DISCUSSION ON THE LOG INTERPRETATION METHOD OF COALBED GAS CONTENT

Using multivariate linear regression,the relationships between coalbed gas content with coal quality parameters and log curve values are established.The regression equation is valid according to the F test,but there still a larger error between the coalbed gas content estimated by the regression equation and the gas content determined by the desorption of coal sample.Thus,their relationships are discussed further by using the BP neural network.The examples show that the prediction accuracy is higher.