GRNN模型在煤与瓦斯突出及瓦斯含量预测中的应用  

Application of GRNN Model in Predicting Coal and Gas Outburst and Gas Content

作　　者：付小平[1] 薛新华[2] 李洪涛[2] 

机构地区：[1]华信邮电咨询设计研究院有限公司 [2]四川大学水利水电学院

出　　处：《中国安全科学学报》2012年第1期24-28,共5页China Safety Science Journal

基　　金：国家自然科学基金资助(51009104)

摘　　要：煤与瓦斯突出的作用机理非常复杂,是诸多因素如地应力、煤层瓦斯、煤体物理力学性质等共同作用的结果。在分析广义回归神经网络(GRNN)的基本原理和算法的基础上,建立煤与瓦斯突出等级以及基于构造复杂程度定量评价的瓦斯含量GRNN模型。然后用收集到的工程实例样本训练和检验该模型。结果表明,GRNN模型具有很好的预测能力和泛化能力,能较好揭示瓦斯含量和诸影响因素间的关系,可用于煤与瓦斯突出判别以及瓦斯含量预测。同时可以看出,光滑因子的合理选取对于提高GRNN模型的预测精度非常重要,因此,在以后的实际应用中需要不断尝试,找出最合理的光滑因子。Coal and gas outburst is a very complicated nonlinear dynamic phenomenon, which is influenced by many factors, such as geo-stress, coal seam gas, physical and mechanical properties of coal. So it is necessary to predict coal and gas outburst with a neural network model. A GRNN model for predicting coal and gas outburst was established based on the basic principles and algorithms, and the GRNN was trained and checked with the collected engineering examples. The results show that the GRNN model pres- ents excellent network performances, high prediction accuracy, and can reveal the internal relationship between the content of gas and the various factors. It is a feasible and credible way to predict coal and gas outburst and the gas content. The conclusions provide a reference to the prediction of coal and gas outburst. At the same time, the smooth factor is import in improving the forecast precision of GRNN model, so it is necessary to find out the reasonable value of smooth factor in the future.

关 键 词：煤与瓦斯突出 构造复杂程度 瓦斯含量 预测 广义回归神经网络(GRNN) 

分 类 号：X936[环境科学与工程—安全科学]