Soft Computation Application: Utilizing Artificial Neural Network to Predict the Fluid Rate and Bottom Hole Flowing Pressure for Gas-lifted Oil Wells

Abstract

Abstract The fluid rates and bottom-hole flowing pressure of the wells are essential parameters in the petroleum industry. The need of accurate readings of these measurements are necessary for many calculations such as gas-lift optimization, field monitoring and depletion plans. Predicting these parameters without running in hole has a good impact on reducing the intervention risk and on organization financials by saving time and money. Huge number of correlations are used to estimate these parameters. These correlations need the values of different parameters that are not accurately found. Therefore, an artificial neural network (ANN) model was built from exported data set of PROSPER1 software, production logging tool (PLT), and test separator data. The ANN model was trained and tested by the PROSPER1 extracted data. Then, a number of test points gathered from the PLT reports validated the ANN model. The developed ANN model results in an accurate prediction of the well flowing bottom-hole pressure and well fluid rate. These readings of each well are used to build an integrated production model (IPM) using GAP2 software to apply different gas-lift optimization scenarios.