Using artificial neural networks and system identification methods for electricity price modeling

Abstract

Electricity price is one of the most important parameters in electricity market. Determining electricity price has always been one challenges in the energy markets. Electricity demand plays an important role in determining electricity prices. In this paper, for electricity price modeling based on electricity demand using Artificial Neural Networks (ANN) and system identification methods are presented. A dataset of Australian energy market is used to model electricity price in this paper. The dataset includes electricity price and demand of Queensland in September 2017. Three scenarios are presented to model electricity price. First and second scenarios are based on system identification methods that include Auto Regressive eXogenous (ARX) model and Nonlinear Auto Regressive eXogenous (NARX) model respectively and third scenario is based on ANNs. All methods are modeled and simulated by MATALB. Results show that ANNs model with 77% fitness is better performance than each other methods.