가변속 냉동시스템의 상태공간 모델링 및 LQR 제어

Abstract

The compressor variable speed (CVS) method used for control of vapor compression refrigeration system is described in detail as main target. This system is basically multi-input multi-output (MIMO) system that controls target temperature and superheat simultaneously. Many control design methods such as PID control based on a transfer function model, fuzzy control without a mathematical model, optimum control based on the state space model have been studied for dealing with the MIMO system. The PID control method is hard to predict the control performance in a design step. Also, this method have to get each transfer function for all control variables. Even though PID controller is designed through this process, it can’t ensure optimal control performance. Therefore, it is desirable that optimum control will be applied to the MIMO system. This paper presents mathematical modeling of an oil cooler system for optimum control based on a state space model. At first, The mathematical model of the oil cooler is built by combining dynamic model of manipulators with differential equations of heat exchangers. Next, the Taylor 1st approximation is used to linearize derived non-linear model. Also, dimension of the linearized model is reduced to design practical controller. The optimum control method on the basis of this model is designed to minimize an evaluation function that includes input energy and control error. The weighted control error matrix and input energy matrix are decided by a designer. The validity and robustness of the proposed optimum controller are reviewed by some computer simulations and real experiments.