Peak load forecasting based on robust regression model

This paper describes the essentiality and the peculiarities of peak load forecasting in power system. The approach is proposed for the daily loads prediction during the peak period, which combines the robust regression model with the pattern clustering; method. Comparing with the traditional regression model, the proposed model has strong robustness and keeps better forecasting accuracy, even on unusual weather conditions. The validity of the approach is shown by forecast results on a real power system