An Application of State Estimation to Short-Term Load Forecasting, Part II: Implementation

A precise short-term forecasting method for estimating the status of systems is required for on-line real-time control of complex power systems. In [11] some state estimation type modelings of load forecasting were introduced and a few practical problems for applying state estimation were discussed. In this paper the identification algorithms of the covariance matrices of system and observation noise using observed data series are developed and their experimental results by simulation model are discussed. Results show that forecasting error by the developed method is quickly converged to minimum error of the ideal state estimation with previously known noise properties.