Optimization of Well Control during Gas Flooding Using the Deep-LSTM-Based Proxy Model: A Case Study in the Baoshaceng Reservoir, Tarim, China

Gas flooding has proven to be a promising method of enhanced oil recovery (EOR) for mature water-flooding reservoirs. The determination optimal well control parameters is an essential step proper and economic development underground hydrocarbon resources using gas injection. Generally, the optimization in requires use compositional numerical simulation forecasting production dynamics, which computationally expensive time-consuming. This paper proposes deep long-short-term memory neural network (Deep-LSTM) as proxy model simulator order accelerate speed. Deep-LSTM was integrated with classical covariance matrix adaptive evolutionary (CMA-ES) algorithm conduct injection flooding. proposed applied Baoshaceng reservoir Tarim oilfield, shows comparable accuracy (with error less than 3%) but significantly improved efficiency (reduced computational duration ~90%) against conventional method.