Permeability prediction method based on improved BP neural network
Online published: 2011-02-20
The traditional BP algorithmhas slowconvergence rate, and is easy to fall into local minimum. It is improved based on Kozeny-Carman equation and the study of Yang Zhengming, and a three-layer feedforward BP neural network model for permeability prediction is established bymeans ofMATLAB neural network toolbox. The simulation training of the improved neural network model is carried out. The result shows that the improved model has faster convergence rate and higher accuracy. The values predicted by the model are consistent with the laboratory test values, and the relative error is less than 10%, so it can completelymeet the accuracy demand ofwell site.
YUAN Jianying , FU Suotang , CAO Zhenglin , YAN Cunfeng , ZHANG Shuichang , MA Dade . Permeability prediction method based on improved BP neural network[J]. Lithologic Reservoirs, 2011 , 23(1) : 98 -102 . DOI: 10.3969/j.issn.1673-8926.2011.01.017
/
〈 |
|
〉 |