Peak and Valley Periods Partitioning Based on Improved K-medoids Algorithm

The time of use (TOU) electricity pricing is one of the electric demand response (DR) strategies and has been widely used in the electrical power systems. The reasonable partitioning of peak, flat and valley periods is the prerequisite for setting the price of TOU. This paper proposes a peak-flat-valley (PFV) period partitioning model based on improved K-medoids algorithm. To address the instability problem of the divided periods caused by the random selection of initial centers in conventional clustering methods like k-means algorithm, the proposed model employs the maximum distance method to determine the initial center of each period. Then, the variation range for the center of each period can be obtained by the continuous nearest neighbor searching (CNNS) strategy, while improving the computation efficiency. A new criterion of partitioning is established based on Hausdorff distance and Euclidean distance, which is used in the particle swarm optimization (PSO) to optimize the period partitioning. Finally, case studies on IEEE reliability test system (RTS) demonstrate the effectiveness of the proposed model.