论文部分内容阅读
矿井突水是矿建与生产过程中最具威胁的自然灾害之一,准确判别突水水源是防治水害的关键。选取6种离子的质量浓度作为突水水源的判别因素,将河南省焦作矿区不同水层的39组水化数据以2种样本设计方案进行Elman神经网络模型的构建与检验。以不同的35组水源样品作为训练样本,运用Matlab软件进行Elman神经网络训练,将所建立的判别模型应用于(相应的)4组待测样本的判别,并与DDA、FDA、Bayes三种判别方法的判别结果进行分析比较。2种方案应用结果表明:将具有非线性动态特征的Elman神经网络应用于突水水源判别,在结合相应的水文地质条件前提下,可以准确判断突水来源;矿井多年的开采促使地下各水层水质呈动态变化,Elman神经网络判别模型能够反映这种变化特性,对探寻地下水运移与演化具有一定的应用价值。
Water inrush from a mine is one of the most threatening natural disasters in the process of mine construction and production. Accurate discrimination of water inrush is the key to preventing and controlling water damage. The mass concentration of six ions was chosen as the discriminant of water inrush. The hydration data of 39 groups in different water layers of Jiaozuo Mining Area in Henan Province were constructed and tested by Elman neural network model with two sample designs. Different samples of 35 water samples were used as training samples, and Matlab software was used to train Elman neural networks. The discriminant model was applied to the discriminant analysis of the four samples to be tested and compared with DDA, FDA and Bayes Method of discrimination results were analyzed and compared. The application results of the two schemes show that the Elman neural network with nonlinear dynamic characteristics can be applied to the determination of water inrush, and the water inrush can be accurately judged on the premise of the corresponding hydrogeological conditions. The water quality shows a dynamic change. The Elman neural network discriminant model can reflect this changing characteristic and has certain application value in exploring the groundwater migration and evolution.