电动汽车电池温度一致性与驾驶行为关联特性研究  

Research on the Association Characteristics Between Battery Temperature Consistency and Driving Behavior of Electric Vehicles

作　　者：张晖[1,2] 李少鹏 丁乃侃[1,2] ZHANG Hui;LI Shao-peng;DING Nai-kan(Intelligent Transportation Systems Research Center,Wuhan University of Technology,Wuhan 430063,Hubei,China;Engineering Research Center of Transportation Information and Safety,Ministry of Education,Wuhan University of Technology,Wuhan 430063,Hubei,China)

机构地区：[1]武汉理工大学智能交通系统研究中心,湖北武汉430063 [2]武汉理工大学交通信息与安全教育部工程研究中心,湖北武汉430063

出　　处：《中国公路学报》2024年第3期216-230,共15页China Journal of Highway and Transport

基　　金：国家重点研发计划项目(2019YFB1600800);国家自然科学基金项目(52072289)。

摘　　要：为实现融合车辆运行状态的电池温度不一致性的精准评估诊断,设计并开展电动汽车自然驾驶试验,利用长周期、精细化的车辆运行数据,从微观运行片段的角度探究了电池温度一致性与驾驶行为的关联特性。通过驾驶人踩/松踏板的驾驶行为将车辆运行过程划分为A、B、C、D四类微观片段。分别针对4类片段,通过计算最大信息系数(Maximum Information Coefficient, MIC)得到了各驾驶行为参数与探针温度变异系数(Variation Coefficient of Probe Temperature, VCPT)的相关性,利用随机森林模型分析了驾驶行为参数对VCPT的重要性及影响机理,利用数据分组统计计算了驾驶行为参数对VCPT的量化影响效应。研究结果表明:驾驶行为参数与电池温度一致性具有弱相关性,且其对电池温度一致性的影响是非线性且非单调的;总体上车速类参数与电池温度一致性的相关性强于加速度和踏板类参数;对VCPT回归预测最重要的4项驾驶行为参数中,4类片段下均包含最大车速,B、C、D三类片段下均包含最大负向加速度;相比于高车速和高车速波动,高减速度驾驶行为引起的温度不一致性增幅是最显著的,4类片段下VCPT促进效应最显著的驾驶行为参数分别是最大负向加速度、平均负向加速度、最大负向加速度、车速标准差,其参数值85%分位点以上对应的VCPT均值比15%分位点以下的分别大9.44%、20.36%、13.05%、16.37%。研究结果可以支撑基于驾驶场景自适应阈值的电池温度不一致性评估诊断方法的提出,进而提高电动汽车电池安全预警准确率。To achieve accurate evaluation and diagnosis of battery temperature inconsistency by fusing vehicle running status,this study designed and conducted the naturalistic driving experiment of electric vehicles(EVs),and the long-term and high-frequency vehicle running data were used to explore the association characteristics between battery temperature consistency and driving behavior from the perspective of microscopic operation segments.Based on the driver's driving behavior of pressing/releasing the pedal,the vehicular running process was divided into four kinds of segments,namely,segments A,B,C,and D.For the four types of segments,the correlation between driving behavior parameters and the variation coefficient of probe temperature(VCPT)was obtained by calculating the maximum information coefficient(MIC),then the importance and influence mechanism on VCPT of driving behavior parameters were analyzed using random forest model,and the quantitative impact on VCPT of driving behavior parameters was calculated by data grouping and statistics.The results show that,the driving behavior parameters are weakly correlated to battery temperature consistency,and their impacts on temperature consistency are nonlinear and non-monotonic.In general,the correlation between battery temperature consistency and vehicle speed related parameters is stronger than that between acceleration and pedal state related parameters.Among the four most important driving behavior parameters for VCPT prediction,maximum speed is included for all four types of segments,and maximum negative acceleration is included for segments B,C,and D.Compared to high vehicle speed and speed fluctuation,the increment of temperature inconsistency caused by high deceleration is the most significant.For the four types of segments,the driving behavior parameters having the most significant promoting effect on battery temperature inconsistency are maximum negative acceleration,average negative acceleration,maximum negative acceleration,and standard deviation of speed res

关 键 词：汽车工程 电池安全 最大信息系数 电池温度一致性 电动汽车 驾驶行为 

分 类 号：U469.72[机械工程—车辆工程]