Prediction of Sandstone Porosity and Permeability Based on Electrical Parameters and GA-BP Neural Network
Taking the Sichuan Basin with complex surface lithology as the sampling area, the neural network can be used to solve the problem of nonlinear mapping between electrical parameters and reservoir physical parameters, and the prediction method of sandstone porosity and permeability in low-porosity and low-permeability tight reser-voirs is discussed. The electrical parameters such as density, resistivity and polarizability were used as the input parameters of the network model, and the weights and thresholds of the BP neural network (BPNN) were optimized by genetic algorithm (GA), and then the GA-BP neural network model was established and trained. Compared with the traditional multiple regression method, the results of the porosity prediction experiment of the GA-BP neu-ral network model are better.
万方数据
维普
