Short-Term Load Forecasting Based on Second-Order Correction of Kalman Filter

For short-term loads,the Kalman filtering theory only makes one-step-ahead forecasts.This paper proposes a method based on a second-order correction to improve the one-step-ahead forecasting.It first uses the estimated covariance matrix to calculate the variance of forecasted value,and then uses the standard error of the calculated variance to correct the forecasted value.Given a confidence level,the envelope of confidence interval,which is the data support of risk analysis and reliability assessment,is estimated.The envelope also helps to construct the final corrected forecast.The proposed method is compared with the basic Kalman filter and Kalman filtering algorithm with moving windows using electricity load data during one week from a local electricity network.Simulation results show that the proposed method is effective and performs better than the other two methods.