论文部分内容阅读
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。
The occurrence of mine gas outburst is a catastrophic behavior of a nonlinear system in the process of space-time evolution. There is a complicated nonlinear mapping relationship between the prominent basic elements and the prominent danger. The traditional mathematical methods have limitations for dealing with such non-linear spatiotemporal evolutionary problems. In order to predict the gas emission of mine well, the gray theory is introduced into the genetic neural network with high prediction accuracy. The gray theory is combined with the genetic neural network. Based on the neural network theory, the genetic algorithm is used to optimize the hidden layer neural Yuan and connections in the network weights and use it to establish a new model of gas emission prediction. Based on the laboratory test data, a set of training and test samples of genetic neural network is established, and the test results are compared respectively with those of the standard BP neural network.