RESEARCH ON PROCESSING OF SHORT-TERM HISTORICAL DATA OF DAILY LOAD BASED ON KALMAN FILTER

The historical data of load is the basis of load forecasting, but the inaccuracy caused by the measurement and artificial factors will lead to the inaccuracy of load forecasting inevitably. Thus, it is indispensable in load forecasting to process the historical data of load. A method to establish basic model of load data by least square linear fitting is presented in which the system parameters of Kalman filter are identified by cubic spline interpolation, then the pre-processing of historical data is performed by Kalman filter to correct the bad data caused by measurement error or human made alteration. After the processing of pseudo-data in historical data of load from practical power system, the posterior error of load forecasting is within 3%.