Model-based control of renewable energy systems in buildings

The paper describes a research project which addresses the problem of supervisory control of systems which include a range of heat sources combined with active and passive thermal storage. The work is based around a prototype building which has a ventilated PV array, solar air and water heating, biomass-fired boiler and a stratified thermal store. The supervisory control problem is, for each source, whether to deploy the energy directly into the building, store for later use or to reject to the environment. These decisions are made by a building management system programmed with a complex, arbitrary set of rules and setpoints. Analysis of the data for the early stages of building operation indicate strongly that it is unlikely that optimal use can be made of the renewable energy sources with this approach. The existing building, plant and control system have been modeled using a commercial simulation environment and calibrated against measured data from the building. Examination of the operation of the existing control scheme has shown that significant improvements are possible. The next stage of the work will be to deploy dynamic optimization methods to update the supervisory control decisions at fifteen minute intervals during the day.