Aiming at the problems existing in polymer flooding development of three types of oil layers (low permeability, strong heterogeneity and thin interbeds), such as uneven displacement, lagging profile control and dependence on experience in parameter optimization, this paper proposes a data-driven integrated intelligent development method of "monitoring-analysis-decision". By constructing a real-time monitoring index system covering micro viscoelastic loss coefficient and macro sweep efficiency, and integrating multi-source data such as distributed fiber optic sensing (DTS/DAS) and inter well tracer, a high-precision downhole twin model is established, Furthermore, the ensemble Kalman filter (EnKF) is used to dynamically modify the numerical simulation model, achieving real-time updates of the permeability field and polymer viscosity. On this basis, a closed-loop intelligent decision-making system based on deep reinforcement learning (DRL) is designed, and the online adaptive optimization of polymer injection concentration and slug size is realized with the goal of maximizing oil recovery. The field test results show that the sweep efficiency is increased from 52.1% to 68.5%, the oil recovery value is increased by 2.9 percentage points, and the polymer utilization rate is increased by 22%, which is significantly better than the traditional empirical parameter adjustment method. The research provides an intelligent technical path that can be popularized for efficient development of three types of reservoirs.
Xuewei Yang (Sun,) studied this question.