摘要
煤矿瓦斯涌出量和瓦斯突出受控于多种因素。如何根据各个影响因素预测计算煤层瓦斯含量是一个非常复杂的问题。近年来迅速发展起来的神经网络具有较高的非线性映射和并行处理能力。广义回归神经网络(GRNN)具有网络结构自适应确定、输出与初始权值无关等优良特性,能够逼近任意连续的非线性函数,可以处理系统内在的难以解析的规律。本文以某矿13-1煤层为研究对象,在分析影响煤层瓦斯含量的各种地质因素和量化定性因素的基础上,应用GRNN神经网络方法建立某矿13-1煤层瓦斯含量预测模型,以达到对井田未开采区域进行瓦斯含量预测的目的。
The coal mine gas emission and gas outbttrst are affacted by a variety of factors. How to predict coal seam gas content according to various factors is a complex issue. Developing rapidly in recent years, neu- ral networks have a high nonlinear mapping and parallel processing capabilities. Generalized regression neural network （GRNN） with a adaptive network structure, the output has nothing to do with the initial weights and other excellent characteristics, which can approximate any continuous nonlinear function, it can handle the in- herent difficulty parsing rule. This paper make a mine 13 - 1 coal seam as the object, in analyzing the impact of coal seam gas content in a variety of geological factors and quantify the qualitative factors, which is based on the application of GRNN neural network method to establish a mine 13 - 1 coal seam gas content prediction model, in order to realize the purposes of making the gas content forecasted.
出处
《中国煤层气》
2010年第1期37-41,共5页
China Coalbed Methane
作者简介
王文才，男，内蒙古伊金霍洛旗人，教授，博士，硕士研究生导师，现从事安全工程及矿业技术经济的教学和研究。
