A Control-oriented Simplified Building Modelling Strategy

Abstract

This paper discusses the development of a controloriented simplified modelling strategy (COSMOS) for model-based predictive control (MPC) in buildings. In MPC, a model of the system is used along with forecast information for optimal planning. A model that is as simple as possible –but accurate enough for the purpose at hand– facilitates the implementation of an MPC strategy. This paper and forecast data (Figure 1) is called model-based predictive control, or MPC (Camacho and Bordons, 2004). MPC has successfully been used for decades in process control and chemical engineering (Qin and Badgwell, 2003). Actual weather Actuator Plant discusses desirable features of models intended for the specific needs of advanced control applications. A path for the creation of such models is presented, based on low-order resistance-capacitance (RC) thermal networks and their equivalent state-space formulation; such an approach provides physical Reference + _ Forecast Error Model Controller Feedback Energy Storage Output insight while facilitating the treatment of the problem. The model parameters are found by applying an optimization to match the output of a building simulation model. Results include an assessment of the uncertainty of the model outputs.