A fuzzy TOPSIS and Rough Set based approach for mechanism analysis of product infant failure
A fuzzy TOPSIS and Rough Set based approach for mechanism analysis of product infant failure
复制标题
基于模糊TOPSIS和粗糙集的产品初期失效机理分析方法
DOI：
10.1016/j.engappai.2015.06.002
复制
发表时间：
2016-01-01
影响因子：
8
通讯作者：
Xie, Min
中科院分区：
文献类型：
--
作者：
He, Yi-Hai;Wang, Lin-Bo;Xie, Min
产品早期失效的根本原因识别是当今产品质量改进的重要课题之一。提出了一种基于公理设计领域映射和产品全生命周期质量可靠性数据关系树的产品早期失效机理分析技术方法。该方法能够将早期故障征兆智能地分解为功能域中的关键功能参数、物理域中的设计参数和工艺域中的工艺参数的根源。该方法综合考虑了质量和可靠性两类定性和定量属性，强调综合应用粗糙集和模糊TOPSIS人工智能技术计算产品早期失效的根本原因权重。为了枚举产品早期失效的潜在根源，首先基于产品生命周期可靠性演化模型给出了产品早期失效的内涵，并基于扩展QR链给出了产品生产过程质量与可靠性数据集成模型。然后，基于公理化设计中的功能域、物理域和过程域，研究了产品早期失效关系树的分解方法。综合应用粗糙集和模糊TOPSIS（Technique for Order Preference by Similarity to Ideal Solution）方法，将根原因（关系树节点）的失效关系权重计算问题视为多准则决策问题（MCDM），其中粗糙集用于质量数据的挖掘，模糊TOPSIS用于失效关系权重计算过程的建模。最后，通过对某轿车车身噪声振动粗糙度投诉故障的分析，验证了该方法的有效性，结果表明，该方法有利于提高复杂产品故障根源识别的智能化水平。(C)2015爱思唯尔有限公司版权所有。
Root causes identification of product infant failure is nowadays one of the critical topics in product quality improvements. This paper puts forward a novel technical approach for mechanism analysis of product infant failure based on domain mapping in Axiomatic Design and the quality and reliability data from product lifecycle in the form of relational tree. The proposed method could intelligently decompose the early fault symptoms into root causes of critical functional parameters in function domain, design parameters in physical domain and process parameters in process domain successively. More specifically, both qualitative and quantitative attributes of quality and reliability types are considered for solving the root causes weight computation problem of product infant failure, this approach emphasizes the integrated application of artificial intelligence techniques of Rough Set and fuzzy TOPSIS to compute the weight of root causes. In order to enumerate the latent root causes of product infant failure, connotation of product infant failure based on the product reliability evolution model in the life cycle and data integration model of quality and reliability in production based on the extended QR chain are presented firstly. Then, a decomposition method for relational tree of product infant failure is studied based on domains of functional, physical and process in Axiomatic Design. The failure relation weight computation of root causes (nodes of relational tree) is considered as multi-criteria decision making problem (MCDM) by integrated application of Rough Set and fuzzy TOPSIS (Technique for Order Preference by Similarity to Ideal Solution), which the Rough Set is used to mining the quality data and fuzzy TOPSIS is adopted to model the computation process of failure relation weight. Finally, the validity of the proposed method is verified by a case study of analyzing a car infant failure about body noise vibration harshness complaint, and the result proves that the proposed approach is conducive to improve the intelligent level of root causes identification for complex product infant failure. (C) 2015 Elsevier Ltd. All rights reserved.