An Efficient and Robust Saturation Pressure Calculation Algorithm for Petroleum Reservoir Fluids Using a Neural Network

摘要

Saturation pressure is one of the key parameters in hydrocarbon reservoir engineering computations that can be obtained by either empirical correlations or equations of state. In the latter case, one of the greatest challenges in calculation is the selection of a good initial value to start the iteration. In this work, a feedforward multilayer neural network model is introduced to predict a good initial value for bubble-point pressure calculation applying iterative methods. The model was developed by using 411 published data samples from fields in the Middle East. This model provides a prediction of the bubble point with a relative average error of 0.532%, an absolute average error of 3.273%, a standard deviation of 3.417%, and a correlation coefficient of 0.999989, which implies great accuracy.