摘要
矿井瓦斯突出的发生是一个非线性系统在时空演化过程中的灾变行为,影响突出的各个基本因素与突出危险性之间存在复杂的非线性映射关系。对于处理这样的非线性时空演变问题,传统的数学方法是有局限性的。为了更好地预测矿井瓦斯涌出量,将灰色理论引入到预测精度高的遗传神经网络,使灰色理论和遗传神经网络有机结合起来,以神经网络理论为基础,利用遗传算法优化隐含层神经元个数和网络中的连接权值,并用其建立瓦斯涌出量的预测新模型。在实验室测试数据的基础上,建立遗传神经网络训练和检验样本集,并且将检验结果分别与标准BP神经网络的预测结果进行比较。
Mine gas emission system is a complex system involving nonlinear change. The traditional methods for gas emission prediction have a certain limitations. Gas gushing forecasting depends on the establishment of a nonliner functional relationship of many factors in which the accuracy of forecasting model for gas emission is determined by the peculiarities of interaction and coupling between all affecting factors. In order to better predict the mine gas emission, a predicting model with GM (Remodel) and Genetic Neural Network with high precision of forecasting was established by introducing grey theory into the genetic neural network. First of all, grey incidence of the original data series of forecasts and then the data sequence associates with a certain relationship in between. It will affect the gas emission of the various factors related to sequence. As a genetic neural network input samples of their training, the genetic algorithm can be combined organically to the neural network theory. Genetic algorithm was used to optimize the number of hidden neurons and the network connection weights and the gas emission prediction model was thus established. Both subsequent training and examination showed that satisfied forecasting results could be obtained by using this method and it is able to meet the requirement of an exact guidance to the practice.
出处
《安全与环境学报》
CAS
CSCD
北大核心
2011年第2期176-178,共3页
Journal of Safety and Environment
关键词
安全工程
煤与瓦斯突出
非线性特征
灰色理论
遗传神经网络
瓦斯涌出量预测
safety engineering
coal and gas outburst
nonlinearcharacteristics
remodel
genetic neural network
minegas emission forecasting