人工神经网络在瓦斯涌出量预测中的应用  

Application of Artificial Neural Network to Forecast of Gas Emission

作　　者：安鸿涛[1] 宋国文[2] 张云中[3] 阳结华[4] 

机构地区：[1]河南理工大学资源环境学院,河南焦作454003 [2]铁法煤业集团公司大兴矿,辽宁调兵山112700 [3]河南省煤田地质四队,河南平顶山467000 [4]河南省地矿局第二地质队,河南焦作454000

出　　处：《河南理工大学学报（自然科学版）》2006年第4期275-278,共4页Journal of Henan Polytechnic University（Natural Science）

基　　金："大兴煤矿突出煤层瓦斯地质规律研究"项目资助

摘　　要：当采掘工作面遇有岩浆岩破坏煤系和煤层时,地质条件尤为复杂,采用常规的矿山统计法和瓦斯含量法预测瓦斯涌出量难以取得理想的结果.作者从矿井地质综合分析入手,采用BP神经网络的方法建立了适用于矿井未采区瓦斯涌出量的预测模型,分别用48个4-2煤层、40个7-2煤层钻孔点的煤层瓦斯质量体积、煤层埋藏深度、煤质、火成岩分布、顶底板砂泥岩比值等数据作为输入层,预测地质条件相对复杂矿井的瓦斯涌出量.经已采区实测值与预测值比较分析认为,预测结果可信.When meeting the magmatic rock destroying coal measures and coal seam in the heading face, geological conditions are particularly complex. It is difficult to make ideal result with mine statistical method and method based gas content to predict the emission amount of gas. In the paper, a prediction model of gas gushing is established, which is suitable for unworked country in mine with BP neural network through geological aggregate analysis. Using 48 and 40 number of gas content, coal bed burying depth, coal quality, igneous rock distributing, ratios of sandstone and mud stone in roof or floor, etc, as inputting layer to predict gas emission of mine of relative complex geological conditions. The difference of forecast result and measured value is extremely small, so the model can be believed.

关 键 词：瓦斯涌出量预测 地质条件 神经网络 

分 类 号：TD712.5[矿业工程—矿井通风与安全]