A fuzzy TOPSIS and Rough Set based approach for mechanism analysis of product infant failure
Introduction
Quality improvement is a routine mission of the product engineering, and the optimization of product infant failure rate is the most important and difficult task (Köksal et al., 2011, Gall et al., 2001). Infant failures have often been referred to the quality problems occurred in the “infant” region of leftmost of bathtub curve of product life (Roesch, 2012), and infant failures are assumed to be caused mainly by design vulnerabilities and manufacturing defects (Jiang and Murthy, 2009). Because of the high occurrence possibility of early failures in infant life, the infant failure rate always keeps unexpectedly high and received extensive attention.
In order to accelerate the decrease of the infant failure rate, the burn-in test is employed for screening potential early life defects. By testing the manufactured items under accelerated stress conditions (that is increased temperature, voltage stress etc.), early failures are weeded out. Therefore, high reliability of the delivered products could be ensured (Kurz et al., 2014). Despite the widespread application of the burn-in test to detect infant failures in product reliability engineering, research on the analysis approach of product infant failure mechanism is few. Most current researches are focused on the estimation of lifetime distributions from perspectives of traditionally statistical inferences.
It is a basic assumption that every failure has explainable physical or functional causes related to hardware, software and service (Nakao et al., 2009). The goal of failure mechanism analysis is to identify the root causes, determine what the sensitive factors including design, manufacturing and usage are involved, and quantify the results (Bernstein, 2014). The root cause is the most basic causal factor or factors that, if corrected or removed, will prevent the recurrence of the failure. Identifying root causes is the key task of failure mechanism analysis, and the correctness of the results is crucial to preventing similar failure occurrences in the future (Kapur and Pecht, 2014).
In the age of big data (Gil et al., 2014, Einav and Levin, 2014), with advances in automation and computer systems, intelligent quality improvement of industrial products and processes requires integration and analysis of big data to solve quality related manufacturing problems (Montgomery, 2014, Meeker and Hong, 2014, Köksal et al., 2011). How to identify and confirm the accurate root causes set of product infant failure from the lifecycle quality and reliability data is a prerequisite to develop continuous in-process quality improvement (Shi, 2013). Traditional methods such as Failure Mode and Effect Analysis (FMEA) and Fault Tree Analysis (FTA) (O’Connor and Kleyner, 2012) are always confined to reliability data and statistical analysis. Still, the general analysis approach for product infant failures in the context of big data and artificial intelligence methodology has not drawn the attention it deserves. Driven by these requirements, the paper puts forward a general technical approach for mechanism analysis for engineering product infant failures by integration the lifecycle quality and reliability data and artificial intelligence techniques of Rough Set and fuzzy TOPSIS for the first time, which could intelligently decompose early fault symptoms into critical functional, design and production parameters in the form of relational tree based on the domain mapping in Axiomatic Design (Suh, 2001). Specifically, this approach emphasizes the integrated application of the artificial intelligence techniques in dealing with this type of problems, in 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 of root causes.
The main contributions of this paper are
- a)
A general decomposition method of root causes based on relational tree for product infant failure is brought forth. This approach could intelligently decompose the early fault symptoms into root causes of critical functional, design and production parameters systematically.
- b)
The failure relation weight computation of root causes is formulated as multi-criteria decision making problem (MCDM) for considering both quantitative and qualitative attributes. Fuzzy is incorporated to remove the vagueness of the decision process. Although, the fuzzy TOPSIS method used to solve the problem is not novel, its application in the mechanism analysis of product infant failure is proposed for the first time here.
- c)
New attributes to compute the weight of root causes are proposed after several rounds of discussions with the industries.
- d)
Rough Set is adopted to mine the value of each criterion directly from historical quality evaluation data. Decision makers are indirectly involved for setting the mining rules for identification of weight values.
The rest of the paper is organized as follows: in Section 2 we give brief literature review of infant failure analysis and fuzzy MCDM. Section 3 expounds the connotation of product infant failure. In Section 4 we focus on preliminary definitions of Rough Set and fuzzy TOPSIS. Section 5 presents the intelligent analysis method of product infant failure mechanism based on the relational tree. Section 6 offers a case study of an infant failure analysis of car NVH. Finally, Section 7 provides conclusions and perspectives.
Section snippets
State of the art
A number of researchers about infant failure have focused on the estimation method of lifetime distributions from perspectives of traditionally statistical inferences (Kuo and Kim, 1999; Gall et al., 2001; Bebbington et al., 2009). However, only a few research works have been done on the mechanism analysis and root causes identification of product infant failure so far. In this section, some important analysis works about infant failure mechanisms and root causes are summarized.
We have used
Product reliability evolution model in the life cycle
There is an evolution process of product reliability proposed by Murthy (2010) in product life cycle, including design reliability, inherent reliability, reliability at sale and field reliability, as shown in Fig. 1.
As shown in Fig. 1, design reliability is determined by the business objectives and reliability specifications and it defines what the ideal reliability state of a product should be. Via the procurement, machining, assembling and delivery inspection processes in the production of
Theoretical background
Although Rough Set and fuzzy TOPSIS are a well discussed concept with several applications in artificial intelligence, the theoretical background of methodology proposed in this paper is highlighted in this section. In many situations where performance rating and weights cannot be given precisely, the Rough Set theory is introduced to model the uncertainty of human judgments and such problems is known as fuzzy multiple criteria decision making (FMCDM). Pawlak (1982) first introduced Rough Set
Relational tree FFPP of product infant failure
Main assignment of the intelligent analysis method is to build a general relational tree of product infant failures based on the collected data, namely to form the hierarchical tree of function requirements, physical parameter and process variables of a product. In view of the following reasons, a new approach of constructing relational tree of product infant failure is put forward: ① Product assembly tree is too large to directly map product infant failure to its corresponding nodes. ② Product
Case study
During the early usage of cars, the problem of noise vibration harshness (NVH) of body stands out in customer complaints. It is a typical infant failure of car, and how to identify the formation mechanism and identify the root causes of this failure is always a dilemma for car producers. Therefore, a case study of a NVH failure analysis with the proposed methodology is provided in this section.
NVH failure descriptions in detail are presented in Table 3.
The main task of this illustrative example
Conclusions and perspectives
In this paper, the connotation and intelligent analysis method of product infant failure mechanism are proposed, and the artificial intelligence techniques of fuzzy TOPSIS and Rough Set are being applied firstly to identify the root causes of the product infant failure. We would conclude with the following remarks:
① Clarification of product reliability evolution in the life cycle and integration model of quality and reliability data in production based on the extended QR chain are beneficial to
Acknowledgments
This research was supported by Grant 61473017 from the National Natural Science Foundation of China. The authors would like to thank the editors and the anonymous referees for their valuable comments and suggestions for improving this paper.
References (49)
Performance measurement model for Turkish aviation firms using the rough-AHP and TOPSIS methods under fuzzy environment
Expert Syst. Appl.
(2011)The inclusion-based TOPSIS method with interval-value dintuitionistic fuzzy sets for multiple criteria group decision making
Appl. Soft Comput.
(2015)- et al.
Fuzzy reliability analysis with only censored data
Eng. Appl. Artif. Intell.
(2014) - et al.
Impact of quality variations on product reliability
Reliab. Eng. Syst. Saf.
(2009) Effects of manufacturing defects on the device failure rate
J. Korean Stat. Soc.
(2013)- et al.
A review of data mining applications for quality improvement in manufacturing industry
Expert Syst. Appl.
(2011) An investigation of a hybrid CBR method for failure mechanisms identification
Eng. Appl. Artif. Intell.
(2004)- et al.
Knowledge-assisted BIM-based visual analytics for failure root cause detection in facilities management
Autom. Constr.
(2014) - et al.
Three typical failure scenarios of the mind process of design from the axiomatic design perspective
CIRP Ann.—Manuf. Technol.
(2009) - et al.
Implementing algorithms of rough set theory and fuzzy rough set theory in the R package “rough sets”
Inf. Sci.
(2014)
Application of fuzzy TOPSIS to scoring the negotiation offers in ill-structured negotiation problems
Eur. J. Oper. Res.
Using a new bathtub curve to correlate quality and reliability
Microelectron. Reliab.
A fuzzy analytic network process model to mitigate the risks associated with offshore wind farms
Expert Syst. Appl.
A fuzzy TOPSIS based approach for e-sourcing
Eng. Appl. Artif. Intell.
A novel general approach to evaluating the reliability of gas turbine system
Eng. Appl. Artif. Intell.
A rough set-based fault ranking prototype system for fault diagnosis
Eng. Appl. Artif. Intell.
Fuzzy theory applied in quality management of distributed manufacturing system: a literature review and classification
Eng. Appl. Artif. Intell.
Fuzzy sets
Inf. Control
Simulation based fuzzy TOPSIS approach for group multi-criteria supplier selection problem
Eng. Appl. Artif. Intell.
Employing fuzzy TOPSIS and SWOT for supplier selection and order allocation problem
Int. J. Adv. Manuf. Technol.
Modelling N- and W-shaped hazard rate functions without mixing distributions
Proc. Inst. Mech. Eng. O: J. Risk Reliab.
Reliability Prediction from Burn-In Data Fit to Reliability Models
Root cause analysis in engineering failures
Trans. Indian Inst. Met.
Failure identification in smart grids based on petri net modeling
IEEE Trans. Ind. Electron.
Cited by (86)
Integrated predictive maintenance approach for multistate manufacturing system considering geometric and non-geometric defects of products
2022, Reliability Engineering and System SafetyArtificial intelligence in industrial design: A semi-automated literature survey
2022, Engineering Applications of Artificial IntelligenceThe use of multi-criteria decision-making methods in business analytics: A comprehensive literature review
2022, Technological Forecasting and Social ChangeCitation Excerpt :The authors indicated that their proposed TOPSIS approach demonstrated the least error accuracy and worked better than standard approaches. The study proposed by He et al. (2016) argued that in the context of big data and artificial intelligence methodology, the general analysis approach to product baby breakdowns did not get the attention it deserved. Based on these needs, they proposed an approach for mechanism analysis for engineering product infant breakdowns by combining the lifecycle quality and reliability data and AI approaches of rough set and fuzzy TOPSIS, which could wisely decompose premature fault indications into critical functional, design and production parameters.
Developing friendlier biodiesel production process via systematic inherent safety interventions
2021, Journal of Cleaner ProductionCitation Excerpt :As a heuristic and holistic procedure, the 5M1E approach has been employed as a guide for identifying the possible sources of quality degradation (Han et al., 2008; Kim et al., 2003). Theoretically, the variations from 5M1E during the process design, operation, and maintenance could be transmitted to the process quality and eventually cripple safety and reliability (He et al., 2016). For incorporating proactive safeguards into a relatively full extent, the SIS framework (see Fig. 3) was adopted to guide the inherent safety interventions.
A bibliometric analysis and cutting-edge overview on fuzzy techniques in Big Data
2020, Engineering Applications of Artificial IntelligenceCitation Excerpt :The proposed approach did better classification on the real-time data as well. He et al. (2016) for the first time applied the artificial intelligence techniques of fuzzy TOPSIS and Rough Set in extracting the actual reasons of the product infant failure. The authors proposed the connotation and intelligent analysis method of product infant failure mechanism.
Combining fuzzy MCDM with Kano model and FMEA: a novel 3-phase MCDM method for reliable assessment
2024, Annals of Operations Research