煤体冲击倾向性指标权系数与冲击危险性研究

Study on Index Weighting Coefficients of Bump Tendency and Bump Hazard of Coal Body

为了准确地反映煤体冲击倾向性,统计分析了大量的实验数据,采用遗传算法优化BP神经网络方法确定了冲击能量指数、弹性能量指数、动态破坏时间的冲击倾向性指标权重系数,并通过专家打分法对其进行修正,根据实验数据的统计规律,分别建立了每项指标的模糊隶属函数,从而形成煤体冲击倾向性的模糊综合评价体系,通过与实际结果进行对比表明:基于遗传算法优化的BP神经网络与专家打分法确定的权重具有较好的可信性。

To accurately reflect the bump tendency of coal specimen,a lot of experimental data had been used to conduct statistical analysis.Based on the genetic algorithm,the BP neural network method was optimized,the index weighting coefficients of bump tendency were determined,including bump energy index,elastic energy index and dynamic damage time.Furthermore,the weighting coefficients were corrected by the expert scoring method.According to the statistic law of experimental data,the fuzzy membership functions were respectively established,and the comprehensive fuzzy evaluation system of bump tendency had formed.Compared with the actual condition,the results showed that this method contained the advantages of BP neural network and expert scoring method,which were optimized by the genetic algorithm,it had good credibility.