论文部分内容阅读
矿井突(涌)水水源的快速识别是矿井水害有效防治的前提条件.为了更有效地区分潘三煤矿B8、C13组煤系突水水源,利用支持向量机(SVM)建立水源判别模型,并将其与模式识别领域发展比较成熟的BP神经网络判别模型对比,发现SVM法能够将煤系B8、C13组混和水源快速、有效地分开.研究结果表明:SVM法的分类函数结构简单,运算速度快,解决了在BP神经网络方法中无法避免的局部极值问题,对于B8、C13组煤系突水水源的区分有更好的适用性和优越性,为矿井水害防治提供了一种辅助决策手段.
In order to distinguish the coal mine water inrush from B8 and C13 in Pansi coal mine more effectively, a water source discriminant model based on support vector machine (SVM) Compared with the BP neural network discriminant model developed more maturely in the field of pattern recognition, it is found that the SVM method can separate the mixed water sources of coal series B8 and C13 quickly and effectively. The results show that the classification function of SVM method is simple in structure and computing speed Fast and solves the local extreme problem that can not be avoided in the BP neural network method and has better applicability and superiority for distinguishing coal mine water inrush from B8 and C13 group and provides a kind of auxiliary decision for mine water damage prevention and control means.