A Hybrid Approach Based on Artificial Neural Network and Integrated Production Modeling for Gas Lift Optimization

Abstract

An artificial neural network model was developed to predict the values of the bottom hole flowing pressure and the total fluid rate per each well using the available field parameters like the water cut samples, static pressure surveys, reservoir gas oil ratio, the well head temperature and pressure in addition to the gas injection rate and gas injection pressure. This developed ANN used in building accurate individual well models on PROSPER and a full field network model gathering all the individuals' models with the surface network. This creates an integrated production model aiming to perform field wide gas lift optimization.