RESUMO
The continuous increase in residential population implies an increase in electricity demand. The demand prediction performed by the utility company (UC) to schedule and plan the next energy purchase from the market may not be efficient since the actual demand consistently exceeds the supply. To use the energy production infrastructure efficiently and decrease the grid system's overload, the UC seeks to match the demand with the supply. Dynamic pricing is a simple yet effective way to influence consumption as price-sensitive consumers may reschedule the operation of some of the appliances. This paper presents a novel, consumption data-driven, selective and dynamic Time-of-Use (ToU) electricity pricing strategy applicable by the UC for residential electricity consumers. The method consists of clustering real consumption data using the k-means clustering technique and heuristically categorizing clusters based only on their consumption data to simplify the design of the ToU tariffs and limit the number of different tariffs in the same population. The method includes a heuristic determination of the time period for the ToU tariff changes for each category. The ToU tariff parameters are determined by minimizing a single cost function using genetic algorithms and considering all consumer clusters such that the consumer behavior model is based on price elasticity. The optimal ToU tariffs result in decreasing the demand in power peaks by targeting the overconsumer categories in the latter. Implementing the ToU tariff results in a profit margin distributed among the UC and the consumer. Moreover, the optimal prices guarantee positive gains for both of them with the presence of a wide range of parameter uncertainties. The proposed pricing strategy necessitates the presence of consumption data only, which implies the conservation of the consumers' privacy. Real consumption data obtained for two towns in Hungary over three years is used to show the efficiency of the proposed pricing strategy.