Predictive control of car refrigeration cycle with an electric compressor

Abstract

Air conditioning in an electric car significantly impairs its driving range. This study intends to save energy consumed by air conditioning in cars while maintaining a similar level of the thermal environment in the cabin. The cooling load is removed by controlling the air temperature by an evaporator. Conventional feedback controllers only counteract errors with phase delay. To control the temperature and also save energy, model predictive control is proposed with an optimization algorithm of quadratic programming. Vehicle tests cost too much to be conducted frequently for the development of predictive control. To replace vehicle tests, a dynamic refrigeration cycle is developed asa high fidelity simulator, and its performance is verified against vehicle tests. The error dynamics of the air temperature are developed as a first order system, and linear approximation of the compressor power with a decision variable is proposed to establish a linear constraint that is necessary for the optimization formulation. The algorithm tested with the simulator leads to an energy savings of 2.65% while maintaining a similar level of temperature control by use of conventional feedback control.