Implementation of multilayer perceptron(MLP)and radial basis function(RBF)neural networks to predict solution gas-oil ratio of crude oil systems

Exact determination of pressure-volume-temperature(PVT)properties of the reservoir oils is necessary for reservoir calculations,reservoir performance prediction,and the design of optimal production conditions.The objective of this study is to develop intelligent and reliable models based on multilayer perceptron(MLP)and radial basis function(RBF)neural networks for estimating the solution gas–oil ratio as a function of bubble point pressure,reservoir temperature,oil gravity(API),and gas specific gravity.These models were developed and tested using a total of 710 experimental data sets representing the samples of crude oil from various geographical locations around the world.Performance of the developed MLP and RBF models were evaluated and investigated against a number of well-known empirical correlations using statistical and graphical error analyses.The results indicated that the proposed models outperform the considered empirical correlations,providing a strong agreement between predicted and experimental values,However,the developed RBF exhibited higher accuracy and efficiency compared to the proposed MLP model.

Data driven prediction of oil reservoir fluid properties

Accuracy of the fluid property data plays an absolutely pivotal role in the reservoir computational processes.Reliable data can be obtained through various experimental methods,but these methods are very expensive and time consuming.Alternative methods are numerical models.These methods used measured experimental data to develop a representative model for predicting desired parameters.In this study,to predict saturation pressure,oil formation volume factor,and solution gas oil ratio,several Artificial Intelligent(AI)models were developed.582 reported data sets were used as data bank that covers a wide range of fluid properties.Accuracy and reliability of the model was examined by some statistical parameters such as correlation coefficient(R2),average absolute relative deviation(AARD),and root mean square error(RMSE).The results illustrated good accordance between predicted data and target values.The model was also compared with previous works and developed empirical correlations which indicated that it is more reliable than all compared models and correlations.At the end,relevancy factor was calculated for each input parameters to illustrate the impact of different parameters on the predicted values.Relevancy factor showed that in these models,solution gas oil ratio has greatest impact on both saturation pressure and oil formation volume factor.In the other hand,saturation pressure has greatest effect on solution gas oil ratio.

A new heuristic model for estimating the oil formation volume factor

The necessity of oil formation volume factor(Bo)determination does not need to be greatly emphasized.Different types of reservoir oil have specific conditions which impart the hydrocarbon's major properties,among which is the oil formation volume factor.Therefore,it seems imperative to construct a model capable of estimating the value of oil formation volume factor.Previous studies have resulted in a number of correlations for oil formation volume factor estimation;however,a large portion of them do not provide an acceptable accuracy(at least in some range of data)and cause a huge error at these points.Some others are not flexible enough to be tuned for a specific type of reservoir oil and a comprehensive piece of work does not exist as well in order to compare the applicability of the new models for estimating the oil formation volume factor.In this research,a model based on simulated annealing(SA)has been built in terms of temperature,solution gas-oil ratio,and gravity of oil and gas to predict the oil formation volume factor.This model is compared with the models proposed in the most recent studies,which shows the greater performance of the new method.In addition,in this paper the models of the recent years were compared with each other and their applicability were discussed.Aiming to compare the models,420 data points were selected and the estimated values of each model for oil formation volume factor were compared with their experimental ones.