Study on Production Pressure Difference Upper Limit of Oil Well in Low Permeability Reservoir Based on Ensemble Learning

For low-permeability reservoirs developed by waterflooding, the production of some wells decreases rather than increases as the production pressure differential increases, so the study to determine the production pressure difference upper limit of wells can help to improve the production of wells and enhance the extraction effect of low-permeability reservoirs. To address the problems of limited applicability of the inflow performance calculation model and the complexity of the model for practical operation and application, an ensemble learning-based flow pressure prediction model corresponding to the maximum production of a low-permeability reservoir well is established. From the point of view of mechanism model, the influence factors of inflow performance are analyzed, the input parameters of the model are determined, and KNN, SVM and BPNN machine learning models are constructed. In order to improve the robustness of the model, three kinds of ensemble learning models are established based on model fusion theory. Through the calculation and analysis of an example, the prediction effects of the three ensemble learning models are better than those of a single model, and the ensemble learning model based on neural network has the highest prediction accuracy, with a mean relative error of 9.63%. It shows that the ensemble learning model can accurately predict the production pressure difference upper limit of oil wells in low permeability reservoirs and provide technical support for the optimal adjustment of production pressure difference of oil wells in the field.