Modelling Of An Automotive Multi-Evaporator Air-Conditioning System

With the arrival of plug-in hybrid electric vehicles (PHEV) and battery electric vehicles (BEV) with significant autonomy, battery cooling becomes a necessity in driving mode to ensure their durability and ability to charge rapidly.  For these vehicles, the refrigerating system may be composed of two evaporators (for front and rear passengers) in order to afford cooled air in the cabin and a chiller or a built-in battery evaporator to cool down the traction battery. This kind of multi-evaporator air-conditioning system has a number of technological barriers that must be undone. They are related to the components sizing in a context of cost reduction and control of such complex systems. The study therefore focuses on understanding the dynamic coupling of the several loop components such as the three evaporators having different cooling capacities. Understanding the behaviour of their respective expansion devices and the choice of these latter is also essential to control properly the transient phase and ensure an optimal operation of the air-conditioning system. In the literature, the effect of battery cooling by means of a chiller on the automotive air-conditioning loop has been already proved by simulation in the Dymola® [1] environment. The simulation results for several driving cycles, refrigerants and ambient conditions emphasize the thermal discomfort caused by the use of the chiller loop. However, no global control strategy has been established. More recently, a first study of an air conditioning system model with three evaporators was carried out [2]. After the validation of their component models, a cool down test was performed to test the performance of their air conditioning system. From a control point of view, a simple PI control on the temperature of air blown at the front evaporator was used to regulate the speed of the compressor. In the building sector [3], the benefits of a supervisory controller to regulate the multi-evaporator air conditioning system was developed. Although this type of decentralized model seems to be robust and applicable to the car, it requires the use of sensors and components currently too costly and subject to a less restrictive environment than in automotive. The challenge of such a cooling loop lies in the dynamic coupling of components as well as their design. The model of a multi-evaporator automotive air conditioning system (two evaporators and a chiller) is thus produced using the 0D AMESim® software. In order to obtain more representative results in the transient state, the majority of components, including the chiller and regulators, are physical models giving a good representation of their internal geometries. These models were validated using experimental test maps. The first results highlight the importance of the regulators choice on the loop stability. A comparison of several types of expansion valves (orifice, thermostatic and electronic) will be conducted in order to select the most suitable to meet the price-performance compromise. Finally, control strategies are studied in transient state to further improve the stability and speed of convergence to the target instructions