Abstract: As an important part of the vehicle development process, the vehicle structure durability test plays a crucial role in validating and evaluating the vehicle structure and vehicle systems. It is particularly critical to establish durability test objectives that are user-centered and align with actual user use. This article proposes a method for determining vehicle durability test objectives based on user driving conditions. Firstly, user typical driving conditions and the mileage distribution within each condition are obtained using machine learning clustering algorithms based on actual user driving data. Secondly, real vehicle road spectrum loads are collected on public roads to obtain the corresponding load damage under different driving conditions. Finally, the user driving conditions are correlated with the damage objectives of real vehicle road spectrum loads, resulting in durability test damage objectives that are linked to the user.

Key words: vehicle durability test; K-means algorithm; operating conditions segmentation; user data; damage target

摘要： 作为整车开发过程中重要的环节之一，整车结构耐久试验具有对整车结构和车辆系统 进行重要验证和考核的功能。制定一个与用户关联，符合用户实际使用的整车耐久试验目标 尤为关键。文章提出了一种基于用户行驶工况的整车耐久试验目标制定方法。首先，基于用 户实际行驶数据利用机器学习的聚类算法获取用户典型行驶工况以及各工况里程占比；其次， 进行了公共道路实车路谱载荷采集，获取对应行驶工况下的载荷损伤；最后将用户行驶工况 与实车路谱载荷损伤相互关联，获得了与用户关联的整车耐久试验损伤目标。

关键词: 整车耐久试验；K-means 算法；工况划分；用户数据；损伤目标