You are on page 1of 12

SPE

-of~Gwinf=a

SPE 15416

Rheology of Oil-Base Muds


by O,H. Houwen, Sch/urnberger Cambridge Research, and T, Geehan, SEDCO FOREX
SPE Members

Copyright 1986, Society of Petroleum Engineers

This paper was prepared for presentation al Ihe 61st Annual Technicsl Conference and Exhibition of Ihe Society of Petroleum Engineers held in New
Orleans, LA October 5-6, 19S6.

This paper waa selected for prasentafion by an SPE Program Committee fotlowing review of information confained in an abstract submitfec by (he
author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the
author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers
presented at SPE meetings are subject fo publication review by Editorial Committees of fhe SOciely of Petroleum Engineers. Permission to copy is
restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment of
where and by whom the paper ia presented. Write Publications Manager, SPE, P.O. Box 833S36, Richardson, TX 760S3-3S36. Telex, 7309S9 SPEDAL.

INTRODUCTION
Although a great deal has been written about the pressure
and temperature behaviour of the viscosity of simple non-
ABSTRACT Newtonian fluids, and an understanding of this behaviour at
The theological behaviour of invert emulsion muds has been the molecular level is emerging, no consensus exists on how
studied at pressures up to 1000 bar and temperatures up to to deal with concentrated suspensions. This can easily be
240°C. Theological parameters were calculated for the Bing- understood, considering the widely different nature of non-
ham, Herschel-Bulkley and Cssson theological models. The Newtonian fluids, Invert emulsion muds are suspensions of
Iierachel-Bulkley and Cesson modeb both give good fits to the solids and emulsions at the same time, and as there is no
experimental rheograms. The Cesson model is more reliable generally accepted theological model that can be applied to
for extrapolation purposes than the Herschel-Bulkley model. emulsions and suspensions, the engineering aspects of invert
A pair of two similar exponential expressions were found to emulsion muds are not always based on very sound scientific
be able to model the pressure and temperature behaviour of principles, Thus, while it is known that at pressures and tem-
the two parameters of the Casson model. The expressions, peratures encountered in the wellbore the rheology of the mud
which are baaed on the relation for pure liquids derived the-
will be different from that measured at the surface, lack of
oretically by Eyring, contain temperature dependent pressure
the ability to quantify the effects involved has perpetuated the
coefficients. The simplifications inherent in the temperature
field practice of using theological parameters measured at at-
and pressure model are dkcussed in the light of the tempera-
mospheric pressure. Traditionally the mud industry has, with
ture and pressure behaviour of the viscosity of common base
a few exceptions, adhered to the uee of the Bingham and power
oils and their constituent hydrocarbons. Field application of
law theological models, which have the advantage that hy-
the model requires measurement of the rheology of the mud at
draulics calculations are available for fluids obeying these mod-
two or more temperatures and knowledge of the pressure co-
els. Hence, it is not surprising that the existing techniques for
efficients relating the behaviour of the plastic viscos~ty to that
prediction of downhole rheology are based on these models.
of the yield point, or the Casson high shear visccgity to that
of the Casson yield stress. Pressure meaauremer !S or other A number of recent publications have dealt with the problem.
Combs and Whitmire (1) showed that the change in the viscos-
information are then not required. Applications can be baaed
ity of the continuous phase is the main factor in controlling the
on Caason or Bingham theological meesurements. The rela-
change in the viscosity of the mud with pressure. Both yield
tionships between the parameters of the Casson and Bingham
point and plastic viscosity seemed to be governed by this effect.
models are disussed.
McMordie et af, (2) concluded that the power law model gives
the beat mathematical description of the viscosity of an oil base
mud at constant temperature and pressure, They require two
sets of constants for shear rates below 200s- 1 and above this
value, which suggests that the choice of the power law model
is not the best one to be made. They found that the loga-
References and illustrations at end of paper.
rithm of the shear stress is proportional to the pressure, giving
2 RHEOLOGY OF OIL BASE MUDS SPE 15416

rise to the exponential law for the pressure dependence that were changed by addition of barite or organophilic clay. Great
also appeared in the appropriate API Bulletin (3). Again, the care wrw taken to use only reproducible results, After viscome-
choice of the power law model necessitated the use of different ter runs at elevated temperatures rheograms were re-recorded,
constants for high and low shear rat es. De Wolfe et al. (4) to see if any irreversible chemical change had occurred. Gen-
studied a number of less toxic oils. They report a close corre- erally this started to happen at temperatures above 140°C.
lation of the results to the Herschel-Bulkley model. Since their As we are then dealing with a problem of a different nature,
results are presented as apparent viscosities, no further infer- the irreversible effect of temperature on mud systems which
ences can be made as to the manner in which the theological can only be alleviated by chemical treatment, we decided to
parameters of this model individually depend on temperature exclude these measurements from our data set.
and pressure. Politte (5) chose to model the invert muds as THEOLOGICAL MODELS
Bingham fluids, which they resembled more closely than power An obvious strategy is to decide upol: a theological model and
law fluids, A multi-term equation with 13 numerical constants to investigate the pressure and temperature variation of the
was presented to model the viscosity behaviour of diesel oil at parameters involved in the model. In order for this strategy to
pressures over 1000 psi. Politte concluded that the plastic vis- work, it is necessary that this model is applicable with the same
cosity could be normalized with the viscosity of the oil, The degree of accuracy at all pressures and temperatures that are
yield point was found to be more of a problem, since it is not relevant to the rig operations. When the rheograms obtained
a true physical parameter, and much more susceptible to ex- with the HTHP rheometer were examined, it became imnledi-
perimental error than the plastic viscosity. It was found to be ately apparent that they all show curvature. It would thus be
a weak function of pressure, the effect of pressure decreasing a simplification to analyse the rheology of the muds in terms
as temperature increases. Bailey et al. (6) used the Bingham
of the Bingham model. The power law model leads to curved
model to describe the rheology of low-toxicity oil mud, but rheogranl& but is not applicable either, as can be seen in Fig-
noted that at higher temperatures departure occurs from this
ure 1. For a power law fluid a plot of log p vs. log ~ should
model.
give a straight line, since
OBJECTIVES
~ = l+, (1)
We decided to reinvestigate the problem. Our first objective
was to find a pressure and temperature dependent model with hence,
a wide app!icabilit y. The second objective was to give prefer- log; =logp=log k+(?t–l)log+. (2)
ence to models which would offer some physical interpretation
of the theological phenomena. This would give hope that ulti- Figure 1 also shows a remarkable similarity of the curves de-
mately a model can be constructed that relates the theological termined at temperatures between 25°C and 110°C, indicat-
parameters to mud composition and changes thereof (our third ing some invariant feature of the rheology as temperature is
objective). It was apparent from previous work that the most changed. The same observation was also made for curves de-
elusive part of the prediction of the temperature and pressure termined at constant temperature and increasing pressure.
behaviour of the viscosity is at the low shear rate end. In order The modified power law model of Herschel and Bulkley has
to study this problem under the best conditions we prepared been advocated (9) as a more realistic description of muds,
muds with appreciable yield points, so that we were able to since it contains a yield stress term:
anal yse numerical values well above the limits set by experi-
mental precision. T = To -i- k~n. (3)
EXPERIMENTAL
We carried out the investigation with a Haake 1000 bar rheome- In order to make some judgement of the closeness of fit of the
ter, essentially a modern version of the instrument described experimental data with theological models, a set of 20 shear
as the BHC Viscometer (7). The nominal shear rate range is stresses was extracted from each complete rheogram at pre-
o-1200 s-l, continuously variable. The bob of this rheometer is selected ~ values, sampling relatively more points from the
magnetically coupled to the drive and torque measuring unit, initial curved portion of the rheogram than at the high shear
rate end. The differences of these experimental shear stress
which can measure shear stresses up to about 60 Pa. In our
values with the predictions based on the model were calculated.
experience an accuracy of about 1 Pain the shear stress read-
Two numerical techniques were used in order to calculate the
ings can be achieved with the system. The shear stress vs.
Herschel-Bulkley parameters. In the first a variable trial value
time profile of the rheometer was controlled by a home made
of TO was subtracted from all 7 values, and a least squares
system, incorporating a micro computer, which also continu-
regression performed on the logarithms of the values of (T – TO )
ously recorded the shear stress readings. Rheograms thus col-
and ~. Maximisation of the correlation coefficient gave the final
lected were reformatted and stored in a main frame database
value of TO. In method 2 the function of TO, k, and n:
system, for later recall and plotting. We also used a more
precise thermostated atmospheric Haake RV1OO system with a
shear rate range to about 2700 s–l. Muds were prepared in (4)
the laboratory, using commercially available products for for-
mulation of a currently popular less toxic invert mud system
used in the North Sea, and also used in the study on solids con-
trol equipment published from our laboratories (8). Viscosities
SPF 15AIA O.H. HOUWEN ANO T,----- .. .
~FFHAN .3

where i runs over the zo data points extracted from the ex- son constants determined in the shear rate interval 1-1200 s-1
perimental rheogramwaa minimized, by considering that at was again judged by comparison with rheograms determined
the optimum values of TO,k, and n the partial derivatives of between 1 and 2700 s-l (Figure 2). The Casson rheogram
E with respect to r., k, and n will be zero. In essence this remains closer to the experimental curve than the Herschel-
method does a least squares regression on the 4, r data points Bulkley rheogram, In fact, if one would compare extrapolated
directly without having to resort to taking logarithms. A con- viscosities at much higher shear rates, those calculated from
sequence of this difference and the way in which we have spaced the Herschel-Bulkley parameters are significantly lower than
out our sampling points is that, if the Herschel-Bulkley model those calculated from the Casson parameters, To put this in
is not the ideal representation of the datapoints, method 2 is perspective, we will summarize here the physical interpretation
expected to give a closer fit than method 1 at the high shear of the two Casson constants.
rate end, at the expense of the fit at the low shear rate end. Squaring both sides of equation 4 gives
It was possible to fit our experimental rheograms to this model
T = koz + 2kok1-j* + klzi. (6)
quite well over the whole range of shear rates (O -1200 s-l),
Table 1 shows an example of the results obtained by both meth-
Equation 6 shows that for ~ approaching zero, the shear stress
ods. A similar closeness of fit was observed for all our HTHP
becomes equal to k.’. Hence, k.’ is identified = a yield stress.
rheograms. It is indeed clear that the predicted discrepancy
The ko2 parameter, which we will call the Casson yield stress
between methods 1 and 2 at high shear rates occurs, method
TY, plays much the same role in the Caason model as the field
2 giving an overall better fit except at shear rates below about
Point (YP) and r. play in the Bingham and Herschel-Bulkley
15 s-l.
model, respectively, and like these should be reported in Pa
The muds were also run in the atmospheric rheometer. A fit
(N/m2) or lb/100sqft. It should also be noted that, as Table
was made to the 0- 12005-1 portion of these rheograms. With
1 shows, the experimentally determined values of the Casson
the fitted parameters an extrapolation was made fro,m 1200-
yield stress are very close to the Herschel-Bulkley yield stresses.
2700 s-1 (Figure 2). It can be seen that at the highest shear
Division of both sides of equation 6 by ~ gives
rates the experimental rheograms depart from the Herschel-
Bulkley model, which consistently gives too low values in this p = ko2~-1 +2kok1~-* + klz. (7)
range, regardless of whether the Herschel-Bulkley parameters
were obtained by method 1 or 2. This is a consequence of the As ~ goes to infinity, only the last term in this equation re-
nature of the model which predicts an ever decreasing viscosity mains, Hence, kl 2 is the viscosity of the fluid at infinitely
With increasing shear rate. However, the behaviour of the log p large shear rate, and its interpretation as such makes it sim-
m, log ~ curves (Figure 1) suggests a leveling off of the viscosity ilar to the Bingham plastic viscosity (~). If r is measured
to some constant Newtonian plateau value, as is generally ob- in Pa, and ~ in s-1, then kl 2 has units of Pas. To make its
served for suspensions (10). Thus, while the Herschel-Bulkley
model is capable of describing the rheology of the muds in the 1 relationship to the Bingham ~ more apparent, we have mul-
-1200 s–l range quite adequately, it cannot be used with confi- tiplied klz by 1000 to give it units of mPas or cp, and call this
dence for extrapolation. Another objection is the fact that the quantity the Caeson high shear viscosity (PC), As far ea muds
power law index n haa no clear physical interpretation other are concerned, the Casson yield stress is always smaller than
than being a convenient curve fitting parameter, which would the ~, and the Casson high shear viscosity is always smaller
stand in the way of reaching our third objective of relating than the ~. See also the Appendix for further comments on
rheology to mud composition. the Casson model.
HTHP RHEOMETRY RESULTS
As an alternative to tlie Herschel-Bulkley model, we decided The Bingham ~ and ~ were calculated from the shear
to investigate the Casson model, which has been claimed by stresses measured at shear rate values of 500 and 1000 see-1,
several authors to give a good represen~ation of the rheology hence at practically the same values as are used in field prac-
Df water based muds (11-15). This is a two-parameter model tice. Casson and Herschel-Bulkley (by methods 1 and 2) pa-
rameters were determined from computer fits to the 20 selected
and can be written as (16) data points in the manner described before. Examples of plots
of theological parameters as a function of pressure at constant
temperature are shown in Figures 3 and 4. The same general
trends as observed by previous workers were observed for the
The constants k. and kl can conveniently be evaluated from ~. Thus, the ~ dropped with rising temperature rmd in-
the experimental data by least squares linearization of the creased with rising pressure. The hypothesis was tested that
square roots of the r and i values. The closeness of fit of the Arrhenius equation
the calculated shear stress values was analysed in the same
way as outlined before and a similar analysis as performed for p= Aexp$, (8)
the Herschel-Bulkley model was made. Within the experimen-
tal errors of the HTHP rheometer, the Caeson and Herechel-
where T is the absolute temperature (K), and the simple rela-
Bulkley models are seen in Table 1 to give about equally satis- tionship
factory results. The reliability of extrapolations based on Cae- p = A exp CP, (9)
h RHEOLOGY OF OIL BASE MUDS <PF 75[,IL

ihat had previously been advanced for the pressure behaviour We analyzed the mud data in a similar way, taking as the vis-
~f the viscosity of muds (2,3), and in which C is a temperature cosity parameters both the _ and Casson high shear viscosi-
ndependent constant, would apply tosome of the theological ties. The comparison between these two parameters was made
>arameterso When for this purpose plots of the logarithms because both should reflect the effect of the continuous phase
>f the theological parameters were studied, it appeared that
;he slopes B of the lines relating log ~ to l/T at constant on the high shear rate viscosity, where inter-particle effects be-

?ressure were dependent on pressure. Also, slopes of log PV come small, The PV data were taken from our own work, in
US.P at constant temperature were dependent on temperature. addition to the data published by Politte (5) and Bailey (6).
I’hisbehaviour suggests thateq. 8andeq, 9can recombined Casson high shear viscosities were deduced from the literature
n a law of the type data by means of the algorithm discussed in the Appendix.
We found that a reasonable fit to the data was achieved if we
(lo) took (a log ~/aP)T and (a log pc/~P)~ to be directly pro-
w = Aexp[~ + C(T) P],
portional to l/T, with proportionality constants CB and CC,
tvhere the pressure coefficient C of eq. 9 is now a function of respectively,
temperature, yet to be specified. Analogous behaviour to the ~ and Casson high shear vis-
Previous workers have shown that the ~ of invert muds is cosity wss displayed by the YP and Casson yield stress, The
proportional to the viscosity of the base oil at elevated tem- slopes of the lines relating the logarithms of these quantities to
perature and pressure (1,5). We therefore turned to viscosity P at constant temperature were found to be smaller than the
corresponding slopes of the ~ and Casson high shear viscos-
iata for the base oils. An extensive set of measurements on ity lines. We found by doing a similar linear regression through
diesel oil is embodied in Figure z of Politte’s paper (5); similar the (8 log YP/aP)T or (~ log ry /ap)T us. l/T plots that the
but less detailed figures showing the viscosities of naphthenic pressure dependence of YP and rY can be represented by two
and solvent based oils can be found in the paper by De Wolfe equations:
et al. (4). We read the experimentally determined viscosities
from these graphs, and replotted their logarithms as functions B’ YBCBP
Df pressure. It became clear that (~ log p/dP)~ had a similar
dependence on T as was displayed by the ~’s of our muds,
~ = A’ exp
( y+-y-
)
(11)

k .g-ure 5, where we have plotted (~ log p/8P)zI as a function and


of (1/2’), shows a least squares fit representing the trend dis-
B/l YCCCP
played by the oils in this way, including the base oil used for Ty = A“ exp ~+— (12)
our work. Because of the experimental uncertainty of some of ( T )
the data, there is still room for the question as to whether this
where YB and YC are multipliers smaller than unity modify-
line should go through the origin of the plot, in which case the
ing the same pressure coefficients CB and Cc that were found
term C(T) in equation 10 bec~,,:es C/T. We have looked at
applicable to _ and Casson high shear viscosity.
this question in the light of what ;s known about the bahaviour
We did not attempt to fit the Herschel-Bulkley parameters to
of pure hydrocarbons, and the experimental data available for
any P, T model. As seen in Figure 4 the three Herschel-Bulkley
muds.
parameters show less regular behaviour than those belonging to
The viscosities of some hydrocarbons of interest have been re- the two-parameter theological models. No doubt this situation
ported at pressures up to several thousands of bars, far ex- could be improved if we would impose the condition that n
ceeding the range over which the available data on invert mud remain constant over the range of pressure and temperature.
base oils extend, and exceeding the range of wellbore pres- The objection against n as a curve fitting parameter remains
wres. At that scale of measurement log p us. P isotherms however.
generally are concave towards the P axis. (But as noted by
several authors, can also become convex at low temperatures; DISCUSSION OF THE P,T MODEL
e.g. Hogenboom et al. (17) show this to occur at pressures of We can rewrite the expressions found for the temperature and
about 2000 to 3000 bar for the 16° C and 38°C isotherms of pressure dependence of the theological parameters in the fol-
ci-decalin.) This is of no practical concern to our problem, lowing generalized form:
however; below 1000 bar the isotherms can be represented by
straight lines without much harm to the accuracy of the rheo- YvaP
VIS(P, T) = A exp ;+= (13)
Iogical predictions. In this way we have calculated the slopes of ( )
the (~ log g/i3P)T isotherms from published data, generally in
the 200-1000 bar range. Results are shown in Figure 5 for the
In this equation VIS stands for ~, Casson high shear viscos-
straight chain alkanea *octane (22) and whexadecane (16),
ity, ~, or Ty. A has the meaning defined before in eq. 8-10.
the naphthenes cyclooctane (22) and cisdecalin (17), and the
El.2 replaces B in eqs. 8 and 10. Va/R takes the value of CB
aromatics butylbenzene (23) and octylbenzene (23), the last
or . .*c. Y is equal to unity if ~ stands for ~ or Casson
three compounds all having 10 carbon atoms. The regression
high shear viscosity, and is equal to YB or YC if VIS stands
lines for the base oils and the n-alkanes have negative inter-
for ~ or TY, respectively.
cepts on the (~ log p/aP)T axis; most of the naphthenes and
The values of A will be different for ~ (or PC) and ~ (or
aromatics have positive intercepts. ry ). We also found that the values of E used in the forms of
SPE 15416 O.H. HOUWEN AND T. GEEHAN 5

eq. 13 describing ~ (~C) ,YP (~Y), or the base oil viscosity An alternative method is to replace the Casson parameters by
are not necessarily equal. Bingham ~ and ~ in steps 2 and 3. Step 5 gives then the
If R is the gas constant, then E has the dimension of energy Bingham rheology at the desired temperature and pressure. If
and V@of volume, We have thus obtained a modified form of a rheogram is to be produced, then, before applying equation
the equation derived theoretically by Eyring (18) 6, the Caason kl and kO are calculated from ~ and ~ by
the method given in the Appendix.

P = #27rmkT)l/2 ~/3 (14) Thus, we have two ways of drawing rheograms based on the
‘f ‘xp(AE”~~+)9 C!ssrjonrheologicalmodel. The fhstusesCaasonconstants
where AEvia is the activation energy for flow and V~ is the throughout, that are directly obtained from measurements at
molar volume. We have used the constant A instead of the surface conditions. The second uses Bingham constants in the
preexponential factor derived by Eyring, on the grounds that initial measurements and the P,T prediction, and then converts
this seemed experimental y justified. Indsed, if for example we these into Casson constants. The question has to be asked if
define a reduced ~ these two routes yield different results, given the rather com-
plicated quadratic equations underlying the conversion. We
–YvaP have applied the two methods using pairs of A and E/R val-
Pv,ed = Pvexp ~ , (17)
-() ues that were obtained separately for the Bingham constants
(determined 500 and 1000 see-l and for the Caason constants
then plots of log ~r.d vs. l/T (Figure 6), using the appropri- (determined by least squares fit to 20 selected points on the
ate values for Y and Va, are straight within experimental error, input rheograms). Results are shown in Figures 9 and 10. The
and introduction of a more rigorous temperature dependence differences are very small and of the order of the experimen-
seems unnecessary. Our numerical choice for Va/R leads to tal error as can be seen by comparison with the experimental
values for the factor n in Eyring’s relation (eq, 14) which are rheograms presented in Figures 7 and 8.
in agreement with n values published for hydrocarbons (18).
A different issue raised by our treatment of YP and TY which CONCLUSIONS
we implicitly consider to be a f~nction of the base oil viscosity, We have found a simple model for the description of the rheol-
is the precise nature of this function. Through the introduc- ogy of invert emulsion muds under downhole conditions, The
tion of our factors YB and Yc, we make clear that ~ and model needs as input parameters four constants which are spe
TY vary less with pressure than PV and the Casson high shear cific for each mud. They are two activation energies and two
viscosity. The details of the mechanism by which the viscosity preexponential factors. These can easily be obtained from Ar-
Df the base oil governs ~ and rY warrants much more inves- rhenius plots of two theological parameters, measured at the
tigation. Firth and Hunter (19) and Van de Ven and Hunter rig site with a conventional rheometer at two or more different
(2o) studied the rheology of mineral suspensions; it is interest- temperatures. Also required are a knowledge of the temperi-
ng to note that they found a linear relation for the viscosity ture dependent pressure coefficients, expressed in this paper as
Df the continuous phase with the Bingham fi. It should also the product of the constants Y and C, which are specific for a
be mentioned that our finding is in qualitative agreement with mud Q’pe”
the work done by Combs and Whitmire (1) and by Politte (5). The pressure behaviour of the muds was found to obey an expo-
nential law. The limitations of this conclusion within the realm
APPLICATIONS OF THE P,T MODEL of pressure dependencies of related hydrocarbons has been dis-
We can use equation 13 for predictions of downhole rheology. cussed, Because base oils are complex mixtures of these classes
According to our model it would be sufficient to determine the of hydrocarbons, the pressure behaviour of flifferent oils shows
Arrhenius constants A and E/Rat atmospheric pressure. This little variation. It is interesting to note thee Politte (5), using a
requires the following steps. different line of reasoning, also tentatively concluded that two
1. Measure the rheology of the mud with a standard oil field of the oils studied by De Wolfe (4), paraffinic oil and solvent oil,
rheometer, at least at two different temperatures. could be modelled by the equation developed by him for diesel
2. Determine the Casson high shear viscosity PC and Cas- oil. Our theory is sufficiently general to allow for radically
son yield stress ry by one of the methods given in the different oil mixtures in a straightforward manner. The acti-
Appendix, vation volumes would then simply be replaced by other values,
3. Plot In WC and In ry as functions of I/T. and even a constant factor (not dependent on temperature)
4. The slopes of these lines give the values of E/R; the inter- can be added to the exponential term. This would correspond
cepts give the values of A. Two sets of these parameters to abandoning the decision to take the functions (tl log /.t/8P)T
will result, to be directly proportional to l/T and to accept that plots of
5, If Cc and YC are given, equation 13 will give the Casson the type presented in Figure 5 have intercepts.
high shear viscosity and rY at the desired combinations of The non-Newtonian character of invert emulsion muds can be
temperature and pressure, described by the two parameter Casson model, which by others
6. If a rheogram is to be produced, convert Casson high shear haa been shown to be applicable to water based muds as well.
viscosity and TY into kl and ko and use equation 6. The two Caason parameters can conveniently be presented as
Of course, these steps can be facilitated by the use of a pro- the viscosity at infinitely high shear rate, and the apparent
grammable hand-held calculator or a computer. An example physical yield stress, which play roles analogously to the ~
of computer drawn rheograrns and experimentally determined and ~ of the Bingham model.
rheograma can be seen in Figures 7 and 8. We have shown in the Appendix that the Caason and Bhgham
6 RHEOLOGY OF OIL BASE MUEJS SPE 15L16

parameters are linked by a set of quadratic equations. Hence 6. Bailey, T. J,, Bern, P. A,, and McEwan, F., “Low-toxicity
the tworheological parameters required for application of the oil muds: A knowledge of downhole theological behavior
P,T model can be taken from either of the two models. We have assists successful field application” , SPE Dril(ing Engi-
also shown that modelling of the rheology by the Herschel- neering (April 1986) 107,
Bulkley model gives shear stresses which are quite close to 7. McMordie, W, C., “Viscometer tests mud to 650°F”, Oil
those calculated from the Casson model. and Gas Journal (May 19 1969) 81,
The advantage of the Caason model over the Bingham model is 8. Froment, T. D., Rodt, G. M., Houwen, O, H., and Titreville,
that it reproduces the curvature of the rheograme at shear rates B., ‘A drilling contractor tests solids control equipment”,
above about 1 s-1, It can therefore serve as a mathematical IADC/SPE 14753 (Dallas, February 10-12, 1986).
basis for advanced hydraulic programs which are based on as 9. Zamora, M., and Lord, D, L,, “Practical analysis of drilling
close an approximation of the real non-Newtonian character of mud flow in pipes and annuli”, SPE 4976 (Houston, Oc-
the fluid as possible. tober 6-9, 1974).
10. Mewis, J., “Rheology of suspensions”, in Rheology, As-
NOMENCLATURE tarita, G., Marucci, G,, Nicolas, L., eds., Vol. 1, (Proc,
A, A’,At’ = Constants in Arrhenius equation VIII Int. Congr. Rheol., Naples, September 1980),
B, B’,B” = Constants in Arrhenius equation (Plenum Press, New York) 149.
c; = Pressure coefficient in eq. 11, K-bar-1 11. Lauzon, R. V., and Reid, K. I. G., “New theological model
cc = Pressure coefficient in eq. 12, K- bar-1 offers field alternative~, Oil and Gas Journal (May 21
E = Least squares sum 1979) 52.
+ = Shear rate, s– 1 12, Lauzon, R. V., and Short, J. S., “The colloidal interac-
k = Consistency index, Pa.sn tion of ferrochrome Iignosulfonate with montmorillonite in
kO . Casson intercept, Pa1i2 drilling fluid applications”, SPE 8225, (Laa Vegas, Sept.
kl = Caason slope, (Pas) 1/2 23-26, 1979).
P = Newtonian viscosity, Pa,s 13. Zhongying, W. and Songran, T., “Casson theological
Pc = Casson high shear viscosity, mPa.s model in drilling fluid mechanics”, SPE 10564 (Beijing,
n = Power law index March 21-23, 1982).
P = Mud pressure, bar
14. Zhongying, W. and Songran, T., ‘The flow effect of mud
Pv = Bingham plastic viscosity mPa.s
in annulus and the selection of theological parameters”,
R = Gas constant
SPE 14871 (Beijing, March 17-20, 1986).
T = Absolute temperature, K
15. Wanneng, S., Jianping, C., and Zhenxue, L., “Compar-
r = Shear stress, Pa
ison of theological models in iiigh shear rate range and
To = Herschel-Bulkley yield stress, Pa
experimental relationship between penetration rate and
ry = Cssson yield stress, Pa
high shear viscosities”, SPE 14858 (Beijing, March 17-20,
YB = Multiplier in eq. 11
1986) .
Yc = Multiplier in eq. 12
16. Caason, N., ‘A flow equation for pigment-oil suspensions
w = Bingham yield point, Pa
of the printing ink typen, in Rhwlogy of disperse systems,
ACKNOWLEDGEMENTS Mill, C. C., cd., (Pergamon, London, 1959) 84.
The authors thank their managements for permission to pub- 17. Hogenboom, D. L., Webb, W., Dixon, J. A., “Viscosity of
lish this paper. They further thank I. Bratchie of SCR for several liquid hydrocarbons as a function of temperature,
making a computer program available to calculate Herechel- pressure, and free volume”, J. G’hem. Phys. (1967) 46,
Bulkley constants by Method 2, and I. Chalmers, N. Alderman, 2586.
and H. Ladva for technical assistance. 18. Ghisstone, S., Laidler, K. J., and Eyring, H., The theory of
rate processes, (McGraw-Hill, New York, 1941) Chapter
REFERENCES Ix.
1. Combs, G.D. and Whitmire, L. D., “Capillary viscometer 19. Firth, B.A. and Hunter, R.J., “Flow properties of coagu-
simulates bottom-hole conditions”, Oil and Gas Journal lated colloidal suspensions III. The elastic floe model”, J.
(Sept. 30 1968) 108. Colloid Interface Sci, (1976) 57,266,
2. McMordie, W. C., Bennett, R. B., and Bland, R. G., “The 20. Van de Ven, T. G.M., and Hunter, R. J., “The energy dis-
effect of temperature and pressure on the viscosity of oil sipation in sheared coagulated SOIS”, Rheol. Acts (1977)
base muds”, SPE 4974 (Houston, Oct. 6-9, 1974). 16,534.
3. API Bulletin on the rhwlogy of oil-well drilling fluids, API 21. Speers, A., “Computer aids analysis of drilling fluids”, Oil
BUL 13D, (First ed. August 1980) 21, and Gas Jcurnaf (Nov. 19 1984) 118.
4. De Wolfe, R. C., Cofiin, G, B., Byrd, R. V., ‘Effects of tem- 22. Gouel, P., “Viscosity of alkanes (C13 to Cle), cycha and
perature and pressure rheology of less toxic oil muds”, alkyl-benzenes”, Bull. Cent. Rech. Ezplor.-Prod. Elf-
SPE 11892 (Aberdeen, Sept. 6-9, 1983). Aquitaine (1978) 2, 439.
5, Politte, M. D.,’’Invert oil mud rheology as a function of 23. Ducoulombier, D., Zhou, H., Boned, C., Peyrehiase, J.,
temperature and pressure”, SPE/IADC 13458 (New Or- Saint-Guirona, H., and Xans, P., “Pressure (1-1000 bars)
leans, March 6-8, 1985). and temperature (20-100 “C) dependence of the viscosity
of liquid hydrocarbons”, J. Chem. Phys. (1986) 90, 1692.
5PE 15416 O.tl. HOUWEN AND -----GFFHAN
T. . .. . .
I

APPENDIX
CALCULATIONOF THE CASSON CONSTANTS
It shouldbe noted that we have adopted a different nomen-
clature from the one used by Lauzon and Reid (11) and in
subsequent papers (13,14,15,21). Our definition ofkl and ko
is as given originally by Caason (16). Our use of Casson high
shear viscosity and TY is designed to remain numerically and
conceptually close to the framework of the Bingham model.
If a conventional six-speed field rheometer is used, then the
most accurate way is to plot the square roots of the shear stress
values against the square roots of the shear rates. The intercept
of the line through the data points is k. and kl is the slope.
This can best be done by a programmable calculator with a
least squares subroutine. Alternatively, but employing only
two of the data points and therefore less accurately, Lauzon
and Reid’s formulae using the 100 and 600 rpm readings may
be used (11).
If a two-speed rheometer is used of the conventional type, then
the Caason constants can be calculated from the 600 and 300
rpm dial readings by using

and
k. = 2.44068:~; – 1.72588:{; (A -1)
where kl is in (Prvs) 112 and k. is in Pa]/z. Of course, if only
~ and ~ are given, a simple conversion gives &jOO and 0600,
from which the Casson constants are then calculated.
[f we calculate the Bingham parameters in the usual way from”
300 and 600 rpm readings, then the following interesting rela-
tions hold:
~ = kl(kl + 0.03665kO),
——
YP = kO(kO+ 26.484kl), (A -2)

where YP is in Pa, —— be shown that k. and kl are


It can
quadratic functions of the PV/YP ratio. For example,

(A -3)

where

( )
~ _ 0,0183 0.000336 0.5147
-13.2+ 175.4+ ~+—— .
– W/v (Pv/YP) pv/yp
(A -4)
SPE 15416

TABLE ]

RNEoLOGICALMODELSANo EXPBNIMENTAL
itNEOWAltS

Sxperi8ent CASSON EBRSCEEL-BUWL8Y(1) BSRSCSIEL-BULKLSY


(2)

sac’-l’ Pa Pa cliff. Pa cliff. Pa cliff.


500 5 93 6UD O 08 5 96 0 03 6 42 049
15:00 7:33 7:30 - G: 03 7:18 -0:15 7:41 :0:08
30.00 8.60 8.68 -0.09 8.58 0,01 8.63 -0.03
45.00 9.88 9.83 0.O6 9.78 0.10 9.70 0.19
60.00 10.79 10.85 -0.05 10.06 -0.06 10.68 0,11
75.00 11.81 11.79 0.02 11.85 -0.05 11.61 0.20
90.00 12.92 12.67 0.25 12.79 0.13 12.49 0.43
150. W 16,09 15.86 0.23 16.14 -0.05 15.72 0.37
200.00 17.98 lB.25 -0.28 18.62 -0,64 18.18 -0.20
250.00 20.42 20.50 -0.08 20.91 -0.48 20.48 -0.06
300.00 22.61 22,65 -0.04 23.06 -0.45 22.68 -0.07
350.00 24.76 24.71 0.05 25.11 -0.35 24.79 -0.03
400.00 26.78 26.72 0.07 27,06 -0.28 26.83 -0.04
450.00 28.68 28.67 0.01 28.94 -0.26 28.80 -0.12
500.00 30! 54 30.58 -0.05 30,76 -0.23 30.73 -0.19
600.00 34.21 34.30 -0.10 34.24 -0.03 34.44 -0.23
700.00 37.89 37.91 -0.02 37.54 0.35 38.01 -0!12
800. 0+3 41.48 41.43 0,05 40.70 0,78 41.45 0.03
900.00 44.84 44.8B -0.04 43.74 1.10 44.79 0.05
1000.00 48.34 48.27 0.07 46.68 1.67 48.04 0.30

!i.
,!:
,[:,
I
,:.
. .,:

I I 1

10 100 1000
Shear Rate, s“’
Fig. l-Vlecoslty VS. shear rate at 800 bar and (from top WIbottom) 25, 50, 80, and 11O°C.
●✎

W 15416

0 I
0 250 500 750 1000 1260 1500 1750 200022502500
Shear Rate, s-’
Fig. 2—Comparlscmof theological models with experimental rheogram.

CASSON MODEL BINGHAM MODEL


High Shear Viscosity Plastic Viscosity

● 25*C
● 50”C
O 80”C
o !.10”C ●

——
b

b



● ■

m

● ■
m 0
■ 0
0 0
0
0 0 c1
0 D ❑ n •1 0 0
0

,,,
I 1 , 1 1
I I I I
0 2004006008001000 20040060080010 10

Pressure (Bar) Pressure (Bar)


Fig. 3–Prnsw9 ●d temperature Lmhsviorof Casson ●nd Rlnghsm parmnetemt
HERSCHEL - B(JLKLEY MODEL
Consistency index Power Law index Yieid Stress
o
0
0
v-
0
0 ● 25°C
m
a 50°c
o m
0 O 80”C 0:
m
o ❑ llo”c
o
* 0 m

‘q : w

● ☛
0: ●

2: I 0
gg c ● ●

?y .]
m
● o– ●
y: . ● c1 =
0
0 ❑
n ✘
0
m 8

o w m o
0

‘L
o. o– .
04 0
0 o
0 ■ c1 El
0 9 D OQ
o q : 0
0 .
0 I I I I 1 I I I
0 200400600 8001( D 0 2004006006001 o 2004006008001000
Pressure (Bar) Pressure (Bar) Pressure (Bar)

F4f, 4—Pre8sum and temperature bhsvior of Hemchel.hlkley pammetcm.

(y
~ octylbenzene
o [ i 1 I I 1 I I I I I

2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3 3.4 : .5
1000/T (K)
F~. S–lm-re dqwxlmccof pressure coeftlclcat for Imu 0118●nd hydrocarbons.
10
4–

:-
Co
g
&_
m
5

q-
0

2.6 2.7 2.6 2.9 3 3.1 3.2 3.3 3.4 3.5


0 100 200 300 400 500 800 700 800 900 10001100
1000/T (K)
Shear Rate, s-’
F@.
6-FMUCNI
PVat1, 200, 400, WO, BOO,●d 1,000 ban four tmnpemlur= at MCh Pr~ura.
Fig. 7-PrediCted and experimental dWOgMM$ at 400 bar ●nd (frOM top to bottom) 25, 50, 80, and 11OQC.

ti
o j
s:
u)
5
%
o :

0
0 100 200 300 400 500 800 700 800 900 10001100
Shear Rate, S-’
Fig. O-PmdktOd and expwlmental ttDOWramsat M-C and (from bottom to top) 200,400,400,800, and 1,000
hr.
$~ 15416

100 200 300 400 500 600 700 800 000 10001100
Shear R@e, s-’
Fig. 9-Comparlbon of modellng vla Elngham ●nd Cas80n pmmetem, conditions of Fig. 7.

0
0 100 200 300 400 500 600 700 800 900 1000 1“
Shear Rate, S-l
F@ 10-ComparisC+I of mod.llng via Bingham ●nd CmaOn pamm.lem at E(I”G 200 ●nd 1,000 Imr.

You might also like