基于SOM神经网络聚类和灰度TOPSIS评价法的岩爆预测

Rock burst prediction based on SOM neural network clustering and grayscale TOPSIS

随着资源量的不断减少,采矿深度的不断增大,岩爆发生概率也逐渐增大。为对岩爆等级进行精确预测,提出了一种聚类-关联度TOPSIS模型预测岩爆的方法。在综合分析岩爆产生条件的基础上,从应力、岩性、能量三个指标对样本进行归类处理,并将灰色关联度和TOPSIS评价法相结合。该方法可以通过自组织特征映射网络将样本准确分类,同时通过灰色关联度计算不同指标的权值,最后通过TOPSIS评价法对岩爆等级进行判据。该方法使得岩爆预测多信息融合更加客观,可操作性强。对比工程实例,发现SOM神经网络聚类-关联度TOPSIS岩爆预测法模拟计算与工程实例情况基本一致。

With the continuous reduction of resources,the depth of mining continues to increase,and the probability of rock burst increases.In order to accurately predict the level of rock burst,a clustering-correlation TOPSIS model for rock burst prediction is proposed.On the basis of comprehensive analysis of rock burst generation conditions,the samples are classified from three indexes:stress,lithology and energy,and the grey correlation method and TOPSIS evaluation method are combined.The method can accurately classify the samples through the self-organizing feature mapping network,and at the same time quantify the importance of different indicators through the gray correlation degree.Finally,the blasting level is judged by the TOPSIS evaluation method.This method makes the prediction result more objective and accurate and operable through multi-information fusion method.Comparing with engineering examples,it is found that the simulation calculation of SOM neural network clustering-correlation TOPSIS rock burst prediction method is basically the same as that of engineering examples.