欢迎访问《汽车安全与节能学报》,

汽车安全与节能学报 ›› 2023, Vol. 14 ›› Issue (2): 173-181.DOI: 10.3969/j.issn.1674-8484.2023.02.004

• 汽车安全 • 上一篇    下一篇

基于DTBN与T-S故障树的电池热管理系统可靠性分析

柳炽伟(), 郭美华   

  1. 中山职业技术学院 机电工程学院,中山市 528403,中国
  • 收稿日期:2022-09-14 修回日期:2023-01-19 出版日期:2023-04-30 发布日期:2023-04-27
  • 作者简介:柳炽伟(1970—),男(汉),广东,副教授。E-mail: luke1011@sohu.com
  • 基金资助:
    广东省教育厅特色创新科研项目(2020KTSCX334);中山市社会公益与基础研究项目(2020B2028)

Reliability analysis of battery thermal management system based on DTBN and T-S fault tree

LIU Chiwei(), GUO Meihua   

  1. Institute of Mechanical and Electrical Engineering, Zhongshan Polytechnic, Zhongshan, 528403, China
  • Received:2022-09-14 Revised:2023-01-19 Online:2023-04-30 Published:2023-04-27

摘要:

提出一种评估电动汽车动力电池热管理系统的可靠性分析算法。构建了电池热管理系统T-S动态故障树,转化为离散时间Bayes网络(DTBN)模型。同时将T-S动态门规则转化为网络节点的条件概率表。依据所建可靠性模型和部件的故障率,计算得到动力电池热管理系统在任务时间内的故障概率值为0.453,并获得各部件的后验概率、概率重要度和关键重要度。结果表明:对比Monte Carlo仿真方法,本方法的故障概率计算值误差小于5%,概率重要度靠前的是单体电池温度传感器、电池冷却液泵、电池冷却液管路等部件。该方法能克服传统故障树分析难以构建Bayes网络条件概率表等问题。

关键词: 动力电池, 热管理系统, 可靠性分析, 离散时间Bayes网络 (DTBN), T-S动态故障树

Abstract:

A reliability analysis algorithm for evaluating the thermal management system of electric vehicle power batteries was proposed. The T-S dynamic fault tree of the battery thermal management system was constructed, which was transformed into a Discrete-Time Bayesian Network (DTBN) model. Meanwhile, the T-S dynamic gate rules were transformed into the conditional probability table of network nodes, and the reliability analysis algorithms were proposed. According to the established reliability model and the failure rates of components, the failure probability value of the power battery thermal management system in the task time was 0.453 by calculated, and the posterior probability, probability importance, and the key importance of each component were calculated. The results show that: Compared with the Monte Carlo simulation method, the error of the calculated value of fault probability were less than 5%. The components with the highest probability importance are the single battery temperature sensor, battery coolant pump, battery coolant pipeline, and other components. This method can overcome the difficulties of traditional fault tree analysis in constructing conditional probability tables for Bayesian networks.

Key words: power battery, thermal management system, reliability analysis, discrete-time Bayesian network (DTBN), T-S dynamic fault tree

中图分类号: