基于改进算法的含水层储气库多周期注采参数优化方法
By constructing a numerical simulation model of gas-water two-phase seepage and a deep reinforcement learning algorithm, and by combining cumulative plastic strain and fatigue damage indices to optimize injection and production parameters, the problems of high cost and low efficiency in multi-cycle injection and production in aquifer gas storage have been solved, and efficient optimization and stable operation of multi-cycle injection and production have been achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA UNIV OF PETROLEUM (EAST CHINA)
- Filing Date
- 2026-05-18
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies for multi-cycle injection and production optimization in aquifer gas storage facilities suffer from several problems, including high cost of high-precision numerical simulation, low efficiency of manual trial calculations, difficulty in reflecting long-term integrity risks due to static pressure constraints, and difficulty in coordinating the optimization of multi-cycle injection and production parameters.
A numerical simulation model of gas-water two-phase seepage was constructed using an improved algorithm. The model was fitted with historical injection and production data, and an operating pressure constraint that evolves with the cycle was established. The dynamic pressure upper limit was determined by combining cumulative plastic strain and fatigue damage indicators. The numerical simulation surrogate model was trained using a convolutional long short-term memory network, and an embedded optimization framework was introduced into the deep reinforcement learning algorithm to construct a reward function to optimize the injection and production parameters.
It achieves strict pressure upper limit constraint during multi-cycle injection and production, reduces the cost of frequent calls to high-precision numerical simulation, realizes second-level prediction of key response quantities, balances short-term output and long-term stability, and optimizes the comprehensive benefits of multi-cycle.
Smart Images
Figure CN122221697B_ABST