مشخصات پژوهش

خانه /Performance of Multi-Layer ...
عنوان
Performance of Multi-Layer Perceptron Neural Network in Estimating the Minimum Miscibility Pressure of CO2 Compared to Conventional Methods
نوع پژوهش مقالات در همایش ها
کلیدواژه‌ها
Gas injection, Multi-Layer Perceptron , minimum miscibility pressure, carbon dioxide gas, artificial neural network
چکیده
The CO2 gas injection process is considered as one of the most important miscible methods of increasing oil production from reservoirs. Different laboratory and theoretical methods have been presented to determine and calculate the minimum miscibility pressure, but considering the cost of laboratory methods, research on theoretical methods to calculate this parameter continues. In this study, the data obtained from previous studies regarding the minimum miscibility pressure (MMP) have been used. After that, the performance of conventional methods such as Glaso, Johns and Orr, as well as Yellig and Metcalfe method and the perceptron multilayer artificial neural network algorithm, have been evaluated in MMP estimation. The statistical parameters of the correlation coefficient and the average relative error have been used to evaluate the performance of the mentioned methods. The results show that among the conventional methods, the Glaso method with the accuracy of R2=0.8749 is the best method in estimating MMP. This is while the accuracy of the artificial neural network method is equal to R2=0.9495. The results show that the artificial neural network method is more reliable and accurate than conventional methods.
پژوهشگران علی اکبری (نفر اول)، علی رنجبر (نفر دوم)، فاطمه محمدی نیا (نفر سوم)