A chaotic model approach for electricity price forecasting in Indian scenario

Abstract

In a deregulated power market, the electricity prices exhibit extreme volatility due to its non-storable nature, supply constraints at peak hours, transmission line congestion at peak hours and seasonal and diurnal variations. The objective of this paper is to model the day-ahead market price of Indian energy exchange (IEX) based on time series modelling approach. The detailed analysis of market clearing price (MCP) collected from IEX clearly indicates the chaotic nature of data. It is because of this chaotic behaviour, the phase space of the time series data is reconstructed using Taken's theorem. With the concept of the add-weighted one-rank multi-step model, the MCP of IEX is modelled as a chaotic model from this reconstructed phase space. Furthermore, the developed chaotic model is verified for forecasting MCP and the simulation results demonstrate that the chaotic model developed outperforms the autoregressive integrated moving average (ARIMA) model and generalised autoregressive conditional heteroscedasticity (GARCH) model in terms of in sample forecasting performances.