计算机集成制造系统 ›› 2016, Vol. 22 ›› Issue (第11期): 2602-2612.DOI: 10.13196/j.cims.2016.11.013

• 产品创新开发技术 • 上一篇    下一篇

基于证据理论和矩阵相似度的工艺失效风险评估

聂文滨1,2,刘卫东1,2,汪建东3,曾天3,胡坤2   

  1. 1.南昌大学机电工程学院
    2.南昌航空大学信息工程学院
    3.中航工业江西洪都航空工业股份有限公司制造工程部
  • 出版日期:2016-11-30 发布日期:2016-11-30
  • 基金资助:
    国家自然科学基金资助项目(71161018,71661023)。

Evaluation approach in process failure risk analysis based on matrix similarity and evidence theory

  • Online:2016-11-30 Published:2016-11-30
  • Supported by:
    Project supported by the National Natural Science Foundation,China(No.71161018,71661023).

摘要: 针对传统工艺失效模式及影响分析进行风险评估时存在的工作量大、容易出现偏差等问题,提出基于证据理论的工艺失效风险群体评估方法。首先,确定风险因子分析所涉及的评估维度、相应的折算系数、各评估维度下的完备等级子块集;其次,收集各评估维度下的专家意见,将其转换为公共模糊识别框架内的专家意见,依据最大熵原理和最大模糊期望效用,确定风险因子的最终专家意见,再用矩阵相似度计算专家权重,集结各专家对风险因子的最终意见,得到风险因子的群体评估意见。用模糊证据推理算法集结各风险因子的群体评估意见,得到工艺失效模式的综合风险群体评估意见,并进行风险排序。通过某型飞机活动舱盖装配工艺失效风险评估的应用研究结果,证明了该方法的可行性与有效性。

关键词: 风险因子分析, 公共模糊识别框架, 证据推理, 矩阵相似度, 最大熵

Abstract: To overcome the shortages of traditional PFMEA which suffered heavy working load and more mistakes,the group evaluation approach was proposed on the basis of evidence theory.The evaluation dimension of risk factors,the corresponding converted coefficients and the corresponding complete sets of grade parts were determined.Various forms of experts' opinions were collected and transformed into the experts' opinions in the basic fuzzy recognition framework.According to maximum entropy principle and maximum value of fuzzy expected utilities,the final opinions of risk factors given by each expert were confirmed,and group opinions of risk factors were obtained by calculating the experts' weights with matrix similarity and aggregating the experts' final opinions on risk factors.Group opinions of risk factors were aggregated by using fuzzy evidential reasoning algorithm,and the comprehensive risk group opinion of process failure mode was obtained and sorted.A risk analysis on the assembly process of aircraft portable cover was given to illustrate the applicability and effectiveness of proposed approach.

Key words: risk factor analysis, basic fuzzy recognition framework, evidential reasoning, matrix similarity, maximum entropy

中图分类号: