An effective hybrid approach for Dynamic State Estimation in power system

Power System Dynamic State Estimation(DSE) considers statistical characters of systemic state variables in past period, has functions of state estimation and forecasting, posses predominance that static estimation hasn't in terms of theory and practicability. On the basis of farther study at DSE theory and method, a general framework for self-adapting dynamic estimator is presented here to improve the forecasting and filtering models. Forecasting model uses ultra-short term multi-node load forecasting technique to increase state forecasting accuracy. Filtering model adopts Least Square Support Vector Machines (LS-SVM) technique, whose nonlinear functions fitting performance is stronger than traditional Artificial Neutral Network (ANN), to find an adaptive dynamic filter. It makes a satisfying result in actual application for power system control center of Shandong province.