Construction of tertiary oil recovery dynamic analysis system based on numerical simulation and data driving

Tertiary oil recovery technology involves complex physical and chemical processes and changeable geological conditions, which brings great challenges to reservoir performance analysis. By combining the accuracy of numerical simulation with the intelligence of data drive, this system realizes a comprehensive and accurate analysis of tertiary oil recovery process. The system architecture includes data input module, numerical simulation module, data drive module, result output module and system integration and control module. The numerical simulation module simulates the fluid flow and pressure distribution of oil reservoir based on partial differential equations, and solves it by finite volume method. The data-driven module analyzes and predicts the simulation results through machine learning technology, and optimizes decision-making by combining the neural network guided by physics and reinforcement learning. The case analysis shows that the system has high accuracy and reliability in predicting water cut and oil increase, and is especially suitable for the simulation and prediction of high permeability homogeneous reservoirs.