基于灰色BP神经网络算法的煤层气直井剩余含气量模型

Residual Gas Content Model of Vertical Coalbed Methane Well Based on Grey BP Neural Network Algorithm

以寺河煤矿西二盘区为例,对3#煤层储层地质条件及煤层气井排采数据进行研究,从地质和工程两方面分析了在渗透率基本相同的情况下,影响煤层剩余含气量的主要因素有原始含气量、煤层厚度、煤层埋深、累计产气量、累计产水量、抽采时间等;利用灰色BP神经网络法建立了剩余含气量预测模型,并把预测结果与参数井3#煤层经过5 a抽采的剩余含气量实测数据进行了对比,结果表明:2种预测结果较为接近。

Takes the west second panel of Sihe mine as an example to study the geological data of 3#coal reservoir and the drainage data of CBM wells,and analyze the influence factors and variation law of residual gas from geological and engineering aspects.It is believed that the governing factors affecting the residual gas content include original gas content,thickness of coal seam,buried depth of coal seam,cumulative gas,cumulative water,extraction time,when the study area has the same permeability.Introduces gray BP neural network method which were used to predict the residual gas content and uses it to establish the residual gas content prediction model,which compared with the measured data of residual gas content in the 3#coal seam of the parameter well after 5 years of extraction,the study result has shown that the two prediction results are close.