Model Predictive Control (MPC) Design for MIMO Direct Expansion Air Conditioning System

Abstract

In both commercial and residential buildings, Direct Expansion (DX) air conditioning (AC) systems widely used, as they are simple in configuration, low maintenance cost and more energy efficient than chilled water-type air conditioning (AC) system in which first water is cooled, then this cold water is used to cool the desired room air. In this research the nonlinear dynamic model of the Direct Expansion (DX) air conditioning (AC) is presented. The nonlinear mathematical model is linearized and then discretized. The discrete state space model is used to design the model predictive control (MPC). Using the variable compressor and volumetric air flow rate as inputs, the model predictive control can effectively control both conditioned room temperature and humidity simultaneously. The simulations are performed in the MATLAB/Simulink. The results demonstrate the efficiency of the predictive control of the proposed model for MIMO DX AC.