Reactive power optimization based on load profile partitioning

This paper presents a method of reactive power optimization based on partitioning of the power system daily load curve. The underlining problem that is being solved is the optimization of elements with defined maximum allowable daily operating times (MADOT). The daily load curve of the power system was partitioned into 24 stationary states which were used as a solution space for determining when to optimize control variables with defined MADOTs. An optimization method, based on a genetic algorithm, was developed for this purpose. The efficiency of the presented method was demonstrated on the Ward-Hale 6-bus system. The proposed method decreased the power losses and improved the system voltage profile. The main disadvantage of this approach is a great computational burden that it requires. Some possible solutions of this problem are discussed in the conclusion.