Affiliation:
1. Key Laboratory of Advanced Manufacture Technology for Automobile Parts (Chongqing University of Technology), Ministry of Education 1 , Chongqing 400054, China
2. Chongqing Tsingshan Industrial Co., Ltd 2 ., Chongqing 402761, China
3. Chongqing SERES New Energy Vehicle Design Institute Co., Ltd. 3 , Chongqing 401335, China
4. Ningbo Shenglong Group Co., Ltd. 4 , Ningbo 315104, China
Abstract
CO2 heat pump air conditioning (HPAC) systems for electric vehicles (EVs) have received widespread attention for their excellent low-temperature heating capabilities. However, the range of EVs is limited by the battery energy storage, which makes the energy demand of the heating system affect the energy use efficiency of the drive battery. In order to measure the thermal economy of the air conditioning (AC) system in terms of heating, the index of coefficient of performance (COP) is often used. Accurate COP prediction can help optimize the performance of heat HPAC systems for EVs to avoid energy wastage and thus improve the range of the vehicle. In this study, we use a backpropagation (BP) neural network combined with the particle swarm optimization (PSO) algorithm to predict and optimize the COP of the CO2 HPAC system for EVs. First, a COP prediction model of the CO2 HPAC system for EVs was established, which can consider a variety of influencing factors, and the key parameters affecting the COP of the AC system were obtained through experiments. Second, a BP neural network is used to predict the COP of the CO2 HPAC system, and in order to overcome the shortcomings of the BP neural network, which is slow and prone to fall into the minimum value, the particle swarm algorithm PSO is introduced to optimize the weights and biases of the BP neural network, so as to improve the accuracy and stability of the prediction. Through this study, we combine the BP neural network with the PSO algorithm to achieve accurate prediction and optimization of the COP of the HPAC system of an EV, which provides a strong support for the improvement of energy use efficiency. Second, we considered a variety of influencing factors, such as outdoor temperature, compressor speed, and EV status, which made the prediction model more accurate and applicable. Finally, the method proposed in this study is validated on a real dataset, and the optimization of the BP neural network using the particle swarm algorithm PSO can improve the accuracy of COP prediction for HPAC systems by 65.8%.
Funder
Natural Science Foundation of Chongqing Municipality
the youth project of science and technology research program of Chongqing Education Commission of China
Chongqing Graduate Student Research Innovation Project
Special Major Project of Technological Innovation and Application Development of Chongqing
Subject
Renewable Energy, Sustainability and the Environment