基于BP神经网络的储层渗透率预测方法研究OACSTPCD

The traditional indirect interpretation method mainly uses linear regression to predict the parameters,but this meth-od has a great disadvantage,because all the data are not necessarily linear,so the linear regression prediction accuracy varies great-ly.Aiming at this problem,this paper proposes a reservoir permeability prediction method based on BP neural network to predict most of the non-linear relationship data existing in the reservoir.It first cleans the selected original logging data and normalizes the data.Then,the BP neural network algorithm is used to analyze and calculate the data characteristics,and finally the lithological pro-file interpretation data is used to verify the prediction results.In this paper,the method of predicting reservoir permeability based on BP neural network and logging curves is used to conduct experiments on a well in the Songliao Basin structural area.The average rel-ative permeability prediction error is reduced,the accuracy is greatly improved,and the logging interpretation is satisfied.Layer pa-rameter accuracy requirements.