A Multi-stage Model Based on Association Rule Analysis for Forecasting and Modifying Electricity Demand and Price in a Smart Grid

摘要

Smart electricity grids are continuously being deployed to enable utilities, customers and third-party providers to monitor and control energy usage. The aggregate reaction of consumers can potentially shift the demand curve. Data collected by the smart grid will provide several advantages to all participants, including better decision-making regarding energy-usage and enhanced understanding of consumer decisions. In this paper, a multi-stage forecasting framework is proposed for forecasting and modifying electricity demand, which contains three stages: Initial Forecasts Manufacturer (IFM), Extracted Rules Manufacturer (ERM) and Modified Forecasts Manufacturer (MFM). In the proposed model, a fuzzy neural network is used to generate the initial price and demand forecasts in IFM. During ERM, the proposed Apriori-AND algorithm is used to detect and extract price-demand dynamics. Finally, the Multi-valued Logic method is used to modify the initial forecasts by using the rules extracted in MFM. The model will further improve the prediction accuracy by using association rule analysis. The simulation results are generated with historical data.