Abstract
Failure modes and effects analysis (FMEA), as a practical and easy-to-use reliability assessment tool, has been widely applied across various fields of researches. At the same time, it also receives criticisms for its limited selection of risk factors as well as its discrete numerical ordinal scales. In an attempt to fill in this gap, this paper proposes an improved FMEA method based on analytic network process (ANP) with probabilistic linguistic term sets (PLTSs). Firstly, the three risk factors of FMEA, namely occurrence, severity and detection, are broken down to more elaborate and specific sub-factors, and a network representing the influential relationship between these sub-factors is constructed. ANP is then utilized to derive the relative weights of factors, sub-factors and failure modes by making pairwise comparisons with the help of PLTS. To verify the rationality and applicability of our proposed method, a case study of hospital information system reliability assessment is carried out. Comparative analyses with other existing FMEA methods are also undertaken to highlight the differences and advantages of our proposed method.
Similar content being viewed by others
References
Stamatis, D.H.: Failure Mode and Effect Analysis: FMEA from Theory to Execution. Quality Press, Milwaukee (2003)
Subriadi, A.P., Najwa, N.F.: The consistency analysis of failure mode and effect analysis (FMEA) in information technology risk assessment. Heliyon 6(1), e03161 (2020)
Wu, Z., Liu, W., Nie, W.: Literature review and prospect of the development and application of FMEA in manufacturing industry. Int. J. Adv. Manuf. Technol. 112(5–6), 1409–1436 (2021)
Asan, U., Soyer, A.: Failure mode and effects analysis under uncertainty: a literature review and tutorial. In: Intelligent Decision Making in Quality Management, pp. 265–325. Springer, Cham (2016)
Lipol, L.S., Haq, J.: Risk analysis method: FMEA/FMECA in the organizations. Int. J. Basic Appl. Sci. 11(5), 74–82 (2011)
Dandachi, E., El Osman, Y.: Application of AHP Method for Failure Modes and Effect Analysis (FMEA) in Aerospace Industry for Aircraft Landing System. Eastern Mediterranean University (EMU)-Doğu Akdeniz Üniversitesi (DAÜ) (2017)
Ying, L.Z.G.: Consideration about the validity of aerospace product FMEA. Spacecr. Eng. 1, 142–146 (2011)
Wu, Z., et al.: Nuclear product design knowledge system based on FMEA method in new product development. Arab. J. Sci. Eng. 39(3), 2191–2203 (2014)
Panchal, D., Kumar, D.: Risk analysis of compressor house unit in thermal power plant using integrated fuzzy FMEA and GRA approach. Int. J. Ind. Syst. Eng. 25(2), 228–250 (2017)
Guimarães, A.C.F., Lapa, C.M.F.: Fuzzy FMEA applied to PWR chemical and volume control system. Prog. Nucl. Energy 44(3), 191–213 (2004)
Baynal, K., Sarı, T., Akpınar, B.: Risk management in automotive manufacturing process based on FMEA and Grey relational analysis: a case study. Adv. Prod. Eng. Manag. 13(1), 69–80 (2018)
Yousefi, S., et al.: HSE risk prioritization using robust DEA-FMEA approach with undesirable outputs: a study of automotive parts industry in Iran. Saf. Sci. 102, 144–158 (2018)
Ramere, M.D., Laseinde, O.T.: Optimization of condition-based maintenance strategy prediction for aging automotive industrial equipment using FMEA. Procedia Comput. Sci. 180, 229–238 (2021)
Chiozza, M.L., Ponzetti, C.: FMEA: a model for reducing medical errors. Clin. Chim. Acta 404(1), 75–78 (2009)
Liu, H.-C.: Improved FMEA Methods for Proactive Healthcare Risk Analysis. Springer, Cham (2019)
Wang, L., et al.: A linguistic risk prioritization approach for failure mode and effects analysis: a case study of medical product development. Qual. Reliab. Eng. Int. 35(6), 1735–1752 (2019)
Liu, Z., et al.: FMEA using the normalized projection-based TODIM-PROMETHEE II model for blood transfusion. Int. J. Fuzzy Syst. 23(4), 1–17 (2021)
Arabian-Hoseynabadi, H., Oraee, H., Tavner, P.: Failure modes and effects analysis (FMEA) for wind turbines. Int. J. Electr. Power Energy Syst. 32(7), 817–824 (2010)
Silva, M.M., et al.: A multidimensional approach to information security risk management using FMEA and fuzzy theory. Int. J. Inf. Manag. 34(6), 733–740 (2014)
Li, X., et al.: Assessing information security risk for an evolving smart city based on fuzzy and Grey FMEA. J. Intell. Fuzzy Syst. 34(4), 2491–2501 (2018)
Saaty, T.L.: Decision making—the analytic hierarchy and network processes (AHP/ANP). J. Syst. Sci. Syst. Eng. 13(1), 1–35 (2004)
Saaty, T.L.: Fundamentals of the analytic network process—dependence and feedback in decision-making with a single network. J. Syst. Sci. Syst. Eng. 13(2), 129–157 (2004)
Saaty, T.L.: Theory and Applications of the Analytic Network Process: Decision Making with Benefits, Opportunities, Costs, and Risks. RWS Publications, Pittsburgh (2005)
Saaty, T.L., Vargas, L.G.: Decision Making with the Analytic Network Process, vol. 282. Springer, Boston (2006)
Pang, Q., Wang, H., Xu, Z.: Probabilistic linguistic term sets in multi-attribute group decision making. Inf. Sci. 369, 128–143 (2016)
Bluvband, Z., Grabov, P.: Failure analysis of FMEA. In: 2009 Annual Reliability and Maintainability Symposium. IEEE (2009)
Lo, H.-W., Liou, J.J.H.: A novel multiple-criteria decision-making-based FMEA model for risk assessment. Appl. Soft Comput. 73, 684–696 (2018)
Lo, H.-W., et al.: A hybrid MCDM-based FMEA model for identification of critical failure modes in manufacturing. Soft Comput. 24(20), 15733–15745 (2020)
Zandi, P., et al.: Agricultural risk management using fuzzy TOPSIS analytical hierarchy process (AHP) and failure mode and effects analysis (FMEA). Agriculture 10(11), 504 (2020)
Khalilzadeh, M., Balafshan, R., Hafezalkotob, A.: Multi-objective mathematical model based on fuzzy hybrid multi-criteria decision-making and FMEA approach for the risks of oil and gas projects. J. Eng. Des. Technol. 18(6), 1997–2016 (2020)
Kumar Dadsena, K., Sarmah, S.P., Naikan, V.N.A.: Risk evaluation and mitigation of sustainable road freight transport operation: a case of trucking industry. Int. J. Prod. Res. 57(19), 6223–6245 (2019)
Kiani Aslani, R., Feili, H.R., Javanshir, H.: A hybrid of fuzzy FMEA-AHP to determine factors affecting alternator failure causes. Manag. Sci. Lett. 4(9), 1981–1984 (2014)
Li, X.-Y., et al.: A novel failure mode and effect analysis approach integrating probabilistic linguistic term sets and fuzzy Petri nets. IEEE Access 7, 54918–54928 (2019)
Kutlu, A.C., Ekmekçioğlu, M.: Fuzzy failure modes and effects analysis by using fuzzy TOPSIS-based fuzzy AHP. Expert Syst. Appl. 39(1), 61–67 (2012)
Başhan, V., Demirel, H., Gul, M.: An FMEA-based TOPSIS approach under single valued neutrosophic sets for maritime risk evaluation: the case of ship navigation safety. Soft Comput. 24(24), 18749–18764 (2020)
Mete, S.: Assessing occupational risks in pipeline construction using FMEA-based AHP–MOORA integrated approach under Pythagorean fuzzy environment. Hum. Ecol. Risk Assess. Int. J. 25(7), 1645–1660 (2019)
Huang, J., et al.: An improved reliability model for FMEA using probabilistic linguistic term sets and TODIM method. Ann. Oper. Res. (2019). https://doi.org/10.1007/s10479-019-03447-0
Yazdi, M.: Improving failure mode and effect analysis (FMEA) with consideration of uncertainty handling as an interactive approach. Int. J. Interact. Des. Manuf. 13(2), 441–458 (2018)
Yucesan, M., Gul, M., Celik, E.: A holistic FMEA approach by fuzzy-based Bayesian network and best–worst method. Complex Intell. Syst. 7(3), 1547–1564 (2021)
Fattahi, R., Khalilzadeh, M.: Risk evaluation using a novel hybrid method based on FMEA, extended MULTIMOORA, and AHP methods under fuzzy environment. Saf. Sci. 102, 290–300 (2018)
Liu, H.-C., Liu, L., Li, P.: Failure mode and effects analysis using intuitionistic fuzzy hybrid weighted Euclidean distance operator. Int. J. Syst. Sci. 45(10), 2012–2030 (2013)
Bhattacharjee, P., Dey, V., Mandal, U.K.: Risk assessment by failure mode and effects analysis (FMEA) using an interval number based logistic regression model. Saf. Sci. 132, 104967 (2020)
Abdelgawad, M., Fayek, A.R.: Risk management in the construction industry using combined fuzzy FMEA and fuzzy AHP. J. Constr. Eng. Manag. 136(9), 1028–1036 (2010)
Sayyadi Tooranloo, H., Saghafi, S.: Assessing the risk of hospital information system implementation using IVIF FMEA approach. Int. J. Healthc. Manag. (2020). https://doi.org/10.1080/20479700.2019.1688504
Qin, J., Xi, Y., Pedrycz, W.: Failure mode and effects analysis (FMEA) for risk assessment based on interval type-2 fuzzy evidential reasoning method. Appl. Soft Comput. 89, 106134 (2020)
Li, G.-F., et al.: Advanced FMEA method based on interval 2-tuple linguistic variables and TOPSIS. Qual. Eng. 32(4), 653–662 (2019)
Ko, W.-C.: Exploiting 2-tuple linguistic representational model for constructing HOQ-based failure modes and effects analysis. Comput. Ind. Eng. 64(3), 858–865 (2013)
Chang, K.-H., Wen, T.-C., Chung, H.-Y.: Soft failure mode and effects analysis using the OWG operator and hesitant fuzzy linguistic term sets. J. Intell. Fuzzy Syst. 34(4), 2625–2639 (2018)
Huang, J.: Research on FMEA Improvement and Application Within Complex and Uncertain Environment, p. 174. School of Management, Shanghai University, Shanghai (2019)
Zadeh, L.A.: The concept of a linguistic variable and its application to approximate reasoning—I. Inf. Sci. 8(3), 199–249 (1975)
Herrera, F., Herrera-Viedma, E., Verdegay, J.L.: A sequential selection process in group decision making with a linguistic assessment approach. Inf. Sci. 85(4), 223–239 (1995)
Xu, Z.: Linguistic Decision Making. Springer, Berlin (2012)
Rodríguez, R.M., Martínez, L., Herrera, F.: A group decision making model dealing with comparative linguistic expressions based on hesitant fuzzy linguistic term sets. Inf. Sci. 241, 28–42 (2013)
Liao, H., Mi, X., Xu, Z.: A survey of decision-making methods with probabilistic linguistic information: bibliometrics, preliminaries, methodologies, applications and future directions. Fuzzy Optim. Decis. Mak. 19(1), 81–134 (2019)
Motevali Haghighi, S., Torabi, S.A.: Business continuity-inspired fuzzy risk assessment framework for hospital information systems. Enterp. Inf. Syst. 14(7), 1027–1060 (2019)
Yucel, G., et al.: A fuzzy risk assessment model for hospital information system implementation. Expert Syst. Appl. 39(1), 1211–1218 (2012)
Lotfi, R., et al.: Viable medical waste chain network design by considering risk and robustness. Environ. Sci. Pollut. Res. (2021). https://doi.org/10.1007/s11356-021-16727-9
Lotfi, R., et al.: Hybrid fuzzy and data-driven robust optimization for resilience and sustainable health care supply chain with vendor-managed inventory approach. Int. J. Fuzzy Syst. 24, 1–16 (2022)
Lotfi, R., et al.: Resource-constrained time–cost–quality–energy–environment tradeoff in project scheduling by considering blockchain technology: a case study of healthcare project. Res. Sq. (2021). https://doi.org/10.21203/rs.3.rs-1088054/v1
Saaty, R.W.: The analytic hierarchy process—what it is and how it is used. Math. Model. 9(3–5), 161–176 (1987)
Forman, E.H., Gass, S.I.: The analytic hierarchy process—an exposition. Oper. Res. 49(4), 469–486 (2001)
Afsharnia, F., and A. Marzban. Risk analysis of sugarcane stem transportation operation delays using the FMEA-ANP hybrid approach. Journal of Agricultural Machinery 9.2 (2019)
Zammori, F., Gabbrielli, R.: ANP/RPN: a multi criteria evaluation of the Risk Priority Number. Qual. Reliab. Eng. Int. 28(1), 85–104 (2012)
Liu, Y., Eckert, C.M., Earl, C.: A review of fuzzy AHP methods for decision-making with subjective judgements. Expert Syst. Appl. 161, 113738 (2020)
Chang, D.-Y.: Applications of the extent analysis method on fuzzy AHP. Eur. J. Oper. Res. 95(3), 649–655 (1996)
Author information
Authors and Affiliations
Corresponding authors
Ethics declarations
Conflict of interest
The authors declare that we have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendices
Appendix 1: Abbreviation and Notation List
Abbreviation list
AHP | Analytic hierarchy process |
ANP | Analytic network process |
BWM | Best–worst method |
COPRAS | COmplex PRoportional ASsessment of alternatives |
FMEA | Failure modes and effects analysis |
GRA | Grey relation analysis |
HLTS | Hesitant linguistic term set |
HIS | Hospital information system |
LTS | Linguistic term set |
MCDM | Multi-criteria decision making |
MOORA | Multi-objective optimization by ratio analysis |
MULTIMOORA | Multiple multi-objective optimization by ratio analysis |
NASA | National Aeronautics and Space Administration |
PLTS | Probabilistic linguistic term set |
RPN | Risk prioritization number |
SAW | Simple additive weighting |
TFN | Triangular fuzzy numbers |
TODIM | An acronym in Portuguese for interactive multi-criteria decision making |
TOPSIS | Technique for order preference by similarity to an ideal solution |
VIKOR | Vlse kriterijumska optimizacija kompromisno resenje |
Notation list
\(O\) | Probability of occurrence |
\(S\) | Severity |
\(D\) | Likelihood of detection |
\({\text{RPN}}\) | Risk prioritization number |
\(t_{\alpha }\) | Linguistic terms |
\(T\) | Additive linguistic term set |
\(\overline{T}\) | Continuous linguistic term set |
\({\text{neg}}(t_{\alpha } )\) | Negation operator of linguistic term set |
\(\max (t_{\alpha } ,\;t_{\beta } )\) | Maximum operator of linguistic term set |
\(\min (t_{\alpha } ,\;t_{\beta } )\) | Minimum operator of linguistic term set |
\(b_{T}\) | Hesitant linguistic term set |
\(L(p)\) | Probabilistic linguistic term set |
\(L^{(k)} (p^{(k)} )\) | Linguistic term \(L^{(k)}\) associated with probability \(p^{(k)}\) |
\(\# L(p)\) | Number of all different linguistic terms in \(L(p)\) |
\(g\left( {L(p)} \right)\) | Numerical score of \(L(p)\) |
\(e_{z} \;(z = 1:Z)\) | The \(z\)th expert |
\({\text{FM}}_{m} \;(m = 1:M)\) | The \(m\)th failure mode |
\(O_{i} \;(i = 1:h_{O} )\) | The \(i\)th sub-factors under the category Occurrence \(O\) |
\(S_{i} \;(i = 1:h_{S} )\) | The \(i\)th sub-factors under the category Severity \(S\) |
\(D_{i} \;(i = 1:h_{D} )\) | The \(i\)th sub-factors under the category Detection \(D\) |
\(A\) | Comparison matrix of risk factors \(O\), \(S\) and \(D\) |
\(L_{ij} (p)\) | Comparing result of the \(i\)th risk factor to the \(j\)th risk factor |
\(\overline{{L_{ij} (p)}}\) | Negation of \(L_{ij} (p)\) |
\(G(A)\) | Crisp-valued matrix obtained by calculating the numerical score |
\({\text{CL}}_{ij}^{r}\) | Consistent preference of the \(i\)th element over the \(j\)th element through the \(r\)th element |
\({\text{Consistency}}\;(G(A))\) | Consistency level of \(A\) |
\(\gamma\) | Required consistency level threshold |
\(w\) | Weighting vector of three main risk factors |
\(A^{O} ,\;A^{S} ,\;A^{D}\) | Comparison matrix of sub-factors |
\(\omega^{O} ,\;\omega^{S} ,\;\omega^{D}\) | Weighting vector of sub-factors |
\(Q\) | Super-matrix representing the influence of sub-factors |
\(A^{{O_{i} }}\) | Comparison matrix of sub-factors regarding their influences on \(O_{i}\) |
\((\dot{\omega }_{O1}^{Oi} ,\;\dot{\omega }_{O2}^{Oi} , \ldots ,\dot{\omega }_{{Oh_{O} }}^{Oi} )^{{\text{T}}}\) | Normalized principal eigenvector of \(G(A^{{O_{i} }} )\) |
\(||Q^{c} - Q^{c - 1} ||_{2}\) | Euclidean norm of the deviation between \(Q^{c}\) and \(Q^{c - 1}\) |
\(\ddot{\omega }\) | Long-term stable weight vector of sub-factors |
\(A_{FM}^{{O_{i} }}\) | Comparison matrix of failure modes regarding the \(i\)th sub-factor \(O_{i}\) |
\(\varpi^{{O_{i} }}\) | Weight of failure modes with respect to the \(i\)th sub-factor \(O_{i}\) |
\(RPN_{m}\) | RPN of the \(m\)th failure mode |
\({\text{RPN}}_{m}^{C}\) | group assessment of the RPN of the \(m\)th failure mode |
\(\varphi (L^{(k)} )\) | Transformation function from linguistic terms to AHP scale |
\(\hat{\varphi }(\rho )\) | Transformation function from PLTS numerical score to TFN scale |
Appendix 2: Computation Process of Comparative Methods
1.1 Classical FMEA Method
As explained before, in classical FMEA method, the expert is instructed to rate the failure modes with a discrete numerical scale of 1–10. To ensure the fairness of comparison, the first step is to convert the entries in failure modes’ comparison matrix from PLTS to the numerical scale.
For example, the comparison of failure mode \({\text{FM}}_{1}\) to failure model \({\text{FM}}_{2}\) under sub-factor \(O_{1}\) is \(\{ t_{5} (0.8),\;t_{6} (0.2)\}\), then via Eq. (3) this entry is defuzzied to \(5 \times 0.8 + 6 \times 0.2 = 5.2\), then rounded to \(5\) because the ratings in classical FMEA is required to be integers. With similar calculations, expert \(e_{1}\)’s comparison matrices of failure modes with respect to sub-factor \(O_{1}\) is obtained, as shown in Table 17.
In Table 17, the entries in the first row represents the criticality of failure mode \({\text{FM}}_{1}\) compared to all other failure modes with respect to sub-factor \(O_{1}\). Thus, the arithmetic mean of the entries in the first row of Table 17 can be seen as the criticality rating of failure mode \({\text{FM}}_{1}\) with respect to sub-factor \(O_{1}\). The same argument goes for other failure modes, and their criticality ratings are list in the first row of Table 18. Corresponding calculation results for sub-factor \(O_{2}\) and \(O_{3}\) are also listed in the second and third row of Table 18, with their calculation process omitted for compactness.
Seeing that in classical FMEA method, only three main risk factors are considered, the criticality ratings of failure modes under sub-factors \(O_{i} \;(i = 1:3)\) need to be further aggregated via the average operator, to derive the ratings under the main risk factor \(O\), the result of which is listed in the first row of Table 19. Similar computations are also carried out for main risk factors \(S\) and \(D\), as shown in the second and third rows of Table 19. Finally, the \({\text{RPN}}\) can be calculated, and the ranking of failure modes via classical FMEA method can be obtained:
1.2 Kiani Aslani et al.’s Method [32]
To counter the problem of equal weight assumption, Kiani Aslani R. et al. took the advantage of AHP to derive the relative weights of main risk factors, and defined the RPN as the weighted sum of these three factors. In [32], TFN is utilized to accommodate the uncertainty of expert judgements. Therefore, the comparison matrix in Table 5 needs to be converted into TFN to ensure the fairness of comparison. This conversion process is carried out in two steps. Firstly, the PLTS is converted into triangular forms by setting the lower and higher bounds as \(L^{{(\min \{ k|L^{(k)} (p^{(k)} ) \in L(p)\} )}}\) and \(L^{{(\max \{ k|L^{(k)} (p^{(k)} ) \in L(p)\} )}}\) respectively, and the center as \(L^{(g(L(p)))}\). For instance, in Table 5, the comparison result of risk factor \(O\) to risk factor \(S\) is \(\{ t_{3} (0.3),\;t_{4} (0.6)\}\). It is easy to see that \(\min \{ 3,\;4\} = 3\), \(\max \{ 3,\;4\} = 4\) and \(g(L(p)) = 3 \times 0.3\)\(+ 4 \times 0.6 = 3.3\). Hence, the triangular form of \(\{ t_{3} (0.3),\;t_{4} (0.6)\}\) is \((t_{3} ,\;t_{3.3} ,\;t_{4} )\).
Secondly, the triangular form of PLTS is transformed to TFN according to their practical meanings. In AHP method [64], the pairwise comparison of two elements are conducted with instructions in Table 20.
Thus, according to the practical meanings in Table 4, the linguistic terms can be translated to the scale of AHP with the following equation:
And the TFN obtained after conversion is:
Continuing with the above instance, via Eq. (17) the linguistic terms \(t_{3}\), \(t_{3.3}\) and \(t_{4}\) are translated to:
Therefore, the PLTS \(\{ t_{3} (0.3),\;t_{4} (0.6)\}\) can be converted to a TFN \((0.33,\;0.42,\;1)\). Identical conversions are also carried out for other entries in Table 5, the results are presented in Table 21.
For two TFNs \(\lambda_{1} = (a_{1} ,\;b_{1} ,\;c_{1} )\) and \(\lambda_{2} = (a_{2} ,\;b_{2} ,\;c_{2} )\), their operation laws [65] are defined as:
-
(1)
\((a_{1} ,\;b_{1} ,\;c_{1} ) \oplus (a_{2} ,\;b_{2} ,\;c_{2} ) = (a_{1} + a_{2} ,\;b_{1} + b_{2} ,\;c_{1} + c_{2} )\);
-
(2)
\((a_{1} ,\;b_{1} ,\;c_{1} ) \otimes (a_{2} ,\,b_{2} ,\;c_{2} ) = (a_{1} a_{2} ,\;b_{1} b_{2} ,\;c_{1} c_{2} )\);
-
(3)
\(\mu \otimes (a_{1} ,\;b_{1} ,\;c_{1} ) = (\mu a_{1} ,\;\mu b_{1} ,\;\mu c_{1} )\);
-
(4)
\((a_{1} ,\;b_{1} ,\;c_{1} )^{ - 1} = (1/c_{1} ,\;1/b_{1} ,\;1/a_{1} )\);
-
(5)
\(\tilde{g}(\lambda_{1} ) = (a_{1} + 4b_{1} + c_{1} )/6\).
Suppose the pairwise comparison matrix is \(\tilde{A} = \left[ {\lambda_{i}^{j} } \right]_{n \times n}\), then the TFN weight of each element is calculated with the following formula [65]:
Thus, according to the comparison matrix in Table 21, the TFN weights of the three risk factors can be calculated via Eq. (19):
Similar aggregation process as described in “Appendix 2.1” is also undertaken to derive the ratings of failure modes with respect to three main risk factors. The only difference is that in “Appendix 2.1” the ratings are given in integers, while here the ratings are transformed to TFNs, as shown in the second to fourth columns of Table 22.
Suppose the ratings of failure modes under risk factors \(O\), \(S\) and \(D\) are \(\lambda^{O}\), \(\lambda^{S}\) and \(\lambda^{D}\), respectively, then the RPN of failure modes are defined as:
The calculation results are shown in the last two columns of Table 22, and finally the ranking of failure modes can be obtained for Kiani Aslani R. et al.’s method [32]:
\(\begin{gathered} {\text{FM}}_{14} \succ {\text{FM}}_{13} \succ {\text{FM}}_{10} \succ {\text{FM}}_{11} \succ {\text{FM}}_{12} \succ {\text{FM}}_{9} \succ {\text{FM}}_{7} \hfill \\ \succ {\text{FM}}_{1} \succ {\text{FM}}_{5} \succ {\text{FM}}_{4} \succ {\text{FM}}_{3} \succ {\text{FM}}_{6} \succ {\text{FM}}_{8} \succ {\text{FM}}_{2} . \hfill \\ \end{gathered}\).
1.3 Abdelgawad and Fayek’s Method [43]
In [43], the authors broke down risk factor \(S\) to sub-factors and utilized classical AHP to derive their relative weights. Similar to the conversion process in “Appendix 2.2”, the entries of comparison matrix in Table 7 are first defuzzied via Eq. (3), then translated to the scale of AHP method via Eq. (17). Take the comparison result of sub-factor \(S_{1}\) to sub-factor \(S_{2}\) as an example, \(\{ t_{5} (0.5),\;t_{6} (0.2)\}\) is first converted to \(5 \times 0.5 + 6 \times 0.2 = 3.7\). Then according to the practical meanings listed in Table 20, linguistic term \(t_{3.7}\) corresponds to \(1/(2 \times (4 - 3.7) + 1) = 1/1.6 = 0.625\) in classical AHP method. The converted comparison matrix is as shown in Table 23. Moreover, the entries of the same column in Table 23 are normalized, then entries of the same row are summed together, which are then normalized to obtain the weights of corresponding sub-factors, as listed in the last column in Table 23.
Similar as in “Appendix 2.1”, the ratings of failure modes under different sub-factors are derived by calculating the arithmetic mean of the entries in the same row of the pairwise comparison matrices like Table 17. Because in [43] only the risk factor \(S\) is broken down to sub-factors while \(O\) and \(D\) remains the same as in classical FMEA method, here the ratings of failure modes under risk factor \(S\) is calculated as the weighted sum of ratings under \(S_{i} \;(i = 1:5)\) with the weighting vector from Table 23. The ratings under risk factors \(O\) and \(D\) are the same as in “Appendix 2.1”, where sub-factors \(O_{i} \;(i = 3)\) and \(D_{i} \;(i = 1:3)\) are assigned equal importance. The calculation results are as listed in Table 24.
With the ratings under \(O\), \(S\) and \(D\) obtained, the RPNs of failure modes can be calculated, as shown in the last column of Table 24. Thus, the ranking of failure modes via Abdelgawad and Fayek’s method [43] is:
1.4 Zandi et al.’s Method [29]
In [29], Zandi P. et al. broke down the risk factor \(S\) to sub-factors and utilized AHP to derive both the weights of three main risk factors and the weights of sub-factors under the category of \(S\). Seeing that in [29] the experts express their assessments in the form of TFNs, to ensure the fairness of comparison, first the comparison matrices in Tables 5 and 7 need to be converted to relevant forms.
As described in Table 25, a special kind of TFN scale is used in [29] where the deviation between lower bound and center of the TFN, as well as the deviation between the center and higher bound of the TFN is 2. In order to conform to this norm, first the PLTS entries in Tables 5 and 7 is defuzzied via the numerical score function in Eq. (3), then rounded to a positive integer. Suppose \(\rho = \left\lfloor {g(L(p))} \right\rfloor\), then the conversion is completed by the following function:
The comparison result of risk factor \(O\) to risk factor \(S\) in Table 5, \(\{ t_{3} (0.3),\;t_{4} (0.6)\}\), is taken once again as an example. It is easy to see that \(g\left( {L(p)} \right) = 3 \times 0.3 + 4 \times 0.6 = 3.3\), then \(\rho = \left\lfloor {g(L(p))} \right\rfloor\)\(= 3\). Via Eq. (20), the TFN scale after conversion is:
Likewise, other entries in Tables 5 and 7 are converted following the above procedures, the results are presented in Tables 26 and 27.
Here, we introduce the principle of comparing TFNs. For any two TFNs \(\lambda_{1} = (a_{1} ,\;b_{1} ,\;c_{1} )\) and \(\lambda_{2} = (a_{2} ,\;b_{2} ,\;c_{2} )\), the possibility degree of \(\lambda_{1} \ge \lambda_{2}\) is defined as:
Moreover, the possibility degree that one TFN \(\lambda = (a,\;b,\;c)\) is greater than \(k\) other TFNs \(\lambda_{i} \;(i = 1:k)\) is defined as:
Next, let us suppose the pairwise comparison matrix is \(\hat{A} = \left[ {\hat{\lambda }_{i}^{j} } \right]_{n \times n}\), then the synthetic extent with respect to the \(i\)th element is defined as:
Then, the normalized possibility degree of \(V_{i} \ge V_{j} \;(j \ne i)\) is regarded as the weight of the \(i\)th element derived from comparison matrix.
For instance, according to the comparison matrix in Table 26, via Eq. (23) we have:
Using Eqs. (21) and (22),
Furthermore, the vector \((0.94,\;1,\;0.35)^{{\text{T}}}\) is normalized to \((0.41,\;0.44,\;0.15)^{{\text{T}}}\), which serves as the weighting vector of the three main risk factors. The same computing process is also performed for the comparison matrix in Table 27, the result of which is shown in the first row of Table 28. Same as in “Appendix 2.3”, the ratings of failure modes are transformed to TFN then defuzzied, as shown in Table 28. Then, the RPN of failure modes can be defined as their weighted sum, as presented in the last column of Table 28. Finally, the ranking of failure modes via Zandi et al.’s method [29] is obtained:
\(\begin{gathered} {\text{FM}}_{13} \succ {\text{FM}}_{14} \succ {\text{FM}}_{10} \succ {\text{FM}}_{11} \succ {\text{FM}}_{12} \succ {\text{FM}}_{1} \succ {\text{FM}}_{9} \hfill \\ \succ {\text{FM}}_{7} \succ {\text{FM}}_{4} \succ {\text{FM}}_{5} \succ {\text{FM}}_{6} \succ {\text{FM}}_{3} \succ {\text{FM}}_{8} \succ {\text{FM}}_{2} . \hfill \\ \end{gathered}\).
Rights and permissions
About this article
Cite this article
Zhang, Zx., Yang, L., Cao, Yn. et al. An Improved FMEA Method Based on ANP with Probabilistic Linguistic Term Sets. Int. J. Fuzzy Syst. 24, 2905–2930 (2022). https://doi.org/10.1007/s40815-022-01302-2
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40815-022-01302-2