Research on adaptive control strategy for CO2 heat pump air conditioning system of electric vehicles based on artificial neural network and genetic algorithm

• Development of a novel ANN–GA-based control framework for electric vehicle CO 2 heat pump air-conditioning systems, enabling adaptive optimization of discharge pressure. • Experimental analysis of key parameters affecting the performance of electric vehicle CO 2 heat pump air-conditioning systems. • Derivation of system control logic and control directions based on the observed parameter-performance relationships. • Simulation model calibrated with experimental data, ensuring high fidelity and reliability of performance predictions. To address the utilization of environmentally friendly refrigerants and enhance the driving range of electric vehicles, this paper proposes an optimal performance prediction model for CO 2 heat pump air conditioning systems in electric vehicles, which is based on the integration of artificial neural networks and genetic algorithms(ANN-GA). Aiming at the long-standing limitations of conventional control strategies under multivariable coupling and highly dynamic operating conditions—particularly the insufficient optimization of energy efficiency and slow dynamic response—this work systematically investigates, through combined experimental and numerical analyses, the influence mechanisms of key operating parameters on system performance and the optimal discharge pressure. By revealing the strong nonlinear relationship between discharge pressure and system coefficient of performance (COP) under varying ambient and load conditions, the control objective of dynamically regulating the discharge pressure to maximize COP is explicitly established. An adaptive control strategy based on ANN-GA is innovatively adopted to realize the online/offline optimization of the optimal discharge pressure,thereby filling the research gap in intelligent, self-adaptive optimization of transcritical CO 2 heat pump systems for electric vehicles. The results show that this ANN-GA strategy, compared with the traditional fixed discharge pressure PI control, significantly improves the energy efficiency in both summer and winter dynamic conditions, with energy savings of 17.5 % in refrigeration conditions and 11.17 % in heating conditions. Furthermore, it is demonstrated that the system can quickly achieve the target cabin temperature and accurately control the fluctuation within ±1.0 °C, indicating excellent thermal comfort control performance. This study confirms that the proposed ANN–GA-based strategy provides an efficient, intelligent, and robust solution for improving both the energy efficiency and thermal comfort of CO 2 heat pump air-conditioning systems in electric vehicles, and offers a new technical pathway for the intelligent control of next-generation automotive thermal management systems.