基于AM-MCMC的瓦斯危险等级测评

Grade Assessment of Gas Evaluation for Mine Based on AM-MCMC Theory

用相对瓦斯涌出量划分矿井瓦斯等级存在着临界级别模糊性和局限性的缺点,将AM-MCMC算法与BP神经网络相结合,建立测评模型,该模型解决了相邻瓦斯危险级别所存在的模糊性以及训练样本少的问题;结合已有的研究成果,选取最能反映瓦斯危害的5项指标作为基本判别指标,以18组典型的实测数据作为训练样本进行分析计算,所得结果与实际情况基本相符;研究结果表明:基于AM-MCMC的BP神经网络可以应用于煤矿瓦斯危险等级测评当中。

Division of mine gas levels with the relative gas emission index exist the shortcoming of critical level fuzzy and limitations, ap- ply AM--MCMC algorithm and the BP neural network, building assessment model. This model is used to solve the problem of adjacent gas dangerous levels in critical ambiguity and the training sample less. Combining with the existing research productions, five main indexes that most reflect Gas hazards were taken as distinguishing factor of the basic, Eighteen groups of measured data from several typical mines were taken as training samples, and the prediction results were completely consistent with the actual. The research results show that, AM--MC- MC of BP neural network can be used in the mine gas danger level assessment.