Investigation on the application of robust model predictive control on air-conditioning systems

This paper investigates how to use robust model prediction control in air-conditioning systems. Air-conditioning processes are described using a first-order plus time-delay model associated with uncertainties in process gain, time constant and time-delay. An uncertainty polytope is developed to describe the corresponding uncertainties when the continuous model is sampled into a discrete one and further reformulated as a state-space model. Offline LMI-based MPC is then employed to design robust controller. In order to improve control performance, an operating mode identification scheme and a gain-scheduling scheme are also developed for application in air-conditioning systems.