上海交通大学学报(自然版) ›› 2012, Vol. 46 ›› Issue (05): 701-705.

• 机械仪表工程 • 上一篇    下一篇

基于小数据集贝叶斯网络建模的偏差源诊断方法

刘银华,金隼   

  1. (上海交通大学 上海市数字化汽车车身工程重点实验室, 上海 200240)
  • 收稿日期:2011-09-15 出版日期:2012-05-28 发布日期:2012-05-28
  • 基金资助:

    国家自然科学基金项目(51175340), 机械系统与振动国家重点实验室开放基金课题资助(MSVMS201006)

Bayesian Networks Modeling Based on Small Samples for VariationSource Diagnosis

 LIU  Yin-Hua, JIN  Sun   

  1. (Shanghai Key Laboratory of Digital Autobody Engineering, Shanghai Jiaotong University, Shanghai 200240, China)
  • Received:2011-09-15 Online:2012-05-28 Published:2012-05-28

摘要:  针对装配过程尺寸偏差的小样本检测条件,提出了基于条件独立性检验的结构学习算法,结合柔性装配偏差关系模型,推导了贝叶斯网络子节点的先验条件概率,将小数据集与先验概率融合并获得贝叶斯网络参数,实现了装配偏差影响因素的贝叶斯网络建模,并用于某车型侧围装配过程的偏差源诊断.结果表明,所提出的偏差源诊断方法具有较高的准确性.

关键词: 装配偏差, 贝叶斯网络, 故障诊断

Abstract: Based on the small samples collected in the assembly process, a new approach based on Bayesian networks was proposed for the variationsource diagnsosis. The conditional independence testing algorithm was proposed to obtain the structure of Bayesian networks. After the prior conditonal probabilities are calculated based on the mapping of the variation simulation model, posterior conditonal probabilities are updated by incorporating the small sample data. The results of the body side case show the method is effective and acurate for fixture fault diangosis.

Key words: assembly variation, Bayesian networks, fault diagnosis

中图分类号: