft) increases at all flowrates other than that at best efficiency.
With a PD pump you can operate the pump on any point of the curve. In fact the volumetric efficiency
actually improves at the high speed part of the curve. This is because the volumetric efficiency is af-
fected by slip, which is essentially constant regardless of speed. At low speed, the percentage of slip
in relation to volumetric displacement is higher than at high speed.
The data presented in these charts is the actual data for a specific application. The centrifugal was
picked at its Best Efficiency Point (BEP) and the PD pump (Internal Gear) selected to match the flow,
viscosity, and pressure. Different applications will have different curves and efficiencies. These
curves are presented as an example of the type of performance behaviour between the two different
principles.
The most obvious reason to use a PD pump is when you have a high viscosity application. It is com-
mon knowledge that a centrifugal becomes very inefficient at even modest viscosity. The acceptable
viscosity ranges for centrifugal pumps tends to be dependent on pump size. Published data is also
variable in this area. We believe that acceptable viscosity limits for centrifugal pumps are as follows.
However, there are other reasons to select a PD pump over a centrifugal other than high viscosity.
These can be summarised as follows:
A simple rule of thumb is that a PD pump should be selected where the smallest available centrifugal
pump needs to operate at a flow less than 50% of best efficiency flow.
Additionally, PD pumps tend to produce higher heads or pressures at a more economical price.
PD pumps may be a more appropriate selection at low flow, high head applications.
A PD pump would be used on applications that have variable pressure conditions. The flowrate from
centrifugal pumps will vary up and down the performance curve which can cause process problems.
A PD pump will give near constant flow that makes it possible to match the flow to the process re-
quirements. The desire to have constant flow is a reason that a PD pump is normally applied for me-
tering applications.
Generally speaking centrifugal pumps tend to shear liquids more as speed is increased and the cen-
trifugal is a high speed pump. This makes the PD pump better able to handle shear sensitive liquids.
Shear rates in PD pumps vary by design but they are generally low shear devices, especially at low
speeds. It is important to contact the manufacturer for specific information on shear rates and appli-
cation recommendations.
By their nature, many PD pumps are self priming i.e. will operate on suction lift without the necessity
of a foot valve. This capability can vary from pump to pump so manufacturers’ recommendations
must be followed.
Pump Standards
There are many pump standards used in the pump industry today. These standards may be national,
international, industry specific, company specific or project specific. This paper attempts to cover the
more commonly used pump specifications within Australia. Undoubtedly, there are other pump
specifications being utilised and the writer would be pleased to hear about them.
Other associated standards such as those defining test standards are also available however we have
not attempted to cover these in this paper.
• Definitions: This area details the organisations that either prepare or certify pump specifications.
There are many national specifications e.g. Japanese, British, Australian etc, however the vast majority
are based on ISO standards or reference other standards such as API. Project specific specifications
would generally reference other standards.
1. Many pumps available on the market today do not comply with any documented standard. This
does not mean that these pumps are not of excellent quality or fit for purpose.
2. If the intention of a buyer is to define compliance with a specific standard, ensure that it is
relevant to the application. For example, specifying API compliance for a building services
application does not make any sense.
3. Any pump specifier should take care that defining compliance with a particular specification does
not disqualify other suitable pumps. For example, specifying ANSI B73.1 may disqualify offers of
very suitable pumps complying with ISO 5199.
4. There have been a number of instances where compliance with more than one pump
specification is detailed. This may appear to be a safe process however this often causes
confusion for a pump supplier as there may be contradictions between the two standards.
DEFINITIONS
Standards Australia – AS
Standards Australia is the nation’s peak non-government Standards organisation. It is charged by the
Commonwealth Government to meet Australia’s need for contemporary, internationally aligned
Standards and related services.
It leads and promotes a respected and unbiased Standards development process ensuring all competing
interests are heard, their points of view considered and consensus reached.
ISO classifies pumps as Class I, II and III with Class I having the most stringent requirements. The
selection of class is determined by the application and the intention is that it is agreed between purchaser
and supplier. It is impossible to standardise the class of technical requirements, however, the criteria for
class determination may include;
- reliability
- required operating life
- operating conditions
- environmental conditions
- local ambient conditions
It is possible that pumps built in accordance with Classes I, II and III may work beside one another in the
same plant.
Hydraulics Institute- HI
The Hydraulic Institute is a non-profit industry (trade) association established in 1917. HI and its
members are dedicated to excellence in the engineering, manufacture, and application of pumping
equipment.
The Institute plays a leading role in the development of pump standards in North America and worldwide.
HI standards are developed within guidelines established by the American National Standards Institute
(ANSI). HI members work through a number of technical committees to develop draft standards. The
Institute involves pump users and other interested parties to reach consensus on published standards. HI
standards are developed to define pump products, installation, operation, performance, testing, and
pump life and quality.
FM Approvals certifies industrial and commercial products and services for companies worldwide. When
a product or service meets the standards of FM Approvals, it is issued the "FM APPROVED" mark to
signify it will perform as expected and support property loss prevention.
It was founded in 1917 as the Normenausschuss der deutschen Industrie (NADI, "Standardisation
Committee of German Industry"). In 1926, the NADI was renamed Deutscher Normenausschuss (DNA,
"German Standardisation Committee"), to reflect the fact that the organisation now dealt with
standardisation issues in many fields; viz., not just for industrial products. Since 1975, the DNA is known
as 'DIN' and is recognised by the German government as the official national-standards body,
representing German interests at the international and European levels.
FIRE
NFPA20 Standard for the Installation of Stationary Fire Pumps for Fire Protection.
This standard deals with the selection and installation of pumps supplying liquid for private fire protection.
The scope of this document shall include liquid supplies; suction, discharge, and auxiliary equipment;
power supplies, including power supply arrangements; electric drive and control; diesel engine drive and
control; steam turbine drive and control; and acceptance tests and operation.
FM Certification
FM Approvals not only evaluates sprinklers and sprinkler system components for compliance with
existing standards, but also work closely with manufacturers to evaluate new products and develop
appropriate standards. FM provides certification for fire pump sets to be utilized in sprinkler systems. It
approves both fire pump package components e.g. pump, driver, panel and also the package assembler.
ISO 13709 Centrifugal Pumps for Petroleum, Petrochemical and Natural Gas Industries
This International Standard specifies requirements for centrifugal pumps, including pumps running in
reverse as hydraulic power recovery turbines, for use in petroleum, petrochemical and gas industry
process services. This International Standard is applicable to overhung pumps, between-bearings
pumps and vertically suspended pumps. It is not applicable to seal/less pumps.
PROCESS
This International Standard includes design features concerned with installation, maintenance and safety
for these pumps including baseplate, couplings and auxiliary piping, but it does not specify any
requirements for the driver other than those related to its rated power output.
Pumps will normally conform to recognised standard specifications (e.g. ISO 5199, explosion protection,
electromagnetic compatibility), except where special requirements are specified herein. This
International Standard includes design features concerned with installation, maintenance and operational
safety of the pumps, and defines those items to be agreed upon between the purchaser and
manufacturer/supplier.
ISO 16330:2003 Reciprocating positive displacement pumps and pump units -- Technical
requirements
This International Standard specifies the technical requirements, other than safety and testing, for
reciprocating positive displacement pumps and pump units. It applies to pumps which utilise
reciprocating motion derived from crankshafts and camshafts and also direct-acting fluid driven pumps.
It does not apply to reciprocating positive displacement pumps, nor pumping water, where the whole
pump is lubricated with the liquid being pumped.
ANSI/ASME B73.1 Specification for Horizontal End Suction Centrifugal Pumps for Chemical
Process
This standard covers centrifugal pumps of horizontal, end suction single stage, centre-line discharge
design. This Standard includes dimensional interchangeability requirements and certain design features
to facilitate installation and maintenance. It is the intent of this Standard that pumps of the same
standard dimension designation from all sources of supply shall be interchangeable with respect to
mounting dimensions, size and location of suction and discharge nozzles, input shafts, baseplates, and
foundation bolt holes.
ANSI/ASME B73.2 Specifications for Vertical In-Line Centrifugal Pumps for Chemical Process
This Standard covers motor-driven centrifugal pumps of vertical shaft, single stage design with suction
and discharge nozzles in line. It includes dimensional interchangeability requirements and certain design
features to facilitate installation and maintenance. It is the intent of this Standard that pumps of the same
standard dimension designation, from all sources of supply, shall be interchangeable with respect to
mounting dimensions and size and location of suction and discharge.
ASME B73.3 Specification for Seal-less Horizontal End Suction Metallic Centrifugal Pumps for
Chemical Process
This Standard covers seal-less centrifugal pumps of horizontal end suction single stage and centre-line
discharge design. This Standard includes dimensional interchangeability requirements and certain
design features to facilitate installation and maintenance. It is the intent of this Standard that pumps of
the same standard dimensional designation from all sources of supply shall be interchangeable with
respect to mounting dimensions, size, and location of suction and discharge nozzles, input shafts,
baseplates, and foundation bolt holes
ASME B73.5M Specification for Thermoplastic and Thermoset Polymer Material Horizontal End
Suction Centrifugal Pumps for Chemical Process (not often used in Australia)
This Standard covers centrifugal pumps of horizontal, end suction single stage, centreline discharge
design, which components are made of thermoplastic and thermo-set polymer materials either reinforced
or non-reinforced. It includes dimensional interchangeability requirements and certain design features to
facilitate installation and maintenance. It is the intent of this Standard that pumps of the same standard
dimension designation from all sources of supply shall be interchangeable with respect to mounting
dimensions, size, and location of suction and discharge nozzles, input shafts, baseplates, and foundation
bolt holes. This Standard does not include lined or non polymer components.
ANSI/HI 5.1-5.6 Seal-less Centrifugal Pump Standards (not often used in Australia)
This standard is for seal/less centrifugal pumps that are driven by canned motors or magnetic couplings.
It includes types and nomenclature; definitions; design and application; installation,
operation/maintenance and testing. The testing section includes:
• Hermetic Integrity Test
• Mechanical Integrity Test
• Winding Integrity Test
ANSI/HI 6.1-6.5 Reciprocating Power Pump Standard (not often used in Australia)
This Standard applies to industrial/commercial reciprocating power pumps. It includes:
• Types and Nomenclature
• Definitions
• Design and Application
- Basic Speeds
- Pump Torque Characteristics
- Calculating Volumetric Efficiency
• Installation
- Protection of Pump Against Seepage or Flood
- Drive Alignment after Piping Installation
• Operation and Maintenance
ANSI/HI 7.1-7.5 Controlled Volume Metering Pumps (not often used in Australia)
This Standard applies to Controlled-Volume Metering Pumps, which are reciprocating power pumps used
to accurately displace a predetermined volume of liquid within a specified time. It contains sections on:
• General Description
• Types and Nomenclature
• Definitions
• Application and sizing
- Typical performance curves
- Materials of construction
- Control methods
• Installation
- Storage recommendations
• Operation and trouble solving
ANSI/HI 10.1-10.5 Air Operated Pumps (to our knowledge the only standard available for this
type of pump and to date not seen in Australia)
This standard is for air-operated pumps and includes those positive displacement reciprocating pumps
used for general fluid transfer, which are driven by means of a compressed gas (usually air) from an
outside source. The pump may be designed with a single diaphragm or double diaphragms connected to
a reciprocating shaft in which one side of the diaphragm is in contact with the liquid being pumped and
the other side is in contact with the compressed gas. The standard includes the following sections:
• Types, configurations, and nomenclature
• Definitions
• Design and Application
• Installation, Operation, and Maintenance
GENERAL PURPOSE
ISO 2858:1975 End-suction centrifugal pumps (rating 16 bar) -- Designation, Nominal duty point
and Dimensions
The standard covers the ISO requirements for general purpose pumps primarily for use with water. The
pump designation consists of three numbers detailing suction and discharge connection sizes and
nominal impeller diameter in mm e.g. 125 x 100-250. The table covers flange sizes from 50 mm up to
and including 200 mm. This is primarily a dimensional standard.
DIN 24255 End-suction centrifugal pumps (normal rating 10 bar although some may claim 16 bar)
This specification details single stage, end suction overhung impeller centrifugal pumps for general
purpose applications design to be used primarily with water up to a maximum of 120 C. This is primarily
a dimensional standard. The majority of DIN 24255 pumps supplied in Australia are manufactured in
China.
Impeller Trimming
Performance curves for centrifugal pumps normally detail the performance for a number of impeller
diameters between a maximum and minimum allowable diameter. Impeller trimming means the
reduction of the impeller diameter from maximum usually to adjust the pump performance to a required
duty point.
Several things can happen when the impeller vane diameter is reduced. The diagram below will be used
throughout this paper.
The change in pump performance with changes in impeller diameter can be predicted similarly to that
with speed change utilising the Affinity laws
a) Pump flow rate (Q) varies directly with the diameter (D)
i.e. Q1/Q2 = D1/D2
b) Pump head (H) varies with the square of the diameter (D)
i.e. H1/H2 = (D1/D2)²
c) Power absorbed varies with the cube of the diameter (D)
i.e. P1/P2 = (D1/D2)³
These relationships (not laws) allow adjustment of the H-Q curve but there is a detrimental impact on
efficiency especially for impeller reductions greater than 10% of maximum.
• The affinity laws assume the impeller shrouds are parallel. This is true only in low specific speed
pumps.
• There is increased turbulence at the vane tips as the impeller is trimmed because the shroud to
casing clearance (Gap "A") is increasing. This is sometimes referred to as "slip".
• The liquid exit angle is changed as the impeller is cut back, so the head/capacity curve becomes
steeper.
Note that mixed flow impellers are more affected than low specific speed, radial vane impellers (high
head/low capacity). Simply put, the greater the impeller reduction and the higher the specific speed of
the impeller, the more the pump efficiency will decrease with impeller trimming.
More accurate information can be obtained from a complete performance chart with different impeller
diameters detailed.
In general, impeller diameter reductions greater than 5% to 10% of the maximum will increase the
NPSHR (net positive suction head required) dependent on impeller specific speed.
The interaction of the physical geometry of the pump inlet, inclusive of the casing, impeller, and all
associated wetted parts within the inlet field of flow determines the NPSHR characteristic of a pump. The
value of NPSHR for any centrifugal pump is determined through performance testing. From NPSH test
data, Suction Specific Speed (S) is calculated using the following equation, where Q represents flow at
the best efficiency point of the pump.
S = rpm x √Q
NPSHR¾
In can be seen in the above equation that NPSHR should not change with changes in impeller diameter
as long as flow and RPM remain constant. There is no factor in the S equation that relates to impeller
diameter. Suction specific speed (S) remains constant, for any defined inlet geometry, as long as the
field of flow into the impeller eye is not disrupted by events taking place downstream of the impeller inlet.
When trimming impellers on pumps that are of a low specific speed (Ns < 30 SI, 1500 US), tests have
shown that there is little effect on NPSHR within the allowable impeller cut range. Beyond the allowable
impeller cut range, recirculation between impeller discharge and the impeller inlet start to disrupt the inlet
field of flow, increasing the NPSHR.
For low Ns applications, full diameter NPSH values may be used for estimating NPSHR for cut impeller
performance. For applications with Ns values above 30 (1500 US), a NPSH test is recommended to
determine the NPSHR for any impeller trim.
3. MECHANICAL IMPACTS
i) Excessive shroud to casing clearance (Gap "A") and the resultant recirculation to the low
pressure side of the pump will produce "eddy flows" around the impeller causing low frequency
axial vibrations that can translate to mechanical seal problems. This can be a real concern in
large pumps with powers over 200 kW or pumps pumping heads in excess 200 metres.
ii) For many years pump people have been machining the vane tips to reduce the vane passing
frequency vibrations (Gap "B") while carefully maintaining Gap "A". The pulsating forces acting
on the impeller can be reduced by 80% to 85% by increasing gap "B" from 1% to 6%.
iii) For impeller diameters up to 355 mm, gap "B" should be at least 4% of the impeller diameter to
prevent "Vane passing syndrome cavitation" problems. For impeller diameters above 355 mm,
Gap "B" should be at least 6% of the impeller diameter to prevent this type of cavitation. This
type of cavitation damage is caused when the outside diameter of the impeller passes too close
to the pump cutwater. The velocity of the liquid increases as it flows through this small passage,
lowering the fluid pressure and causing local vaporisation. The bubbles then collapse at the
higher pressure just beyond the cutwater. This is where volute damage occurs. Unless damage
iv) Although both the vanes and shrouds are often cut in end-suction, volute-type centrifugal pumps,
it is not a good idea to do this in double suction designs. With these types of pumps you can
reduce the vane diameters, but the shrouds should remain untouched.
v) Structural strength is a consideration when deciding how much to reduce the vane diameter in
double ended pumps because you could leave too much unsupported shroud. Some
manufacturers recommend an oblique cut that will improve the vane exit flow and add some
strength to the shrouds.
vi) Machining a radius where the trimmed vane meets the shroud is another good idea to add
strength to the assembly. Square corners are never a good idea.
4. NOISE
When writing a pump specification, many practicing engineers limit the impeller diameter to 85% of its
maximum diameter. Such a limitation is actually a misunderstanding of a design concept known as
"quiet pump operation." This misunderstanding may force the selection of a larger pump for the
application. The idea here is not that the impeller diameter should be 85% of the maximum published
diameter, but 85% of cutwater diameter (0.85 cutwater ratio). To fully understand the quiet pump
operation design concept, refer to figure below.
In designing a pump casing, a design engineer first determines the volute scroll (A) necessary to handle
the desired volume of water. This volute scroll terminates at the volute cut water (B) at the base of the
discharge nozzle (C). The volute scroll is drawn around a base circle (D), which is sufficiently large
enough to allow insertion of the impeller. The distance from the shaft centerline to the volute cut water is
called the cutwater radius and twice this distance is the cutwater diameter
Hydraulic noise becomes a factor when the periphery of the impeller passes too close to the cutwater. In
designing a pump, the distance between the impeller and the cutwater is a compromise between the
pump efficiency and pump noise. Typically, cutwater ratios (D/F) of 0.9 and above produce higher noise
and cutwater ratios of 0.8 and below produce significantly lower pump noise. Cutwater ratio of 0.85 is
commonly specified by practicing engineers, thereby realising a minimum reduction in efficiency with a
mean reduction in noise level.
From the above, it may be understood that a specification should more properly read "impeller diameter
not to exceed 85% of the volute cutwater diameter” rather than "impeller diameter shall not exceed 85%
of the maximum impeller diameter capable of being installed in the pump casing."
Specifying the later statement is safer since the impeller diameter would be even smaller than the
desired maximum. In some cases this may force selection a larger pump than necessary.