Adjustment Oriented Electricity Tariff Recommendation Based on Typical Customers

Abstract

Tariff recommendation is an important marketing means for retailers in electricity market. Based on the power consumption data acquired by the advanced metering infrastructure, tariff recommendation can not only match customers with appropriate tariff but also guide customers to adjust power consumption strategy. In this paper, an adjustment oriented electricity tariff recommendation method based on typical customers is proposed. Fuzzy clustering algorithm is applied to power consumption data of appliances to extract the consumption feature of each kind of appliance. Then the decision tree is constructed, which correlate the class features of appliances with clusters of customers. Pre-pruning decision tree is deployed together with use of depth-first search to exclude meaningless nodes of the decision tree, so that the typical customer set is determined. The tariff recommendation model is constructed aiming at minimum electricity bill, which credits to optimize benchmark strategy of power consumption for typical customers. The model is solved by particle swarm optimization. Simulations show that the proposed method enables customers to understand the adjustment signal revealed by the tariff, and results in electricity bill reduction due to participating in demand response as retailers expected.