A robust ANN-based price forecasting model for deregulated electricity market

World Congress on Industrial Automation

July 20-22, 2015 San Francisco, USA

Abdollah Ahmadi1, Hani Mavalizadeh1, Adel M Sharaf3 and Ali Esmaeel Nezhad1

Posters-Accepted Abstracts: Adv Robot Autom

Abstract :

This paper proposes an Artificial Neural Network (ANN) approach to forecast the next week electricity market prices. In restructured power systems a lot of factors affect the electricity prices. Provided that market participants have to know the future prices, they can model their risk management strategies. In this paper, among different forecasting tools, ANN is used to perform this task due to its flexibility and simplicity. Historical load data as well as historical price data are used to train the neural network. In this work a three-layered feed-forward neural network is selected to forecast the next week prices. Besides, a new incremental neural network is employed applying the initial knowledge to adapt the learning process and modify the weights in each step. The proposed approach has been implemented in mainland Spain and California markets to assess the accuracy of the forecasted values obtained from the proposed neural network.