Design and comparison of model-based controllers for an automotive air conditioning system in an electric vehicle.

The air conditioning system in an electric vehicle has to fulfill multiple tasks, which are ensuring thermal comfort for the passengers and tempering the battery. In this paper four model-based control approaches for the air conditioning system are developed. The approaches are compared regarding their ability to track desired references, reject disturbances and avoid windup effects. Two approaches are feedback controllers, a decentralized proportional-integral control strategy and a centralized linear-quadratic-integral control strategy. The other two approaches combine the feedback controllers with an inversion based feedforward controller in a two degree of freedom control structure. In addition, the four concepts are complemented by the Hanus-Conditioned anti-windup mechanism. Three of the four proposed controllers explicitly consider the couplings of the multiple-input-multiple-output system, which allows for a high-performance control. Additionally, all approaches have the benefit of being based on a physical model of the system to be controlled. The development and identification of this physical model is part of the paper. The physical basis for the controllers ensures a high level of reusability and therewith an efficient controller design process. The proposed control concepts are validated and compared by measurement data from a testbench. • Complementation of the proposed FF approach from Göltz and Sawodny (2022) by two FBs. • Implementation of the resulting 2DOF control structure on a testbench. • Development of a split observer in quasi-diagonal system representation. • Experimental validation of a MIMO anti-windup accounting for the system couplings. • Comparison of controller structures w.r.t. reference tracking & disturbance rejection. [ABSTRACT FROM AUTHOR]