汽车工程 ›› 2025, Vol. 47 ›› Issue (8): 1634-1645.doi: 10.19562/j.chinasae.qcgc.2025.08.019

• • 上一篇    

基于用户数据的整车结构耐久“试验场-用户”关联特性分析

赵礼辉1,2,3,徐晓宇1,翁硕1,2,3,刘东俭4,张东东1,2,3()   

  1. 1.上海理工大学机械工程学院,上海 200093
    2.机械工业汽车机械零部件强度与可靠性评价重点实验室,上海 200093
    3.上海市新能源汽车可靠性评价公共技术平台,上海 200093
    4.中汽研汽车试验场股份有限公司,盐城 224100
  • 收稿日期:2024-11-29 修回日期:2025-01-09 出版日期:2025-08-25 发布日期:2025-08-18
  • 通讯作者: 张东东 E-mail:dongdongzhang@usst.edu.cn

Research on Vehicle Structural Durability Based on User Data and the Correlation Between “Proving Ground and User”

Lihui Zhao1,2,3,Xiaoyu Xu1,Shuo Weng1,2,3,Dongjian Liu4,Dongdong Zhang1,2,3()   

  1. 1.School of Mechanical Engineering,University of Shanghai for Science and Technology,Shanghai 200093
    2.CMIF Key Laboratory for Strength and Reliability Evaluation of Automotive Structures,Shanghai 200093
    3.Public Technology Platform for Reliability Evaluation of New Energy Vehicles in Shanghai,Shanghai 200093
    4.CATARC Automotive Proving Ground Co. ,Ltd. ,Yancheng 224100
  • Received:2024-11-29 Revised:2025-01-09 Online:2025-08-25 Published:2025-08-18
  • Contact: Dongdong Zhang E-mail:dongdongzhang@usst.edu.cn

摘要:

基于现有试验场结构耐久规范难以全面捕捉实际用户运行条件下的复杂载荷特性,导致试验场工况与用户工况之间存在差异,本文以用户大数据为基础,分析了用户结构耐久工况特征并进一步研究了“试验场-用户”关联特性。首先,对用户数据进行预处理并采用多元时序自适应分割方法构建用户运行片段,从时域、频域和损伤3个角度构建片段特征参数。其次,采用线性和非线性两种方法进行特征参数降维,基于K-Means聚类算法将用户的运行片段划分为6类典型工况,进一步探讨了各典型工况的运行特征、损伤贡献及不同城市用户数据的差异性。最后,通过试验场强化路面和典型工况的映射关系进行相似性分析;通过损伤目标、功率谱密度、外推雨流进行差异性分析,结合“损伤等级-车速”联合分布对产生差异的原因进行探讨。研究成果为用户使用条件下试验场整车结构耐久性试验工况的选取、全寿命周期目标的确定和现有试验场规范的优化提供了参考和依据。

关键词: 用户运行大数据, 结构耐久, 工况特征, 试验场与用户关联

Abstract:

Current structural durability specification of the proving ground fail to fully capture the complex load characteristics under the actual user operating conditions, which results in a discrepancy between the proving ground conditions and the actual user use. In this study the characteristics of user structural durability are analyzed based on user big data. Furthermore, the characteristics of the 'proving ground-user' correlation are studied. Firstly, the user data is subjected to preprocessing, and a multivariate time-sequence adaptive segmentation method is employed to construct user-operating segments. The characteristic parameters of operating segments are constructed from three perspectives: time domain, frequency domain and damage. Subsequently, linear and nonlinear methods are employed for feature parameter dimensionality reduction. K-Means clustering is employed to categorize the data into six typical operating conditions. For each typical operating condition, operating characteristics, damage contributions, and inter-city variations in user data are analyzed. Finally, a similarity analysis is conducted to map the relationship between the proving ground reinforced pavement and typical operating conditions. Differences in damage targets, power spectral densities, and extrapolated rain flow are analyzed, and the reasons for the differences are explored through the joint ‘damage level-speed’ distribution. The findings provide a foundation for selecting structural durability test conditions, defining life cycle objectives, and improving proving ground specifications under user conditions.

Key words: user-operated big data, structural durability, condition characterization, proving ground and user association