Kalman Filter Algorithm Based On Unscented Transform of Short-term Load Forecasting

Power system short-term load data along with time change presents certain range,nonlinear wave. In the nonlinear short term load forecasting method,Kalman filter difficult to achieve satisfactory results. To accurately and efficiently predict the nonlinear load,put forward Unscented Kalman filter based on short term load forecasting method,input to the historical load data、the random disturbance factor as. Modeling of a summer 9 days of power load data by using this algorithm,the impulse response of Hankel matrix method to identify the model sequence based on order. The Unscented filter predict load data and the real load data and traditional Kalman filter forecast data analysis,simulation results show that Unscented Kalman filter method based on nonlinear prediction load data is adaptive and effective,not only has high prediction accuracy,and the model convergence speed,high stability of filter. The method provides a new approach for modeling of high nonlinear power system.