复杂不确定环境下基于FMEA的产品质量与可靠性管理方法研究

Quality management plays an important role in manufacturing and services, and the quality of products can be a decisive factor for consumers to make decision about their purchases. Failure mode and effect analysis (FMEA) is one of the most popular quality and reliability analysis tools for identifying, assessing and eliminating potential failure modes in a wide range of industries. However, the conventional risk priority number (RPN) method has been considerably criticized for various reasons. First, it only considers the occurrence (O), severity (S) and detection (D) risk factors, disregarding other important risk factors which may also influence failure modes. Moreover, the dependency between risk factors is not taken into account. Second, the traditonal FMEA assumes that all of the risk factors have equal importance. In practice, this assumption is unrealistic since they are generally differ on the weights of risk factors. Moreoer, the information concerning risk factor weights is usually incompletely known or even completely unknown due to time pressure, lack of knowledge or expert’s limited expertise about the FMEA problem.Third, the risks of failure modes are difficult for FMEA team members to precisely determine. The FMEA team often demonstrates different opinions and produces different types of assessment information because of its cross-functional and multidisciplinary nature. Fourth, the mathematical formula for the calculation of RPN is questionable and lacks a complete scientific basis. For example, interrelationships among failure modes and effects of a system are not considered. Thus, the risk of a failure mode may be underestimated when it has multiple effects. To slove these problems, in this project, we will develop a risk factor conceptual framework through literature review and questionnaire survey and analyze the dependency of risk factors combinging DEMATEL and ISM methods; propose two multiple objective optimization models based on TOPSIS and Shannon Entropy for determing the weights of risk factors with incomplete and unknown weight information; manage various uncertainties in the risk assessment information using hesitant 2-tuple linguistic term sets and hybrid averaging operator; and propose a risk ranking model intergraing graph therory and matrix approach and modified ELECTRE method. Finally, based on the above researches, an integrated FMEA model and its reliability management system will be deveoped, whch can be used to analyze the overall risk of failure modes under complex and uncertain environments, and provide more effective information for decision makers to enhance the reliability and safety of products and services.

产品质量是现代制造和服务业赢得市场竞争的关键因素。本项目针对故障模式及影响分析(FMEA)技术在可靠性管理实践中所面临的彼此相关的四个重要问题加以研究。首先，考虑风险因子的多维性，利用文献分析和问卷调查的方法构建因子层次框架模型，并集成DEMATEL和ISM方法分析其关联性。其次，考虑风险因子权重信息的不完全性，建立基于TOPSIS方法的多目标规划模型，提出基于信息熵的因子权重计算方法。第三，由于故障模式评价的不确定性和多样性，提出基于犹豫语言集和混合加权平均算子的风险评价信息获取策略。第四，考虑故障模式相互影响关系，提出基于图-矩阵的影响关系分析算法和基于ELECTRE的故障模式风险排序模型。最后，在上述研究的基础上建立综合FMEA模型和可靠性管理原型系统，从而帮助工程设计人员有效运用FMEA技术识别系统中关键故障模式，提前采取预防改进措施，提高复杂产品和服务的质量与可靠性。

产品质量是现代企业赢得市场的关键，也是国家竞争力的综合体现。质量与可靠性管理理论与方法是产品质量与可靠性管理的前提与基础，对提高中国制造在国际市场的竞争能力具有重要意义。本项目针对故障模式及影响分析(FMEA)技术在可靠性管理实践中所面临的重要问题，开展了以下研究：(1)通过研究风险因子的多维性，构建了风险因子层次框架模型及其关联关系分析方法；(2)通过考虑风险因子权重信息的不完全性，研究建立了一种主观客观权重相结合的风险因子权重计算方法；(3)通过研究故障模式评价的不确定性和多样性，设计了一种不确定风险评价信息获取策略；(4)在考虑故障模式相互影响关系的基础上，构建了故障模式影响相互关系分析算法和风险排序模型；(5)最后开发了一个集成FMEA模型的可靠性管理原型系统，并推广应用到企业的实际风险管理中，有效提高了复杂产品和服务的质量与可靠性。本项目完成了预定的主要研究内容，取得了超过预期的研究成果。在IEEE Transactions、RESS、IJPE、IJPR等国际知名学术期刊上累计发表论文60篇，在国内重要期刊上发表论文32篇，出版英文专著1部。项目成果得到了国内外同行与业界的关注与积极评价，其中5篇论文入选ESI高被引论文，企业应用成果获得上海市自然科学二等奖奖1项。项目执行期间，培养博士和硕士研究生16名，邀请10余名国内外学者来校合作交流，项目组成员出国合作交流9人次。此外，通过本项目的研究积累，项目负责人入选国际质量科学院IAQ准院士，荣获“全面质量管理推进40周年杰出推进者”等称号。