基于LSTM模型的汽车乘员舱内温度预测方法研究  

Research on Predicting Vehicle Cabin Temperature Based on the LSTM Model

作　　者：赵东鹏 席椿富 杨静 徐磊[1] 王庆洋 龙海生 谭文林 ZHAO Dongpeng;XI Chunfu;YANG Jing;XU Lei;WANG Qingyang;LONG Haisheng;TAN Wenlin(China Automotive Engineering Research Institute Co.,Ltd.,Chongqing 401122,China;School of Mechanical Engineering,Tianjin University,Tianjin 300350,China)

机构地区：[1]中国汽车工程研究院股份有限公司,重庆401122 [2]天津大学机械工程学院,天津300350

出　　处：《汽车工程学报》2024年第S1期59-65,共7页Chinese Journal of Automotive Engineering

基　　金：电动汽车产业技术创新战略联盟共性技术课题(ZDD-20230919-127)。

摘　　要：准确获取乘客周围区域的温度,对乘员舱内热舒适性的控制有非常重要的作用。提出一种基于长短期记忆网络(LSTM)模型的乘员舱内温度预测方法,利用环境温度、车速、太阳辐射强度、送风量、送风口温度等7种可测量的参数对乘客头部、腹部和足部区域温度进行实时预测。结果表明,训练后的LSTM模型能利用7种可测参数对乘员舱内温度进行实时预测,稳定状态下温度的预测偏差在2℃以内,能满足汽车自动空调的控制要求。Precisely obtaining the temperature around passengers is crucial for regulating thermal comfort within the vehicle cabin.This paper proposes a novel method for predicting the cabin temperature based on the long short-term memory(LSTM)network model.By utilizing seven measurable parameters,including ambient temperature,vehicle velocity,solar radiation intensity,air flow rate and air outlet temperature,this approach predicts the real-time temperatures around the passenger's head,abdomen,and feet.The results indicate that the trained LSTM model can employ these seven measurable parameters to predict cabin temperatures accurately in real-time.Under stable conditions,the temperature prediction deviation remains within 2℃,meeting the control requirements of the automatic A/C's system.

关 键 词：汽车乘员舱 热舒适性 LSTM 温度预测 

分 类 号：U462[机械工程—车辆工程]