Prediction of brittleness index based on neural network method——taking part of wells in Area A of x oilfield as an example
The tight sandstone is mainly developed in Chang 62 of Well A,and the brittleness index is directly related to the effect of fracturing and construction,how to establish the brittle index model accurately and efficiently is an urgent problem to be solved in fracturing of oilfield reservoir reconstruction.The traditional experimental mineral analysis method needs a lot of experimental data to establish the model,and the common logging parameter prediction method has a low accuracy.Therefore,this paper chooses the log parameters with high correlation coefficient as the original data of the input layer of the modeling through the linear regression between each logging curve parameter and the brittle index,the weight coefficient of the original data is calculated by the gray correlation method,and the final input layer parameter of the neural network is taken as the prod-uct of the weight coefficient and the weighted average of the original data,the prediction accuracy of the method is 90.24%,which is better than that of the elastic parameter method.It is necessary for the subsequent development and fracturing of the target reservoir in this block.
neural networkbrittle Index predictionlog interpretation
NETL
NSTL
万方数据
