煤层气含气量的DBN预测方法研究

Study on DBN Prediction of CBM Content

为了解决目前煤层气含气量测井定量计算中存在的问题,引入了DBN的方法,利用与含气量灰色关联度大于0.8的测井数据及其对应的实验室分析含气量结果,通过讨论限制玻尔兹曼机个数与损失函数的选择,训练了含气量预测神经网络。

In order to solve the problems in the logging quantitative calculation of CBM content,introduces the method of DBN,and uses the logging data and its corresponding CBM content of laboratory analysis that their grey relational grade 0.8,to train DBN.In the process,the selections of restricted Boltzmann machines quantity and the loss function are discussed.