Spe 26604 Ms

Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

Society of Petroleum Engineers

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


SPE 26604

Thermodynamic Measurement and Prediction of Paraffin


Precipitation in Crude Oil
D.O. Erickson, V.G. Niesen, and T.S. Brown, Conoco Inc.
SPE Members

Copyright 1993, Society of Petroleum Engineers, Inc.


II
This paper was prepared for presentation at the 68th Annuai Technical Conference and Exhibition of the Society of Petroleum Engineers held in Houston, Texas, 3-6 October 1993.

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the 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 to pUblication review by Editorial Committees of the Society
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 is presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A. Telex, 163245 SPEUT.

ABSTRACT tuning of the model using binary systems; comparison of the


model to the microscopic cloud points and wax contents; and
Economically and technically viable methods for controlling the usefulness of the model for estimating the effects of light
paraffin buildup in pipelines and wellbores are critical to the ends and mixed fluids.
production of crude oil in deep water and cold environments.
Conoco has identified laboratory test methods and developed
a paraffin prediction program to better quantify the paraffin EXPERIMENTAL TECHNIQUES
(wax) deposition potential of a crude oil and to evaluate or
develop methods of preventing or controlling paraffin
buildup. This paper will describe the thermodynamic paraffin Cloud Point Methods
model and the complementary laboratory test methods. A
companion paper will describe the kinetic deposition model One of the most important measurements that characterize
and the associated laboratory and field data. the waxing potential of fluids is the cloud point. As long as
operating temperatures remain above the cloud point, wax
deposition will not occur. Typical cloud point measurements
INTRODUCTION used in industry are inadequate. The standard ASTM D2500-
88 or IP 219/82 cloud point methods are only applicable for
The inadequacy of the traditional experimental methods for clear fluids since they rely on visual observation of the wax
determining the likelihood of wax deposition in a pipeline or crystals. For dark crude oils, cloud points are typically
wellbore became apparent to Conoco when these methods inferred from viscometry of Differential Scanning Calorimetry
failed to identify the potential for paraffin deposition in a (DSC). Viscometry and DSC estimate the cloud point from
Gulf of Mexico pipeline. Consequently, more accurate properties of the crude oil that change (viscosity or apparent
laboratory and modeling techniques were developed. This heat capacity) as a result of wax precipitation. Often times,
paper will first discuss the inadequacies of traditional the viscosity or DSC data are misinterpreted. A more serious
experimental cioud point method and the comparison of a problem with these methods is that sufficient paraffin must
microscopic cloud point measurement to field occurrences of crystallize before the properties of the crude oil change
wax. Second, the paper will discuss a thermodynamic model enough for a cloud point to be detected. This means that
for predicting the solid-liquid-vapor (SLV) equilibria; the cloud points determined by these methods are almost always

References and illustrations at end of paper.


933
2 THERMODYNAMIC MEASUREMENT AND PREDICTION OF SPE 26604
PARAFFIN PRECIPITATION IN CRUDE OIL

below the true cloud point. It also means that for low-wax- cloud point determined via CSM. Further discussions

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


content oils, a cloud point may go completely undetected, regarding this particular pipeline will be presented in the
leaving engineers unaware of a potential waxing problem. companion paper on the kinetic model for wax deposition3•
Also shown in Figure 2 is the relative estimate of the wax as
Figure 1 is a plot of viscosity versus reciprocal temperature a function of temperature. Wax contents via acetone
for a crude that has a wax content less than 1 weight percent precipitation at different temperatures have been shown to
at 50°F. Typically, the cloud point would be taken to be as agree with wax contents via CSM.
the temperature where the data begins to deviate from
linearity. At the point of deviation, sufficient wax has Table 1 provides comparisons of cloud points obtained by the
precipitated to change the rheology of the crude oil from three methods for a variety of crudes. In almost all cases, it
Newtonian to non-Newtonian (ie., the viscosity becomes has been found that the microscopic cloud point is higher
dependent upon the shear rate). As shown in Figure 1, the than the cloud points by DSC and viscometry method where
data never deviates from linearity so a cloud point is not the microscopic cloud points represent the maximum
detectable. Field operating experience has demonstrated temperature at which wax deposition can occur. Field
conclusively, that this crude generates a significant buildup of verification of the CSM cloud point method is presented in
wax in a pipeline at temperatures greater than 55°F. Table 2. In this case, the cloud points of the oils from a
number of wells were measured. Independently, G/O/W
Conoco has spent considerable effort to review cloud point flowrate data, well geometry, fluid properties, wireline
techniques to identify a more accurate cloud point technique. reports, and measured or simulated temperature profiles were
Based upon a comparison of cloud points with field used to model the wells. The wireline reports gave the depth
observations of wax buildup and with laboratory deposition at which the first wax deposits occurred and the modeling or
measurements, the method identified by Conoco is Cold- measured temperature profiles provided the estimated
Stage Microscopy (CSM). The use of a microscope has been temperature at that depth within 5-1O°F. Table 2 shows that
reported in the literature previouslyl,2. CSM uses a the microscopic cloud points are in excellent agreement with
temperature controlled microscope to visually determine the the depth at which the wax deposition first begins in the wells
temperature at which paraffin first begins to crystallize in the (within lOOF). Only in one case (A-5) were solid deposits
oil. The sample is preheated and transferred to a microscope found in the well far below the microscopic cloud point. In
slide and covered with a cover Slip. The sample is viewed that case one must consider the possibility of the solids being
under polarized light so that only crystalline material (wax) either asphaltenes or hydrates. Other wells in this field have
is visible. known occurrences of asphaltenes and hydrates.

Once the sample has been transferred to the slide, the


temperature of the slide is dropped at a rate of 1.0°F/min. DESCRIPTION OF PARAFFIN THERMODYNAMIC
Hardware and software have been developed to scan and MODEL
record the images at 5 minute intervals to determine the area
percent that shows up white. The cloud point is determined Thermodynamic Model of the Solid-Liquid Phase Equilibria
as the temperature at which the first wax crystal appears. for Paraffin Formation
Typically, crystal sizes on the order of 1 micron are detectable
at the cloud point. The area percent of white has been The thermodynamic model used in the Conoco program is a
shown to provide a good, relative estimate of the amount of modification of the model first proposed by Won. 4 The
solid wax formed as a function of temperature. Figure 2 is a principle improvements are: the differentiation between
plot of the cloud point and wax content by CSM for the same normal and iso-paraffin, an identification of the need for
crude oil presented in Figure 1. The clOUd point via CSM is accurate, detailed compositional analyses, a more accurate
74°F in contrast to the the cloud point being non-detectable thermodynamic representation of a heterogeneous solid
via the viscometry method. Cloud points for the same crude mixture and a changing crude oil compOSition as paraffin
were measured by 5 different service companies, one service deposits, and tuning of the model with binary phase behavior
company failed to detect a cloud point, 3 companies reported data. These modifications to the model will be discussed as
cloud points in the range 62-66°F and one reported a cloud well as comparison of the model with laboratory data for
point of 104°F. All but one company could not measure wax crude oils.
deposition at the field operating conditions. Field data for
this crude indicate that wax deposition does begin near the

934
SPE 26604 D.D. ERICKSON, V.G. NIESEN, T.S. BROWN 3

The Canoco model is a modification of the model first temperatures for all of the non-normal paraffins, the
proposed by Won4• Won's model calculates equilibrium following formula was used as an average. The exact formula

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


solid/liquid k-values, ~, as follows: is largely unimportant except in the unusual cases where the
wax content/cloud point is determined by non-normal
paraffins. In order to model the effect that branching has on
Sj l I::.H! T
= _, exp[--' (1--»)
(1) the melting temperature, the model uses the following
function for non-normal components (iso-paraffins, aromatics,
xj ..d RT T{ and naphthenics), where Cj is the carbon number.

80*C.
where Xi is the molar concentration of component i in the
f
Ti,non-normal = Tfnonnal - (90 - I ) (4)
100 + C.I
liquid phase, Si is the molar concentration of component i in
the solid phase, T is the temperature, aH fj is the heat of
fusion of pure component i, Yi L and YiS are the activity
coefficients of component i in the liquid and solid phases
respectively, and R is the ideal gas constant. T j is the
freezing/melting point of the pure component i and, for The need to differentiate between normal and non-normal
normal paraffins; is calculated using the following equation. paraffin is also illustrated in the modeling of Blandol.
Blandol is marketed on the basis of its low cloud point (32°F)
and pour point «32°F). Based on detailed compositional
T,! = 374.5 + 0.02617*MW. - 20172 (2) analysis of this oil out to C:,o, the average carbon number
I MW.I present in Blandol is 21 with only 3 weight percent being
normal paraffin and the rest consisting of cyclic and iso-
paraffin components. Normal paraffin ~l has a melting
Where MWj is the molecular weight of the component i. By point of 105°F. As a result, Won's model predicts a cloud
using the melting point temperature, the heat of fusion is point temperature of 107°F and that Blandol is 100% solid at
calculated as follows: 45°F. Fredenslund, et al. 2 presented a paper that stated that
the problem with Won's model is that aromatic components
are responsible for depressing the cloud point. Fredenslund
(3)
developed a complex activity coefficient model for aromatic/n-
paraffin interaction. However, since Blandol contains almost
no aromatic components, Fredenslund's model predicts
The constant 1.14 is a fitting parameter used to obtain a essentially the same behavior as Won's model. The Canoco
better fit with the experimental data. Won's original model model is able to more correctly predict the waxing behavior
uses a value of 1.0. of Blandol (model cloud point=18°F) becauses it uses the
fact that only 3 weight percent is normal paraffin with the
balance consisting of cyclic and iso-paraffinic components.
Differentiation Between Normal and Iso Paraffins

A critical improvement in the model is the identification of You Must Have Accurate, Detailed Compositional Analyses
the need to differentiate between the normal and iso-
paraffins. Won assumed that all components with the same The critical need to differentiate between the normal
carbon number have the same melting temperature Tf. Table paraffins and the iso-paraffins, aromatics, and naphthenics
3 shows that the melting point of several paraffin compounds. necessarily requires a detailed compositional analysis of the
The melting pOint drops significantly when a branch is added crude oil. Accurate compositions at high carbon numbers (at
to a normal paraffin. s For example, the melting point of least 35) are essential since the solubility of normal paraffins
C43H ss is 185°F and one would expect that a Cs2H 104 would decreases exponentially with increasing carbon number;
have an even higher melting point. However, due to the consequently, the cloud point is determined by the higher
branching in the Cs2H 104 compound, the melting point is carbon number, normal paraffins. Other models attempt to
almost 100°F lower than the melting point for C43 Hss. use compositional analyses without isolating the n-paraffins
Because there is such a wide variation in melting point and without differentiation of the Cw+ fraction.

935
4 THERMODYNAMIC MEASUREMENT AND PREDICTION OF SPE 26604
PARAFFIN PRECIPITATION IN CRUDE OIL

Considerable effort was expended to identify an analytical solvent/wax systems (the same systems as listed in Tables 5
and 6) were obtained from the literature and plotted' in

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


technique capable of providing accurate, reproducible,
detailed compositions. Gas Chromatography/Mass Figure 37'9. These data show experimentally measured cloud
Spectroscopy was initially identified as a passable method for points for binary solvent/wax mixtures as a function of the
model development; however, as the model became more concentration of wax molecules in the liquid phase, in other
fully developed, the need for an even more accurate words, the solubility of wax in solvent. As can be seen, the
compositional analysis became apparent. Recently, sample data lie on straight lines indicating the ideal nature of the
preparation and analytical techniques have been identified solvent/waxsystems. Figure 3 also demonstrates that the
which have been shown to provide reproducible and accurate solubility of n-C;6 is more than a factor of 10 less than that
results. This technique has been tested for approximately 15 of n-~ regardless of the solvent type. The wax components
crudes and condensates of wide ranging wax compositions. (n-~ n-C;2' and n-~) have melting temperatures, 142, 158,
Table 4 provides an example of the compositional analysis for and 169°F respectively. These small differences in the
the crude oil presented in Figure 1. melting point temperature have a much greater effect on wax
solubility than the type of solvent used.
Even with the more accurate techniques, the analyses typically
end around C;s to C4Q because the analytical instrument is not As a consequence of the strong dependence of the ratio of
sensitive enough to accurately detect the presence of higher the temperature to the melting point temperature (as shown
molecular weight compounds present at low concentrations. in Equation 1), the use of an activity coefficient model has
Unfortunately, minute quantities of those higher molecular only a small effect on the cloud point predictions. Tables 5
weight n-paraffin compounds have a profound effect on the and 6 present cloud points for the binary systems predicted
cloud point. The Conoco model estimates the concentration by Won's model (Table 5) and by the Conoco model (Table
of the higher molecular weight n-paraffin compounds by 6). The Conoco model has a lower error using only the heat
addition an n-paraffin tail. It has been shown that the of fusion/melting temperature model from equation 1.
concentrations of the high molecular weight n-paraffins Figures 4 to 6 show that the Conoco model works well for
decrease logarithmically with carbon number. Consequently, most solvent/wax systems with the greatest errors occurring
the Conoco model uses the concentrations of the last known for xylene. The Conoco model cloud point predictions for
carbon numbers to estimate the concentrations of the higher xylene systems at low concentrations are 2 to 3°C too low.
molecular weight n-paraffins via an exponential decay. For This implies that the Conoco model is predicting that xylene
the majority of crudes, an extrapolation out to Cso provides is a better solvent than it has been shown to be
excellent agreement between the Conoco model and the experimentally.
laboratory cloud point measurements. In a few cases for
highly paraffinic crudes with pour points in the range of 80-
100°F, the extrapolation was taken out to ~s. The Conoco Model Has Only A Single Fitted Parameter

The binary experimental data were used to determine the


Why the Conoco Model Does Not Use Activity Coefficients constant (1.14) in equation 3. Unlike other models proposed
in the literature, this is the only adjustable parameter in the
Won used a solubility parameter model to calculate activity Conoco model. The most recently published model by
coefficients. Solubility parameters account for differences in Narayanan and Leontaritis 1o uses 36 binary interaction
attractive forces between the solvent and wax molecules. One coefficients which are determined by tuning the model to
important contribution that is not in Won's model is the experimental cloud point data for crude from a particular
Flory Huggins6 contribution which accounts for the effect of reservoir. Since the Conoco model does not have adjustable
the large size differences between the wax and solvent parameters which are are crude oil specific, predicting the
molecules. Both of these activity coefficient terms (solubility phase behavior is possible for mixtures of crude oils,
parameters and Flory Huggins) are demonstrated to be much condensates, solvents and light ends can be readily predicted.
smaller than the dominant heat of fusion/melting temperature
effect.
Real World Paraffin Deposition Modeled by Staged Equilibria
To demonstrate the dominance of the heat of fusion/melting
temperature effect versus the activity coefficient terms (effect The next modification to the model was to more accurately
of solvent type), cloud point data for various binary represent the heterogeneous solid paraffin mixture and

936
SPE 26604 V.G. NIESEN, T.S. BROWN, D.D. ERICKSON 5

changing crude oil composition as paraffin deposits in a to ~s) can make up to 200 P difference. Figures 1 and 7
pipeline. Normally, thermodynamics assumes that the liquid present comparisons between the laboratory data and the

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


phase is in equilibrium with all of the solid phase and that model prediction for crudes with very different wax contents.
the composition of the solid phase is homogeneous. This As can be seen, not only is there good qualitative agreement
assumption would be valid if all of the wax formed at the between the laboratory data and the model for the cloud
same temperature; however for a crude oil, wax first begins point but also for the solid wax content. Since the
to form 'at the cloud point. The wax precipitating at the microscope is only a relative indicator of the wax content, we
cloud point will have a different composition than wax only expect the model to match the trends and magnitudes of
precipitating at lower temperatures. Normally, the microscopic data.
thermodynamics always assumes that the two phases
(solid/liquid) are in thermodynamic equilibrium. However,
for a pipeline situation, because of the low diffusion Conoco Model Predictions for the Effects of Light Ends Loss
coefficient in solids, it is not likely that the the solid phase
will be homogeneous. Consequently, a staged All of the preViously presented information is based on dead
thermodynamic concept was built into the model which crude samples. However, since the binary experimental data
makes the assumption that once solid forms it does not remix has demonstrated that the SOlid-liquid equilibria is a relatively
with additional solid that precipitates at lower temperatures. ideal system requiring only one fitted parameter, the effects
The staged thermodynamics is a series of flashes and of light ends loss can be easily modeled. Figure 8 is a phase
separations at 1.8°P temperature increments. Staged diagram for a dead crude combined with 10 weight percent
thermodynamics does not change the predicted cloud point, gas. The cloud point decreases as the pressure increases
but at lower temperatures, staged thermodynamics results in because gas dissolves in the liquid phase. This continues
lesser amounts of solid wax forming than for normal until the bubble point is reached. At the bubble point, all of
thermodynamics. the gas has gone into solution and therefore, further increases
in pressure do not add any more solvent to the liquid phase.
There will be some changes in the cloud point at elevated
MODEL PREDICTIONS AND COMPARISONS WITH pressures. The Conoco model does not incorporate the
EXPERIMENTAL DATA effects of pressure at this time.

Comparisons between Predictions and Laboratory Data for Conoco Model Predictions for the Mixing or Dilution of
Real Crudes Crudes

The previous comparisons between the Conoco model and In many cases, pipelines do not transport a single crude;
experimental data were for laboratory prepared binary consequently, the cloud point and wax deposition potential
mixtures. This section compares predictions for real fluids must be measured for possible crude oil blends. Because the
(crudes and condensates) with experimental data using the Conoco model does not need to be tuned to individual crudes
Cold-stage microscope described earlier. It is estimated that via a number of fitted parameters, the mixing of crudes,
the microscope is able to detect cloud points to within +/- solvents, and condensates can easily accomplished. As an
3°P and is reproducible to within +/- 30% in estimating the example, laboratory measurements of cloud points and wax
amount of wax formed as a function of temperature. Table contents were performed for a crude oil diluted at different
1 presents the comparison between experimentally concentrations with xylene. Those results were compared to
determined cloud point measurements and the model cloud the model predictions for the same crude oil and xylene
point predictions for a number of crudes and condensates. mixtures. Pigure 9 compare the laboratory results to the
Using only a compositional analysis, no tuning parameters model.
and a n-paraffin tail out to Cso, in general, the Conoco model
agrees with the microscopic cloud points within lOoP for a Xylene reduces the amount of wax formed for two reasons.
wide variety of condensates and crude oils. In two cases, the The first reason is best illustrated by example. If we had a 10
cloud points are up to 15°P degrees greater than the kg sample containing 10 percent (1 kg) of wax in it and we
microscopic cloud points. For most fluids changing the added 10 kg of xylene, the dilution would be 1 kg of wax in
length of the n-paraffin tail results in only a 0_8°P difference 20 kgs of the mixture resulting in a solid concentration of 5
in cloud point. The exception is for highly paraffinic crudes percent. In other words, the reduction of solid wax is simply
(high pour point) where the length of the tail (say from Cso

937
6 THERMODYNAMIC MEASUREMENT AND PREDICTION OF SPE 26604
PARAFFIN PRECIPITATION IN CRUDE OIL

because of an increase in the overall volume. However, = freezing/melting point of the normal paraffin

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


Ttnorma'
dilution alone would not reduce the cloud point. component i
= molar concentration for component i in the
Xylene also acts as a solvent for the paraffin. This is best liquid phase
illustrated by Figure 10 which shows the amount of wax = heat of fusion of component i
formed on a xylene free basis. In other words, the quantity activity coefficient of component i in the liquid
of wax formed based only on the original 10 kg sample. In phase
this case, the difference between the upper curve for the pure = activity coefficient of component i in the solid
crude oil and the mixture curves is the quantity of wax phase
solvated by the xylene. This figure also shows that as more
xylene is added, the cloud point decreases. If enough xylene
(97% ) were added, then it would be possible to depress the ACKNOWLEDGEMENTS
cloud point to below 40°F. The large amount of xylene
required to keep all the wax in solution helps point out that The authors thank Conoco, Inc. for allowing publication of
wax is relatively insoluble and that solvent washes to remove this work. Thanks to Roger Hudgins, Norman Brown, and
paraffin from pipelines may not always be very effective. Leroy Grimsley for providing volumes of test data required
for the development of the model. Special thanks go to the
field personnel (Fid Maurin and Dave Pharis) who facilitated
CONCLUSIONS the validation of the model by providing us with crude oil
samples and field operating information. Roger Woods
Complementary laboratory analyses and a thermodynamic collaborated in the development of the recent sampling and
model have been developed which provide an excellent analytical techniques for obtaining crude oil compositional
method for predicting cloud points and solid-liquid-vapor analyses. Thanks to Michael Mai for the initial work on this
phase behavior for hydrocarbon systems based solely upon a project and for his and Elwyn Shimoda's continuing support.
detailed compositional analysis. Specifically, the model
differentiates between normal and non-normal paraffins and
consequently, the necessity for detailed compositions out to REFERENCES
at least ~5. By identifying the differences between normal
and non-normal paraffins the model was greatly simplified 1) Ronningsen, H.P. and Bjorndal B., "Wax
which resulted in no adjustable parameters. Lastly, a staged Precipitation from North Sea Crude Oils", Energy &
thermodynamic equilibrium calculation was incorporated to Fuels (1991) 5:895-908.
more accurately represent solid wax formation in the field.
In conjunction with the kinetic model described in a 2) Fredenslund, A, Hansen, J. H., Pederson, K S., and
companion paper, this work provides a means for an Ronnningsen, H. P., "A Thermodynamic Model for
engineering estimate of wax deposition rates, locations, and Predicting Wax Formation in Crude Oil", AIChE
quantities based solely upon a detailed compositional analysis Journal (1988) 34:1937-1942.
of the hydrocarbon system.
3) Brown, T.S., Niesen, V.G., and Erickson, D.D.,
"Measurement and Prediction of the Kinetics of
NOMENCLATURE Paraffin Deposition", Paper SPE 26548 to be
presented at the 1993 SPE Annual Convention,
Cj = carbon number of component i Houston, October 1993.
~ = solidlliquid equilibrium ratio for component i
MWj = molecular weight of component i 4) Won, K W., "Thermodynamics for Solid Solution-
R = ideal gas constant Liquid-Vapor Equilibria: Wax Phase Formation
Sj = molar concentration for component i in the from Heavy Hydrocarbon Mixtures", FLUID PHASE
solid phase EQ (1986) 30:265.
T = temperature
Tf I
= freezing/meeting point of component i 5) Dunstan, AE., Science of Petroleum, Oxford
Ttnon-n,ormal = freezing/melting point of the non-normal University Press., (1938).
paraffin components i

938
SPE 26604 V.G. NIESEN, T.S. BROWN, D.D. ERICKSON 7

6) Flory, P.J., Principles of Polmer Chemistry, Cornell

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


University Press, Ithaca, NY (1953).

7) Madsen, H. E. L. and Boistelle, R "Solubility of


Long-Chain n-Paraffins in Pentane and Heptane," J
Chern. Soc. Faraday Trans1 (1975) 72: 1078.

8) Boistelle, R. and Madsen, H. E. L., "Solubility of


Long Chain n-Alkanes in Petroleum Ether", J. Chern.
Eng. Data (1978) 23:28-29.

9) Belny, D. H. M. and Mullin, J. W., "Solubilities of


Higher Normal Alkanes in m-Xylene," J. Chern. Eng.
Data, (1987) 32:9-10.

10) Narayanan, L., Leontaritis, KJ., and Darby, R,


"Model for Predicting Wax Deposition from Crude
Oils," Paper #55a presented at the 1993 AIChE
Spring National Convention, March 28 - April 1.

939
8 THERMODYNAMIC MEASUREMENT AND PREDICTION OF SPE 26604
PARAFFIN PRECIPITATION IN CRUDE OIL

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


TABLE 1: CLOUD POINTS

DSC (5° C/MIN) VISCOSITY MICROSCOPE MODEL


OIL # (OF) (OF) (OF) (OF)

1 79 77 95-99 110

2 92 86 100 114

3 70 <30' 30
4 88 99 137 128

5 77 81 97 102
6 74 74 77
7 92 119 117
8 77 104 108

9 ND 1 53 63

10 120 122 134

11 70 125 119

12 80 107 104

13 153 170
14 53 ND 1 100 99
15 104 105
16 101 100

* No wax detectable down to 30°F -

1 ND = not detected

940
SPE 26604 D.D. ERICKSON, V.G. NIESEN, T.S. BROWN 9

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


TABLE 2: COMPARISON OF CLOUD POINTS WITH WIRELINE REPORTS

MAXIMUM APPROXIMATE
NUMBER OF REPORTED MAXIMUM DEGREE
WELL REPORTS DEPTH OF TEMPERATURE CLOUD OF WAX
NUMBER OF PARAFFIN DEPOSITS OF DEPOSITS POINT PROBLEM

A-1 4 3000' 111°F 101°F MEDIUM


A-3 10 3500' 101°F 100 of HIGH
A-5 5 2000' 98 of 43 of MEDIUM
A-7 10 3600' 87 of 91°F HIGH
A-9D 12 3000' 95 OF 100 of HIGH
A-lOD 9 3500' 75 of 84 OF HIGH
A-13 2 1500' 78 OF 77 of LOW
A-16 12 4200' 97 of 100 of HIGH
A-19D 9 3000' 83 of 80 of HIGH
A-20 2 2010' 67 of 64 of LOW

TABLE 3: COMPARISON OF MELTING POINTS

HYDROCARBON MELTING-
I I POINT COF) I
Normal tritetracontane, C43H gg
CH3-(CH2)41-CH2 185
Methyl-di-heneicosyl-methane, C44 H 9o
CH3-(CH2)2o-CH-(CH2)2o-CH3 143
I
CH3
Nonyl-di-heneicosyl-methane, Cs2H 104
CH3-(CH2ho-CH-(CH2)20-CH3 91
I
C9H 19

941
10 lliERMODYNAMIC MEASUREMENT AND PREDICTION OF SPE 26604
PARAFFIN PRECIPITATION IN CRUDE OIL

TABLE 4: COMPOSITIONAL ANALYSIS FOR OIL #6

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


CARBON CARBON NUMBER WEIGHT n-PARAFFIN
NUMBER FRACTION FRACTION

3 0.0004 1.0000

4 0.0223 0.2831

5 0.0596 0.3528

6 0.1280 0.2823

7 0.1987 0.2466

8 0.1557 0.3155

9 0.1139 0.2772

10 0.0759 0.3688

11 0.0535 0.4516

12 0.0418 0.4187

13 0.0320 0.4651

14 0.0129 0.8339

15 0.0128 0.8064

16 0.0209 0.4035
17 0.0170 0.4065
18 0.0141 0.3786

19 0.0055 0.5817
20 0.0065 0.6118
21 0.0037 0.5958

22 0.0040 0.4906

23 0.0030 0.5542

24 0.0026 0.5701

25 0.0018 0.6941

26 0.0021 0.4344

27 0.0017 0.4301
28 0.0013 0.5065
29 0.0013 0.4587
30 0.0012 0.5538
31 0.0008 0.4701
32 0.0008 0.5746
33 0.0004 0.7081

34 0.0009 0.6038

35 0.0006 0.7616
36 0.0021 0.0977

942
SPE 26604 D.O. ERICKSON, V.G. NIESEN, T.S. BROWN 11

TABLE 5: COMPARISON BE:lWEEN EXPERIMENTAL AND PREDICTED CLOUD POINTS· WON'S MODEL

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


ERROR IN CLOUD POINT TEMPERATURE PREDICTIONS "F

SOLVENT SOLUTE #POINTS BIAS RMS MEAN

nCS nC28 9 1.60 2.02 1.66

nC7 nC28 14 0.40 1.19 1.04

nC10 nC28 20 1.57 1.84 1.57

nC12 nC28 19 1.21 1.37 1.22

XYLENE nC28 7 3.33 3.82 3.33

nCS nC32 12 -0.38 0.79 0.59

nC7 nC32 11 -0.99 1.10 1.01

XYLENE nC32 6 2.39 3.47 2.88

nCS nC36 17 1.49 1.60 1.49

nC6 nC36 17 0.94 1.13 0.94

nC7 nC36 20 0.68 0.86 0.74

nCB nC36 20 -0.29 0.47 0.38

nC10 nC36 20 0.13 0.36 0.29

nC12 nC36 21 -0.14 0.43 0.29

XYLENE nC36 5 1.44 3.15 2.65

TOTAL 218 0.88 1.64 1.33

TABLE 6: COMPARISON BE:lWEEN EXPERIMENTAL AND PREDICTED CLOUD POINTS· PARAFFIN MODEL WITH NO ACTIVllY COEFFICIENT

ERROR IN CLOUD POINT TEMPERATURE PREDICTIONS "F

SOLVENT SOLUTE #POINTS BIAS RMS MEAN

nCS nC28 9 1.03 1.57 1.26

nCi nC28 14 ·0.09 1.04 0.95

nC10 nC28 20 1.06 1.40 1.15

nC12 nC28 19 0.77 0.99 0.77

XYLENE nC28 7 2.77 3.29 2.77

nCS nC32 12 ·1.01 1.19 0.99

nCi nC32 11 -1.60 1.64 1.60

XYLENE nC32 6 1.76 2.99 2.45

nCS nC36 17 0.76 0.94 0.83

nC6 nC36 17 0.23 0.63 0.45

nC7 nC36 20 -0.02 0.49 0.36

nCB nC36 20 -0.97 1.04 0.97

nC10 nC36 20 -0.65 0.74 0.67

nC12 nC36 21 -0.77 0.88 0.77

XYLENE nC36 5 0.77 2.81 2.43

TOTAL 218 0.27 1.44 1.22

943
12 THERMODYNAMIC MEASUREMENT AND PREDICTION OF SPE 26604
PARAFFIN PRECIPITATION IN CRUDE OIL
1O.,.----:--.,----,----.---r---r---.,-----,------,
I I I I I I I I

========~=======~========~========~========1========1========t========r--------

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


I I I I I I I

--------~-------~--------i--------i--------i--------+--------,- -----~--------
I 1 1 I I I I
1 I I 1 1 I 1 1
--------~-------~--------~--------~--------~--------
I
1
1 1
I
I I
I 1 1
1
I
-------+--------r--------
I
1
I
1

-------~-------~--------~--------~----.
• I I 1 --~--------~--------t--------~--------
1 I I 1
1 I 1 " I I
I 1 I I I 1 1 •
I I I 1 I 1 1
--------r-------i------

l--------1--------1--------1--------t--------~--------
I 1 1 I I 1
1 1 I I 1 I I
I I I • I 1 I
I I I I I I I

en ---~-------~--------~--------i--------1--------t--------t--------r--------
o 1
I
: :
I
I
:
I
1
I
I
:
I
I
:
I
1
:
I
I
:

I
:
o I • I • I I I I
en ________L
I

J•
I
J1
1 1
I
I
I
I
I
1
LI

•L _

:>
~ ~ ~

1 121.0 F .102.8 F 85 .8F


1 69.7F .69.7F
3.1 3. 5 3.2 3.25 3.3 3.35 3.4 3.45 3.5 3.55
1000/T (K)
Figure 1 : Viscosity of Oil #6 at 170/sec
0.4 ,-------r---...,----r----.,...----r-----.

-Model
X
CO • Experiment
;:0.3
....c:
(I)
0

CiiO. 2
a.
....
.c:

C)
-;0.1

;: iII
1
I
I
I
I
1

0.0 -t----+---+----;--'--~~----4--=:::::---1
50 55 60 65 70 75 80
Temperature (OF)
Figure 2: Cloud Point and Wax Content for Oil #6
944
SPE 26604 D.D. ERICKSON, V.G. NIESEN, T.S. BROWN 13

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


-
U.
0
104
95 • C 28
A
A • ...

....c:
'-"
• C 32 • • iii

.-
86
AC36
••
; .. .1)
0
a. 77
+.,.., .. .a-..,JiI
• ••
.
III.
68
••• • .,::
-0

-
:::s
0 59
• •••
(.)
50
•• •
0.00009 0.0009 0.009 0.09
Mole Fraction Wax
Figure 3 : Experimental Cloud Points of Wax in Various Solvents

....r::
.-
o 5.4 - + • Pentane
+
a. + Xylene
-0 3.6 - !A
T Heptane
+
::J_ • 0 •• 0 0 + + • 0 Decane
1.8 -
0
..2 u.. 0 T 'ta o 0 0 . !A n-C 12
(.)0 !AlA tv AI~~~6"'6~!,
:V
c: '-" 0 ·.1/\ ~ "
1\
<Q
IT •T
<II
0

-1.8 - T
T

0.002 0.02 0.2


Mole Fraction Wax
Figure 4 : Predictions for Solvent / n-C28 using Conoco Model

945
14 THERMODYNAMIC MEASUREMENT AND PREDICTION OF SPE 26604
PARAFFIN PRECIPITATION IN CRUDE OIL
.....r::

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


.-o 3.6
+ • Pentane
a. 2.7 - +
1.8 - + Xylene
"0 +
::s v Heptane
0"-"
_U- 0.9 -
O~ o • • •
c -0.9 - •v • • v •... •
v
'-
o'-
-1.8 -
.... ....
v •
" .. -... ....
v
'-
-2.7
w 0.001 0.01
Mole Fraction Wax
Figure 5 : Predictions for Solvent / n-C32 Using Conoco Model

.....c
.-o 3.6 • Pentane
a. + Xylene
"0 +
1.8 o "Iff Heptane
::s
0"-"
_U- ¢ Decane
• ~ n-C
O~ 12
x Octane
.-c'- -1.8
o Hexane
o
'-
'- 0.00007 0.0007 0.007
w
Mole Fraction Wax
Figure 6 : Prediction for Solvent / n-C36 using Conoco Model

946
SPE 26604 D.D. ERICKSON, V.G. NIESEN, T.S. BROWN 15
40 I
I
I
I
I • Oil #13 Microscope
35 ~,~~-=~-----~------------~------------~--

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


>< :I :I :I --Oil #13 Model
ca
* Oil #4 Microscope
I I

3: 30 ____________ I
~------L-----------L-
I I
I I I

"C ~I _Oil #4 Model


-0
-- : :
I I I
I I I

25 I I
------------r------ - . -----------T-- I
X Oil #5 Microscope
en I I

....c I
:
I
I
I
: ··_·_·"Oil #5 Model
Q)
20 ------------r------------l ~----1:--L--------~------------~------~
0
J-
:
I
:.:
I . I I I
Q)
15 -----~------------~------ -Jl-~-----------~-----------~------------
a. I I • I
:
I I
-: :
....
.c
:
I : ..
_______I~___________ IL
:
~
I
:
:
~
I _

0) 10 I I I

--
Q)
I
I ~.
:
I
: . :
I I
:
I

3: 5 ------------~------~**-! ----------+--- • ~~ FL ~------------


~[...=':-"~..~.~. . ~-~~:~~~~~~~~.,;. ::~ :
o40
• • 1.....

6 80
I
I. • _ _
");-,",...,;*~
~ >'~'-*

1do 1 0
: -,...
.. _
:-.....:.... I

160
Temperature ( OF )
Figure 7: Comparison of Model Predictions with Microscope

Figure 8: Crude Oil #1 with 10 Weight Percent Gas


947
16 TI-lERMODYNAMIC MEASUREMENT AND PREDICfION OF SPE 26604
PARAFFIN PRECIPITATION IN CRUDE OIL

Downloaded from http://onepetro.org/SPEATCE/proceedings-pdf/93SPE/All-93SPE/SPE-26604-MS/1983892/spe-26604-ms.pdf/1 by Universidad Industrial De Santander user on 10 November 2022


8
....C • Experimental 25%
A Experimental 50%
Q)>< 6 • Experimental 75%
~ea - Model by Whole Oil 25%
-Model by Whole Oil 50%
Q)3: . ······'Model by Whole Oil 75%
D. 4
....."'C-
.c-
OlO
.-
Q)
en 2
3:
45 55 65 75 85 95 105
Temperature (OF)
Figure 9: Oil #7 with 25, 50, 75% Xylene

.....~20 - , - - - - - - - - - - - - - - - - - - - . - - - - - - - - - - .
Ccn - - 0 % Xylene
Q)ea
Om •••,._,,",W>·k" 25 % Xylene
CD Q) 15
D.Q) ."'-."'-."'-."'-.~ 50 % Xylene
.... 1-
.c LL . - - - 7 5 % Xylene
0l1-10 .
- - G) -..,........
Q) .... ••.".... ~.
;:::.. •.c-
::> -.. .
"'C:E 5 ~ .
•- s:::: ~ .
"0 - ........... .....~=" .... -
en oc 0 J~----,--__r--~-=-:-:;.:-~..:..:;:.~~
-..~..-~. . ~.......
-;;.~-~~~~~--r---._J
..:a.... ~."' .....

35 45 55 65 75 85 95 105 115
Temperature ( OF )
Figure 10 : Calculations for Oil #7 / Xylene Mixtures

948

You might also like