Short-term electric load forecasting using Extreme Learning Machine - a case study of Indian power market

In a deregulated electricity market there is a huge necessity of prediction of electric loads as well as the electricity prices keeping in view of challenges and competitions faced by the market participants i.e. the power generating companies and the grids. In India deregulation of the electricity markets is not fully functional rather some of the market participants are into such environment. The market clearing price (MCP) is based on the load demands. So it is highly essential to predict the day-ahead load in the deregulated market. In this paper, we propose a day-ahead load forecasting method which is based on Extreme Learning Machine (ELM). The structure of ELM is referred to generalized single-hidden-layer feed-forward networks (SLFNs), where the hidden layer need not require to be tuned. The load data of energy exchange of India has been considered for testing the proposed model. The forecasting has been done for the periods which comprise of spikes in loads and the accuracy of forecasting is measured by Mean Absolute Percentage Error (MAPE).