Model selection mechanism of Interactive Multiple Load Modeling

Accurate load modeling with fluctuate loads has been one of the most challenging problems of modern power system. In this paper, a model selection mechanism is proposed for the Interacting Multiple Load Modeling (IMLM), to achieve the adaptive adjustment of best model set. A large number of models with different characteristics are chosen to compose the model set in order to completely cover the load characteristics of transformer substation, which brings severe competitions between these models and increases the complexity of load modeling. Equivalent asynchronous machine parallel ZIP load model is adopted as the interacting load model structure, and the Normalized Innovation Square Sum is chosen as the test statistic of model selection. Under the system of IMLM, this paper exhibits the processes to choose the best-matched models from the model set to participate in model mixing. Finally, a simulation case and the results illustrate the effectiveness and advantages of the proposed model selection mechanism.