Dynamic Forecasting of Power System Load Based on Kalman Filter

As a basic work of power system planning, medium-term load forecasting of power system provides necessary basic data for the development of power grid planning and plays a key role in the quality of planning. To face the difficulty that the error of traditional prediction algorithms will diverge with time, this paper establishes a Kalman filter (KF) load prediction model and uses it to predict the load data of a region at the same moment within six months. The obtained forecasting results are compared with those of the least squares (LS) load forecasting model in terms of both model accuracy and stability, and the residuals between them are calculated and analyzed. The results show that the KF model has a good prediction effect and verify the feasibility of the KF model in medium-term load forecasting.