Professional Documents
Culture Documents
Conference, Houston, Texas, May 47, 1998, of THE AMERICAN SOCIETY OF ME-
CHANICAL ENGINEERS. Manuscript received by the Petroleum Division, March 23, P 0.00045 1.62
1999; revised manuscript received March 3, 2000. Associate Technical Editor: C. C (2)
Sarica. L D 1.2
Journal of Energy Resources Technology Copyright 2000 by ASME JUNE 2000, Vol. 122 71
Downloaded 23 Aug 2011 to 200.174.154.60. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
Table 1 Speculations regarding the origin of API RP 14E erosional velocity equation
where P/ Lpressure drop per unit length psi/ft DPipe di- tests were conducted at a velocity corresponding to a C-factor
ameter in. CAPI C-factorV m m V m mixture velocity between 220 and 260 in a seawater flow loop containing fiberglass
ft/s m mixture density lb/ft3 pipes and pipe bends of CuNi and stainless steel 18. The tests
were concluded without any erosion damage in the fiberglass,
CuNi, or stainless steel.
Application of the API RP 14E Erosional Velocity Single distilled water and two-phase water and nitrogen flow
loop test results on simulated tubular joints 19 showed that,
Equation providing corrosion can be suppressed, a C-factor of 400 can be
Although the source and validity of the API 14E erosional ve- used without any concern for erosion. The results show that there
locity equation is being questioned by many, its use within the oil is no difference between erosion/corrosion rate for a C-factor of
industry is widespread. However, many companies are using 40 and that of 400. The results also show that at a C-factor of 400,
higher values for the C-factor than suggested in the API RP 14E carbon steel showed no signs of erosion when corrosion was sup-
document. Deffenbaugh and Buckingham 9 reported that Mobil pressed by cathodic protection. High corrosion rates were, how-
does not limit flow velocities, and Arco uses a C-factor of 200 for ever, observed when the steel joints were not cathodically pro-
continuous service and C-factor of 250 for intermittent service tected. This high corrosion rate was unexpected because the
when corrosion is controlled and if sand can be avoided. Deffen- oxygen level was very low. However, experimental results 20
baugh and Buckingham also presented data developed by Arco on have confirmed that corrosion rates in a deaerated system can be
velocity effect of inhibited systems with and without solids on high when the pH value is low, which was the situation in this
carbon steel and 316 stainless steel for pipes, elbows, and chokes. case.
The results showed that for a straight pipe section, no erosion/ Since corrosion rates can be influenced by flow velocity,
corrosion was observed for C-factors up to 500. For other com- C-factor values higher or lower than 100 are possible, depending
ponents, no erosion/corrosion was reported for C-factors up to on the operating condition. Even for systems that rely on inhibi-
300, even with sand. tors to suppress corrosion, the use of a C-factor of 150 to 200 as
Heidersbach 5 reported that Phillips does not use API 14E to suggested by API RP 14E 2 can be risky unless the inhibitor is
determine production rates. Erichsen 16 reported that one North evaluated using a flow loop testing at the operating C-factor. In
Sea operator produced from a condensate well at a velocity of 286 many cases, inhibitors that provide good protection under station-
ft/s C-factor of 726 for 1050 days @2.9 yr until a failure oc- ary conditions lose their effectiveness at higher velocities, even at
curred in the AISI 4140 carbon steel tubing at the flow coupling. C-factors lower than 100 21. However, there are inhibitors that
The failure of the coupling was attributed to liquid impingement maintain their effectiveness even at a C-factor of 400 21. There-
caused by the fluids exiting the 2-in. downhole safety valve into fore, extreme care must be taken in selecting inhibitors for sys-
the 3.9-in. tubing. The flow coupling was replaced by L80-13 Cr tems operating at high C-factors.
material and no failure was reported, but the velocity was also
reduced. Erichsen also reported that another North Sea operator
has used a C-factor of 300 as upper limit for Gullfaks subsea
water injectors completed with L80-13 Cr tubing. One should not, Sand Erosion
however, be surprised if corrosion failure occurs in this system at Unlike erosion in sand-free systems where erosion rate is re-
the joints because of the susceptibility of 13 Cr to crevice corro- lated to two parameters, i.e., mixture density and flow velocity,
sion and pitting. erosion due to sand is influenced by several factors including fluid
Results by Camacho 17 showed no erosion damage, in most characteristics flow rate, composition, density, viscosity, sand
experiments, for N-80 steel after repeated impact by liquid slug at characteristics concentration, impact velocity, impact angle,
a velocity of 100 ft/s, which corresponds to a C-factor of 800. number of particles hitting the surface, shape/sharpness, hardness,
When erosion damage was observed, it was attributed to the pres- size distribution, density, component geometry bend, tee, choke,
ence of microscopic solid particles in the liquid. Three-month joint, and material properties hardness, microstructure. There
Downloaded 23 Aug 2011 to 200.174.154.60. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
exists an extensive database that can be used to characterize ero- erosion prediction model. The Tulsa model relies on empirical
sion rate of different materials. These data are generally presented formulas to account for particle tracking, while the DNV model
using the following equation: allows actual calculations, though simplified, of the trajectories of
the sand particles. While all other models predict a single value
E r AV np F (3) that corresponds to the maximum erosion rate, the DNV model
where E r is erosion ratio measured as the ratio between the mass predicts erosion rate distribution along a pipe bend based on cal-
of metal loss and the mass of sand hitting the target material. A culations of impact velocity and angle at all locations.
and n are experimentally determined constants that depend on the The models developed by Salama and Venkatesh 4, DNV
material properties. For ductile materials the value of n is in the 22, and Tulsa University 23 incorporate the standard erosion
range of 2 to 3. For brittle material, n can be as high as 6. V p is the equation, Eq. 3, but the values of the constants are different.
impact velocity of the sand particle on the metal surface. This While the value of n in Salama and Venkateshs model is 2, the
velocity depends on the flow conditions, the geometry of the com- value is 1.73 in the Tulsa Universitys model and 2.6 in the
ponent, and sand properties density and size. F( ) is a function DNVs model. Although each model claims to be verified based
whose value varies between 0 and 1, depending on the impact on experimental data, their predictions for the same case can vary
angle. The function depends on the target material ductile/brittle by two orders of magnitude. Resolution of these differences is
behavior. The value of F( ) is maximum for ductile materials critical because, while one model shows that certain operating
such as steel at impact angles of 20 to 40 deg, and for brittle conditions are acceptable, another model shows them unaccept-
materials such as ceramics at 90 deg. able, which makes it necessary to reduce production rate.
The difficulty in calculating erosion rates is in predicting the
proper values of particle impact angle, , and velocity, V p , whose Proposed Sand Erosion Model
values depend on: fluid density, fluid viscosity, sand particle di-
Extensive effort has been devoted to develop an approach for
ameter, sand density, pipe diameter, and pipe geometry elbow,
establishing velocity limits for sand-laden fluids. Unfortunately,
tee, choke, etc.. Also, the amount of sand hitting the target is
none of these proposals have been adopted as a standard practice
influenced by the flow conditions, sand concentration, and the
because of their complexity. There is a need for a reliable, yet
geometry of the component; therefore, it may not be the same as
simple, equation, as simple as the API RP 14E equation, to estab-
the total amount of sand in the flow. One can account for these
lish erosion rate or erosional critical velocity for fluids containing
factors through the use of computational fluid dynamic CFD
sand. Although the equation proposed by Salama and Venkatesh
analyses and particle tracking simulation models.
4 is simple, it is not very accurate when applied to two-phase
There are four main models that have been developed within
gas-liquid flow systems. When proposing their equation, Eq. 4,
the industry for predicting sand erosion in piping systems. These
they suggested that the fluid properties have an effect on erosion
models are based on work done by Salama and Venkatech 4,
rate, but they selected the constant of the equation based on cali-
Kvernvold of DNV 22, Shirazi et al. of Tulsa University 28,
bration with sand erosion in air. Not surprising that their equation
and Lockett et al. of AEA 24. All models are limited to erosion
becomes increasingly conservative as the liquid-gas ratio in-
predictions in simple pipe geometries, such as pipe bends and
creases, i.e., as the mixture density ( m ) increases. In addition,
tees.
they did not account for the sand particle size which is known to
Salama and Venkatechs model 4 is a closed-form equation
have an effect on erosion rate for particles less than 400 microns.
whose predictions are accurate for mainly gas systems. The model
Above 400 microns, the effect of sand particle size becomes neg-
was verified using sand erosion data in air flow. This model sug-
ligible because of the different competing parameters of particle
gests the following equation for erosion prediction in steel with
weight, drag and number of particles per unit sand weight.
yield strength of 50 to 80 ksi:
The new equation being proposed in this paper is based on
WV 2 modifying Eq. 4 by incorporating the effect of fluid mixture
ERS m (4) density and particle diameter as follows:
D2
2
1 Vm d
where ER is erosion rate in mpy, W is sand flow rate in lb/day, V E p (6)
is fluid flow velocity in ft/s, D is pipe internal diameter in in., S m Sp D m
2
Downloaded 23 Aug 2011 to 200.174.154.60. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
Table 2 Measured and predicted erosion rates using Eq. 6
uncertainty associated with the forecast of production data, uncer-
1
ER 3
M dV m (9) tainty of fluid properties, uncertainty of sand size distribution, and
Sc physical uncertainty of pipe diameter and wall thickness. This is
Typically, a tolerable sand concentration of 5 ppm is specified and in addition to the obvious uncertainty in the predictive model
a sand size of 250 micron is considered. Considering a tolerable itself. It is critical to properly account for these uncertainties, par-
erosion rate of 0.1 mm/yr 4 mpy, the critical erosional mixture ticularly, when erosion predictions are used as basis for placing
velocity for elbows is: V e 11.7 m/s. limitations on production rate and/or serving as the sole input in
Downloaded 23 Aug 2011 to 200.174.154.60. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
Table 2. (Continued)
Downloaded 23 Aug 2011 to 200.174.154.60. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
Table 3 Values of sand erosion constants
Bourgoyne 30 appear to be high, and therefore cannot be used to express his appreciation to Oddmund Kvernvold of DNV, Tim
without further validation. In the proposed equation, the effect of Locket of AEA, and Sia Shirazi of Tulsa University for their
pipe bend radius was not considered because test results did not input.
show a major difference between erosion in 1 1/2 and 5 D elbows.
For plugged tee, both CFD analysis and limited experimental
work suggest that the erosion rate is lower than that for elbows.
But the effect decreases as the liquid to gas ratio increases. This Nomenclature
observation is also illustrated by Bourgoynes work 30. Using
Eq. 8 as the basis, the following is the recommended equation Units are as stated here, unless noted otherwise in the text of
for establishing erosional velocity limits for oil and gas produc- the paper.
tion:
Erosion measurements
D m E r erosion ratio in kg/kg, which is ratio between mass
V e (10)
20W of metal loss and mass of sand hitting target material
E p erosion parameter in mm/kg, which is ratio between
Conclusions and Recommendations penetration in metal and mass of sand hitting target
material
1 For solid-free, noncorrosive fluids, providing pressure drop is ER erosion rate in mm/y, which is rate of penetration in
not a concern, the maximum flow rate can be established using the metal by erosion
following form of API RP 14E equation:
Sand
400 W sand flow rate in kg/day
V
m M sand concentration ppm by weight, which is the
ratio of mass of sand to mass of fluid
where V is the maximum fluid velocity limit in ft/s, m is the d sand size in micron typical value 250 micron.
gas-liquid mixture density at flowing pressure and temperature in Note: The effect of d on ER becomes negligible
lb/ft3. above 400 micron. Therefore, for d400, the limit
2 For sand-free, corrosive fluids, inhibitors exist that are effec- of 400 is used.
tive at flow velocities corresponding to C-factors higher than 300. s sand density in kg/m3 typical value 2650 kg/m3
However, it is very important that the effectiveness of the inhibi-
tor be evaluated in a flowloop at these high velocities. For multi- Fluids
phase pipelines, the effectiveness of the corrosion control program V l liquid superficial velocity in m/s
depends on the proper transport of the inhibitors in the pipeline. V g gas superficial velocity in m/s
3 For sand-laden fluids, the maximum flow rate limit can be V m fluid mixture velocity in m/s
established using the following equation: V l V g
V e erosional velocity limit, m/s
D m l liquid density in kg/m3
V e
20W g gas density in kg/m3
m fluid mixture density in kg/m3
4 At high flow rates, the presence of sand enhances the corro- ( l V l g V g )/V m
sion of steel in both uninhibited and inhibited solutions due to
erosive wear of protected corrosion product and/or depolarization Pipe geometry
of anodically/cathodically controlled corrosion process by plastic D pipe internal diameter in mm
deformation of the metal surface. At low flow rates where sand
Constants
settling occurs, sand has no effect on corrosion rates in uninhib-
ited solutions, but it can have a profound effect on the rates in C empirical constant specified by API 14RP 14E to
inhibited solutions. predict the erosional velocity limit, V e in ft/s
S geometry-dependent constant, specified in this paper
for typical operating conditions tolerable erosion rate
of 0.1 mm/yr 4 mpy and sand size of 250 micron
Acknowledgment to predict erosional velocity limit, V e m/sec
The author would like to thank the management of Conoco for S p geometry-dependent constant, specified in this paper
their permission to publish this paper. The author would also like and used to predict E p in terms of flow parameters
Downloaded 23 Aug 2011 to 200.174.154.60. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
S m geometry-dependent constant, specified in this paper Corrosion in Oil and Gas Production, Corrosion 91, Paper 468, NACE.
15 Patton, C. C., 1993, Are We Out of the Iron Age Yet?, Corrosion 93, Paper
and used to predict ER given sand rate W and other
No. 56, NACE.
flow parameters 16 Erichsen, H., 1988, Nipple, Lock, and Flow Coupling Recommendations and
S c geometry-dependent constant, specified in this paper Subassembly Description for North Sea Wells, private communications, Nor-
and used to predict ER given sand concentration M ske Conoco a.s.
and other flow parameters 17 Camacho, R. A., 1988, The Design, Construction, and Testing of a Liquid
Impingement Apparatus and a Study of Metal Surfaces Eroded by Liquid
Note that the parameters S,S p ,S m ,S c are all related. As an ex- Impingement, M. S. thesis, The University of Tulsa, Tulsa, OK.
ample: S m equals 365/S p to convert E p to ER. 18 Saetre, O., 1991, Testing of Composite Pipes in High Velocity Seawater,
Volume III, Part B Proc., 10th International Conference on Offshore Mechan-
ics and Arctic Engineering, ASME, eds., M. M. Salama et al., pp. 577583.
References 19 Salama, M. M., 1996, Velocity Limits for Multi-Phase Piping, unpublished
1 API, 1981, API RP 14E Recommended Practice for Design and Installation internal report, Conoco Inc..
of Offshore Production Platform Piping Systems, 3rd Edition, American Pe- 20 Salama, M. M., 1993, Erosional Velocity for Water Injection Systems,
troleum Institute, Washington, DC, p. 22. Materials Performance, NACE, 32, No. 7, pp. 4449.
2 API, 1991, API RP 14E Recommended Practice for Design and Installation 21 Greving, D., 1991, Effect of Flow Velocity on the Performance of Selected
of Offshore Production Platform Piping Systems, 5th Edition, American Pe- Oil Field Corrosion Inhibitors in Vertical Tubing, Under Two-Phase Flow
troleum Institute, Washington, DC, p. 23. Conditions, M. S. thesis, Mechanical Engineering, University of Tulsa,
3 Salama, M. M., 1998, An Alternative to API 14 E Erosional Velocity Limits Tulsa, OK.
for Sand Laden Fluids, Proc., Annual Offshore Technology Conference, 22 Kvernvold, O., 1998, ERBEND 2Erosion in Pipe Bends, Det Norske
OTC Paper 8898. Veritas DNV, Norway.
4 Salama, M. M., and Venkatesh, E. S., 1983, Evaluation of API RP 14E 23 Shirazi, S. A., McLaury, B. S., Shadley, J. R., and Rybicki, E. F., 1995,
Erosional Velocity Limitations for Offshore Gas Wells, Proc., 15th Offshore Generalization of the API RP 14E Guidelines for Erosive Services, J. Pet.
Technology Conference, Paper OTC 4485. Technol., 47, No. 8, pp. 693698.
5 Heidersbach, R., 1985, Velocity Limits for Erosion-Corrosion, Proc., 17th 24 Lockett, T. J., Beech, P. M., Birchenough, P. M., McCarthy, P., Dawson, S. G.
Offshore Technology Conference, Paper OTC 4974. B., and Worraker, W. J., 1997, Erosion/Corrosion in Sweet Multiphase Sys-
6 Beggs, H. D., and Brill, J. P., 1973, A Study of Two-Phase Flow in Inclined tems, AEAT Report No. 1174, AEA Technology plc., UK.
Pipes, J. Pet. Technol., 25, No. 5, pp. 607, 617. 25 Svendeman, S. J., and Arnold, K. E., 1994, Criteria for Sizing Multi-phase
7 Rybicki, E., 1987, private communication, University of Tulsa, Tulsa, OK.
Flow Lines for Erosive/Corrosive Services, SPE Prod. Facil.
8 Engel, O. G., 1955, Water Drop Collisions with Solid Surfaces, J. Res.
26 Birchenough, P. M., Dawson, S. G. B., Lockett, T. J., and McCarthy, P., 1995,
Natl. Bur. Stand., 54, No. 5, pp. 281298.
Critical Flow Rates Working Party, Report No. AEA-TSD-0348, AEA
9 Gipson, F., 1989, Petroleum Production Engineering, Pits and Pieces,
Technology, UK.
Manual of Southwest Petroleum Short Course, Texas Tech University, April
1720. 27 Kvernvold, O., and Sandberg, R., 1993, Production Rate Limits in Two-
10 Deffenbaugh, D. M., and Buckingham, J. C., 1989, A Study of the Erosional/ Phase Flow Suystems: Erosion in Piping Systems for Production of Oil and
Corrosional Velocity Criterion for Sizing Multi-Phase Flow Lines, South- Gas, Technical Report No. 93-3252, Det Norske Veritas DNV, Norway.
west Research Institute Final Report, Project No. 04-2433, prepared for the 28 Tolle, G. C., and Greenwood, D. R., 1977, Design of Fittings to Reduce
Minerals and Management Service, U.S. Department of the Interior. Wear Caused by Sand Erosion, API OSAPER Project No. 6, American Pe-
11 Smart, J. S., 1990, A Review of Erosion Corrosion in Oil and Gas Produc- troleum Inst., Texas A&M Research Foundation.
tion, Corrosion 90, Paper 10, NACE. 29 Weiner, P. D., and Tolle, G. C., 1976, Detection and Prevention of Sand
12 Coker, A. K., 1990, Understand Two-Phase Flow in Process Piping, Chem. Erosion of Production equipment, API OSAPER Project No. 2, American
Eng. Prog., 86, No. 11, Nov. pp. 6065. Petroleum Inst., Texas A&M Research Foundation.
13 Coulson, J. M., and Richardson, J. F., 1977, Chemical Engineering, 1, 3rd 30 Bourgoyne, A. T., 1989, Experimental Study of Erosion in Diverter Systems
Edition, p. 91, Pergamon Press. Due to Sand Production, Proc., SPE/IADC Drilling Conference, New Or-
14 Smart, J. S., 1991, The Meaning of the API RP 14E Formula for Erosion/ leans, LA, SPE/IADC 18716.
Downloaded 23 Aug 2011 to 200.174.154.60. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm