Clustering of Uncertain Load Model Parameters with K-medoids Algorithm

Abstract

Load model is difficult to build due to the uncertain property of power load. Using ambient signal based load model parameter identification method, load model parameter identification can be performed very frequently and then many different identification results at different time points can be obtained. To deal with these uncertain load model parameters, a load model parameter clustering method is proposed to pick up the representative load model parameters from the identification results. The distances of models used for clustering are based on the post-fault response curves to get better clustering results. K-medios clustering algorithm is applied and the cluster number is decided by the radius of the clusters. The simulation results have shown the effectiveness of the proposed load model parameter clustering method.