Enhanced decomposition and hybrid Heuristic methods for training the Neural networks for multi step ahead Electricity price forecasting

Abstract

Electric power generation recently has undergone a massive shift from traditional fossil fueled generation to sustainable and environmental friendly renewable energy generation. This technological advancement not only created various power generation options, perhaps instigated the need for flexible electricity pricing. Another advancement in electric power technology enables restructuring of electric power generation thereby providing opportunity for the consumers to choose their service providers (here generating companies, in short, GenCo) confirming reliable and economic perspectives. The electricity policy of various countries enables several players to come up with electric power generation using indigenous technologies. Ultimately, the prime objectives of all these GenCos will be to maximize their profit by providing reliable and quality power supplies to its consumers. The power producers generally expected to establish power generation plants at micro level, which generates electric power at the capacity ranging 1 to 10MW. Practically such generations are viable only with renewable energy sources like solar PV generations, wind energy generations and micro hydro power plants. In particular, the solar and wind energy are highly volatile and strictly non-stationary in nature. In addition, these are weather relying variables with uncertain and abnormal characteristics. Hence the power companies has to prepare a comprehensive report for deciding the generation of electric power and thereby ensuring reliable and continuous power to their consumers. The Electricity prices are characterized as non-stationary time series data that entails vigorous earning model for predicting the future electricity price from past data. As elaborated earlier, due to nature dependent functionality of renewable energy generation it necessitates the development of reliable prediction tools for scheduling power generation and thereby the electricity pricing. Perhaps, the consumer usage statistics will also influence the electricity pricing largely. newline