Applications of Adaptive CKF Algorithm in Short-term Load Forecasting of Smart Grid

Short-term load forecasting plays a very important role in the operation of power system. Because of the strong uncertainty and nonlinear variations of smart grid, ordinary Kalman filtering algorithm used in the short-term load forecasting is of low precision and the forecasting results are not very ideal. Aiming to solve this problem, adaptive Cubature Kalman Filter(ACKF) had been proposed by introducing the noise estimator into the newly-proposed CKF filter. Combine ACKF with the bilinear models, in which daily loads in adjacent days are defined to be the input signals and daily loads at the same day in adjacent weeks are defined to be the output signals. This method can be used to forecast short-term load of smart grid. Finally, this paper takes the load data of European Intelligent Technology Network(ENUNITE) as an example. Simulation results prove that this method is effective and practical in short-term load forecasting of smart grid, which has a greater precision and wider application value comparing with CKF and traditional UKF methods.