中国机械工程 ›› 2021, Vol. 32 ›› Issue (11): 1307-1314.DOI: 10.3969/j.issn.1004-132X.2021.11.006

• 机械基础工程 • 上一篇    下一篇

基于非参数核密度估计法的车辆大数据服役载荷外推方法

于佳伟1;郑松林2;赵礼辉2;井清3   

  1. 1. 上海机动车检测认证技术研究中心有限公司,上海,201805
    2. 上海理工大学机械工程学院,上海,200093
    3. 上汽集团商用车技术中心,上海,200438
  • 出版日期:2021-06-10 发布日期:2021-06-25
  • 作者简介:于佳伟,男,1990年生,工程师、博士。研究方向为车辆强度耐久性与可靠性评价、车辆服役载荷特征统计与数据挖掘。E-mail:jia_wei_yu@126.com。
  • 基金资助:
    上海汽车工业科技发展基金(1740);
    上海机动车检测认证技术研究中心有限公司科研课题(KY-2020-18-整,KY-2020-19-整,KY-2020-39整)

Extrapolation Method of Vehicle Big-data Service-loads Based on Non-parametric KDE Method#br#

YU Jiawei1;ZHENG Songlin2;ZHAO Lihui2;JING Qing3   

  1. 1. Shanghai Motor Vehicle Inspection Certification & Tech Innovation Center Co., Ltd., Shanghai, 201805
    2. College of Mechanical Engineering, University of Shanghai for Science and Technology, Shanghai, 200093
    3. SAIC Commercial Vehicle Technology Center, Shanghai, 200438
  • Online:2021-06-10 Published:2021-06-25

摘要: 提出了基于非参数核密度估计法的用户使用习惯特征统计方法和车辆全寿命周期服役载荷外推方法。以某型宽体轻客的全国大范围用户使用习惯数据调研和道路载荷谱大数据采集为基础,运用非参数一维核密度估计法,研究了90%分位下用户使用习惯典型特征的统计方法。运用非参数二维核密度估计法结合蒙特卡罗仿真,研究了基于用户道路实测载荷的车辆全寿命周期服役载荷外推方法。此外,研究了不同核密度估计法多次外推结果的变化特性,研究结果表明多次载荷外推保持了较好的载荷损伤与分布的一致性。

关键词: 服役载荷, 核密度估计, 雨流矩阵外推, 耐久性试验

Abstract: A customer-usage-behavior characteristic statistical method and vehicle life cycle service-load extrapolation method were presented based on non-parametric KDE method. After the national wide investigation of customer-usage-behavior data survey and big-data collection of road load spectra, a statistical analysis method for 90% quantile customer usage typical behavior characteristics was studied by applying non-parametric one-dimensional KDE method. Combining non-parametric two-dimensional KDE method with Monte Carlo simulation, a vehicle life cycle service-load extrapolation method was carried out based on the collected customer road loads. Furthermore, the variation characteristics of multiple extrapolations with different KDE methods were studied. Results show good consistency of load spectra damage and distribution of multiple load extrapolations. 

Key words: service-load, kernel density estimation(KDE), rain-flow matrix extrapolation, durability test

中图分类号: