神经网络在矿井突水水源判别中的应用

Application of Neural Network in Water Source Distinguishing of Mine Water Inrush

摘要: 提出了一种采用改进的SOM神经网络对矿井突水水源进行判别的方法。该方法把水质中的Na+、K+、Ca2+、Mg2+、Cl-、SO2-4和HCO-3等7种离子的含量作为判断因素，结合改进的SOM神经网络模型，对20个水源样品进行分类。实验结果表明，该方法的误判率为0，能够准确地判别矿井突水水源。

Abstract: The paper proposed a method of using improved SOM neural network to distinguish water source of mine water inrush. The method takes seven irons contents of Na+, K+, Ca2+, Mg2+, Cl-, SO2-4 and HCO-3 in water as distinguished factors and combines with improved SOM neural network model to classify 20 samples of water source. The experiment result showed that mistake rate of the method is 0 and can distinguish water source of mine water inrush accurately.