A neural network approach to day-ahead deregulated electricity market prices classification

This paper proposes a day-ahead electricity price classification that could be realized using three-layered feed forward neural network (FFNN), cascade-forward neural network (CFNN) trained by the Levenberg–Marquardt (LM) algorithm and generalized regression neural network (GRNN). The electricity price classification method is as an alternative to numerical electricity price forecasting due to high forecasting errors in various approaches. These electricity price classifications are important because all market participants do not know the exact value of future prices in their decision-making process. In this paper, various electricity market price classification classes with respect to pre specified electricity price thresholds are used. The simulation results show that the proposed CFNN method provides a robust and accurate method for day-ahead deregulated electricity market price classification classes. The proposed neural network classification models of electricity prices are tested on the electricity markets of mainland Spain and New York.

► We presented neural network classification models of electricity prices are tested on the markets of Spain and New York.
► The key factors impacting electricity price classification are historical prices.
► Past 42 days were trained and the next 7 days were classified.
► The proposed approach has a simple and better neural network (NN) structures.
► Cascade-forward NN classification model is a good tool for price classification in terms of accuracy as well as convenience.