Model researching on daily peak load and climate factors of Hubei grid in summer

Short-term load prediction is an important part of power grid dispatching, which is effective to unit commitment, economic dispatching and optimal power flow. With electrification level higher and higher, especially improvement of air conditioning ownership and usage, peak load fluctuation is more obvious for cooling in summer. In view of the typical characteristics of summer climate in Hubei, the model of daily peak load prediction based on climate factors is built and applied for actual production, which includes the historical factors of load, temperature and holidays. The relation and correlation between climate and load have been mined deeply. After evaluation, the goodness of fit highly reaches 95% and the mean relative error is only 4.40%. The model could be used for prediction of daily peak load in summer. This model has been used to predict peak load significantly in Hubei grid for 3 years.