基于Elman神经网络的煤层气储层产气量动态变化预测研究

Productivity Prediction on Dynamic Change in Coalbed Methane Reservoir Based on Elman Neural Network

以沁水盆地某研究区的排采资料为基础,围绕煤层气储层产气量动态变化预测技术开展研究。煤层气开采过程是一个非常复杂的系统,其产气量受煤层气储层自身特性和排采制度等多种因素的影响,是各种影响因素综合作用的结果,是开采过程中系统内在变化的反映。因此,可以通过研究煤层气产量的历史数据,挖掘产气量随时间变化的内在规律,并对未来的产气量进行预测。基于煤层气储层产气量的历史数据,通过建立Elman神经网络模型预测煤层气储层产气量的动态变化,并通过实际产气量数据验证,该模型具有很好的预测效果。

Based on the productivity data of a research zone in Qinshui Basin,the prediction on productivity dynamic change in the coalbed methane reservoir is systematically studied.Coalbed methane exploitation process is a very complex system,which is affected by many factors such as the characteristics of coalbed methane reservoir and drainage.It is the result of comprehensive actions of various influencing factors and reflects the internal changes of the system in the exploitation process.Therefore,through the study of the historical data of coalbed methane productivity,we can find the inherent law of gas productivity change over time,and predict the future gas productivity.Based on the historical data of coalbed methane productivity,the Elman neural network model is established to predict the dynamic change of gas productivity.And the predicted results of the model are highly interrelated to the actual results.