Charging pile group metering quality evaluation index construction method based on error matrix

Abstract

The invention relates to a charging pile group metering quality evaluation index construction method based on an error matrix, and the method comprises the following steps: determining factors influencing the metering quality performance of an electric vehicle public charging pile group, and collecting and filtering the metering data and data abnormal values under the influence factors; the collected data is compared with standard measurement data, the error value of each influence factor is calculated, and an error data matrix is established according to the error value; and in combination with metering error characteristics and scene factors, the charging pile group is classified by using an intelligent clustering technology. And based on the error data matrix and the classification result, constructing first-level and second-level evaluation indexes under each classification, and formulating an evaluation standard. An expert scoring method is adopted to distribute initial weights for the second-level evaluation indexes, and then a random forest algorithm is utilized to dynamically adjust the weights of the second-level evaluation indexes according to the influence of the second-level evaluation indexes on the first-level indexes under each classification category. And finally, integrating the scoring standard of the second-level evaluation index to the first-level index, thereby establishing a complete measurement quality performance evaluation index system.

Images