Mid-term load power forecasting considering environment emission using a hybrid intelligent approach

The forecasting of electricity load is considered as an essential instrument, especially in countries with a restructured electricity market. The mid-term prediction is performed for the period within 1 month to 1 or 2 years and it is important for mid-term planning, including planning of repairs and economic exploitation of power systems, which are related to the reliability of the system directly. The forecast horizon in this paper is monthly and on a daily basis (peak load). The combined method of the neural network and the particle optimization algorithm were used to predict the load, and then the maximum amount of environmental pollution caused by the production of electricity required to supply the predicted load was calculated. The applied method was tested on the data of a North American electric company for four months (four seasons) and in comparison to the other methods presented in previous studies, it had an acceptable accuracy.