控制策略对新能源汽车CO₂热泵空调性能的影响研究

Research on the Impact of Control Strategies on Performance of CO₂ Heat Pump Air Conditioning System in New Energy Vehicles

摘要: 采用AMESim搭建了新能源汽车CO₂热泵空调仿真模型，并通过台架试验数据验证了模型的准确性，最大误差为4.4%。利用Simulink建立了开关控制策略、PID控制策略和基于SAC算法的强化学习控制策略，并研究了三种控制策略对CO₂热泵空调性能的影响。结果表明：在制冷模式下，开关控制响应最快，但温度波动较大，波动范围为±0.5℃；PID控制稳定性较好，但存在温度超调；SAC控制较稳定，稳态误差在0.8%以内。在制热模式下，开关控制的温度波动较为频繁；PID控制响应快，但超调明显；SAC控制的温度波动较小，但在CLTC循环工况下存在一定的温度超调。研究结果揭示SAC控制策略在温控方面具有一定优势，可为CO₂热泵空调的智能控制提供技术支持。

Abstract: A simulation model of a CO₂ heat pump air conditioning system for new energy vehicles was established by employing AMESim software. The accuracy of the simulation model was validated by bench test data, with a maximum relative error of 4.4%. The on-off control, PID control and reinforcement learning control based on the Soft Actor-Critic (SAC) algorithm were developed in Simulink to comparatively investigate their impact on the performance of the CO₂ heat pump air conditioning system. The results show that under cooling mode, the on-off control strategy exhibits the fastest response, but results in substantial temperature fluctuations with a range of ±0.5˚C; the PID control strategy offers better stability, but introduces a temperature overshoot; the SAC-based control strategy achieves superior stability, maintaining a steady-state error within 0.8%. Under heating mode, the on-off control strategy results in relatively frequent temperature oscillations; the PID control strategy responds rapidly, but exhibits significant overshoot; the SAC-based control strategy maintains minimal temperature fluctuation, yet a certain degree of overshoot is observed under the CLTC driving cycle. The findings reveal that the SAC control strategy offers distinct advantages in temperature management, thereby providing a promising technical pathway for the intelligent control of CO₂ heat pump air conditioning systems.