Evaluation of the Adjustable Potential of Industrial User Load Based on Electricity Data Mining.

Demand-side resources have the characteristics of flexibility and economy. In order to give full play to the adjustable potential of load, it is necessary to study the evaluation method of the adjustable potential of load. In this paper, we first use affinity propagation (AP) -K-means clustering algorithm to divide the load curve after cleaning into three modes: normal working, interrupted working and transfer working. Secondly, the multi-dimensional index system of interrupt and error time is defined, and the potential evaluation matrix is established. Thirdly, the entropy weight method is used to get the weight of each index. Finally, the adjustable potential of different industrial users is evaluated based on the method of approaching ideal solution sorting.