基于MCD-AHP的预装式变电站城市化属性指标评价

为提高城市轨道交通预装式变电站建设的城市化属性效果，需要对其城市化属性指标做出准确评价. 首先，构建了包含经济属性、社会属性、节能属性3类指标的预装式变电站城市化属性指标评价体系，并给出了评价指标的量化模型；然后，引入最小协方差行列式 （minimum covariance determinant，MCD）稳健分析法对层次分析法 （analytic hierarchy process，AHP）中专家组所给出的判断矩阵进行检测，去除由于专家组成员之间主观认知的差异性而产生的离群矩阵；最后，应用改进后的层次分析法（MCD-AHP）对预装式变电站城市化属性指标权重进行计算并评价. 结果表明:去除离群矩阵样本后，上、下限矩阵计算所得指标权重差值的绝对值为0～0.0103、相对变化比率为0～9.31%，明显优于全样本状态下权重差值的绝对值率0.0138～0.1355和相对变化比15.99%～106.23%，该方法可有效提高判断矩阵的逻辑一致性和权重值的聚合性. 

Urbanization Attribute Evaluation for Prefabricated Substation Based on MCD-AHP

In order to improve the urbanization attribute effect of the prefabricated substation construction for urban rail transits, it is necessary to make an accurate evaluation of its urbanization attribute index. First, an evaluation system of prefabricated substation urbanization attribute indexes is built, including economic attributes, social attributes, and energy-saving attributes, and a quantitative model of the evaluation indexes is presented. Then, the robust analysis method with the minimum covariance determinant (MCD) is used to detect the judgment matrix provided by the expert group in the analytic hierarchy process (AHP) and remove the outlier matrix due to the difference in subjective cognition among the members of the expert group. Finally, the improved MCD-AHP is used to calculate and evaluate the weights of the urbanization attribute indexes for the prefabricated substation. The results show that after removing the outlier matrix samples, the absolute value of the index weight difference calculated by the upper and lower limit matrices ranges from 0 to 0.0103, and the the relative change ratio is from 0 to 9.31%, which is significantly better than the corresponding ranges of 0.0138?0.1355 and 15.99%?106.23% under the full sample state. This method can effectively improve the logical consistency of the judgment matrix and the aggregation of weight values.