采空区风险评价的改进灰靶模型

Improved Grey Target Model for Risk Evaluation of Goaf

采空区风险评价是矿山安全管理的重要工作之一,灰靶模型是风险评价的有效方法。为克服灰靶模型中单一主观或客观赋权法带来的局限性、未考虑评价指标相关性的问题,采用层次分析法、序关系分析法、熵权法和离差最大化法,构建基于矩估计理论的改进主客观权重集成赋值方法;采用相关系数矩阵替换马氏距离中的协方差矩阵,引入到传统灰靶模型中,准确表征评价指标之间的相关性。将改进的灰靶模型应用到采空区风险评价中,结果表明,改进主客观权重集成赋值方法得到的最优组合权重是以4种赋权方法的偏差最小为目标进行优化集成,兼顾人为和实践获取的经验因素和评价指标的固有属性特征,评价结果更加全面;采用相关系数矩阵替换马氏距离中的协方差矩阵,考虑评价指标之间相关性的影响,提高了评价结果的准确性,验证了改进的灰靶模型在采空区风险评价中的合理性和有效性。

In order to solve the problems of limitations caused by the single subjective or objective weighting method in the gray target model and the problems of correlation of evaluation indicators being out of consideration,the analytic hierarchy process,order relationship analysis method,entropy weight method and dispersion maximization method are used to build moment estimation theory to improve the subjective and objective weight integration assignment method.Then,the correlation coefficient matrix is used to replace the covariance matrix in the Mahalanobis distance,which is introduced into the traditional gray target model to accurately characterize the correlation between evaluation indicators.What’s more,applying the improved gray target model to the goaf risk assessment,the results show that the optimal combination weight obtained by improving the subjective and objective weight integration assignment method is to optimize integration based on the minimum deviation of the four weighting methods.This means that considering the human and empirical factors obtained in practice and the inherent attributes of the evaluation indicators.Finally,the evaluation results are more comprehensive.The correlation coefficient matrix is used to replace the covariance matrix in the Mahalanobis distance and the influence of the correlation between the evaluation indicators is considered.Then,the accuracy of the evaluation results is improved.Hence,the rationality and effectiveness of the improved gray target model in the goaf risk assessment are verified.