Potential Jet-Space Based Practical Identifiability and Integral Backstepping Sliding Mode Control of Heat Exchanger

Abstract

This work develops the minimal modeling methodology for modeling and control of the liquid-liquid heat exchanger system. An application of the potential jet space theory is firstly presented, where the cross-convection model of the heat exchanger dynamics is prolonged onto the potential jet space to generate the input-output equations that are written in terms of integrals of the measured data. This integral based input-output equation then facilitates the parameter estimation without extensive computational loading demands. The linear potential jet space model was first applied to the measured data, and extended to capture the time delay effect in the measurement. An integral backstepping sliding mode controller was also designed. The delay model was shown to give a tracking temperature error to within 0.0004 degrees, with a mean integral absolute error (ITAE) of around 0.27, even with heavy parameter changes.