Application of MATLAB-based of BP Neural Network in Reservoir Parameters Prediction
Wang Feng · Well Logging Technology · 2009
On the basis of expounding structure of the BP neural network model,and considering the relationship between reservoir the physical properties and well logging data,the authors iterate the constructed BP neural network model by choosing corresponding well logging data and using functions of neural network toolbox offered by MATLAB.This process can establish better nonlinear mapping relationship between reservoir physical parameters(porosity and permeability) and well logging data,with which we can predict the physical parameters of unknown samples in the studied area.Compared with real samples,the predictive accuracy is higher and the BP neural network has potential advantage and application value in the reservoir parameter prediction.