Case-based reasoning for inert systems in building energy management

Abstract

Energy management systems are a typical example for inert systems where an event or action causes an effect with a delay. Traditional solutions for energy management, such as PID controllers (PID=proportional-integral-derivative loops), control target values efficiently but are sub-optimal in terms of energy consumption. The paper presents a novel, case-based reasoning approach for inert energy management systems that aims to reduce energy wastage in over heating and over cooling for buildings. We develop a case representation based on time series data, taking environmental impact factors into consideration, such as weather forecast data. This includes a post-mortem assessment function that balances energy consumption with comfort for the users. We briefly discuss retrieval and reuse issues. We report on an experimental evaluation of the approach based on a building simulation, including 35 years of historical weather data.