Reduced multiple model predictive control of an heating, ventilating, and air conditioning system using gap metric and stability margin

Abstract

In this paper, a reduced multiple-model predictive controller based on gap metric and stability margin is presented to control heating, ventilating, and air conditioning (HVAC) systems. To tackle the strong nonlinearity and large number of degrees of freedom in HVAC system, two approaches, called Reduced Order Model Bank-Multiple Model (ROMB-MM) and Multiple Model-Reduced Order Model (MM-ROM), are introduced. In the first approach, the order reduction is performed prior to multiple models selection and in the second one multiple models selection is implemented before the model order reduction. Furthermore, soft switching is employed to enhance the closed-loop performance as well as to gain optimal energy consumption. In order to evaluate the proposed approaches, a sliding mode controller is also simulated to compare the results in terms of energy and cost savings, transient and steady-state responses, and robustness against external disturbances and model uncertainties. Practical application: HVAC control systems are in charge of control indoor air temperature with energy optimization so that the thermal comfort is maintained. But how to model HVAC systems in each weather conditions is a significant challenge. A simpler and more accurate model will provide more efficient control of indoor air temperature. This paper suggests model order reduction and multiple model to select the simple linear model(s) in extreme weather conditions. The effectiveness of the proposed method can be implemented on nonlinear HVAC system.