Forecasting market-clearing price in day-ahead market, using SOM-ANN

Forecasting the market-clearing price (MCP) is the most essential task for any decision-making. In this paper, a three-layer backpropagation (BP) network model is proposed firstly. Three kinds of factors: the historical loads and MCP, the behavior of the participants and the reliability level of the power system, are considered as the inputs of the model. Secondly, based on the characteristic of self-organizing and clustering of self-organizing map (SOM), a method is proposed to deal with the sample dataset of the BP model, which can perform a consistency analysis, and then form a new training dataset. Historical loads and prices of 2002 in the Queensland of Australia day-ahead energy market are used for training and forecasting test. The results show that our methods are promising, and that by using the BP network on the new dataset analyzed by SOM for the prediction, the precision and the efficiency are advanced remarkably.