A New Correlation for Gas-condensate Dewpoint Pressure Prediction

A New Correlation for Gas-condensate Dewpoint Pressure Prediction

A.A. Humoud, SPE, Saudi Aramco, and M.A. Al-Marhoun, SPE, King Fahd U. of Petroleum and Minerals

Abstract

The paper presents a new empirical correlation to predict the

dewpoint pressure of gas-condensate fluids from readily

available field data. The new correlation relates the dewpoint

pressure of a gas-condensate fluid directly to its reservoir

temperature, pseudoreduced pressure and temperature, primary

separator gas-oil ratio, the primary separator pressure and

temperature, and relative densities of separator gas and

heptanes-plus fraction. The correlation was developed based

on field and laboratory PVT analysis data of several gascondensate

fluid samples representing different gas reservoirs

in the Middle East. Additional data sets, not included in the

development of this correlation, were used to validate the new

model’s accuracy. Based on the error statistical analysis

results, the new model outperforms the existing correlations.

Introduction

The dewpoint of a gas-condensate fluid occurs when a gas

mixture containing heavy hydrocarbons is depressured until a

liquid is formed. The dewpoint is defined by a substantial

amount of gas phase that exists in equilibrium with an

infinitesimal amount of liquid phase. The determination of

gas-condensate dewpoint pressure is essential for fluid

characterization, gas reservoir performance calculations, and

for the design of production systems.

Traditionally, the dewpoint pressure of a gas-condensate

fluid is experimentally determined in the laboratory in a

process called constant mass expansion (CME) test using a

visual window-type PVT cell. The laboratory measurement of

the dewpoint pressure provides the most accurate and reliable

determination. However, due to the following economical and

technical reasons, quite often this information cannot be

obtained from laboratory measurements: (1) the laboratory

analysis can be expensive and cumbersome, (2) inability to

obtain a representative sample, (3) sample volume is

insufficient to obtain complete analysis, and (4) laboratory

analyses are in error. In this case, the dewpoint pressure values

can be predicted from empirically derived correlations.

The dewpoint pressure correlations, proposed in the

literature, are considered very limited. In addition, these

correlations were developed based on gas-condensate fluid

samples obtained from certain reservoirs of specific regions in

the world. Due to varying compositions of gas-condensate

fluids from reservoirs of different regions, different empirical

correlations may not provide good predictions of dewpoint

pressures when they are applied to gas-condensate fluids

behaving differently from the fluids based on which they were

developed. Most of these empirical correlations are strongly

relating the dewpoint pressure to the gas-condensate fluid

composition. Therefore, there is a great interest to evaluate the

accuracy of these empirical correlations in predicting the

dewpoint pressures of the Middle East gas-condensate fluids.

This paper presents a new empirical correlation for

predicting dewpoint pressure of gas-condensate systems

exclusively for the Middle East gas-condensate fluids using

multiple linear/nonlinear regression procedures. In addition,

the accuracy of some empirically derived dewpoint pressure

correlations was evaluated to determine their applicability for

the gas-condensate fluids prevailing in the Middle East.

Review of Literature

In 1947, Sage and Olds 1 studied experimentally the behavior

of five paired samples of oil and gas obtained from wells in

San Joacuin fields in California. Their investigations resulted

in developing a rough correlation relating the retrograde

dewpoint pressure to the gas-oil ratio, temperature and stocktank

API oil gravity. The results of this correlation were

presented in tabulated and graphical forms. This correlation is

applicable only for gas-oil ratio of 15,000-40,000 scf/STB, for

temperature of 100-220oF, and for API oil gravity of 52o-64o.

In 1952, Organick and Golding 2 presented a correlation to

predict saturation pressures, which could be a dewpoint or a

bubble point pressure, for gas-condensate and volatile oil

reservoir fluids. Saturation pressure is related directly to the

chemical composition of the mixtures with the aid of twogeneralized

composition characteristics: (1) the molal average

boiling point ( B ) in oR, and (2) the modified average

equivalent molecular weight (Wm). These parameters can be

calculated from the composition of the gas mixture. The

correlation was given in the form of 14 working charts, and on

each chart the saturation pressure is plotted against

temperature. Each chart is for a specific value of Wm and gives

a set of curves representing different values of B .

In 1967, Nemeth and Kennedy 3 developed a correlation in

the form of an equation, which relates the dewpoint pressure

of a gas-condensate fluid to its chemical composition,

temperature and characteristics of C7+. The final form of the

equation contains eleven constants; See the Appendix. The

dewpoint pressure and temperature ranges varied from 1,270-

10,790 psi, and 40-320oF respectively. The average absolute

error for the 579 experimental data points used to develop this

correlation was found to be 7.4%.

In 1996, Potsch and Braeuer 4 presented a graphical

method for determining the dewpoint pressure as a backup for

the laboratory visual reading of dewpoint pressure during a

CME test. The key idea of this method is to plot the number of

moles, calculated as a function of single-phase compressibility

factor (Z-factor), versus pressure. Above dewpoint pressure,

the plot yields a straight line, and below dewpoint pressure the

plot shows a curve. The point of intersection marks the

dewpoint pressure.

Data Acquisition

Seventy-four (74) data sets, representing different gascondensate

fluids of the Middle East, were acquired and made

available for this research. Each data set included field

production data, fluid compositional data, and the results of

CME test. The field production data included the reservoir

pressure and temperature, the operating pressure and

temperature of the primary separator, the primary separator

gas-oil ratio, the separator gas specific gravity and the

heptanes-plus specific gravity. The results of compositional

analysis were utilized to estimate the pseudocritical properties

of the gas-condensate fluids. All the dewpoint pressure values

used in this study were experimentally determined from the

CME tests performed on the gas-condensate fluids.

Table 1 lists the ranges of main parameters of the Middle

East gas-condensate fluids. Wide ranges of dewpoint pressure,

temperature and gas-oil ratio were covered. Lean and rich

gases with high concentration of acid gas were also covered in

this study.

Development of Correlation

Multiple least-square linear/nonlinear regression was utilized

to develop this new dewpoint pressure correlation. Statistical

regression programs were developed to build the models,

investigate the behavior of many regression models and

evaluate each combination of the dependent variable with the

independent variables.

The first step in developing the correlation was to select

the parameters that are anticipated to influence the dewpoint

pressure behavior. Two types of independent variables were

used to develop the new model. The first type of independent

variables was based on the pressure and temperature

conditions, while the second type of independent variables was

related to the gas-condensate fluid composition.

Comparison of Correlations

The data sets used to develop the new correlation were utilized

to evaluate the accuracy of two existing dewpoint pressure

correlations: (1) the Organick and Golding correlation, and (2)

the Nemeth and Kennedy correlation. Both statistical and

graphical means were used in this comparative evaluation.

Validation of New Correlation

In order to examine the applicability and reliability of the

newly developed correlation, it was validated using 20 data

sets for the Middle East gas-condensate fluids that were not

used in the development of the new correlation. The new

correlation was also compared against the two existing

correlations.

Table 3 summarizes the results of error analyses of the

three correlations. The table shows that the Organick and

Golding correlation has the highest average absolute and

maximum relative errors. The Nemeth and Kennedy

correlation shows a better accuracy than the Organick and

Golding correlation.

The accuracy of the new correlation was tested using two

different methods for estimation of pseudocritical pressure and

temperature, based on the available fluid composition (Eqs. 6

and 7) and the gas specific gravity correlation (Eqs. 8 and 9)

for unknown fluid composition. The predictions of the new

correlation for the dewpoint pressure were found within less

than 4% average absolute error, and with a maximum relative

error of less than 10% as shown in Table 3.

Conclusions

1. A new empirical correlation to predict the gascondensate

dewpoint pressure has been developed.

2. The developmental approach of this new correlation

is based on some readily available gas-condensate

properties and parameters. Most of these parameters

were not considered in the existing correlations.

3. Comparative evaluation of existing correlations was

made using statistical and graphical error analyses. It

showed that the new correlation outperforms the

existing correlations.

4. The empirical correlation developed in this study was

validated by data, which were not used in the

development of this correlation, and the model

provided a better accuracy than the existing

correlations.

5. The new correlation is considered more relevant and

applicable to the Middle East gas-condensate

properties and conditions than the existing

correlations since the new model has been developed

based on a large number of data sets for the Middle

East gas-condensate fluid samples.

6. New correlations to estimate pseudocritical properties

as a function of gas relative density were developed

based on Middle East gas-condensate data. Higher

accuracies were obtained when these correlations are

used instead of Standing’s ones.