Muhammad Ali Al-Marhoun

Prediction of Oil PVT Properties Using Neural Networks

Reservoir fluid properties are very important in reservoir engineering computations such as material balance calculations, well test analysis, reserve estimates, and numerical reservoir simulations. Ideally, these properties should be obtained from actual measurements. Quite often, however, these measurements are either not available, or very costly to obtain. In such cases, empirically derived correlations are used to predict the needed properties. All computations, therefore, will depend on the accuracy of the correlations used for predicting the fluid properties. This study presents Artificial Neural Networks (ANN) model for predicting the formation volume factor at the bubble point pressure. The model is developed using 803 published data from the Middle East, Malaysia, Colombia, and Gulf of Mexico fields. One-half of the data was used to train the ANN models, one quarter to cross-validate the relationships established during the training process and the remaining one quarter to test the models to evaluate their accuracy and trend stability. The results show that the developed model provides better predictions and higher accuracy than the published empirical correlations. The present model provides predictions of the formation volume factor at the bubble point pressure with an absolute average percent error of 1.789%, a standard deviation of 2.2053% and correlation coefficient of 0.988. Trend tests were performed to check the behavior of the predicted values of Bob for any change in reservoir temperature, Gas Oil Ratio (GOR), gas gravity and oil gravity. The trends were found to obey the physical laws.

Evaluation of empirically derived PVT properties for Pakistani crude oils

This study evaluates the most frequently used pressure-volume-temperature (PW) empirical correlations for Pakistani crude oil samples. The evaluation is performed by using an unpublished data set of 22 bottomhole fluid samples collected from different locations in Pakistan. Based on statistical error analysis, suitable correlations for field applications are recommended for estimating bubblepoint pressure, oil formation volume factor (PVF), oil compressibility and oil viscosity.

Artificial Neural Networks Models for Predicting PVT Properties of Oil Field Brines

Knowledge of chemical and physical properties of formation water is very important in various reservoir engineering computations especially in water flooding and production. Ideally, those data should be obtained experimentally. On some occasions, these data are not either available or reliable; then, empirically derived correlations are used to predict brine PVT properties. These correlations offer a handy and an acceptable approximation of formation water properties. However, the success of such correlations in prediction depends mainly on the range of data at which they were originally developed. These correlations were developed using linear, non-linear, multiple regression or graphical techniques. Recently, researchers utilized artificial neural networks (ANN) to develop more accurate oil PVT correlations. The developed models outperformed the existing correlations. However, there is no similar research done so far to utilize the power of ANN in developing similar models for formation waters. In the present study, two new models were developed to predict different brine properties. The first model predicts brine density, formation volume factor (FVF), and isothermal compressibility as a function of pressure, temperature and salinity. The second model is developed to predict brine viscosity as a function of temperature and salinity only. An attempt was made to develop a comprehensive model to predict all properties in terms of pressure, temperature and salinity. The results were satisfactory for all other properties except for viscosity. This was attributed to the fact that viscosity depends only on temperature and salinity. The models were developed using 1040 published data sets. These data were divided into three groups: training, cross-validation and testing. Radial Basis Functions (RBF) and Multi-layer Preceptor (MLP) neural networks were utilized in this study. Trend tests were performed to ensure that the developed model would follow the physical laws. Results show that the developed models outperform the published correlations in terms of absolute average percent relative error, correlation coefficient and standard deviation. Introduction PVT properties of oil field brines are very important in several reservoir engineering computations. These properties include formation volume factor (FVF), isothermal compressibility, density and viscosity. These properties are used in material balance calculations, water flooding, enhanced oil recovery and numerical reservoir simulations. The compressibility of water is a component of the reservoir fluid effective compressibility which is used in material balance calculations. Water formation volume factor is used in both material balance calculations and en evaluating wateroil ratios. Density and viscosity are widely used in mobility ration determination in water flooding and in reservoir simulation. Ideally, brine PVT data should be obtained from laboratory studies on bottom-hole collected samples. However, in some instances, these data are either not available or reliable; then, empirically derived correlations are used to predict brine PVT properties. There are many empirical correlations for predicting different brine PVT properties, they were developed using linear or non-linear multiple regression or graphical techniques. These correlations offer a handy and an acceptable approximation of formation water properties. However, the success of such correlations in prediction depends mainly on the range of data at which they were originally developed. Recently, researchers utilized artificial neural networks (ANN) to develop more accurate oil PVT correlations. The developed models outperformed the existing correlations. However, there is no similar research done so far to utilize the power of ANN in developing similar models for formation waters. The objective of this study is therefore, to develop new predictive models for brine density, formation volume factor (FVF), isothermal compressibility and viscosity based on Artificial Neural Networks (ANN). Two new models were developed to predict different brine properties. The first model predicts brine density, formation volume factor (FVF), and isothermal compressibility as a function of pressure, temperature and salinity. The second model is developed to predict brine viscosity as a function of temperature and salinity only. Also, developing a comprehensive model to predict all properties in terms of pressure, temperature and salinity was attempted. However, except for viscosity, results were satisfactory for all other properties. This is because viscosity depends only on temperature and salinity.

The Coefficient of Isothermal Compressibility of Black Oils

This paper presents a new correlation for the coefficient of isothermal compressibility of black oils at pressures above the bubble point. The correlation is expressed as an empirical function of oil relative density at bubble point pressure, reservoir temperature, bubble point pressure and the reservoir pressure. A total of 3412 data points from 186 laboratory PVT analyses from Middle East fields were used to develop the oil compressibility correlation. The data encompassed a wide range of gas-oil ratios, oil and gas relative densities, reservoir pressure, and reservoir temperature. Multiple linear and nonlinear regressions were used to develop this model. This model is chosen from large number of models tested. The evaluation of the correlation includes comparative studies of the new and existing mathematical models of oil compressibility correlation. The model is validated by using three different data sets from other geographical region of the world not used in the development of the model. The newly developed mathematical model and correlation outperforms the existing mathematical models and correlations for oil compressibility based on low value of average absolute percent relative error and standard deviation. Background The isothermal oil compressibility is an important physical property in the design of high-pressure surface equipment and reservoir calculations. Higher accuracy of oil compressibility estimates will improve the accuracy of the design of highpressure surface equipment and material balance calculations. The isothermal oil compressibility is defined as the unit change of volume with pressure. It is a point function as shown:

Evaluation of empirical correlations for bubblepoint oil formation volume factor

Abstract This paper evaluates several empirical correlations for estimating the bubblepoint oil formation volume factor (FVF) for worldwide application. A total of 674 experimentally obtained pressure-volume-temperature (PVT) data gathered from different published sources is used for analysis of correlated parameters of physical properties and for comparison of the accuracy of correlations. A literature survey of empirical correlations for predicting bubblepoint oil FVF is provided along with their limit of applicability. The statistical and graphical results for the data used in this study show that some of the correlations were violating the physical behavior of bubblepoint oil FVF as a function of the gas relative density. Also, all the available correlations do not adequately represent the contribution of solution gas-oil ratio and temperature at their higher values.

Waterflooding in a tarmat reservoir laboratory model

Abstract The existence of natural barriers such as tar deposits in oil reservoirs can create problems in primary oil recovery as well as in the application of EOR methods. Significant reduction in oil recovery is reported from this type of reservoirs due to isolation of the oil zone from the adjacent water aquifer. In this study, the effects of tar viscosity and thickness of a tar zone on oil recovery as well as the pressure variation and average water saturation in the tar and oil zones were studied in a tarmat reservoir laboratory model. Waterflooding experiments were conducted, whereby the three adjacent oil, tar and water zones were simulated by means of a berea composite core saturated with kerosene, a blend of asphalt and crude oil and KCl brine, respectively. In every experiment, brine was injected at a constant rate in the water zone and was forced to penetrate the tar zone to flood the oil zone. The results show a slight decrease in oil recovery as the product of the viscosity and the thickness of the tar zone increases. An opposite and more pronounced trend was found for the average water saturation in the tar zone. The injection pressure was found to go through a maximum shortly after commencement of injection and the maximum value increased with both tar viscosity and tar zone thickness. On the other hand, the effective permeability to water was found to be smaller in tests where the product of the tar viscosity and tar zone thickness is higher. Finally, the water saturation distribution in the oil zone combined with the pressure behavior points to the development of water fingers in both tar and oil zones.

Treatment of drilling fluid to combat drill pipe corrosion

Abstract The rate of corrosion of drill pipe due to different drilling fluid compositions has been studied. Polymers used to formulate different fluid samples are: cellulose ethers, lignosulfonates and synthetic copolymers. Samples are stressed chemically by adding gypsum and lime and heated up to a temperature of 230°C. Both electrochemical and weight-loss techniques were used to study the corrosion rate of the steel coupons. Results have shown that drilling fluids containing cellulose ethers and lignosulfonates exhibit severe corrosion rate of the steel coupons under bottom hole conditions. However, synthetic copolymers such as, vinyl sulfonate vinyl amide and styrene sulfonate maleic acid maintain the corrosion rate of the steel coupons to an acceptable level.

Prediction of Bubble Point Pressure From Composition of Black Oils Using Artificial Neural Network

In the present study, an artificial neural network (ANN) constitutive model was developed to predict bubble point pressure for the case of Canadian data. The accuracy of prediction of bubble point pressure was compared using two sets of inputs to the model. One was based on composition of the oil and the other based on easily available parameters such as solution gas-oil ratio, reservoir temperature, oil gravity, and gas relative density. The performance of bubble point pressure prediction with ANN was compared with that of equation of state (EOS) and other available empirical correlations. It was found that ANN models can produce a more accurate prediction of bubble point pressure than the existing empirical correlations and EOS calculations.

Application of Neural Network for Two Phase Flow Through Chokes

This study shows the utilization of the Artificial Neural Network (ANN) as a practical engineering tool for estimating the flow rate and selecting the optimal choke size. In this study, the existing choke correlations available in the literature were reviewed, evaluated and compared with the newly derived ANN. The new method can be used to predict the required choke size and can also be used to provide a quick and accurate evaluation of the well performance, by considering wellhead conditions and pressure-volume-temperature (PVT) parameters. Two models were developed based on 4,031 data points: 80% for training, 10% for validation and 10% for testing. The new models were found to outperform all the existing correlations and have provided the lowest error, with an average absolute percent error of 3.7% for the choke size prediction and 6.7% for the flow rate estimation. The new models can estimate with a higher accuracy the optimal choke size and flow rate. Therefore, the new models can help advance reservoir management and production operations in the following ways: producing the reservoir at the optimal rate; preventing water or gas coning; maintaining back pressure; and protecting formation and surface equipment from unusual pressure fluctuation.

Isothermal Oil Compressibility Curve Crossing

Abstract Oil compressibility above bubble point pressure is important in reservoir simulation, material balance calculations, design of high-pressure surface-equipment and the interpretation of well test analysis. Accurate calculation of oil compressibility is very important for reservoir evaluation. The oil compressibility above bubble point pressure increases with increasing temperature and decreasing pressure, therefore, curves at the same pressure and different temperatures should not cross each other. The conventional method of obtaining oil compressibility is from individual temperature measurement of pressure-volume data of constant composition expansion test. The oil compressibility determination involves pressure-volume function and its derivative at each temperature. This individual estimation with derivative calculation makes oil compressibility evaluation very sensitive to small derivative change that may lead to an invalid and non physical behavior of curve crossing. This paper presents a special constrained multiple linear regression optimizations that ensures non-crossing of oil compressibility curves above bubble point pressure for multiple temperatures.