Research on cost differentiation model of power grid multi-service demand based on data drive

In order to optimize the cost control strategy of the whole process of power grid maintenance operation and improve the cost allocation accuracy, a differentiated classification study on multi-service demand based on fine operation level is carried out in view of the large differences in business characteristics, regional characteristics and cost characteristics among regions in China. Firstly, considering the power grid service difference factor, regional difference factor and cost deviation disturbance factor, the multi-service demand hierarchical evaluation index system is formed. Then, multi-factor system analysis method is used to construct the Gradient Boosting Decision Tree (GBDT) model, and the evaluation criteria and mechanism of hierarchical differentiation reflecting multi-service requirements are set up by the model decision conditions. Finally, taking the maintenance operation of a province as an example, according to the actual data, the differentiation classification of multi-service requirements is realized, and the performance parameters and classification accuracy of the model are verified. The research shows that the proposed multi-service demand differential classification technology is effective and reasonable, and provides theoretical and practical basis for cost demand quantitative classification and allocation optimization.