Department of Electrical and Computer Engineering, University of Birjand, Birjand, Iran.
Accurate and effective electricity price forecasting is critical to market participants in order to make an appropriate risk management in competitive electricity markets. Market participants rely on price forecasts to decide on their bidding strategies, allocate assets and plan facility investments. However, due to its time variant behavior and non-linear and non-stationary nature, electricity price is a complex signal. This paper presents a model for short-term price forecasting according to similar days and historical price data. The main idea of this article is to present an intelligent model to forecast market clearing price using a multilayer perceptron neural network, based on structural and weights optimization. Compared to conventional neural networks, this hybrid model has high accuracy and is capable of converging to optimal minimum. The results of this forecasting method for Market Clearing Price (MCP) of Iranian and Nord Pool Electricity Markets, as well as Locational Marginal Price (LMP) forecasting in PJM electricity market, verify the effectiveness of the proposed approach in short-term price forecasting.