Model-Based Control of Renewable Energy Systems in Buildings

Abstract

This paper describes a research project concerned with the problem of supervisory control of systems that combine a range of heat sources with active and passive thermal storage. The work is based on a prototype building that includes a ventilated photovoltaic array, solar air and water heating, a biomass-fired boiler, and a stratified thermal store. The supervisory control problem is, for each source, whether to deploy the received energy directly into the building, store it for later use, or reject it to the environment. These decisions are currently made by a building management system programmed with a complex, arbitrary set of rules and setpoints. The considerable number of operational states means that such a control system is very difficult to commission. Analysis of the data for the early stages of building operation indicate strongly that it is unlikely optimal use is being made of the renewable energy sources with this approach. The objective of this research is to investigate the potential effectiveness of replacing the rule-based control scheme with one based on a combination of a building/system model with an optimization algorithm. The existing building, plant, and control systems were modeled using a commercial simulation environment and calibrated against measured data from the building. A period of swing-season operation was used to compare the existing control with that based on an optimal approach. Results indicate that significant improvements in system operation are possible but also that significant improvement in execution time will be needed for any future online deployment.