A predictive model for the critical desorption pressure of coalbed methane(CBM)has been successfully devel-oped,which employed a Genetic Algorithm(GA)and Elman neural network.The integrated six reservoir parameters in this model,including minimum horizontal principal stress,reservoir pressure,reservoir temperature,burial depth,gas content and daily water yield before gas breakthrough,were used as input data.And the critical desorption pressure of CBM was used as output result.It is shown that the predictive accuracy of this model is substantiated by a correlation coefficient of 0.99 be-tween predicted and the actual values,coupled with mean absolute percentage error of 10.8% .
维普
万方数据