T.C BAHÇEŞEHİR UNIVERSITY A METHOD FOR MANAGING THE RISK: A HYBRID APPROACH TO FMEA Master’s Thesis ALİ KAAN PASTIRMACI İSTANBUL, 2014 T.C BAHÇEŞEHİR UNIVERSITY The Graduate School of Natural and Applied Sciences Industrial Engineering A METHOD FOR MANAGING THE RISK: A HYBRID APPROACH TO FMEA Master’s Thesis ALİ KAAN PASTIRMACI Supervisor: Asst. Prof. Dr. Ahmet BEŞKESE İSTANBUL, 2014 T.C BAHÇEŞEHİR UNIVERSITY THE GRADUATE SCHOOL OF NATURAL AND APPLIED SCİENCES INDUSTRIAL ENGINEERING Title of the Master’s Thesis : A Method for Managing the Risk: A Hybrid …Approach to FMEA Name/Last Name of the Student : Ali Kaan PASTIRMACI Date of Thesis Defense : 11.06.2014 The thesis has been approved by the Graduate School of Natural and Applied Sciences. Assoc. Prof. Dr. Tunç BOZBURA Director ................................... I certify that this thesis meets all the requirements as a thesis for the degree of Master of Science. Assoc. Prof. Dr. BarıĢ SELÇUK Program Coordinator …………………….. This is to certify that we have read this thesis and that we find it fully adequate in scope, quality and content, as a thesis for the degree of Master of Science. Examining Committee Members: Asst. Prof. Dr. Ahmet BEġKESE (supervisor) : Prof. Dr. Selim ZAĠM : Assoc. Prof. Dr. BarıĢ SELÇUK : ACKNOWLEDGEMENTS First of all I would like to thank Asst. Prof. Ahmet BEġKESE, my thesis supervisor, for his guidance, patience, help, and comments in improving this thesis. Without his instruction, guidance and encouragements, this thesis could not be accomplished. I would also thank ġafak UTAġ for her help and comments about occupational health and safety concept. Secondly, I would also like to express my love and gratitude to my mother for her encouragements, motivation and support during my master program and master thesis study. Lastly, to Ece: Thank you for being with me and for your love. Istanbul, 2014 Ali Kaan PASTIRMACI iv ABSTRACT A METHOD FOR MANAGING THE RISK: A HYBRID APPROACH TO FMEA Ali Kaan PASTIRMACI Industrial Engineering Master Program Supervisor: Asst. Prof. Dr. Ahmet BeĢkese June 2014, 76 pages In today’s competitive business world, companies from all sectors are aiming maximum profit with minimum effort. In order to accomplish this goal, companies can ignore or do not give importance to the potential failures that may give devastating harm to their reputation. These potential failure modes do not only endangering the health/safety of workers and their working conditions but also the reputation of the company. Therefore, companies should also take precautions for potential failure mode to fulfill the obligations of occupational health and safety (OHS) and maintain the reputation of the company. OHS concept requires a detailed and renewable risk analysis methodology. A risk analysis methodology that is used for eliminating the risk related to OHS concept should be simple, straightforward and easy to apply. Indeed, a vast majority of these failure modes can be prevented in advance. A risk analysis methodology which is suitable to the characteristic of the company may help. Risk analysis should be applied by experienced and knowledgeable analysts. In this thesis, one of the most widely used risk analysis methodologies, Failure Mode and Effect Analysis, is proposed. Failure mode and effects analysis is a widely used engineering technique for designing, identifying and eliminating known and/or potential failures, problems, errors and so on from system, design, process, and/or service before they reach the costumer (Schneider & Stamatis 1996). Unfortunately, traditional FMEA methodology has several shortcomings. This work has been planned to eliminate these shortcomings with the help of Fuzzy Analytical Hierarchy Process (Fuzzy AHP) and Grey Relational Analysis (GRA). Fuzzy AHP method is used to determine the importance weights for Decision Makers (DMs) and to calculate the criteria weights of decision factors both for First Risk Priority Number (RPN1) and Second Risk Priority Number (RPN2). RPN1 is calculated by using decision factors, occurrence, severity and detectability (O, S and D) with the help of GRA methodology. According to RPN1 values, FMEA team prioritizes the failure modes and determines the proper corrective actions. The team also determines the threshold intervals for corrective actions to make the work more realistic. Then, RPN2 is calculated by means of five additional decision factors (criteria). Criteria consider the cost, time, regulatory obligations, prevention policy of company, and reputation of company. Thanks to RPN2, a contribution to the literature is made. After that, RPN3 is obtained from the summation of RPN1 and RPN2 that are multiplied with their corresponding coefficients. Finally, FMEA team reprioritizes the corrective actions according to RPN3 values and the corrective actions are performed according to this prioritization. vi To demonstrate the effectiveness of the proposed methodology, a case study is applied in a yarn manufacturing company from Turkey. Keywords: Risk Analysis, Occupational Health and Safety, FMEA, GRA, fuzzy AHP. vii ÖZET RĠSK YÖNETĠMĠ ĠÇĠN BĠR YÖNTEM: HTEA’NE HĠBRĠT BĠR YAKLAġIM Ali Kaan PASTIRMACI Endüstri Mühendisliği Yüksek Lisans Programı Tez DanıĢmanı: Doç. Dr. Ahmet BeĢkese Haziran 2014, 76 sayfa Günümüz rekabetçi iĢ dünyasında, tüm sektörlerdeki Ģirketler en az çaba ile en fazla karı elde etmeyi hedeflemektedirler. Bu gayelerini gerçekleĢtirebilmek için Ģirketin itibarına önemli derecede zarar verebilecek olan potansiyel hataları ya görmezden gelirler ya da önem vermezler. Bu potansiyel hata türleri sadece iĢçilerin sağlığını/güvenliğini ve çalıĢma ortamlarını tehlikeye atmakla kalmayıp, Ģirketin itibarını da tehlikeye atmaktadırlar. Bu yüzden, Ģirketler, iĢ sağlığı ve güvenliliğinin gereklerini yapabilmek için ve de Ģirketin itibarını devam ettirebilmek için potansiyel hata türleri için önleyici tedbirleri almalıdırlar. ĠSG hususu, detaylı ve yenilenebilir bir risk analizi metodolojisi gerektirir. ĠSG hususu ile ilgili riskleri gidermek için kullanılacak risk analiz metodolojisi, basit anlaĢılır ve uygulaması kolay olmalıdır. Aslında bu hata türlerinin çok büyük bir bölümü önceden engellenebilecektir. ġirketlerin karakteristik özelliklerine uygun risk analiz yöntemi bu konuda yardımcı olabilir. Risk analizi tecrübeli ve bilgili analistler tarafından yapılmalıdır. Bu tez, çok geniĢ kullanıma sahip olan bir risk analiz yöntemi olan Hata Türü ve Etkileri Analizi (HTEA) öne sürmektedir. Hata türü ve etkileri analizi, potansiyel hata türlerini, problemleri, aksaklıkları, benzerlerini sistem, tasarım, süreç ve/veya servis üzerinden müĢteriye ulaĢmadan önce tasarlayan, tanımlayan ve ortadan kaldıran çok geniĢ kullanıma sahip bir mühendislik tekniğidir (Schneider & Stamatis 1996). Ne yazık ki, geleneksel HTEA metodolojisi birçok eksikliği içermektedir. Bu çalıĢma, Bulanık Analitik HiyerarĢi Proses (Bulanık AHP) ve Gri ĠliĢkisel Analiz (GĠA) yardımıyla bu eksiklikleri gidermeyi planlamıĢtır. Bulanık AHP, Karar Vericilere önem ağırlıkları vermek ve de birinci risk öncelik sayısının (RÖS1) ve ikinci risk öncelik sayısının (RÖS2) karar faktörlerini ağırlıklandırmak amacıyla kullanılmıĢtır. RÖS1, karar faktörlerini (ortaya çıkma durumu, Ģiddet, tespit edilebilirlik) kullanarak ve GĠA yardımıyla hesaplanır. RÖS1 değerlerine göre HTEA takımı hata türlerini önceliklendirir ve bunlara uygun düzeltici faaliyetleri belirler. Takım ayrıca yapılan çalıĢmanın daha gerçekçi olabilmesi için sıralanan bu düzeltici faaliyetler için eĢik sayısı belirler. Sonrasında RÖS2 beĢ ek karar faktörleri (kriterler) ile hesaplanır. Kriterler, maliyeti, zamanı, kanuni gereklilikleri, iĢçilerin sağlığını/güvenliğini, ürünün/ servisin kalite artırımını, müĢteri memnuniyetini ve Ģirketin itibarını değerlendirmektedir. RÖS2 sayesinde, literatüre katkı yapılmıĢtır. Daha sonar RÖS3 değeri, uygun katsayılarla çarpılmıĢ olan RÖ1 ve RÖS2 nin toplamıyla elde viii edilir. Son olarak takım, düzeltici faaliyetleri RÖS3 değerlerine göre tekrardan önceliklendirir ve düzeltici faaliyetleri bu sıraya göre gerçekleĢtirir. Öne sürülen bu metodolojinin etkinliği ispat edebilmek için Türkiye’den bir iplik üretim Ģirketinde örnek çalıĢma gerçekleĢtirilmiĢtir. Anahtar Kelimeler: Risk Analizi, ĠĢ Sağlığı ve Güvenliği, HTEA, GĠA, Bulanık AHP. ix CONTENTS TABLES……………...……………………………………………………………….......xv FIGURES………………………………………………………………………………..xvii ABBREVIATIONS………………………………………………………………….....xviii LIST OF SYMBOLS………………………………………………………………….…xix 1. INTRODUCTION……………………………………………………………….....1 2. LITERATURE REVIEW………………………………………………………….5 2.1 RISK ANALYSIS AND OCCUPATIONAL SAFETY AND HEALTH…....5 2.1.1 Risk Assessment and Risk Analysis………………………...……...........5 2.1.2 Occupational Health and Safety (OHS).................................................10 2.2 REVIEW OF THE MODELS..........................................................................14 2.2.1 Failure Mode and Effects Analysis (FMEA).........................................14 2.2.1.1 Types of FMEA...........................................................................19 2.2.1.1.1 System FMEA..............................................................19 2.2.1.1.2 Design FMEA..............................................................19 2.2.1.1.3 Service FMEA..............................................................20 2.2.1.1.4 Software FMEA...........................................................20 2.2.1.1.5 Process FMEA.............................................................20 2.2.1.2 Shortcomings of FMEA..............................................................21 3. METHODOLOGY..................................................................................................28 3.1 ANALYTICAL HIERARCHY PROCESS (AHP) AND FUZZY ANALYTICAL HIERARCHY PROCESS (FAHP)....................28 3.1.1 Analytical Hierarchy Process (AHP).......................................................28 3.1.2 Fuzzy Analytical Hierarchy Process (FAHP).........................................29 3.1.2.1 Assigning different weights to decision makers with Fuzzy AHP..........................................................................34 3.1.2.2 Assigning different weights to decision factors (both for RPN1 and RPN2) with Fuzzy AHP..........................38 3.2 GREY RELATIONAL ANALYSIS...............................................................39 3.2.1 Application of Grey Relational Analysis to FMEA..............................41 3.3 THIRD RISK PRIORITY NUMBER (RPN3)...............................................44 x 4. CASE STUDY..........................................................................................................51 4.1 ADOPTATION THE PROPOSED METHODOLOGY TO THE CASE COMPANY............................................................................51 4.1.1 Decision Makers Weighting.....................................................................52 4.1.2 Determining RPN1 values.......................................................................57 4.1.3 Determining RPN2 values.......................................................................65 4.1.4 Determining RPN3 values.......................................................................69 5. RESULTS AND CONCLUSION...........................................................................71 REFERENCES....................................................................................................................77 APPENDICES.....................................................................................................................84 APPENDIX A1 : RPN2 Calculation of CA.2..........................................................85 APPENDIX A2 : RPN2 Calculation of CA.3..........................................................86 APPENDIX A3 : RPN2 Calculation of CA.4..........................................................87 APPENDIX A4 : RPN2 Calculation of CA.5..........................................................88 APPENDIX A5 : RPN2 Calculation of CA.6..........................................................89 APPENDIX A6 : RPN2 Calculation of CA.7..........................................................90 APPENDIX A7 : RPN2 Calculation of CA.8..........................................................91 APPENDIX A8 : RPN2 Calculation of CA.9..........................................................92 APPENDIX A9 : RPN2 Calculation of CA.10........................................................93 APPENDIX A10 : RPN2 Calculation of CA.11......................................................94 APPENDIX A11 : RPN2 Calculation of CA.12......................................................95 APPENDIX A12 : RPN2 Calculation of CA.13......................................................96 APPENDIX A13 : RPN2 Calculation of CA.14......................................................97 APPENDIX A14 : RPN2 Calculation of CA.15......................................................98 APPENDIX A15 : RPN2 Calculation of CA.16......................................................99 APPENDIX A16 : RPN2 Calculation of CA.17....................................................100 APPENDIX A17 : RPN2 Calculation of CA.18....................................................101 APPENDIX A18 : RPN2 Calculation of CA.19....................................................102 APPENDIX A19 : RPN2 Calculation of CA.20....................................................103 APPENDIX A20 : RPN2 Calculation of CA.21…………………………...…….104 APPENDIX A21 : RPN2 Calculation of CA.22………………………................105 APPENDIX A22 : RPN2 Calculation of CA.23…………………………...…….106 APPENDIX A23 : RPN2 Calculation of CA.24……………………………...….107 APPENDIX A24 : RPN2 Calculation of CA.25……………………………...….108 xi APPENDIX A25 : RPN2 Calculation of CA.26…………………………...…….109 APPENDIX A26 : RPN2 Calculation of CA.27……………………………...….110 APPENDIX A27 : RPN2 Calculation of CA.28……………………………...….111 APPENDIX A28 : RPN2 Calculation of CA.29……………………………...….112 APPENDIX A29 : RPN2 Calculation of CA.30……………………………...….113 APPENDIX A30 : RPN2 Calculation of CA.31……………………………...….114 APPENDIX A31 : RPN2 Calculation of CA.32…………………………...…….115 APPENDIX A32 : RPN2 Calculation of CA.33…………………………...…….116 APPENDIX A33 : RPN2 Calculation of CA.34………………………...……….117 APPENDIX A34 : RPN2 Calculation of CA.35…………………………...…….118 APPENDIX A35 : RPN2 Calculation of CA.36……………………………...….119 APPENDIX A36 : RPN2 Calculation of CA.37…………………………...…….120 APPENDIX A37 : RPN2 Calculation of CA.38…………………………...…….121 APPENDIX A38 : RPN2 Calculation of CA.39…………………………...…….122 APPENDIX A39 : RPN2 Calculation of CA.40…………………………...…….123 APPENDIX A40 : RPN2 Calculation of CA.41……………………...………….124 APPENDIX A41 : RPN2 Calculation of CA.42……………………...………….125 APPENDIX A42 : RPN2 Calculation of CA.43……………………...………….126 APPENDIX A43 : RPN2 Calculation of CA.44………………………...……….127 APPENDIX A44 : RPN2 Calculation of CA.45………………………...……….128 APPENDIX A45 : RPN2 Calculation of CA.46………………………...……….129 APPENDIX A46 : RPN2 Calculation of CA.47………………………...……….130 APPENDIX A47 : RPN2 Calculation of CA.48………………………...……….131 APPENDIX A48 : RPN2 Calculation of CA.49………………………...……….132 APPENDIX A49 : RPN2 Calculation of CA.50……………………………...….133 APPENDIX B1: Photos of failure modes as to RPN3 prioritization (Pr.1).......134 APPENDIX B2: Photos of failure modes as to RPN3 prioritization (Pr.2).......135 APPENDIX B3: Photos of failure modes as to RPN3 prioritization (Pr.3).......136 APPENDIX B4: Photos of failure modes as to RPN3 prioritization (Pr.4).......137 APPENDIX B5: Photos of failure modes as to RPN3 prioritization (Pr.5).......138 APPENDIX B6: Photos of failure modes as to RPN3 prioritization (Pr.6).......139 APPENDIX B7: Photos of failure modes as to RPN3 prioritization (Pr.7........140 APPENDIX B8: Photos of failure modes as to RPN3 prioritization (Pr.8).......141 APPENDIX B9: Photos of failure modes as to RPN3 prioritization (Pr.9).......142 xii APPENDIX B10: Photos of failure modes as to RPN3 prioritization (Pr.10)...143 APPENDIX B11: Photos of failure modes as to RPN3 prioritization (Pr.11)...144 APPENDIX B12: Photos of failure modes as to RPN3 prioritization (Pr.12)...145 APPENDIX B13: Photos of failure modes as to RPN3 prioritization (Pr.13)...146 APPENDIX B14: Photos of failure modes as to RPN3 prioritization (Pr.14)...147 APPENDIX B15: Photos of failure modes as to RPN3 prioritization (Pr.15)...148 APPENDIX B16: Photos of failure modes as to RPN3 prioritization (Pr.16)...149 APPENDIX B17: Photos of failure modes as to RPN3 prioritization (Pr.17)...150 APPENDIX B18: Photos of failure modes as to RPN3 prioritization (Pr.18)...151 APPENDIX B19: Photos of failure modes as to RPN3 prioritization (Pr.19)...152 APPENDIX B20: Photos of failure modes as to RPN3 prioritization (Pr.20)...153 APPENDIX B21: Photos of failure modes as to RPN3 prioritization (Pr.21)...154 APPENDIX B22: Photos of failure modes as to RPN3 prioritization (Pr.22)...155 APPENDIX B23: Photos of failure modes as to RPN3 prioritization (Pr.23)...156 APPENDIX B24: Photos of failure modes as to RPN3 prioritization (Pr.24)...157 APPENDIX B25: Photos of failure modes as to RPN3 prioritization (Pr.25)...158 APPENDIX B26: Photos of failure modes as to RPN3 prioritization (Pr.26)...159 APPENDIX B27: Photos of failure modes as to RPN3 prioritization (Pr.27)...160 APPENDIX B28: Photos of failure modes as to RPN3 prioritization (Pr.28)...161 APPENDIX B29: Photos of failure modes as to RPN3 prioritization (Pr.29)...162 APPENDIX B30: Photos of failure modes as to RPN3 prioritization (Pr.30)...163 APPENDIX B31: Photos of failure modes as to RPN3 prioritization (Pr.31)...164 APPENDIX B32: Photos of failure modes as to RPN3 prioritization (Pr.32)...165 APPENDIX B33: Photos of failure modes as to RPN3 prioritization (Pr.33)...166 APPENDIX B34: Photos of failure modes as to RPN3 prioritization (Pr.34)...167 APPENDIX B35: Photos of failure modes as to RPN3 prioritization (Pr.35)...168 APPENDIX B36: Photos of failure modes as to RPN3 prioritization (Pr.36)...169 APPENDIX B37: Photos of failure modes as to RPN3 prioritization (Pr.37)...170 APPENDIX B38: Photos of failure modes as to RPN3 prioritization (Pr.38)...171 APPENDIX B39: Photos of failure modes as to RPN3 prioritization (Pr.39)...172 APPENDIX B40: Photos of failure modes as to RPN3 prioritization (Pr.40)...173 APPENDIX B41: Photos of failure modes as to RPN3 prioritization (Pr.41)...174 APPENDIX B42: Photos of failure modes as to RPN3 prioritization (Pr.42)...175 APPENDIX B43: Photos of failure modes as to RPN3 prioritization (Pr.43)...176 xiii APPENDIX B44: Photos of failure modes as to RPN3 prioritization (Pr.44)...177 APPENDIX B45: Photos of failure modes as to RPN3 prioritization (Pr.45)...178 APPENDIX B46: Photos of failure modes as to RPN3 prioritization (Pr.46)...179 APPENDIX B47: Photos of failure modes as to RPN3 prioritization (Pr.47)...180 APPENDIX B48: Photos of failure modes as to RPN3 prioritization (Pr.48)...181 APPENDIX B49: Photos of failure modes as to RPN3 prioritization (Pr.49)...182 APPENDIX B50: Photos of failure modes as to RPN3 prioritization (Pr.50)...183 APPENDIX C1: Risk analysis and FMEA test………………………………....184 CURRICULUM VITAE…………………………………………………………...........188 xiv TABLES Table 2.1 : Risk assessment methodologies comparison…………………………..... ..8 Table 2.2 : An action plan made by National Occupational Health and Safety Council (2006-2008)…………………………………………...………... .11 Table 2.3 : An action plan made by National Occupational Health and Safety Council (2009-2013)…………………………………………………….. .12 Table 2.4 : The reasons for applying FMEA as a risk management methodology….. .13 Table 2.5 : Classification of risk evaluation methods in FMEA……………………. 26 Table 3.1 9-point ratio scales for AHP……………………………………………. 29 Table 3.2 : Literature review of fuzzy FMEA approaches………………………….. 30 Table 3.3 : Triangular fuzzy numbers for linguistic terms………………………….. 32 Table 3.4 : Criteria of decision makers……………………………………………… 35 Table 3.5 : Criterion-1, job relevance……………………………………………….. 35 Table 3.6 : Criterion-2, experience………………………………………………….. 36 Table 3.7 : Criterion-3, education level……………………………………………... 37 Table 3.8 : Criterion-4, test result…………………………………………………… 37 Table 3.9 : Criterion-5, repetition number…………………………………………... 38 Table 3.10 : Threshold interval for corrective actions………………………………... 44 Table 3.11 : Criteria for RPN2 evaluation…………………………………………..... 45 Table 3.12 : Criterion-1, additional cost……………………………………………… 46 Table 3.13 : Criterion-2, time loss…………………………………………………..... 46 Table 3.14 : Criterion-3, regulation obligatory……………………………………….. 47 Table 3.15 : Criterion-4, workers health & safety…………………………………..... 47 Table 3.16 : Criterion-5, reputation of company……………………………………... 48 Table 4.1 : A list of failure modes…………………………………………………... 51 Table 4.2 : A pairwise comparison matrix of DM1…………………………………. 54 Table 4.3 : A pairwise comparison matrix of DM2…………………………………. 54 Table 4.4 : A pairwise comparison matrix of DM3…………………………………. 55 Table 4.5 : Geometric mean values of fuzzy comparison values…………………… 55 Table 4.6 : Fuzzy weights matrix…………………………………………………..... 55 Table 4.7 : The crisp values of DMs criteria weights……………………………….. 56 Table 4.8 : Criteria score table of DMs……………………………………………... 56 Table 4.9 : The final crisp values of DMs’ weights………………………………… 57 Table 4.10 : DM1 pairwise comparison matrix……………………………………..... 58 Table 4.11 : DM2 pairwise comparison matrix………………………………………. 58 Table 4.12 : DM3 pairwise comparison matrix……………………………………..... 58 Table 4.13 : Geometric mean values of fuzzy comparison values…………………… 58 Table 4.14 : Fuzzy weights matrix…………………………………………………..... 58 Table 4.15 : The crisp values of DMs criteria weights……………………………….. 59 Table 4.16 : FM scores given by three DMs………………………………………….. 59 Table 4.17 : Degree of grey relationship for FMs…………………….……………… 62 Table 4.18 : Final RPN1 calculation………………………………………………….. 63 xv Table 4.19 : Performing C.A. according to the threshold interval……………………. 64 Table 4.20 : Prioritization and categorization of C.A. according to RPN1 values….. 65 Table 4.21 : A pairwise comparison matrices of RPN2’s criteria (DM1)…………..... 66 Table 4.22 : A pairwise comparison matrices of RPN2’s criteria (DM2)……………. 66 Table 4.23 : A pairwise comparison matrices of RPN2’s criteria (DM3)…………..... 67 Table 4.24 : Geometric mean values of fuzzy comparison values (RPN2’s criteria)... 67 Table 4.25 : Fuzzy weights matrix of RPN2’s criteria……………………………….. 67 Table 4.26 : The crisp values of RPN2’s criteria weights……………………………. 68 Table 4.27 : Criteria score table of RPN2 criteria……………………………………. 68 Table 4.28 : RPN2 values CA.1………………………………………………………. 69 Table 4.29 : RPN3 value calculation of CA.1………………………………………... 70 Table 5.1 : RPN1, RPN2, RPN3 and two prioritizations……………………………. 72 Table 5.2 : C.A.s’ priority changes according to the proposed methodology………. 73 xvi FIGURES Figure 2.1 Risk management phases………………………………………………....…..6 Figure 2.2 Suggested 10 point scales for Occurrence, Severity, and Detectability……..16 Figure 2.3 FMEA application process…………………………………..........................18 Figure 3.1 A triangular fuzzy number,M………………………………………………..32 Figure 3.2 FMEA report paper(Sample)…… ……………………….............................50 xvii ABBREVIATIONS AHP : Analytical Hierarchy Process ANP : Analytical Network Process CF : Cause of failure CA. : Corrective action DEA : Data Envelopment Analysis DF : Decision factor DM : Decision Maker D : Detectability ETA : Event Tree Analysis ER : Evidential Reasoning FE : Failure effects FM : Failure mode FMEA Failure Mode and Effects Analysis FMECA : Failure Mode, Effects and Criticality Analysis FTA : Fault Tree Analysis FRPN : Fuzzy Risk Priority Number CA.1 : First corrective action RPN1 : First Risk Priority Number GRA : Grey Relational Analysis HAZOP : Hazard and operability Analysis JSA : Job Safety Analysis MIL-STD : Military Standard MCDM : Multiple Criteria Decision Making O : Occurrence OHS : Occupational Health and Safety PHA : Process Hazard Analysis RPN : Risk Priority Number RPN2 : Second Risk Priority Number S : Severity RPN3 : Third Risk Priority Number TFN : Triangular Fuzzy Number xviii SYMBOLS Piarwise comparision matrix that established by decision maker k : Ak TFN that demonstrates the comparison of criteria : aij Fuzzy multiplication : Geometric mean of fuzzy comparison value of criterion i to each criterion : ri Fuzzy weight of criterion I : wi A crisp weight number : wr Comparative series : Xi(k) Standard series : X0(k) Differences between comparative series and standard series : Grey relational coefficient : Γ An identifier coefficient : The weighting coefficient of the risk factors (O, S, and D) : Β RPN1 coefficient : ρ1 RPN2 coefficient : ρ2 Degree of grey relation : τi 1 1. INTRODUCTION 1.1 MOTIVATION OF THE RESEARCH All companies want to make maximum profit with minimal capital, minimum number of workers and without any accident or event that hinder the process. Unfortunately, it is not possible in the normal condition. Risk and uncertainty are associated with all projects undertaken by individuals and organizations, regardless of their size, nature, and place of execution (Abdelgawad et al. 2010). Instead of trying to deal with a problem that is appeared, it makes more sense to apply an effective risk analysis methodology. Although there are so many methodologies about risk management, each of them has some limitations together with their advantages. Risk analysis methodology should be applied by talented and experienced experts. Experts should also have detailed information about the process so they can determine the most appropriate risk analysis method. One of the most widely used risk analysis method is Failure Mode and Effects Analysis FMEA. Failure mode and effects analysis is a widely used engineering technique for designing, identifying and eliminating known and/or potential failures, problems, errors and so on from system, design, process, and/or service before they reach the costumer (Schneider & Stamatis 1996). In this method, the failure modes are identified and ranked with help of Risk Priority Number (RPN) by FMEA team. RPN is the product of occurrence (O), severity (S) and detection (D) of failures. That is, RPN = O*S*D (Yang et al. 2011). Based on RPNs, corrective actions of failure modes are prioritized and then proper corrective actions are performed by responsible person. FMEA is chosen because it can be applied in all workplaces and from all sectors. In detail, FMEA is the only risk analysis methodology that considers detectability of failure modes. In this study, occupational health and safety requirement, risk analyzing, is fulfilled with the help of FMEA because of its decision factors and detectability property. 2 Despite the effectiveness of the method, calculations of RPNs have been criticized for many reasons because the crisp RPN calculation method shows some important weaknesses when FMEA is applied in the real-world cases (Liu et al. 2013). A List of these shortcomings is stated in the second section of the research. There are significant efforts which have been made in FMEA literature to overcome the shortcomings of the traditional RPN. Several new approaches have been made. Fuzzy approach, Multiple Criteria Decision Making (MCDM) technique, group-based evidential reasoning (ER), DEA/ Fuzzy DEA, ANP/AHP, Grey theory, Cost based model are most used approaches. A detailed literature review that contains more approach is stated in the second part of this study. Traditional FMEA is basic and practical risk analysis methodology. It is easy to renew traditional FMEA application. Although firms and experts want to apply FMEA or another risk analysis methodology in a short time period and with minimum cost, the proven effects of the shortcomings of traditional RPN calculation force experts to make contribution to the current methodology. In this research, firstly decision makers and decision factors (severity, occurrence and detectability) will be weighted with the help of Fuzzy Analytical Hierarchy Process (FAHP). Secondly, Grey Relational Analysis (multi criteria decision making method) is used to calculate and determine the proper first risk priority numbers (RPN1) for failure modes. RPN1 values are used for prioritizing the corrective actions of corresponding failure modes. GRA is mainly used to incorporate the weights of decision factors to RPN1 calculation. Then, RPN2 is calculated. RPN2 is derived from the need for reprioritization of corrective actions. Even though failure modes are prioritized with the help of Fuzzy AHP and GRA methodologies, it is not enough to perform the corresponding corrective actions. Additional five criteria are added to calculate RPN2. By multiplying criteria’s scores with their weights and then the summation of all five, RPN2 value is calculated. In the last step, RPN3 values are calculated. RPN1 and RPN2 values are multiplied with their different coefficients and then summation of these two RPN gives us RPN3. Reprioritization is done based on RPN3 and the corrective actions are performed according to this reprioritization. A threshold interval values should be determined by decision makers because firms have several limitations such as cost, time, and capacity of workers. 3 The point that makes this research different is fact of the new approach to FMEA. This approach gives importance to what the corrective action is in essence. In this study, FMEA is performed to overcome occupational health and safety drawbacks in the workplace. FMEA should be applied by employees of company. According to their education level, experience, job relevance, risk analyzing test result, and experiences on risk analysis methodology (repetition number of risk analysis methodology), these employees, decision makers, are given weights. Traditional RPN, in this study RPN1, is calculated by using three decision factors O, S and D. These three decision factors have equally importance in the traditional RPN calculation but in this research, fuzzy AHP is used to determine reasonable weights for them. Detailed explanations of these methods are stated in the third part of this research. 1.2 RESEARCH OBJECTIVES The objectives of this research are as follows: i. To understand risk analysis, occupational health and safety (OHS) concept. ii. Review of FMEA researches, new calculation approaches of RPNs. iii. Review of fuzzy approaches and grey relational analysis used in FMEA application. iv. To introduce the most common FMEA application of which can be used in wide range of working condition. v. To obtain more realistic corrective actions and prioritization of corrective actions with the help of RPN2. vi. To obtain more realistic applicability of corrective action by means of threshold interval values. vii. To weight decision makers and decision factors with the help of prepared criteria charts. viii. To apply more consistent and reasonable FMEA application with the help of weighted decision factors and decision makers. ix. To generate a FMEA application that is easy to make and renew periodically. x. Quality and reputation contribution to the literature by means of RPN2 which considers the corrective actions in detail. xi. To take the corrective actions into action which satisfies customer’s demands? xii. To calculate all calculations in this proposed method faster and to spend little 4 time consumption, thanks to excel templates. 1.3 THESIS ORGANIZATION In this study, the fuzzy AHP approach is adopted to determine the weights of decision factors (both for RPN1 and RPN2) and decision makers. Then, the optimization and the calculation of RPN1s are made by using grey relational analysis. After that, the RPN2 values are calculated by means of five additional criteria. RPN2 considers the real life needs and the feasibility of corrective actions. Finally, RPN3s are calculated and the reprioritization is made based on these values. This thesis is organized as follows: In section II literature review is presented. A brief information is presented about risk, risk analysis, risk analysis methodologies. Then, reasons for choosing FMEA as a risk analysis methodology are explained. Another subheading of section II is occupational health and safety (OHS). This topic clarifies the contribution to OHS with the help of weighted corrective actions. Fuzzy approaches and grey relational analysis adoption to FMEA are also stated in section II. In section III, the proposed methodology is explained. Detailed information about Fuzzy AHP, GRA, and proposed method are presented step by step. In section IV, a case study is presented. A yarn manufacturing company from Turkey is used. The case study considers the requirements of the company’s OHS concept. Thus, the proposed FMEA methodology is used to meet the OHS needs of the company. The thesis ends with a discussion and conclusion given in section V. 5 2. LITERATURE REVIEW 2.1 RISK ANALYSIS AND OCCUPATIONAL HEALTH AND SAFETY 2.1.1 Risk Assessment and Risk Analysis In today’s competitive world, a firm’s success does not only depend on its performance, but also many other factors, contributing this process. One of these factors is risk management. In engineering contexts, risk is often linked to the expected loss so that a proper risk management process should be applied by the firms (Lirer et al. 2001). Risk is defined in many ways. More common definitions are below: i. Risk is the measure of the probability and severity of adverse effects. ii. Risk is the combination of probability of an event and its consequences. iii. Risk refers to uncertainty of outcome, of actions and events. iv. Risk is a situation or event where something of human value (including human themselves) is at stake and where the outcome is uncertain. v. Risk is an uncertain consequences of an event or an activity with respect to something that humans value. vi. Risk is equal to the two-dimensional combination of events/consequences and associated uncertainties. vii. Risk is uncertainty about and severity of the consequences (or outcomes) of an activity with respect to something that humans value (Aven 2010). viii. A risk is a future event that may or may not occur. ix. A risk must also be an uncertain event or condition that, if it occurs, has an effect on, at least, one of the project objectives, such as scope, schedule, cost or quality. x. The impact or consequence of the future event must be unexpected or unplanned (Nieto-Morote & Ruz-Vila 2011). People talk about risk when there is the chance, but not the certainty, that something they don’t want may happen (Covello & Merkhoher 1993). As to these definitions, risk is usually an undesirable and uncertain event. Nowadays, to protect firms from encountering unexpected situations, systematic well-designed precautions are taken by experts. To reduce the effects of the risks, risk management and risk assessment strategies 6 should be applied by the firms. Size of the firm is not the decision factor while experts are considering the risk analysis necessity. The scope of the risk analysis application is normally less detailed in small sized firms. Risk management is the identification, assessment, prioritization of risks and finally eliminating them. Risk assessment provides a mechanism for identifying which risks represent opportunities and which represent potential pitfalls. If the risk assessment performed right, a risk assessment provides organizations a clear view of variables of which they may be exposed. Thus, it is really crucial for the organizations to determine the right and proper risk assessment methodology. Firstly, a detailed risk management plan is prepared by experts. After risk management plan proper risk assessment methodology should be applied by experts. Risk management is divided into four phases: a. Identification b. Quantification c. Decision d. Reduction and Control (Ale 2002) Figure 2.1: Risk management phases Source: Ale 2002 Risk assessment process is crucial for the firms at least profit performance. Some unheeded risky activities may cause huge damage and loss. Risk assessment includes both risk estimation (identifying hazard and estimating their outcomes and probabilities) and risk evaluation (determining the significance or value of risks to those concerned with or affected by the decision). Risk estimation is about situations, and risk evaluation Risk Management Phases 7 about the effect on people (Cohen 1984). When experts are estimating, evaluating risks or deciding the risk assessment methodology, there should be some criteria about firm. For instance, firm’s old data and attendance of workers should be considered while experts are analyzing the risks. Moreover, experts should have ability on risk assessment. In addition, experiences and education level of experts who know the firm’s operational goal are also crucial for this process. Many risk assessment methodologies are applied by the firms. Some of them are required an experienced team others do not require. Moreover, some of them are suitable for comprehensive risk analysis and need for long time for application. Comparisons of these methodologies can help decision makers to choose the most appropriate methodology. Table 2.1 shows the comparisons of mostly used risk analysis methodologies. 8 Table 2.1: Risk assessment methodologies comparison aaaCriteria Risk Assessment Type Short Description & Advantages of Method Experience of Team Leader & Teamwork Quantitative or Qualitative & Need For Documentatio n Appropriat e Business Sector The Success Rate of Application What if? A structured brainstorming method of determining what thing can go wrong and judging the likelihood and consequences of those situation occurring. Mid-level experience & Can be done by an analyst Qualitative & Very little Simple procedure works Not enough alone while the analyst determining the risks. Success of application is based on team leader's success and experience. PHA A risk analysis that is performed to identify all potential hazard and accidental event, rank them according to their severity and finally identify required hazard controls and follow-up actions. Mid-level experience & Can be done by an analyst Qualitative & Medium Fits all sectors Not enough alone while the analyst determining the risks. Success of application is based on team leader's success and experience. JSA The JSA is very effective tool for helping to reduce incidents, accidents, and injuries in the workplace. So much experience & Teamwork Qualitative & Too much Oil and gas industry It is an excellent tool if it is used during new employee orientations and training and also can be used to investigate “near misses” and accidents.. Checklist A type of informational job that helps to reduce failure by compensating for potential limits of human memory and attention. Mid-level experience & Teamwork Qualitative & Medium Fits all sectors Success proportion is based on the checklist preparation. HAZOP The basic principle of HAZOP is to have full process description and to ask in each node what deviations to the design purpose can occur, what causes produce them, and what consequences can be presented. So much experience & Teamwork Qualitative & Too much Chemical industry This method is quite tough one to apply and need for high level of experience and performance of the team. FMEA/FMECA Failure mode and effects analysis is a widely used engineering technique for designing, identifying and eliminating known and/or potential failures, problems, errors and so on from system, design, process, and/or service before they reach the costumer. So much experience & Teamwork Can be both of them & Too much Fits all sectors If done by experienced analysts, the success rate increases. 9 Criteria Risk Assessment Type Short Description & Advantages of Method Experience of Team Leader & Teamwork Quantitative or Qualitative & Need For Documentatio n Appropriat e Business Sector The Success Rate of Application Safety audit Safety audit is a systematic and independent examination to determine whether activities and related results conform to planned arrangements and whether these arrangements are implemented effectively and are suitable to achieve the organization’ policy and objectives. Mid-level experience & Can be done by an analyst Qualitative & Very little Fits all sectors Not enough alone while the analyst determining the risks. Success of application is based on team leader's success and experience. FTA A systematic safety analysis tool that proceeds deductively from the occurrence of accident to the identification of the failure cause or accident cause of that event. So much Experience & Teamwork Can be both of them & Too much Fits all sectors This method needs for high level of experience and performance of the team. FTA is very effective method for determining the risks. ETA ETA is an established risk analysis tool to assess likelihood of an accident. The aim of this technique to estimate the likelihood of event that often missing with the help of collected data. So much experience & Teamwork Can be both of them & Too much Fits all sectors This method needs for high level of experience and performance of the team. Eta is very effective method for determining the risks. L -type matrix L-type matrix is based on cause and effect relationship. It a simple and can be done quickly. Mid-level experience & Can be done by an analyst Qualitative & Very little Simple procedure works This method can be applied for simple procedure and urgent works. Success of application is based on team leader's success and experience. X-type matrix This technique makes a research on accidents that have occurred in the past in the workplace. It also considers the corrective actions’ cost. So much experience & Teamwork Qualitative & Too much Fits all sectors This method can be applied for all works. Success of application is based on team leader's success and experience. Cause-effect analysis This diagram based technique, which depend on brainstorming, pushes you to consider all possible causes of a problem, rather than just ones that are most obvious. So much experience & Teamwork Can be both of them & Too much Fits all sectors, especially chemical industry This method needs for high level of experience and performance of the team. Cause-effect analysis is very effective method for determining the risks. Source: Özkılıç 2005, Diberardinis 1998, Ferdous et al. 2009 and Gharahasanlou 2014 10 2.1.2 Occupational Health and Safety (OHS) Every morning in Africa, a Gazelle wakes up. It knows it must run faster than the fastest lion or it will be killed. Every morning a Lion wakes up. It knows it must outrun the slowest Gazelle or it will starve to death. It doesn't matter whether you are a Lion or a Gazelle. Nowadays, as implied in the fable above, companies may lose everything just because of instant error. Financial losses can be compensated by means of extra effort. In particular, it is really hard to provide former prestige if the company makes a mistake regarding with safety (of product, process), health of workers or any other mistake that can harm to reputation of company. Companies aim to get maximum profit by spending minimum money. Reputable and long-term companies not only focus on making money but also some other institutionalization requirements. One of the most important considerations in this regard is occupational health and safety activities. It is meaningless to see the OHS activities as a procedure that the law obligates. Attention to OHS in the construction industry has increased dramatically over the past decades. The time for OHS awareness has arrived and that OHS is no luxury, it is a necessity (Geminiani et al. 2013). OHS is similar to health insurance. For instance, a person does not need to make health insurance because he/she feels good. Unluckily, he/she is 65 years old. It means that a risky situation is present because of his/her age. If this person faced with health problem that requires a lot of money means that he/she has not much chance surviving. If he/she had insurance, insurance company pays the treatment cost. OHS activities are so similar to that situation. If a boss ignores some potential failures due to financial burden or apply OHS activities just on paper, he/she will most likely to encounter an excess losses, health problems of workers and/or reputation loss. Another benefit of OHS is that OHS makes workers feel comfortable while they are working. Workers give an instinctive reaction by means of good working conditions. For instance, think about a baby. If a baby lives in clean, warm and relaxing condition and the baby food is of good in quality and also if the parents take preventative measure to reduce the likelihood of any accident (for example, the parent provide something that prevents baby from reaching the stove or provide safe electrical outlets) the baby would 11 be feel happy and his/her behavior will show this happiness to the parents. This example is compatible with OHS activities. If the boss provides workers a working environment where all safety measures and personal protective equipment are provided besides of good quality food then workers concentrate on their job and feel confidence about the company. In Turkey, firms had not fulfilled the occupational health and safety essentials until the recent years. After establishment of National Occupational Health and Safety Council, some radical changes made thanks to governmental sanctions (Karabulut 2012). In 2006, council prepared the “National Occupational Health and Safety Policy” and determined the action plan. Table 2.2 contains requirements of plan. Table 2.2: An action plan made by National Occupational Health and Safety Council (2006-2008) Targeted Issue Fulfillment Status Putting the occupational health and safety law that compatible with EU norms. X To apply occupational health and safety regulations to all employees. X Occupational health and safety regulations will be extended to all sectors. X Occupational health and safety services would be more efficient units. X The number of occupational accidents will be reduced by at least 20 percent. X Occupational disease diagnosis system will be developed. X Thanks to the technical support services that carried out by public, health and safety status of people will be increased by 20 percent. X Source: Karabulut 2012 Two-year action plan (2006-2008) could not be realized as we see in Table 2.2 because the government did not put law about OHS. Furthermore, no occupational disease diagnosis system had been developed. Thus, the number of occupational accidents was not reduced by at least 20 percent (Karabulut 2012). 12 A four-year action plan was held by Ministry of Labour and Social Security (2009- 2013). Table 2.3 contains requirements of plan. Table 2.3: An action plan made by Ministry of Labour and Social Security (2009- 2013) Targeted Issue Fulfillment Status Providing the occupational health and safety law come into force and completion of the relevant legislation.  To ensure the implementation of the new legislation giving the information to the interested groups and the public.  Increasing the number of the employees who work in OHS laboratory by 20 percent.  Increasing the detecting of occupational disease by 500 percent. X The number of occupational accidents will be reduced by at least 20 percent. X Increasing the number of projects, education and publicity that related to OHS by 20 percent.  Source: Karabulut 2012 As presented in Table 2.3, the detailed law come into force but can be partially fulfilled by companies. On this subject, Turkey is in the transition period. More clearly, although there is a law in force, most of the small workshops cannot meet the obligations of the law yet. Government noticed this case and extended the law’s obligation due time to a later date. During this time interval, numerous education program, publicity activities and projects related OHS are provided. Turkey shows development about the OHS necessity. Companies began to employ OHS specialist who is suitable for their job and law requirements. Three levels (class-A, class-B and class-C) are available for specialist in Turkey. There are some criteria that determine danger level of work process. Companies should employ suitable class specialist considering the danger level of work process. 13 It is impossible to apply a risk analysis method that fits to all work process because danger level of work process, materials used in this process, working place, working hours and other criteria affects the risk analysis method. A detailed and renewable risk analysis which is compatible with the current business process provides convenience to the company to impose OHS. One of the fundamental steps of OHS is risk analyzing. There are many risk analysis method in the literature as mentioned in the above. With few additions, FMEA methodology can help OHS specialists to make risk analysis of their responsible work process. FMEA considers the possibility, severity and detectability of failure modes. These decision factors help risk analysts to evaluate risks more consistently. Especially the decision factor of detectability makes the FMEA differ from the other risk analysis methodologies. Detectability plays important role while identifying the risks so that this factor helps us in OHS concept. FMEA also provides experts to take corrective actions for failure modes, in addition to identify responsible person. As presented above, risk management requires identification, quantification, decision, reduction and control. In table 2.4 the reasons for applying FMEA as a necessity of OHS and risk management methodology are explained by comparing these two concepts. Table 2.4: The reasons for applying FMEA as a risk management methodology Risk Management FMEA Introduction Identifying the failure modes Quantification RPN calculation. Decision Deciding to corrective actions Reduction Corrective actions Control Determining control period of eliminated failure modes In table 2.4, it is presented that the steps of risk management are parallel with FMEA. In this research, additional steps are added for deciding on the corrective actions. One of these steps is the calculation of RPN2 values. RPN2 values mostly consider occupational health and safety concept with the help of additional five decision criteria. By considering these criteria and then adding RPN2 values to the calculation of final RPN values (RPN3), the requirements of OHS concept are met. 14 2.2 REVIEW OF MODELS 2.2.1 Failure Mode and Effects Analysis (FMEA) Companies aim to make profit without spending a lot of money. All works that companies carry out should be completed without any occupational accident, machine breakdown or anything that hinder the process. Unfortunately, in real life, this is just a dream. There cannot be a perfect worker, so it is normal to encounter with failure modes. In fact, it is normal that failure modes are under the determined limits. At this point, talented and experienced experts or occupational health and safety specialists should determine the failure modes by means of previous year’s data, foreman’s’, workers’ and their own experiences. Failure mode and effects analysis (FMEA) methodology helps experts to see failure modes visually and make risk analysis easily. FMEA is used to anticipate and mitigate risk (Cody 2006). Failure mode and effects analysis is intended to provide information for making risk management decisions (Pillay & Wang 2003). While the FMEA team is applying FMEA as a risk analysis methodology they should be aware of several definitions. These definitions are presented below: FMEA team: FMEA team should be formed up three or more people. The team members should be experienced persons. They should have been active in FMEA or another risk analysis methodology. His/her education level and knowledge on the FMEA are other important criterion. Failure mode: Failure mode is the failure that hinders process success. It describes what could go wrong in the process. A failure mode can be the result of another failure mode. On the other hand, a failure mode can be the cause of another failure. It is crucial that the FMEA team have to identify all potential failure modes that may hinder the process. Potential failure effect: Potential failure effect can be defined by answering to this question: How does the failure affect the function of the step? Usually, failure affects customers. Customers may be the internal customer or the last user of the product. Team 15 should consider the customer’s situation and determine the potential failure effect properly. Cause of failure: It is inevitable that all failure modes have root cause or reason. When the team determines the causes of failure, this process can save time effectively. If the cause of failure affects the failure mode in a special way, the team should eliminate the cause directly. Therefore, there is no need to maintain successive FMEA steps for this failure mode. Severity: If a failure were to occur, what effect would that failure have on the product quality and on the customer (if any)? Probability (of occurrence): How likely is it for a particular failure to occur? (Kahraman et al. 2013). Detectability: Chance of detection of failure mode before it causes an accident. RPN: The three values of severity, frequency and detectability of each potential failure are multiplied, forming a risk priority number (RPN) (Sant’Anna 2012). Corrective action: All failure modes should have corrective actions. Corrective actions are performed as to the RPN values. In traditional RPN calculation, RPN value can take on values between 1-1000. The prioritization is made as to these values. The bigger RPN value implies that the higher priority of corrective actions should be performed. Responsible: A person who is responsible for performing the corresponding corrective action. Responsibility can be given to a group of people. Due time (Target completion date): Due time is the last day to perform the corrective action completely. Next control time/ control period: FMEA should be applied periodically even if there is no undesirable risky event occurred or nearly occurred. Thus, the next control time or control period of the current failure mode should be determined by FMEA team. 16 The main objective of FMEA is to identify the potential failure modes, evaluate the causes and effects of different component failure modes, and determine what the chance of failure could eliminate and reduce. FMEA allows the analysts to identify and prevent known and potential problems from reaching the customer (Liu et al. 2011). The first work in establishing a procedure for performing FMEA was created by the U.S. military in 1949. In the early 1960s, the U.S. military established a military standard (MIL-STD-1629a) for systematically evaluating the potential impact of functional or hardware failures on mission success, system performance, maintainability, and maintenance requirements. In the early 1980s, the automotive industry began to incorporate FMEA into the product development process (Abdelrahman & Abdelgawad 2011). In 1993, firms created FMEA to meet the requirement of QS-9000 (Chang et al. 2013). Nowadays, FMEA is a method applied in various industries such as the machinery, medical, aviation, automotive, food industry, OHS. A careful construction of decision factors’ scales should be formed. Scale of values from 1 to 10 is for the classification of events involving possible causes and effects of failures from three independent points of view: severity, frequency (occurrence) and detectability of the event. Sample scales are represented in Figure 2.2. Traditional FMEA works by a series of successive steps. To understand FMEA steps more clearly, whole FMEA application can be divided into several steps as shown in Figure 2.3. Figure 2.2 Sample 10 point scales for Occurrence, Severity, and Detectability Source: Barends et al 2012 17 The importance and the meaning of the scores change according to the structure of company and also expectations of customer (Sofyalıoğlu & Öztürk 2012). By using these 10-point scales, experts give score to the each decision factors. After that, RPN values of each failure modes are calculated by multiplying these three. As the RPNs increase, risk of failure modes increase too. Experts then determine the proper corrective actions and follow these actions. The FMEA application should be done in periods after the corrective actions are performed and new RPNs should be calculated for current failure modes again. Here is the graphical presentation and list of the FMEA steps shown in Figure 2.3. 18 : 1. Develop a good understanding of what the system is supposed to do when it is operating properly. 2. Divide the system into sub-systems and/or assemblies in order to localize the search for components. 3. Use blue prints, schematics and flow chart to identify components and relations among components. 4. Develop a complete component list for each assembly. 5. Identify operational and environmental stresses that can affect the system. Consider how these stresses might affect the performance of individual components. 6. Determine failure modes of each component and the effects of failure modes on assemblies, sub-systems, and entire system. 7. Categorize the hazard level (severity) of each failure mode. 8. Estimate the probability. 9. Estimate the detectability. 10. Calculate the risk priority number (RPN). 11. Determine if action needs to be taken depending on the RPN. 12. Develop recommendations to enhance the system performance. 13. Prepare FMEA report by summarizing the analysis. 19 One of the beneficial aspects of the FMEA that makes it widely accepted is the group dynamic in which personnel from different functional groups are gathered to evaluate the risk associated with a system from various standpoints (Abdelrahman & Abdelgawad 2011). It is not forgotten that the FMEA application would be successful if only made by a group of people instead of only one person. The optimal size of the FMEA team is 3 to 6, and the maximum is 10 (Goel & Graves 2007). One of the most important factors affecting the success of FMEA is timing. It should be done before any problem occurred. There is a Turkish saying like “Önlemek, ödemekten daha ucuzdur”. It means that taking precaution for a problem is cheaper than paying the cost of event that derives from that problem. 2.2.1.1 Types of FMEA In the literature, there are several types of FMEAs, which are system, design, process, service, and software, are available (Kahraman et al. 2013). 2.2.1.1.1 System FMEA This type of FMEA aims to identify potential failure modes in the system functions by analyzing the systems, sub-systems and interactions of these systems with each other (Sofyalıoğlu & Öztürk 2012). System FMEA concerns the failure modes that are resulted from design FMEA. Potential failure causes differentiate from other types because failures are occurred due to direct customer contact. Broken pieces by accident, technical errors, and electrical errors can be good examples. The aim of the System FMEA is to develop the quality, reliability and the maintainability of the system. 2.2.1.1.2 Design FMEA This type of FMEA aims to identify and prevent possible failures, their causes and effects during a product design. Design FMEA should start before the concept design (Sofyalıoğlu & Öztürk 2012). For instance in the automotive industry, design FMEA is a procedure to identify that the right materials are being used, to conform the customer specifications, and to ensure that government regulations are being met, before finalizing the product design. 20 2.2.1.1.3 Service FMEA A service failure occurs when customers’ needs are not met and or service performance falls below a customer’s expectation (Goldstein et al. 2002). Service failure is essentially a flawed outcome that reflects a breakdown in reliability. It compasses any problematic situation during service while service is delivered to a customer, causing significant damage to customer satisfaction (Reichheld 1996). 2.2.1.1.4 Software FMEA Software base FMEA worksheet helped us to utilize entered data to ensure the consistency of analyzing the worksheet of FMEA with minimum time requirements. First of all, each data of various functions and failures entered. Then the causes and effects on each failure were assigned. Moreover, Risk Priority Numbers (RPNs) of each failure were calculated automatically right after allocating rating scales for severity, occurrence and detection (Feili et al. 2013). 2.2.1.1.5 Process FMEA All of the possible problems that may arise during the product creation can be overcome with help of process FMEA. Process FMEA deals with the manufacturing and assembly processes. It identifies any potential failures that could be caused by manufacturing/assembly processes, machines, fixtures, and production. Before the application, it is important to determine which part of the production process will be taken into account. Process FMEA traditionally begins when the design FMEA report is available. The purpose of process FMEA is to determine and correct the weak points of production. In this research, the combination of system, design, service and mostly the process FMEA is used in case study. The combination of four types meets the needs of OHS in workplace. 21 2.2.1.2 Shortcomings of FMEA It is true that the FMEA is the simple risk analysis methodology to apply, instruct and understand (Chang et al. 2013). FMEA helps experts to detect failure modes, cause of failures quickly and also diminish the frequency of failures. On the other hand, traditional calculation of RPN (occurrence* severity* detectability = RPN) has several shortcomings that presented below (Liu et al 2013; Liu et al. 2012; Chang et al. 2013; Helvacioglu & Ozen 2014; Kutlu & Ekmekcioglu): a. The acceptance of three decision factors equally important. If there are no weights determined for each of them, it is ⅓. Every company or even each process has failure modes that have different number of frequency, importance level of severity importance and importance level of detectability. b. Since all of three decision factors have 10-point scale it is possible that the different combinations of decision factors may produce same RPN value, but their hidden risk implications may be totally different. For example, three failure modes with O, S, D values of 8, 9, 1; 6, 6, 2 and 4, 3, 6 have the same RPN and same prioritization too. c. Although FMEA assumes the RPN is distributed 1 to 1000 it not true exactly because only 120 numbers can generated. Thus, traditional FMEA gives high duplication rate. d. The mathematical form adopted for calculating the RPN is strongly sensitive to variations in risk factor evaluations. Small variations in one rating may lead to vastly different effects on the RPN, depending on the values of the other risk factors. For instance, assume that occurrence and severity ratings are both 9. One point difference in detectability rating results in an 81 point differences. On the other hand, if occurrence and severity ratings are both 2, then one point differences in detectability rating results in a 4 point differences. e. The RPN values are heavily distributed at the bottom of the scale from 1 to 1000. 22 f. The RPN does not consider cost of corrective actions, workers health and safety, regulatory obligations aspects. The RPN only considers three decision factors. g. The calculation of RPN is not accepted completely because there is no meaningful explanation as to why three decision factors are multiplied instead of summation. h. It is really hard and mostly subjective to determine the ratings for decision factors. Much information in FMEA is often uncertain or vague and can be expressed in a linguistic way such as likely, important, high and so on. i. It is hard to interpret the RPNs which are closed in numbers. For instance, one failure mode has the RPN equal to 80. Another failure mode has 81. The second one should be resolved first because of greater RPN but some other factors may need to be considered to make the right and reasonable decision. j. Traditional FMEA methodology fails to consider the direct and indirect relationship between failure modes and cause of failures. Interdependencies among the various failure modes and effects on the same level and different levels of hierarchical structure of an engineering system are not taken into account. k. Traditional FMEA methodology also fails to consider customer satisfaction. l. Traditional RPN calculation does not take decision makers’ experiences, education levels, and academic knowledge levels into account. m. The RPN cannot be used to measure the effectiveness of corrective actions. n. The traditional RPN method is only measuring from the risks viewpoint while ignoring the importance of corrective actions. o. FMEA is a time-consuming and very tedious activity; hence it is highly prone to errors. To overcome the above shortcoming of traditional FMEA, many approaches have been suggested in FMEA literature. Most of the approaches have used fuzzy approach. The studies about FMEA that considering fuzzy approach describe risk factors O, S, D by using fuzzy linguistic terms. The linguistic variables were used for evaluating three decision factors (Kutlu & Ekmekçioğlu 2012). Distinct studies about fuzzy approach used FMEA are presented in this research. For instance, Wang et al. (2009) use fuzzy risk priority numbers (FRPN) to prevent false 23 prioritization. False prioritization occurs because of direct judgment of decision makers. It is crucial to consider real applications to determine relevant prioritization of failure modes. In this research, the decision factors occurrence, severity and detectability are treated as fuzzy variables and evaluated by using fuzzy linguistic terms and fuzzy grades. Wang et al. (2009) also treat the decision factors O,S and D as fuzzy variables and evaluate them using fuzzy linguistic terms and fuzzy ratings. After proposing FRPNs, FRPNs are defined as fuzzy weighted geometric means of the fuzzy ratings for O,S and D, and can be computed using alpha-level sets and linear programming models. Xu et al. (2002) proposed a fuzzy logic based method for FMEA to address the problem of interdependencies among various failure modes with uncertain and imprecise information. The validity of the results may be questionable just because of this uncertainty. Thus, the potential problems in sharing information among experts are resolved with the help of a platform. This platform for a fuzzy expert assessment and it is integrated with proposed system. Garcia et al. (2005) present a data envelopment analysis approach for determining prioritization of failure modes in which the typical FMEA decision factors are modeled as fuzzy sets. Inference rules of the IF THEN can be bypassed by means of this approach. Chin et al. (2009) also use data envelopment analysis to determine the risk priorities of failure modes. The maximum and the minimum score of each failure mode are measured and two score are then geometrically averaged to measure the overall risk of failure modes. Braglia (2000) occurred a multi-attribute failure mode analysis (MAFMA) approach based on AHP technique. Braglia incorporate the decision factors O, S, D and expected cost as decision criteria. Alternatives are possible causes of failures and the decision goal is the selection of cause of failure. Chang, Wei, and Lee (1999) used fuzzy method and grey theory for FMEA, where linguistic variables were used to evaluate to decision factors O, S, and D and grey relational analysis was applied to determine the risk priority of potential causes. 24 Chang et al. (2013) try to eliminate four chief shortcomings of FMEA: High duplication rate, assumption of equal importance of O, S, and D, not following the ordered weighted rule and failure to consider the direct and indirect relationship between the failure mode and the cause of failure. To resolve these drawbacks, they propose a novel approach, integrating GRA and decision making trial and evaluation laboratory (DEMATEL) approach to rank the risk of failure. GRA is used to modify RPN values to lower duplications. By integrating grey relational analysis and the decision making trial evaluation laboratory (DEMATEL) method, to rank the risk of failure, wherein GRA is used to modify RPN values to lower duplications and the ordered weighted rule is followed; then, the DEMATEL method is applied to examine the direct and indirect relationship between failure modes and cause of failures. This approach gives higher priority when a single cause of failure causes failure modes to occur multiple times. Liu et al. (2011) used fuzzy evidential reasoning approach and GRA. Fuzzy evidential reasoning approach is used to solve the problem of the acquirement of FMEA team members’ diversity opinions. GRA is used to solve the problem of prioritization of the modes that have been identified and then calculated based to this new approach. Unluckily, this research does not concern the DMs ability, experiences, education level and their job relevance with FMEA application. FMEA should be made by team and it is based on group decision function and cannot be done on an individual basis. The FMEA team often demonstrates different opinions and knowledge with each other. To resolve this issue Chin et al. (2009a) proposed a new model that allows FMEA team members to assess risk factors independently. They presented an FMEA that uses the evidential reasoning (ER) approach to model the diversity and uncertainty of the assessment information in FMEA. Mandal & Maiti (2014) tries to overcome the drawbacks of RPN calculation. According to study, it is hard to interpret failure modes by using crisp numbered scales. To overcome this drawback they developed a new methodology integrating the concepts of similarity value measure of fuzzy numbers and possibility theory. Similarity value measure has been applied to group together failure modes having similar amount of risk value. 25 Su et al. (2014) try to improve the reliability of the thin-film transistor liquid crystal display products with the help of a combined approach integrating failure mode and effect analysis (FMEA) and Taguchi methods. Considering the issues during design and manufacturing phases simultaneously, the proposed approach provides a systematic framework to find the causes of failure, to conduct experiments, and to optimize the process output. Liu et al. (2014) a new approach to RPN calculation is presented. This study based on a more effective representation of uncertain information, called D numbers, and an improved grey relational analysis method, grey relational projection (GRP), a new risk priority model is proposed for the risk evaluation in FMEA. Helvacioglu & Ozen (2014) criticizes the RPN calculation. Another point that is criticized is that the differences among the decision makers are not mentioned enough. According to the study, RPN method cannot emphasize the nature of the problem, which is multi-attributable and has a group of experts’ opinions. Furthermore, attributes are subjective and have different importance levels. Therefore, they propose a framework to overcome the shortcomings of the traditional method through the Fuzzy Multi-Attribute Group Decision Making (FMAGDM), which helps to solve the selection of risky failure modes. Fuzzy sets are utilized for expressing fuzziness of crisp/linguistic knowledge coupled with the well-known TOPSIS methodology for decision making. So many approaches applied and important contributions were made to the FMEA literature. Liu et al. (2013) made a literature review on FMEA, especially RPN calculation. As to this study methods used in FMEA literature may be divided into five main categories: Multi-criteria decision making (MCDM), mathematical programming (MP), artificial intelligence (AI), hybrid approaches and others. Table 2.5 represent these methods in detail and gives the proportion of usage. 26 Table 2.5: Classification of risk evaluation methods in FMEA Categories Approaches Total number MCDM (22,50%) ME-MCDM 1 Evidence theory 2 AHP/ANP 4 Fuzzy TOPSIS 1 Grey theory 7 DEMATEL 1 Intuitionistic fuzzy set ranking technique 1 VIKOR 1 Mathematical programming (8,75%) Linear programming 4 DEA/ Fuzzy DEA 3 Artificial intelligence (40,00%) Rule-base system 1 Fuzzy rule-base system 29 Fuzzy ART algorithm 1 Fuzzy cognitive map 1 Integrated approaches (11,25%) Fuzzy AHP- Fuzzy rule-base system 1 WLSM-MOI-Partial ranking method 1 OWGA operator DEMATEL 1 IFS-DEMATEL 1 Fuzzy OWA operator-DEMATEL 1 2-tuple-OWA operator 1 FER-Grey theory 1 Fuzzy AHP-fuzzy TOPSIS 1 ISM-ANP-UPN 1 Other approaches (17,50%) Cost based model 6 Monte Carlo simulation 1 Minimum cut sets theory (MCS) 1 Boolean representation method (BRM) 1 Diagraph and matrix approach 1 Kano model 1 Quality functional deployment (QFD) 2 Probability theory 1 Source: Liu et al. (2013) 27 In this thesis research, fuzzy AHP is used for weighting of decision factors O, S, and D. After that, weights of each decision factors are incorporated to RPN calculation by means of GRA. Fuzzy AHP technique is also used for weighting process of decision factor. By determining weights for each decision makers, FMEA can be applied more consistently. Each decision maker applies FMEA by himself. At the final step, weighted arithmetic means are calculated and then weighted team decisions appears. Each decision maker applies FMEA individually and also team decisions give us more convenient results. Fuzzy AHP technique is finally used to weight RPN2 criteria weights. 28 3. METHODOLOGY 3.1 ANALYTICAL HIERARCHY PROCESS (AHP) AND FUZZY ANALYTICAL HIERARCHY PROCESS (FAHP) 3.1.1 Analytical Hierarchy Process (AHP) One of the crucial approaches that underlie the proposed methodology of this thesis is Analytical Hierarchy Process. AHP is one of the well-known multi-criteria decision making techniques that was first proposed by Saaty (1980). AHP is used in various decision-making areas such as planning, R&D, choosing the best policy alternative, predicting outcomes, measuring performance, and optimizing and resolving decision conflicts (Saaty 1986). This method mainly based on human experiences and ability to make judgments about small problems. “It facilitates decision making by organization perceptions, feeling, judgment, and memories into framework that exhibits the forces that influence a decision.” (Çakır 2009). After introduction of AHP, it has become one of the most used multiple-criteria-decision-making techniques. Thanks to AHP, decision makers structure a complex problem in the form of a simple hierarchy and to evaluate a factors, criteria and alternatives (Lee et al. 2008). After constructing the hierarchy, the DMs begin to determine the relative importance of the elements in each level. DMs make judgments to determine the dominance of one element over another with the help of 9-point ratio scaling. Scale is presented in Table 3.1. Pairwise comparisons can be done by asking decision makers how valuable a criterion (C1) when compared to another criterion (C2) with respect the overall goal (Oztaysi 2014). The decision makers are not forced to determined crisp number. They can express their opinion by means of linguistic terms such as equally important, moderately more important, strongly more important, very strongly more important, and extremely strongly more important. These linguistic terms would then be translated into numerical values as presented in Table 3.1. 29 Table 3.1: 9- Point ratio scale for AHP Intensity of importance Definition Explanation 1 Equal importance Two activities contribute equally to the objective 3 Moderate importance Experience and judgment slightly favor one activity over another 5 Strong importance Experience and judgment strongly favor one activity over another 7 Very strong importance An activity is favored very strong over another, its dominance demonstrated in practice 9 Extreme importance The evidence favoring one activity over another is of the highest possible order of affirmation 2,4,6,8 For compromise between the above values Sometimes one needs to interpolate a compromise judgment numerically because there is no good word to describe it Source: Oztaysi 2014 DMs make judgment to compare criteria and then they form the comparison matrix. The prioritization of criteria is then calculated. As stated above, AHP can be used to solve MCDM problem. Similar procedure can be applied to calculate the weight of alternatives. For each criterion respectively, the pairwise comparison matrix for all alternatives should be formed. After that the value of each alternative is multiplied by the weights of the corresponding criterion. Finally, to find out the importance and weight of each alternative, all alternative values are summed up. To make this technique successful, there should be a little subjectivity of judgment of criteria or alternative comparison. Comparison scores should be determined by precise explanation so that the experts make definite judgments. If the experts make a decision on criteria in a subjective way, this method cannot help perfectly. 3.1.2 Fuzzy Analytical Hierarchy Process (FAHP) Fuzzy set and fuzzy logic, first proposed by Lotfi A. Zadeh in 1965, are useful for modeling the uncertain system in industry, nature and humanity. They have crucial role when applied to complex phenomena not easily described by traditional mathematic 30 methods, especially when the goal is to find a good approximate solution. A set is the extended crisp set. While crisp sets only allow full membership or non-membership, fuzzy sets allow partial membership. This property helps FMEA team to evaluate failure modes and give score to them easily. Zadeh proposed to use values ranging from 0 to 1. 0 means complete non-membership whereas 1 means complete membership. Values between 0 and 1 represent intermediate degrees of membership. Fuzzy approach is used in FMEA by so many researches. Kahraman et al. (2013) represented a literature review on this topic (see in Table 3.2). In this thesis research, fuzzy AHP and the GRA approach are used. Table 3.2 Literature review of fuzzy FMEA approaches Authors and year FMEA type Computational technique Application area Pelaez and Bowles 1995 Design Fuzzy cognitive maps knowledge representation Pressure tank system Xu et al. 2002 Design Fuzzy inference system (FIS) Fuzzy expert assessment system Mechanical system Guimares and Lapa 2004a System Fuzzy inference system (FIS) Nuclear reliability engineering Guimares and Lapa 2004b System Fuzzy inference system (FIS) Knowledge-based fuzzy system Reactor nuclear problem Garcia et al. 2005 System Fuzzy data envelopment analysis Nuclear safety systems Guimares and Lapa 2007 System Fuzzy inference system (FIS) Knowledge-based fuzzy system Nuclear engineering systems Yang et al.2008 Design Fuzzy rule-based Bayesian reasoning Offshore engineering systems Chin et al. 2008 Design Fuzzy knowledge based system Knowledge based product design system Tay et al. 2008 Process Fuzzy inference system based occurrence updating model Semiconductor manufacturing environment Nepal et al. 2008 Design Rule-based fuzzy logic Product architecture (PA) capturing interaction failures between different modules Huadong and Zhigang 2009 Design Fuzzy inference system Risk evaluation of boiler tube Rivera and Nunez 2009 Design Fuzzy inference system Discontinuous distillation plant of biofuel 31 Abdelgawad and Fayek 2010 Design Fuzzy inference system, fuzzy AHP and fuzzy expert systems Risk management in the construction industry Hakeem et al. 2010 System Fuzzy logic, fuzzy rule base, ANN model Batch reactor system Yang et al. 2010 Design Fuzzy FMEA integrated with fuzzy linguistic scale method CNC machine tool Chang and Cheng et al. 2010 Process Intuitionistic fuzzy set (IFS) and decision making trial and evaluation laboratory (DEMATEL) DRAM etching process Zhang and Chu 2011 Design Fuzzy-RPNs-based method integrating weighted least square method, the method of imprecision and partial ranking method New product development of a new horizontal directional drilling (HDD) machine Kutlu and Ekmekcioglu 2011 Process Fuzzy TOPSIS-based fuzzy AHP Manufacturing facility of a SME in automotive industry Source: Kahraman et al. 2013 Although AHP is practical and includes the opinions of experts and the use of AHP does not involve cumbersome mathematics, it is criticized because this technique is not capable of reflecting human’s vague thoughts about complicated issue (Seçme et al. 2009). As it is known, AHP involves decomposition, pairwise comparisons, and priority vector generation and synthesis. Though the purpose of AHP is to capture expert’s knowledge, the traditional AHP cannot reflect the experts thinking style. It is hard to determine the values of AHP scale. Experts may prefer intermediate judgments rather than certain judgments. For this reason, a method based on fuzzy approach, fuzzy AHP, make the comparison process more flexible and capable to explain experts’ preferences (Kahraman et al. 2003). There are so many methods presented for the fuzzification of AHP technique. In this study, Buckley’s fuzzy AHP method is used because this method has not been criticized heavily or refuted for its mathematical calculations (Buckley 1985; Kahraman & Çebi 2009). In this method problems should be presented by a hierarchical structure so as to make the solutions of these problems. Then, decision makers determine the linguistic terms for pairwise comparisons of criteria. These linguistic terms should then be converted to fuzzy numbers with the help of Table 3.3 Triangular fuzzy numbers are defined by real numbers, expressed as (l, m, u). l, m and u respectively indicate the smallest, most promising and the largest possible value that describe the fuzzy event (see in Figure 3.1). 32 Figure 3.1 A triangular fuzzy number, M Table 3.3: Triangular fuzzy numbers for linguistic terms Linguistic Scale TFN Just equal (1,1,1) Equally important (1,1,3) Weakly important (1,3,5) Essentially important (3,5,7) Very strongly important (5,7,9) Absolutely important (7,9,9) Source: Kahraman & Çebi 2009 The steps of Buckley’s fuzzy AHP method can be given as in the following: Step 1: The importance of the weights is obtained by pairwise comparisons. Linguistic variables are used to evaluate criteria. These variables convert into fuzzy numbers by using Table 3.1. Then, the pairwise comparison matrices for each decision makers obtained. Ak = (3.1) 1 a12 . . a1n a21 1 . . a2n . . . . . . . . . . . . . . . . . . . . an1 an2 . . 1 33 Ak: Piarwise comparision matrix that established by decision maker k. aij: TFN that demonstrates the comparison of criteria i: criterion i j: criterion j n: number of criteria k: number of decision makers An aggregated pairwise comparison matrix is established to incorporate all decision makers’ pairwise comparison matrices. Aggregate matrix is calculated with the help of geometric mean and brings each decision maker evaluations together. aij = K Kijijij aaa ...21 : fuzzy multiplication Step 2: The fuzzy weight matrix is calculated by Buckley’s method (Kahraman & Çebi 2009) as follows: n iniii aaar ~...~~~ 21 121 )~...~~(~~ nii rrrrw : fuzzy addition :~ir geometric mean of fuzzy comparison value of criterion I to each criterion. :~iw weight of criterion i Step 3: The final step requires defuzzification and normalization of calculated fuzzy weights (wi) in order to convert them into crisps values (Eq. (5)). For the defuzzification process, centroid method is selected since it is the most commonly used method (Opricovic & Tzeng 2004) n i i rurmrl n i i r r w www w ww 11 ~~ ~ (3.4) (3.5) (3.2) (3.3) 34 wr : a crisp weight number 3.1.2.1 Assigning different weights to decision makers with Fuzzy AHP Decision makers have an important role in FMEA application. Their judgments on decision factors determine and bring about the values of RPN1s, RPN2s and RPN3s. Therefore the prioritization of failure modes depends on DMs’ judgment. To reduce the responsibility of DMs and the difficulty of making decision, a fuzzy approach is added to this research. Thanks to fuzzy approach, DMs have chance to make their judgment by using linguistic terms. Unfortunately it is not enough for getting satisfied results from FMEA. Each firm employs different decision makers who have different ability on risk analyzing. FMEA team often demonstrates different opinions and knowledge from one team member to another and produces different types of assessment information because they differ from each other with their experience, knowledge, education level. FMEA team members cannot possess the same feature about evaluating the risks. For instance, FMEA team consists of four people. One of them is C-class OHS specialist, another one is A-class OHS specialist and the others are company’s maintenance workers. Maintenance workers are always be in the same place (a place that the FMEA applied in) so they know the workplace very well. On the other hand, OHS specialists are good at laws and regulations. For these reasons, DMs should be weighted and their judgment should be associated with corresponding weights. In this research, five criteria are presented in Table 3.4. The criteria are objective so at least two experienced and knowledgeable risk analysts/experts/occupational health and safety specialists can determine these five criteria’s importance weights. They make pairwise comparisons of criteria and then calculate the crisp number of criteria weights by using Buckley’s fuzzy AHP approach as presented above. Once the analysts/experts/occupational health and safety specialists determine the criteria’s weights, the weights are not change anymore. They also determine the importance weights for decision makers of current FMEA application. Risk analysts/experts/occupational health and safety specialists give scores to FMEA team’s member (decision makers) by using Table 3.5-3.9. It is easy because they are not forced to make subjective judgment. 35 Each DM’s weight is calculated by multiplying each criterion score with its criterion weight and then the summation of all five gives the weights of DM. Remember that the criteria weights derive from the fuzzy AHP calculations. Finally, we normalized these five weights and obtain the crisp values of DMs’ weights. Table 3.4 Criteria of decision makers Table 3.5: Criterion-1, job relevance Criteria 1- Decision maker’s job relevance with the process 2- The number of years of experience in the current sector 3- Education level and the relevance of decision maker’s job with his/her education 4- The result of test which is about FMEA and risk analyzing 5- The number of times that the decision maker has took part in FMEA application or another risk analysis methodology and the similarity between the current process and the process that have been already experienced Criterion-1 Decision maker’s job relevance with the current process of FMEA application Score The application of FMEA is taken place in DM’s workplace and DM is a member of this process already 5 The application of FMEA is taken place in DM’s workplace and DM was a member of this process already 4 The application of FMEA is taken place in DM’s workplace and DM has never be a member of this process 3 The application of FMEA is not taken place in DM’s workplace but DM had been worked similar process in the past 2 The application of FMEA is not taken place in DM’s workplace and DM has never be a member of similar process but DM is able to make judgment 1 36 Criterion-1, decision maker’s job relevance with the current process of FMEA application, provides us an ability to evaluate FMEA team member’s routine job relevance with the current FMEA process. For instance, a person who is responsible for maintenance work in the workplace can easily evaluate the potential failure modes. On the other hand, an OHS specialist may not be aware of the failure modes in the current workplace because of his/her lack of knowledge about the process. It is normal that the maintenance worker knows about the workplace’s failures and possible risks. It is also normal that the OHS specialist does not know about the workplace’s failures like maintenance person does. Table 3.6: Criterion-2, experience Decision maker should be an experienced one. We cannot learn everything by reading books or being educated in school. If the knowledge is not used in the workplace arena, it is almost nothing. Criterion-2 is about the decision maker’s experiences. Experiment can contribute beneficial features to the DMs. Criterion-2 The number of years of experience in the current sector Score DM has/had worked in this sector for more than 10 years 5 DM has/had worked in this sector for more than 6-10 years 4 DM has/had worked in this sector for more than 3-6 years 3 DM has/had worked in this sector for more than 1-3 years 2 DM has/had worked in this sector for more than 0-1 year 1 37 Table 3.7: Criterion-3, education level Education makes people think more analytically. The more education means the more knowledgeable people. Decision makers should be educated persons because they determine the failure modes, risks, and so the health and safety of people in general. Besides the level of education, it is also important that the relevance is between the education of decision maker and his/her current job. Table 3.8: Criterion-4, test result FMEA team member should be given information about risk analyzing and especially FMEA methodology before the application of FMEA. After that a test should be applied to decision makers (FMEA team). The test should evaluate the knowledge level of risk analyzing and FMEA methodology. If the DM gets the grade that is under the 50 points, Criterion-3 The education level and the relevance of decision maker’s job with his/her education Score Bachelor’s degree/ Master degree or better and the current process is parallel with DM’s education 5 Bachelor’s degree/ Master degree or better and the current process is partially or less related with DM’s education 4 DM graduated from high school and the process is parallel with DM’s education 3 DM graduated from high school and the process is partially or less related with DM’s education 2 DM graduated from any school that is below the high school and the process is parallel with DM’s education 1 Criterion-4 The result of test which is about FMEA and risk analyzing Score DM’s test result is between 91-100 points 5 DM’s test result is between 81-90 points 4 DM’s test result is between 66-80 points 3 DM’s test result is between 56-65 points 2 DM’s test result is between 50-55 points 1 38 he/she will be re-educated and then re-evaluated. Sample test is presented in Appendix C-1. Table 3.9: Criterion-5, repetition number In the criterion-2, the importance of experiment is mentioned. Criterion-5 is also about experience but it considers the repetition number of risk analyzing that the DM has been taken part in. The relevance of risk analysis methodology and DM’s routine job is important for this criterion. For example, it makes no sense that the industrial engineer applies FMEA in the architecture office to meet the OHS needs. 3.1.2.2 Assigning different weights to decision factors (both for RPN1 and RPN2) with Fuzzy AHP Remember that RPN1 is calculated by using decision factors O, S, and D. Weighted DMs give scores for each of decision factors. After that, GRA is applied to obtain RPN1 values. RPN1 values are more consistent and convenient from traditional RPN because DMs and decision factors have been weighted and GRA is applied to incorporate decision factors’ weights to the calculation. Criterion-5 The number of times that the decision maker has took part in FMEA application or another risk analysis methodology Score DM has took part in FMEA application or another risk analysis methodology that is relevant with his/her job 5 or more times 5 DM has took part in FMEA application or another risk analysis methodology that is relevant with his/her job 1-5 times 4 DM has took part in FMEA application or another risk analysis methodology that is not relevant with his/her job 3 or more times 3 DM has took part in FMEA application or another risk analysis methodology that is not relevant with his/her job 1-3 times 2 DM has took part in FMEA application or another risk analysis methodology 0-1 time. 1 39 RPN2 values are obtained because of the need for reprioritization of corrective actions. Corrective actions should be reprioritized because RPN1 values do not consider the cost of C.A., time loss due to C.A. performing, regulations, firm’s prevention policy and so on. Thus, five additional criteria are added to calculate RPN2. RPN2 calculated by means of summation of decision factors (criteria) scores that scores are multiplied with corresponding weights. Both RPN1 and RPN2 have their own decision factors and these factor need for weights. These weights are derived from pairwise comparison of Fuzzy AHP approach in the calculation of RPN1 values and added to RPN1 calculation with the help of GRA. While RPN2 is calculating, the weight is added to calculation directly by multiplying corresponding score of decision factor. While decision makers make pairwise comparisons for both RPN1 and RPN2 calculation, they evaluate factors such as cost, time, worker’s ability, capability, obligation, in addition to their experiences about OHS concept. Then, they determine the comparison terms with respect to overall goal. 3.2 GREY RELATIONAL ANALYSIS Another approach that underlies the proposed methodology of this thesis is grey relational analysis. Firstly, the grey theory should be understood to comprehend the aim of grey relational analysis. There exists any system that is capable of capturing all information perfectly without uncertainty (Chang et al. 2013). Deng (1982) first proposed the Grey theory to deal with the analysis of systems that are struggle with incomplete information. Grey theory asserts that the system is a white system when the required information is entirely available. On the other hand, the system is black if the required information is entirely unavailable. A system with partially available information is called as a grey system (Chang et al. 2013). Deng (1989) presented that the grey theory consist of six major components: grey generating, grey relational analysis, grey prediction, grey model, grey decision making and grey control. Grey relational analysis is the part of grey theory, dealing with the multiple criteria decision making problem, and it is consists of a complicated interrelationship between multiple factors. GRA can be adapted to FMEA especially to RPN calculations because GRA is suitable for solving problems with complicated relationship between multiple factors and variables. The use of grey theory and so the GRA within the FMEA framework is practical and can be accomplished (Liu et al. 2011). GRA also fits FMEA because its 40 decision factors have all of these characteristics of being existent, countable, extensible and independent. These characteristics derive from a grey theory. Grey theory provides for experts a measure to analyze relationship between discrete quantitative and qualitative series. As stated above, under the sub-title of “Shortcomings of FMEA”, the traditional RPN calculation method is criticized because of several shortcomings. To handle with these drawbacks, incorporating GRA method to the FMEA may help in many ways and these ways are presented below: i. GRA eliminates high duplication rate problem. Because the combinations of S, O and D are a discrete series and the statistical distribution is unknown. Some of the duplication problems that the GRA improves are listed below: a. Traditional RPN calculation includes 120 unique RPN values; whereas GRA integrated RPN calculation include 220 unique GRA values. b. Traditional RPN calculation can be have 24 times duplication; whereas GRA integrated RPN calculation can be have 6 times duplication. c. Almost all RPN values are less than 500 in the traditional RPN calculation; whereas GRA values are more uniformly distributed (Chang et al. 2013). ii. By grey relational analysis, the deviation in traditional RPN calculation due to excessive classified levels is eliminated and the accuracy is improved. iii. GRA has competitive advantages in terms of the effective processing of uncertainty, multi-input, discreet data, and data incompleteness. Especially, the flexibility that enables to assign different weighting coefficients to the decision factors O, S, and D. iv. GRA is not requiring utility function of any form. v. GRA can be adapted to FMEA especially to RPN calculations because GRA is suitable for solving problems with complicated relationship between multiple factors and variables. 41 3.2.1 Application of Grey Relational Analysis to FMEA GRA can be applied FMEA easily. To carry out the GRA, firstly fuzzy linguistic variables are defuzzified as crisp values. These fuzzy linguistic variables are about the decision factors O, S, and D. The crisp value for them can be incorporated to the GRA. The lowest levels of the three decision factors are defined as standard series (1,1,1). The evaluation information of the three decision factors for each failure mode is viewed as comparative series. After that the grey relational coefficient and degree of relational with standard series are computed in terms of the grey theory. By incorprating weights of O, S, and D to the calculation, we finally get the RPN values. The greater RPN value means that the smaller effect of the current failure mode. In the following part, the steps of grey relational analysis to the FMEA are presented (Chang et al. 1998). Step 1: Establish the comparative series An information series with n components or decision factors, such as chance of occurrence, chance of undetection and severity of failure is the comparative series. Comparative series can be expressed as, )3(),2(),1()( iiii XXXkX k= 1,2 or 3 (number of decision factors) i= 1,2,3,….n (n is the number of failure modes) Here k has the value of 1,2 and 3; meanings )3(),2(),1( iii XXX are the scores of each decision factors respectively. If all series are comparative series, the n information series can be defined as followin matrix, equation 1. )3( )3( )3( )2( )2( )2( )1( )1( )1( )( )( )( )( 2 1 2 1 2 1 2 1 nnn n i X X X X X X X X X kX kX kX kX    (3.6) 42 Step 2: Establish the standard series Degree of relation can describe the relationship of two series, thus, an objective series called as standart series can be expressed as the following; Series notation: ),3(),2(),1()( 0000 XXXkX Matrix notation: )3()2()1()( 0000 XXXkX When applying the traditional FMEA, the smaller the score, the less the risk, therefore the standart series should be the lowest score of occurrence, detectability and severity factors which is shown as following: Matrix notation is: 1,1,1)3(),2(),1()( 0000 XXXkX 111)(0 kX Step 3: Obtain the differences between comparative series and standard series To discover the degree of grey relationship, the difference between the scores of decision factors and scores of standard series must be calculated and expressed as the matrix shown below. )3( )3( )3( )2( )2( )2( )1( )1( )1( )( )( )( )( 0 02 01 0 02 01 0 02 01 0 02 01 nnn n i k k k k    Here, i= 1,2,3….n ( n is the number of failure modes) )(ki is calculated as the following; )()()( 00 kXkXk ii (3.7) (3.8) (3.9) 43 Step 4: Compute the grey relation coefficient To compute the relation coefficient, the decision factors of the failure model are compared with the standard series, and the relation coefficient is expressed as: max maxmin 0 )()(),( kkXkX ii )(0 kX : standard series )(kX i : comparative series i = 1,2,3…..n (n is the number of failure modes) k= 1,2 or 3 (number of risk factors) min = Minimum value of all )(ki max = Maximum value of all is an identifier coefficient and only affects the relative value of the risk without changing priority. It is equals to 0,5. Step 5: Determine the degree of relation To find the degree of relation, weighting coefficient of the decision factors must be first decided. The relative weights of decision factors will be used in the following formulation. 3 1 )()( k iki kk k = the weighting coefficient of the risk factors and 3 1 1k k In the proposed methodology 3 1 1k k . Each of the risk factors (decision factors O, S, and D) has different weight. A numerical example is represented in the case study. (3.10) (3.11) 44 3.3 THIRD RISK PRIORITY NUMBER (RPN3) Traditional FMEA uses the old-fashioned RPN calculation that requires the multiplication of O, S, and D scores. Although decision makers and decision factors are weighted by using Fuzzy AHP and calculation of RPN values (RPN1) by the help of GRA, there should be last step before performing the corrective actions. So far, we try to remove the shortcomings of RPN calculations. There exists almost no study that considers the need for corrective actions as to in real life situation. More clearly, the prioritization by using RPN1 scores is not enough to realize the corrective actions. This research is presented a new approach, a new logic. “Is it really worth to do the corrective action of failure mode 12?”. All firms want to get rid of all failure modes but there are some limitations such as cost, time, and obligation of laws. Unfortunately, a firm may contain hundreds of failure modes and the decision makers should make reasonable decisions about corrective action realization. Thus, a new RPN (RPN2) is added to realize this notion. The steps of this approach are stated below: Step 1: Weighting of criteria Five criteria are determined as decision factors of RPN2 calculations. First, these 5 criteria are weighted by using Fuzzy AHP. Criteria are presented in Table 3.11. Step 2: Determining the threshold integral for corrective actions Although the determination of threshold for traditional has criticized a lot in so many study, for instance in the study of Abdelgawad et al. (2010). This approach determines a threshold interval. The threshold is not only a crisp RPN value or a number of corrective actions. In this study, a threshold interval is asserted. Table 3.10 represent threshold interval. Table 3.10: Threshold intervals for corrective actions of failure modes Threshold interval for corrective actions of failure modes RPN ≤ 0,600 Corrective action should be performed immediately RPN > 0,600 Corrective action should be performed according to the company's prevention policy within a reasonable time 45 We determine the two parts. In the first part, corrective actions should be performed as soon as possible. In the second part, corrective action should be performed again but it should be done after the first part’s corrective action. Every company has prevention policy to meet the needs for OHS concept. In this policy, a reasonable time should be identified clearly. For each corrective action which is in the second part, a reasonable time to perform the corrective action should be determined under the leadership of OHS specialist. The threshold should be determined based on the cost, number, frequency and time consuming criteria of corrective actions. Moreover, the capabilities of firm (cost, labor force, time limitation and so on) and also time period of FMEA renewing are other important factors. Therefore the value of 0,6000 may differ from company to company. Step 3: Determining scores for decision factors (criteria) Each criterion has 5-point decision scale and scales are presented in Table 3.12-3.16. Decision makers determine scores for each corrective action. DMs use the linguistic terms in that criteria scales while giving scores. Each DM make pairwise comparison matrix and evaluate criteria individually. At the final step, each DM’s weights are incorporated into the calculation so that the weighted RPN2 values are occurred. Table 3.11: Criteria for RPN2 evaluation Criteria 1- Additional cost due to corrective action 2- Time loss due to the performing of corrective action 3- Regulation obligatory 4- Prevention policy 5- Customer satisfaction and reputation of firm 46 Table 3.12: Criterion-1, additional cost When we talk about the applicability of corrective actions, it is important additional cost due to corrective action. Although safety and health issues are more crucial, cost of C.A. should be considered when there are big differences in cost of C.A. For example, the absence of fire extinguisher can be eliminated for a few dollars. On the other hand, buying a good quality earplug for 1000 workers require more money to handle it. Therefore, if these two corrective actions have very closed RPN1 values, it is possible that the fire extinguisher will be bought before the earplugs. Table 3.13: Criterion-2, time loss Corrective action can cause time loss in two ways. First, corrective action needs new machine or new operator. Finding new operator or supplying new machine may take too much time. Consider that a failure mode is detected and, as to management decision, the machine stopped working. As far as the time that the corrective action is performed there would be time loss and it causes additional cost. In clear, time loss may cause Criterion-1 Additional cost due to corrective action Score Corrective action causes very little additional cost 0 Corrective action causes little additional cost 0,25 Corrective action causes ordinary additional cost 0,50 Corrective action causes serious additional cost 0,75 Corrective action causes huge additional cost 1 Criterion-2 Time loss due to the performing of corrective action Score Corrective action causes very little additional time loss 0 Corrective action causes little additional time loss 0,25 Corrective action causes acceptable additional time loss 0,50 Corrective action causes serious time loss 0,75 Corrective action causes huge time loss 1 47 additional cost. Second, machine may need to be repaired and this process also takes some time. Table 3.14: Criterion-3, regulation obligatory Criterion-3 is more about the OHS concept. Almost all countries have laws and regulation to provide the welfare and satisfaction for workers and the customers. Regulations force companies to take some precautions. Companies have to meet the requirement of regulations because of worker’s and customer’s health and safety. Moreover, they have to meet the requirements to be able to avoid from governmental punishment. Table 3.15: Criterion-4, prevention policy Criterion-3 Regulation obligatory Score Corrective action have to be done because of the regulation obligatory 0 Corrective action must be done because of the regulation obligatory 0,25 Corrective action ought to be done because of the regulation obligatory 0,50 Corrective action can be postponed to a later date 0,75 Managers do not have to do corrective action just because of the regulation obligatory 1 Criterion-4 Prevention policy of the company Score Corrective action have to be done because of the prevention policy of the company 0 Corrective action must be done because the prevention policy of the company 0,25 Corrective action ought to be done because of the prevention policy of the company 0,50 Corrective action can be postponed to a later date 0,75 Managers do not have to do corrective action just because of the prevention policy of the company 1 48 In Turkey and other developed/developing countries, firms have to form their own prevention policy to meet the requirements of OHS. Prevention policy helps FMEA team member to apply the methodology faster. Furthermore, prevention policy provides old data in some way. Decision makers can use them effectively. Table 3.16 Criterion-5, customer satisfaction and reputation The reputation of the firm is one of the fundamental properties that affect the firms in many ways. For instance, the reliability, prestige of firm and sales ratio of the products are affected by reputation of the firm. If a failure mode is missed out by risk analysts and then this failure mode cause to big accident, poisoning, etc. that affects the reputation of the firm negatively. Another point is that once a time the reputation is exposed to reputational harm; it is really hard to gain past reputation and prestige of the firm again. Criterion-5 considers customer satisfaction status of firms. Responds and complaints from customer play important role. Step 4: Determining RPN2 value For each decision factors, decision makers give scores. Multiplication of each corrective actions’ scores with corresponding weights of criteria (decision factors) together with weights of decision makers and then the summation of all what we calculate the value of RPN2 for current corrective action. Finally, the RPN2 values are normalized. For example, we have 5 criteria and the weights are 0,25, 0,15, 0,20, 0,10, 0,30 by DM1; 0,25, 0,20, 0,10, 0,30, 0,15 by DM2 and 0,40, 0,25, 0,10, 0,15, 0,10 respectively. And Criterion-5 Customer satisfaction and reputation of firm Score Corrective action have to be done because of customer satisfaction and/or reputation of firm 0 Corrective action must be done because of customer satisfaction and/or reputation of firm 0,25 Corrective action ought to be done because of customer satisfaction and/or reputation of firm 0,50 Corrective action can be postponed to a later date 0,75 Managers do not have to do corrective action just because of customer satisfaction and/or reputation of firm 1 49 the scores given by each DMs are 0,25, 0,50, 0,50, 0, 0,75 by DM1; 0,50, 0,75, 0,75, 0,25, 0,25 by DM2 and 0,50, 0,25, 0,75, 0,50, 0,50 by DM3. The weights of DMs are 0,40, 0,25, 0,35 respectively. RPN2 value for that failure mode is presented below: RPN2= (0,25*0,25+0,15*0,50+0,2*0,50+0,10*0+0,30*0,75)*0,40 + (0,25*0,50+0,20*0,75+0,10*0,75+0,30*0,25+0,15*0,25)*0,25 +(0,40*0,50+0,25*0,25+0,10*0,75+0,15*0,50+0,10*0,50)*0,35 = 0,247 Step 5: Calculation of RPN3 RPN1 is calculated by means of three decision factors O, S, and D and RPN2 is calculated by means of five decision factors (criteria). As to 3 occupational health and safety specialists (One of them is class-A, one of them class-B and the other one is class-C) the reasonable coefficient for RPN2 is 0,05. Thus, RPN1 has a coefficient equal to 0,95. They determined these number as to their experiences (the application number of FMEA in different firms) and academic knowledge. After calculating RPN3 values, FMEA team should prepare report paper to follow the actions, check the calculation and keep the record of failure modes. A sample paper is presented in Figure 3.2. 50 Date: 28.06.2014 FAILURE MODE AND EFFECTS ANALYSIS (FMEA) RISK ANALYSIS FORM FMEA Type: Proses Proses: Fuel Tanks FMEA NO. 1 FMEA Team Member: Chemist Ece ESEN, C-Class OHS Specialist Fulya DURU, Maintenance worker Ahmet TAYLAN, No. System/ Part FM FE. CF. DM. NO. & Weight s Weights of DF (O,S, and D) DF scores for each DMs RPN1 RPN2 RPN3 Pr. Corrective Action Respon sibility Due Time Next Control Time/ Control Period DF scores for each DMs after C.A. RPN1 1 Tank Liquid Burst Poisoning Measurement Error 1 (0,44) O= 0,25 O= 6 0 ,5 4 0 0 ,2 0 7 0 ,2 0 7 0 ,5 5 2 6 New measurement machine should be bought. Labora tory (Ece ESEN) 08.07 2014 08.10 2014 O= 2 0 ,9 2 0 S=0,45 S=7 S=1 D=0,30 D=6 D=1 2 (0,18) O= 0,25 O= 7 O= 1 S=0,45 S=6 S=2 D=0,30 D=5 D=1 3 (0,38) O= 0,25 O= 5 O= 3 S=0,45 S=7 S=2 D=0,30 D=4 D=2 51 4. CASE STUDY 4.1 ADOPTATION THE PROPOSED METHODOLOGY TO THE CASE COMPANY The application of the proposed FMEA approach has been applied in a yarn manufacturing company. The company is in the Turkey, covered 35000 m2 area. Its annual production is 6000 tons. Target market is European countries (%70) and the remaining is for domestic market. In recent years, occupational health and safety concept is rapidly growing in Turkey. Thus, the company fulfills the obligations which are about the OHS. One of the fundamental steps of OHS activities is risk analyzing. The risk analysis is made with the help of proposed FMEA methodology. Three decision makers are assigned to apply the proposed FMEA methodology. Thanks to the visual research in all departments of the company, risk analysis, accidents and near misses event records from previous years, in addition to getting feedback from engineers, foremen and workers, 50 failure modes are determined. The list of failure modes is presented in Table 4.1. Table 4.1: A list of failure modes FM NO. System/ Part Failure Mode 1 Husks storage Too high stacking 2 Waste landfill area Fluorescent dust is breathed by workers 3 Blow room section Working in high places 4 Blow room section Dusty air condition in workplace 5 Manufacturing section It seems hard to reach fire extinguisher in emergency 6 Manufacturing section Non-ergonomic way of working 7 Manufacturing section Working without steel consolidated shoes 8 Manufacturing section The absence of fire extinguisher 9 Manufacturing section Working in dusty workplace 10 Manufacturing section Workers do not use earmuff 11 Welding workshop Workers use no vise while they use drill machine 12 Mess hall Workers in mess hall do not wear protective shoes 13 Packaging section Working without machine protector 14 Packaging section There is no automatic stopper mechanism and any sensor when the door is opened 52 15 Packaging section Inconvenient conversation for Solvent 16 Packaging section Manometer’s limits is not identified 17 Packaging section Workers enter the machine and then carry out materials from the inside 18 Mess hall Electric plug is very close to LPG connection cable 19 Operation-1 section The possibility of deactivation of safety cover of machine 20 Operation-1 section There are unnecessary materials inside the fire cabinet 21 Operation-1 section Opening the iron bale package by using iron scissors 22 Operation-1 section The workplace is not isolated against the risk of cylinder falling 23 Operation-1 section Workers put order the items one by one by using their naked hands. 24 Operation-1 section The machine door is throwing back too fast 25 Operation-2 section The possibility of fire in BOX machine 26 Operation-2 section The absence of middle-railing 27 Operation-1&2 section The safety wire of Bobbin machine is worn out 28 Administrative building Machine of drinking water is placed on the wet floor 29 Administrative building Workers monitor the PC screen for a long time period without break 30 Administrative building The possibility of existence of Legionella bacteria inside air conditioner water 31 Administrative building The emergency door is opening to the inside of workplace 32 Husks pressing section Workers put their naked hands inside the press machine 33 Husks pressing section The usage of falcate 34 Husks pressing section Working in high workplace without preventative measurement 35 Husks pressing section The absence of fire extinguisher 36 Items winding section Workers use their naked hands to put items into rotating parts of machine 37 Items winding section Too much stacking of items 38 Carder machine There some snacks, drinks etc. exist inside the machine 39 Carder machine Broken safety switches on the control panel 40 Carder machine Workers insert cotton into machine by their naked hands 41 Carder machine Workers clean the moving parts of machine with naked hand 42 Ring machine There is no dust absorber mechanism in the workplace 43 Ring machine Rotating parts of the machine have no protective cover 44 Bobbin machine Coiling up the reel too much forces the reel to fly off 45 Bale opener machine The possibility of deactivation of safety chains and safety sensor 46 Fuel tank (Across the Box storage) The tank is plastic and has no earthling 47 Diesel forklift (Warehouse) Using diesel forklift in closed workplace 48 Diesel forklift (Warehouse) Driving forklift uncontrollable and fast 49 Diesel forklift (Warehouse) Safety lock which is for forklift basket is broken 50 Laboratory The rotating parts of the machine has no preventative cover 4.1.1 Decision Makers Weighting Firstly, 3 experienced and knowledgeable occupational health and safety specialists (One of them is class-A, one of them class-B and the other one is class-C) determine weights for FMEA team members (decision makers, DMs) with the help of interpersonal comparisons and fuzzy AHP. Fuzzy AHP is used for determining the criteria weights of DMs and the interpersonal comparison is used for determining scores to DMs. In this case, three decision makers are employed to apply proposed FMEA 53 methodology for the case company. One of the decision makers (FMEA team member) is experienced maintenance worker; the other one is C-Class OHS specialist and last one is one of supporting worker. DMs’ features are presented below: Maintenance worker (DM1): Maintenance worker has worked for that firm for 8 years. He graduated from industrial vocational high school. He always tries to eliminate the failures in the workplace. He tries to repair the machines, provide maintenance of machines, machine parts and so on. He has 54 from the test risk analyzing test. He has been a member of risk analysis application for 4 times (2 of them has taken part in his workplace the other two has not). C-Class OHS specialist (DM2): OHS specialist has worked for that firm for 2 years. He graduated from industrial engineering department. He is aware of workplace failures but not as maintenance worker does. He has deep knowledge about the regulations and laws regarding with OHS concept. He has 96 from the test risk analyzing test. He has been a member of risk analysis application for 3 times (none of them has taken part in his workplace before). Supporting worker (DM3): Supporting worker is the worker who is employed by employer for preventing, protecting, evacuating and firefighting issues regarding with OHS. He/she is specifically assigned for first-aid and other similar matters. He/she has proper equipment and adequate knowledge. These definitions are taken from the Turkish OHS law. In the case study, supporting worker has worked for that firm for 5 years. He graduated from university, philosophy department. Unluckily he could not find a job that is related his education so that he began to work for that firm as a production worker. He partially knows and is unaware of failure in the workplace because he works in the same place. He has 70 from the test risk analyzing test. He has been a member of risk analysis application for 4 times (1 of them has taken part in his workplace the other three has not). After three experienced and knowledgeable occupational health and safety specialists determine weights for DMs’ criteria by using fuzzy AHP, they give scores to DMs for each criterion. By multiplying each criterion weight with the decision makers’ criterion 54 score we obtain the DMs’ weights. Finally, the normalization of calculated weights gives us the crisp values of DMs’ weights. A numerical study is presented below: Step 1: Determination of DMs’ criteria weights Pairwise comparison matrices are established by three OHS specialists as shown in Table 4.2-4.4. Each OHs specialist makes pairwise comparisons for five criteria. Table 4.2: A pairwise comparison matrix of DMs (OHS-1) OHS-1 Job relevance Experience Education Level Test Result Repetition Number Job relevance (1.00,1.00,1.00) (0.20,0.33,1.00) (3.00,5.00,7.00) (5.00,7.00,9.00) (3.00,5.00,7.00) Experience (1.00,3.00,5.00) (1.00,1.00,1.00) (3.00,5.00,7.00) (7.00,9.00,9.00) (5.00,7.00,9.00) Education Level (0.14,0.20,0.33) (0.14,0.20,0.33) (1.00,1.00,1.00) (1.00,3.00,5.00) (0.33,1.00,1.00) Test Result (0.11,0.14,0.20) (0.11,0.11,0.14) (0.20,0.33,1.00) (1.00,1.00,1.00) (0.20,0.33,1.00) Repetition Number (0.14,0.20,0.33) (0.11,0.14,0.20) (1.00,1.00,3.00) (1.00,3.00,5.00) (1.00,1.00,1.00) Table 4.3: A pairwise comparison matrix of DMs (OHS-2) OHS-2 Job relevance Experience Education Level Test Result Repetition Number Job relevance (1.00,1.00,1.00) (0.33,1.00,1.00) (1.00,3.00,5.00) (5.00,7.00,9.00) (5.00,7.00,9.00) Experience (1.00,1.00,3.00) (1.00,1.00,1.00) (3.00,5.00,7.00) (5.00,7.00,9.00) (5.00,7.00,9.00) Education Level (0.20,0.33,1.00) (0.14,0.20,0.33) (1.00,1.00,1.00) (3.00,5.00,7.00) (0.14,0.20,0.33) Test Result (0.11,0.14,0.20) (0.11,0.14,0.20) (0.14,0.20,0.33) (1.00,1.00,1.00) (0.11,0.14,0.20) Repetition Number (0.11,0.14,0.20) (0.11,0.14,0.20) (3.00,5.00,7.00) (5.00,7.00,9.00) (1.00,1.00,1.00) 55 Table 4.4: A pairwise comparison matrix of DMs (OHS-3) OHS-3 Job relevance Experience Education Level Test Result Repetition Number Job relevance (1.00,1.00,1.00) (0.33,1.00,1.00) (5.00,7.00,9.00) (7.00,9.00,9.00) (3.00,5.00,7.00) Experience (1.00,1.00,3.00) (1.00,1.00,1.00) (5.00,7.00,9.00) (7.00,9.00,9.00) (1.00,3.00,5.00) Education Level (0.11,0.14,0.20) (0.11,0.14,0.20) (1.00,1.00,1.00) (3.00,5.00,7.00) (0.14,0.20,0.33) Test Result (0.11,0.11,0.14) (0.11,0.11,0.14) (0.14,0.20,0.33) (1.00,1.00,1.00) (0.11,0.14,0.20) Repetition Number (0.14,0.20,0.33) (0.20,0.33,1.00) (3.00,5.00,7.00) (5.00,7.00,9.00) (1.00,1.00,1.00) Step 1.1: Calculation of ir and iw values After obtaining the pairwise comparison matrix, geometric mean values of fuzzy comparison values and fuzzy weight matrix is calculated by using Eqs. (3.3) and (3.4) as follows in the Table 4.5 and 4.6: Table 4.5: Geometric mean values of fuzzy comparison values OHS-1 OHS-2 OHS-3 l m n l m n l m n r1 1,5518 2,2552 3,3798 1,5281 2,7131 3,3227 2,0362 3,1598 3,5540 r2 2,5365 3,9363 4,9036 2,3714 3,0049 4,4273 2,0362 2,8529 4,1392 r3 0,3686 0,6544 0,8891 0,4146 0,5818 0,9510 0,3505 0,4592 0,6223 r4 0,2181 0,2814 0,4911 0,1813 0,2255 0,3056 0,1813 0,2039 0,2671 r5 0,4366 0,6118 1,0000 0,7137 0,9349 1,2030 0,8441 1,1847 1,8384 Table 4.6: Fuzzy weights matrix OHS-1 OHS-2 OHS-3 l m n l m n l m n w1 0,1455 0,2914 0,6612 0,1497 0,3637 0,6379 0,1954 0,4020 0,6523 w2 0,2379 0,5086 0,9593 0,2323 0,4028 0,8499 0,1954 0,3629 0,7597 w3 0,0346 0,0846 0,1739 0,0406 0,0780 0,1826 0,0336 0,0584 0,1142 w4 0,0205 0,0364 0,0961 0,0178 0,0302 0,0587 0,0174 0,0259 0,0490 w5 0,0409 0,0791 0,1956 0,0699 0,1253 0,2309 0,0810 0,1507 0,3374 56 Step 1.2: Calculation of rw values The defuzzification and normalization is calculated by using Eq.(3.5) as presented in Table 4.7: Table 4.7: The crisp values of DMs criteria weights OHS-1 Weights OHS-2 Weights OHS-3 Weights Arithmetic mean w1 0,3080 0,3317 0,3638 0,3345 w2 0,4784 0,4279 0,3837 0,4300 w3 0,0822 0,0868 0,0600 0,0763 w4 0,0429 0,0307 0,0269 0,0335 w5 0,0885 0,1228 0,1657 0,1257 To see the criteria weights more visually; Job relevance = 0,3345 Experience = 0,4300 Education level = 0,0763 Test Results = 0,0335 Repetition Number = 0,1257 Step 2: Giving scores to DMs by using criteria scales The criteria scores that are given by three OHS specialists are shown in Table 4.8. Table 4.8: Criteria score table of DMs Job relevance Experience Education Level Test Result Repetition Number DM1 5 4 3 1 4 DM2 2 2 4 5 3 DM3 3 3 4 3 4 57 Step 3: Calculation of the final crisp values of DMs’ weights And finally the crisp values of DMs’ weight are obtained as shown in Table 4.9. Table 4.9: The final crisp values of DMs’ weights Job Relevance x 0,3345 Experience x 0,4300 Education Level x 0,0763 Test Result x 0,0335 Repetition Number x 0,1257 Weights of DMs Normalized Weights of DMs DM1 5 4 3 1 4 4,1577 0,4269 DM2 2 2 4 5 3 2,3788 0,2442 DM3 3 3 4 3 4 3,2020 0,3288 4.1.2 Determining RPN1 values Then, each decision maker makes judgments on decision factors (O, S and D) and determines their weights separately with the help of fuzzy AHP. At this point, GRA is applied to calculate RPN1 values. The first prioritization of failure modes and the prioritization of corresponding corrective actions are made according to RPN1 values. Step 1: Assigning different weights to decision factors O, S, and D At this point each decision makers makes judgments on decision factors (O, S and D) and determine their weights separately with the help of fuzzy AHP. To be able to incorporate these weights to the calculations GRA is used. First, to assign weights to decision factors O, S, and D we should establish a pairwise comparison matrix of decision factors for each of the decision makers as shown in Table 4.10-4.12. Then, DMs apply Buckley’s Fuzzy AHP method to obtain the weights. Steps are similar to DMs’ weighting process. 58 Table 4.10: DM1 pairwise comparison matrix DM1 O S D O (1.00,1.00,1.00) (0.11,0.14,0.20) (0.20,0.33,1.00) S (5.00,7.00,9.00) (1.00,1.00,1.00) (3.00,5.00,7.00) D (1.00,3.00,5.00) (0.14,0.20,0.33) (1.00,1.00,1.00) Table 4.11: DM2 pairwise comparison matrix DM2 O S D O (1.00,1.00,1.00) (0.14,0.20,0.33) (0.20,0.33,1.00) S (3.00,5.00,7.00) (1.00,1.00,1.00) (1.00,3.00,5.00) D (1.00,3.00,5.00) (0.20,0.33,1.00) (1.00,1.00,1.00) Table 4.12: DM3 pairwise comparison matrix DM3 O S D O (1.00,1.00,1.00) (0.14,0.20,0.33) (0.33,1.00,1.00) S (3.00,5.00,7.00) (1.00,1.00,1.00) (1.00,1.00,3.00) D (1.00,1.00,3.00) (0.33,1.00,1.00) (1.00,1.00,1.00) Step 1.1: Calculation of ir and iw values After obtaining the pairwise comparison matrix, geometric mean values of fuzzy comparison values and fuzzy weight matrix is calculated by using Eq. (3.3) and Eq. (3.4) as follows in the Table 4.13 and 4.14: Table 4.13: Geometric mean values of fuzzy comparison values DM1 DM2 DM3 l m n l m n l m n r1 0,3625 0,5848 0,6934 0,4055 0,6934 1,0000 0,6934 1,0000 1,4422 r2 1,4422 2,4662 3,2711 1,0000 1,4422 2,4662 1,0000 1,0000 2,0801 r3 0,5848 0,6934 1,4422 0,6934 1,0000 1,4422 0,4807 1,0000 1,0000 Table 4.14: Fuzzy weights matrix DM1 DM2 DM3 l m n l m n l m n w1 0,0670 0,1562 0,2902 0,0826 0,2211 0,4765 0,1533 0,3333 0,6634 w2 0,2668 0,6586 1,3689 0,2037 0,4600 1,1750 0,2211 0,3333 0,9568 w3 0,1082 0,1852 0,6036 0,1413 0,3189 0,6872 0,1063 0,3333 0,4600 59 Step 1.2: Calculation of rw values The defuzzification and normalization is calculated by using Eq.(3.5) as follows in Table 4.15: Table 4.15: The crisp values of DMs criteria weights DM1 Weights DM2 Weights DM3 Weights w1 (occurrence) 0,1386 0,2072 0,3230 w2 (severity) 0,6193 0,4882 0,4244 w3 (detectability) 0,2421 0,3046 0,2526 Step 2: Giving scores to decision factors O, S, and D After determining the decision factors’ weights, each DM gives scores to decision factors from 10-point scales. According to these scores and determined weights of decision factors, the application of GRA is performed individually for all failure modes and with each DMs individually. The score that are given by DMs is presented in Table 4.16. At the end of GRA methodology, we obtain RPN values for each of DMs. Finally, the weighted arithmetic means is calculated to obtain RPN1 values of failure modes. Table 4.16: FM scores given by three DMs No System/ Part Failure Mode DM1 DM2 DM3 O S D O S D O S D 1 Husks storage Too high stacking 3 2 3 7 5 3 4 3 4 2 Waste landfill area Fluorescent dust is breathed by workers 3 5 3 6 4 5 4 4 3 3 Blow room section Working in high places 3 5 3 6 6 2 4 5 3 4 Blow room section Dusty air condition in workplace 3 5 6 7 2 4 4 5 6 5 Manufacturing section It seems hard to reach fire extinguisher in emergency 6 6 3 6 7 3 7 7 3 6 Manufacturing section Non-ergonomic way of working 3 5 4 7 5 4 4 5 4 7 Manufacturing section Working without steel consolidated shoes 7 3 5 9 4 2 7 4 3 8 Manufacturing section The absence of fire extinguisher 6 7 4 5 8 3 6 8 4 9 Manufacturing section Working in dusty workplace 3 5 5 7 3 4 6 3 5 10 Manufacturing section Workers do not use earmuff 3 5 5 7 6 3 5 4 5 11 Welding workshop Workers use no vise while they use drill machine 6 7 5 5 9 6 6 8 5 12 Mess hall Workers in mess hall do not wear protective shoes 3 4 6 5 2 3 3 3 5 13 Packaging section Working without machine protector 6 5 5 4 7 5 6 5 5 14 Packaging section There is no automatic stopper mechanism and any sensor when the door is opened 3 5 6 6 7 6 5 6 6 15 Packaging section Inconvenient conversation for Solvent 5 7 5 8 8 6 7 8 6 16 Packaging section Manometer’s limits is not identified 3 6 5 2 6 8 3 6 6 60 17 Packaging section Workers enter the machine and then carry out materials from the inside 3 3 4 7 7 4 5 5 5 18 Mess hall Electric plug is very close to LPG connection cable 3 7 4 8 7 6 5 6 4 19 Operation-1 section The possibility of deactivation of safety cover of machine 1 3 7 3 7 3 2 4 2 20 Operation-1 section There are unnecessary materials inside the fire cabinet 3 3 4 8 2 5 4 3 4 21 Operation-1 section Opening the iron bale package by using iron scissors 3 3 5 4 3 4 3 3 4 22 Operation-1 section The workplace is not isolated against the risk of cylinder falling 3 2 5 6 7 5 3 3 5 23 Operation-1 section Workers put order the items one by one by using their naked hands. 3 6 3 7 7 5 3 6 3 24 Operation-1 section The machine door is throwing back too fast 3 3 6 6 7 6 3 5 4 25 Operation-2 section The possibility of fire in BOX machine 3 6 7 3 7 5 3 6 4 26 Operation-2 section The absence of middle-railing 3 4 4 6 2 2 3 3 3 27 Operation-1&2 section The safety wire of Bobbin machine is worn out 3 5 7 3 7 5 3 6 6 28 Administrative building Machine of drinking water is placed on the wet floor 3 6 5 8 5 2 4 5 3 29 Administrative building Workers monitor the PC screen for a long time period without break 6 5 4 9 3 3 7 3 3 30 Administrative building The possibility of existence of Legionella bacteria inside air conditioner water 3 6 6 2 7 9 2 6 8 31 Administrative building The emergency door is opening to the inside of workplace 3 6 5 5 6 5 4 6 5 32 Husks pressing section Workers put their naked hands inside the press machine 5 5 4 3 7 6 3 6 5 33 Husks pressing section The usage of falcate 3 3 4 8 4 4 6 3 4 34 Husks pressing section Working in high workplace without preventative measurement 4 6 3 6 8 3 4 7 3 35 Husks pressing section The absence of fire extinguisher 3 6 5 4 7 3 4 6 3 36 Items winding section Workers use their naked hands to put items into rotating parts of machine 4 5 4 6 7 4 4 5 4 37 Items winding section Too much stacking of items 6 2 3 7 4 3 6 3 3 38 Carder machine There some snacks, drinks etc. exist inside the machine 7 3 4 8 2 4 7 2 4 39 Carder machine Broken safety switches on the control panel 3 5 6 3 5 5 3 5 5 40 Carder machine Workers insert cotton into machine by their naked hands 6 5 6 7 6 4 6 5 6 41 Carder machine Workers clean the moving parts of machine with naked hand 4 5 4 6 7 4 4 4 4 42 Ring machine There is no dust absorber mechanism in the workplace 6 3 5 8 5 5 6 5 5 43 Ring machine Rotating parts of the machine have no protective cover 4 4 5 6 7 3 4 8 6 44 Bobbin machine Coiling up the reel too much forces the reel to fly off 3 2 5 5 6 5 4 4 4 45 Bale opener machine The possibility of deactivation of safety chains and safety sensor 2 6 7 2 7 7 2 6 6 46 Fuel tank (Across the Box storage) The tank is plastic and has no earthling 6 6 7 7 8 7 7 7 7 47 Diesel forklift (Warehouse) Using diesel forklift in closed workplace 5 5 4 6 6 3 5 5 4 48 Diesel forklift (Warehouse) Driving forklift uncontrollable and fast 3 6 7 6 7 3 4 5 6 49 Diesel forklift (Warehouse) Safety lock which is for forklift basket is broken 6 6 6 2 6 8 7 6 5 50 Laboratory The rotating parts of the machine has no preventative cover 3 3 5 4 7 3 4 4 3 61 Step 3: Incorporating the weights of decision factors (O, S, and D) to the RPN1 calculation and determining threshold At this point, GRA is used to incorporate the weights of decision factors and then to calculate RPN1 values. GRA is applied by DMs individually. To demonstrate the proposed methods, second decision maker’s (DM2) calculations are obtained as following steps: Step 3.1: Establishing the comparative series To reduce the potential risk, all decision factors should be as small as standard series. Step 3.2: Establishing the standard series X0(k) = [X0(1), X0(2), X0(3)] = [1,1,1] ∆01(1) ∆01(2) ∆01(3) 5 4 2 ∆02(1) ∆02(2) ∆02(3) 5 3 4 . . . . . . . . . . . . ∆49(1) ∆49(2) ∆49(3) 1 3 7 ∆50(1) ∆50(2) ∆50(3) 3 6 2 Step 3.3: Computing the grey relation coefficient To compute grey relation coefficient, O, S, and D are compared with the corresponding standard series by using Eq. (3.10). = 62 To be able to use Eq. (3.10) the minimum value of all ∆i (k) and the maximum value of all ∆i (k) should be obtained. ∆min = 1, ∆max =8 And the == 0,5 Step 3.4: Computing the DM1’s grey relation coefficients: γ01 γ01 γ01 0,556 0,625 0,833 γ02 γ02 γ02 0,556 0,714 0,625 . . . . . . . . . . . . γ49 γ49 γ49 1,000 0,714 0,455 γ50 γ50 γ01 0,714 0,500 0,833 Step 3.5: Computing the degree of relation At the final stage, decision factors’ (O, S, and D) weights are incorporated into calculation of degree of relation for all failure modes by using Eq. (3.11). τi = i th degree of relationship For all failure modes, 50 degree of relationship is calculated and then the relational series for each DM are presented in table 4.17. Table 4.17: Degree of grey relationship for FMs FM NO DM1’s relational series DM2’s relational series DM3’s relational series 1 0,800 1,000 0,800 0,500 0,625 0,833 0,692 0,818 0,692 2 0,800 0,571 0,800 0,556 0,714 0,625 0,692 0,692 0,818 3 0,800 0,571 0,800 0,556 0,556 1,000 0,692 0,600 0,818 4 0,800 0,571 0,500 0,500 1,000 0,714 0,692 0,600 0,529 5 0,500 0,500 0,800 0,556 0,500 0,833 0,474 0,474 0,818 6 0,800 0,571 0,667 0,500 0,625 0,714 0,692 0,600 0,692 7 0,444 0,800 0,571 0,417 0,714 1,000 0,474 0,692 0,818 8 0,500 0,444 0,667 0,625 0,455 0,833 0,529 0,429 0,692 9 0,800 0,571 0,571 0,500 0,833 0,714 0,529 0,818 0,600 10 0,800 0,571 0,571 0,500 0,556 0,833 0,600 0,692 0,600 11 0,500 0,444 0,571 0,625 0,417 0,556 0,529 0,429 0,600 12 0,800 0,667 0,500 0,625 1,000 0,833 0,818 0,818 0,600 13 0,500 0,571 0,571 0,714 0,500 0,625 0,529 0,600 0,600 14 0,800 0,571 0,500 0,556 0,500 0,556 0,600 0,529 0,529 15 0,571 0,444 0,571 0,455 0,455 0,556 0,474 0,429 0,529 16 0,800 0,500 0,571 1,000 0,556 0,455 0,818 0,529 0,529 17 0,800 0,800 0,667 0,500 0,500 0,714 0,600 0,600 0,600 = 63 18 0,800 0,444 0,667 0,455 0,500 0,556 0,600 0,529 0,692 19 1,333 0,800 0,444 0,833 0,500 0,833 1,000 0,692 1,000 20 0,800 0,800 0,667 0,455 1,000 0,625 0,692 0,818 0,692 21 0,800 0,800 0,571 0,714 0,833 0,714 0,818 0,818 0,692 22 0,800 1,000 0,571 0,556 0,500 0,625 0,818 0,818 0,600 23 0,800 0,500 0,800 0,500 0,500 0,625 0,818 0,529 0,818 24 0,800 0,800 0,500 0,556 0,500 0,556 0,818 0,600 0,692 25 0,800 0,500 0,444 0,833 0,500 0,625 0,818 0,529 0,692 26 0,800 0,667 0,667 0,556 1,000 1,000 0,818 0,818 0,818 27 0,800 0,571 0,444 0,833 0,500 0,625 0,818 0,529 0,529 28 0,800 0,500 0,571 0,455 0,625 1,000 0,692 0,600 0,818 29 0,500 0,571 0,667 0,417 0,833 0,833 0,474 0,818 0,818 30 0,800 0,500 0,500 1,000 0,500 0,417 1,000 0,529 0,429 31 0,800 0,500 0,571 0,625 0,556 0,625 0,692 0,529 0,600 32 0,571 0,571 0,667 0,625 0,556 0,625 0,818 0,529 0,600 33 0,800 0,800 0,667 0,455 0,714 0,714 0,529 0,818 0,692 34 0,667 0,500 0,800 0,556 0,455 0,833 0,692 0,474 0,818 35 0,800 0,500 0,571 0,714 0,500 0,833 0,692 0,529 0,818 36 0,667 0,571 0,667 0,556 0,500 0,714 0,692 0,600 0,692 37 0,500 1,000 0,800 0,500 0,714 0,833 0,529 0,818 0,818 38 0,444 0,800 0,667 0,455 1,000 0,714 0,474 1,000 0,692 39 0,800 0,571 0,500 0,833 0,625 0,625 0,818 0,600 0,600 40 0,500 0,571 0,500 0,500 0,556 0,714 0,529 0,600 0,529 41 0,667 0,571 0,667 0,556 0,500 0,714 0,692 0,692 0,692 42 0,500 0,800 0,571 0,455 0,625 0,625 0,529 0,600 0,600 43 0,667 0,667 0,571 0,556 0,500 0,833 0,692 0,429 0,529 44 0,800 1,000 0,571 0,625 0,556 0,625 0,692 0,692 0,692 45 1,000 0,500 0,444 1,000 0,500 0,500 1,000 0,529 0,529 46 0,500 0,500 0,444 0,500 0,455 0,500 0,474 0,474 0,474 47 0,571 0,571 0,667 0,556 0,556 0,833 0,600 0,600 0,692 48 0,800 0,500 0,444 0,556 0,500 0,833 0,692 0,600 0,529 49 0,500 0,500 0,500 1,000 0,556 0,455 0,474 0,529 0,600 50 0,800 0,800 0,571 0,714 0,500 0,833 0,692 0,692 0,818 Step 4: Calculation of final RPN1 values The RPN1 values of failure modes then obtained by arithmetic means of each decision makers RPN values. RPN values of each DMs and the final RPN2 values and prioritization are presented in Table 4.18. Table 4.18 Final RPN1 calculation FM NO RPN values of DM1 Weights of DM1 RPN values of DM2 Weights of DM2 RPN values of DM3 Weights of DM3 RPN1 Prioritizati on 1 0,924 0,4269 0,663 0,2442 0,746 0,3288 0,801 50 2 0,658 0,4269 0,654 0,2442 0,724 0,3288 0,679 37 3 0,658 0,4269 0,691 0,2442 0,685 0,3288 0,675 36 4 0,586 0,4269 0,809 0,2442 0,612 0,3288 0,649 31 5 0,573 0,4269 0,613 0,2442 0,561 0,3288 0,579 11 6 0,626 0,4269 0,626 0,2442 0,653 0,3288 0,635 29 7 0,695 0,4269 0,740 0,2442 0,653 0,3288 0,692 38 8 0,506 0,4269 0,605 0,2442 0,528 0,3288 0,537 4 9 0,603 0,4269 0,728 0,2442 0,670 0,3288 0,655 32 10 0,603 0,4269 0,629 0,2442 0,639 0,3288 0,621 24 11 0,483 0,4269 0,502 0,2442 0,504 0,3288 0,495 3 12 0,645 0,4269 0,872 0,2442 0,763 0,3288 0,739 41 13 0,562 0,4269 0,582 0,2442 0,577 0,3288 0,572 9 14 0,586 0,4269 0,528 0,2442 0,552 0,3288 0,561 7 15 0,493 0,4269 0,485 0,2442 0,469 0,3288 0,483 2 16 0,559 0,4269 0,617 0,2442 0,623 0,3288 0,594 15 17 0,768 0,4269 0,565 0,2442 0,600 0,3288 0,663 33 18 0,548 0,4269 0,508 0,2442 0,593 0,3288 0,553 6 64 19 0,788 0,4269 0,671 0,2442 0,869 0,3288 0,786 48 20 0,768 0,4269 0,773 0,2442 0,746 0,3288 0,762 45 21 0,745 0,4269 0,772 0,2442 0,786 0,3288 0,765 46 22 0,869 0,4269 0,550 0,2442 0,763 0,3288 0,756 44 23 0,614 0,4269 0,538 0,2442 0,696 0,3288 0,622 25 24 0,727 0,4269 0,528 0,2442 0,694 0,3288 0,668 35 25 0,528 0,4269 0,607 0,2442 0,664 0,3288 0,592 14 26 0,685 0,4269 0,908 0,2442 0,818 0,3288 0,783 47 27 0,572 0,4269 0,607 0,2442 0,623 0,3288 0,597 16 28 0,559 0,4269 0,704 0,2442 0,685 0,3288 0,636 30 29 0,585 0,4269 0,747 0,2442 0,707 0,3288 0,664 34 30 0,542 0,4269 0,578 0,2442 0,656 0,3288 0,588 13 31 0,559 0,4269 0,591 0,2442 0,600 0,3288 0,580 12 32 0,594 0,4269 0,586 0,2442 0,641 0,3288 0,607 19 33 0,768 0,4269 0,660 0,2442 0,693 0,3288 0,717 40 34 0,596 0,4269 0,591 0,2442 0,631 0,3288 0,606 18 35 0,559 0,4269 0,646 0,2442 0,655 0,3288 0,612 21 36 0,608 0,4269 0,577 0,2442 0,653 0,3288 0,615 22 37 0,882 0,4269 0,706 0,2442 0,725 0,3288 0,787 49 38 0,718 0,4269 0,800 0,2442 0,752 0,3288 0,749 43 39 0,586 0,4269 0,668 0,2442 0,670 0,3288 0,634 27 40 0,544 0,4269 0,592 0,2442 0,559 0,3288 0,561 8 41 0,608 0,4269 0,577 0,2442 0,692 0,3288 0,628 26 42 0,703 0,4269 0,590 0,2442 0,577 0,3288 0,634 28 43 0,644 0,4269 0,613 0,2442 0,539 0,3288 0,602 17 44 0,869 0,4269 0,591 0,2442 0,692 0,3288 0,743 42 45 0,556 0,4269 0,604 0,2442 0,681 0,3288 0,609 20 46 0,487 0,4269 0,478 0,2442 0,474 0,3288 0,480 1 47 0,594 0,4269 0,640 0,2442 0,623 0,3288 0,615 23 48 0,528 0,4269 0,613 0,2442 0,612 0,3288 0,576 10 49 0,500 0,4269 0,617 0,2442 0,529 0,3288 0,538 5 50 0,745 0,4269 0,646 0,2442 0,724 0,3288 0,714 39 Step 5: Determining the threshold intervals After establishing the RPN1 values, three OHS specialists determine the threshold interval values. They consider the factors such as budget, work-time, and labor force of company and decide to perform the corrective actions which have the RPN1 value smaller than or equal to 0,600 immediately. The failure modes that have a RPN1 value greater than the 0,600 should be performed with in a reasonable time. Table 4.20 shows the two parts separately by highlighting the part one with light grey color. Table 4.19: Performing C.A. according to the threshold intervals Threshold interval for corrective actions of failure modes Number of C.A. RPN1 ≤ 0,600 Corrective action should be performed immediately 16 RPN1 > 0,600 Corrective action should be performed according to the company's prevention policy within a reasonable time 34 65 Table 4.20: Prioritization and categorization of C.A.s according to RPN1 values FM No. Priority Corrective action No. RPN1 value 46 1 CA.1 0,480 15 2 CA.2 0,483 11 3 CA.3 0,495 8 4 CA.4 0,537 49 5 CA.5 0,538 18 6 CA.6 0,553 14 7 CA.7 0,561 40 8 CA.8 0,561 13 9 CA.9 0,572 48 10 CA.10 0,576 5 11 CA.11 0,579 31 12 CA.12 0,580 30 13 CA.13 0,588 25 14 CA.14 0,592 16 15 CA.15 0,594 27 16 CA.16 0,597 43 17 CA.17 0,602 34 18 CA.18 0,606 32 19 CA.19 0,607 45 20 CA.20 0,609 35 21 CA.21 0,612 36 22 CA.22 0,615 47 23 CA.23 0,615 10 24 CA.24 0,621 23 25 CA.25 0,622 41 26 CA.26 0,628 39 27 CA.27 0,634 42 28 CA.28 0,634 6 29 CA.29 0,635 28 30 CA.30 0,636 4 31 CA.31 0,649 9 32 CA.32 0,655 17 33 CA.33 0,663 29 34 CA.34 0,664 24 35 CA.35 0,668 3 36 CA.36 0,675 2 37 CA.37 0,679 7 38 CA.38 0,692 50 39 CA.39 0,714 33 40 CA.40 0,717 12 41 CA.41 0,739 44 42 CA.42 0,743 38 43 CA.43 0,749 22 44 CA.44 0,756 20 45 CA.45 0,762 21 46 CA.46 0,765 26 47 CA.47 0,783 19 48 CA.48 0,786 37 49 CA.49 0,787 1 50 CA.50 0,801 4.1.3 Determining to RPN2 values As mentioned in the methodology, DMs should calculate one more RPN value to calculate RPN3 value. This is RPN2. While RPN2 values are calculating, five decision factors (criteria) are weighted. Each decision makers makes judgments on decision 66 factors (criteria) and determine their weights separately with the help of fuzzy AHP. By multiplication of each corrective action’s scores with corresponding weights of criteria (decision factors) together with weights of decision makers and then the summation of all, we calculate the value of RPN2 for current corrective action. The steps of RPN2 determination for the case company are as follows: Step 1: Assigning different weights to criteria Pairwise comparison matrices are established by DMs and presented in Table 4.21-4.23. Table 4.21: A pairwise comparison matrices of RPN2’s criteria (DM1) DM1 cost time loss obligations prevention policy customer satisfaction and reputation cost (1.00,1.00,1.00) (5.00,7.00,9.00) (0.14,0.20,0.33) (0.14,0.20,0.33) (3.00,5.00,7.00) time loss (0.11,0.14,0.20) (1.00,1.00,1.00) (0.14,0.20,0.33) (0.14,0.20,0.33) (3.00,5.00,7.00) obligations (3.00,5.00,7.00) (3.00,5.00,7.00) (1.00,1.00,1.00) (1.00,3.00,5.00) (1.00,1.00,3.00) prevention policy (3.00,5.00,7.00) (3.00,5.00,7.00) (0.20,0.33,1.00) (1.00,1.00,1.00) (1.00,1.00,3.00) customer satisfaction and reputation (0.14,0.20,0.33) (0.14,0.20,0.33) (0.33,1.00,1.00) (0.33,1.00,1.00) (1.00,1.00,1.00) Table 4.22: A pairwise comparison matrices of RPN2’s criteria (DM2) DM2 cost time loss obligations prevention policy customer satisfaction and reputation cost (1.00,1.00,1.00) (1.00,3.00,5.00) (0.11,0.14,0.20) (0.14,0.20,0.33) (1.00,3.00,5.00) time loss (0.20,0.33,1.00) (1.00,1.00,1.00) (0.14,0.20,0.33) (0.14,0.20,0.33) (3.00,5.00,7.00) obligations (5.00,7.00,9.00) (3.00,5.00,7.00) (1.00,1.00,1.00) (1.00,1.00,3.00) (3.00,5.00,7.00) prevention policy (3.00,5.00,7.00) (3.00,5.00,7.00) (0.33,1.00,1.00) (1.00,1.00,1.00) (1.00,3.00,5.00) customer satisfaction and reputation (0.20,0.33,1.00) (0.14,0.20,0.33) (0.14,0.20,0.33) (0.20,0.33,1.00) (1.00,1.00,1.00) 67 Table 4.23: A pairwise comparison matrices of RPN2’s criteria (DM2) DM3 cost time loss obligations prevention policy customer satisfaction and reputation cost (1.00,1.00,1.00) (3.00,5.00,7.00) (1.00,1.00,3.00) (1.00,3.00,5.00) (5.00,7.00,9.00) time loss (0.14,0.20,0.33) (1.00,1.00,1.00) (0.20,0.33,1.00) (0.14,0.20,0.33) (0.14,0.20,0.33) obligations (0.33,1.00,1.00) (1.00,3.00,5.00) (1.00,1.00,1.00) (1.00,3.00,5.00) (1.00,3.00,5.00) prevention policy (0.20,0.33,1.00) (3.00,5.00,7.00) (0.20,0.33,1.00) (1.00,1.00,1.00) (1.00,1.00,3.00) customer satisfaction and reputation (0.11,0.14,0.20) (3.00,5.00,7.00) (0.20,0.33,1.00) (0.33,1.00,1.00) (1.00,1.00,1.00) Step 1.1: Calculation of ir and iw values According to Buckley’s Fuzzy AHP method the aggregate pairwise comparison matrix is the next step but we want to incorporate the DMs’ importance weights to the calculation. Thus, ri, wi and crisp weight values of criteria are calculated by each of the DMs individually. To demonstrate the proposed model, Table 4.24 is presents ri values, Table 4.25 presents wi values and Table 4.26 presents the crisp values of RPN2’s criteria weights. Table 4.24: Geometric mean values of fuzzy comparison values (RPN2’s criteria) DM1 DM2 DM3 l m n l m n l m n r1 0,7892 1,0696 1,4758 0,4366 0,7621 1,1076 1,7188 2,5365 3,9363 r2 0,3686 0,4911 0,6893 0,4146 0,5818 0,9510 0,2255 0,3056 0,5173 r3 1,5518 2,3714 3,7433 2,1411 2,8094 4,2103 0,8027 1,9332 2,6265 r4 1,1247 1,5281 2,7131 1,2457 2,3714 3,0049 0,6544 0,8891 1,8384 r5 0,2959 0,5253 0,6444 0,2412 0,3385 0,6444 0,4670 0,7505 1,0696 Table 4.25 Fuzzy weights matrix of RPN2’s criteria DM1 DM2 DM3 l m n l m n l m n w1 0,0852 0,1787 0,3573 0,0440 0,1110 0,2473 0,1721 0,3954 1,0175 w2 0,0398 0,0820 0,1669 0,0418 0,0848 0,2123 0,0226 0,0476 0,1337 w3 0,1675 0,3962 0,9063 0,2159 0,4093 0,9400 0,0804 0,3014 0,6790 w4 0,1214 0,2553 0,6569 0,1256 0,3455 0,6709 0,0655 0,1386 0,4752 w5 0,0319 0,0878 0,1560 0,0243 0,0493 0,1439 0,0468 0,1170 0,2765 68 Step 1.2: Calculation of rw values Table 4.26: The crisp values of RPN2’s criteria weights DM1 DM2 DM3 w1 0,1684 0,1098 0,3993 w2 0,0783 0,0924 0,0514 w3 0,3985 0,4270 0,2672 w4 0,2802 0,3115 0,1712 w5 0,0747 0,0593 0,1109 Step 2: Giving scores to decision factors O, S, and D Now, DMs give scores to criteria for each corrective action. Table 4.27 presents the first three corrective actions’ criteria score table. Table 4.27: Criteria score table of RPN2 criteria CA.1 Cost Time Loss Obligations Prevention Policy Customer Satisfaction and Reputation DM1 0,75 0,50 0,50 0,50 0,75 DM2 0,50 0,75 0,50 0,50 0,75 DM3 0,75 0,50 0,25 0,50 0,75 CA.2 Cost Time Loss Obligations Prevention Policy Customer Satisfaction and Reputation DM1 0,00 0,00 0,25 0,25 0,50 DM2 0,00 0,00 0,50 0,00 0,25 DM3 0,25 0,00 0,25 0,25 0,25 CA.3 Cost Time Loss Obligations Prevention Policy Customer Satisfaction and Reputation DM1 0,25 0,25 0,00 0,25 0,25 DM2 0,25 0,50 0,00 0,25 0,25 DM3 0,25 0,00 0,25 0,00 0,50 69 Step 3: Calculation of final RPN2 values And finally the RPN2 values of corrective actions are obtained as shown in the Table 4.28. The table consists of the first corrective action RPN2 calculation. A Full list of RPN2 values of all failure modes and so the corrective actions are presented in Appendix A. Table 4.28: RPN2 values CA.1 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.1 DM1 (0,4269) Cost 0,75 0,1684 0,5608 0,5551 Time Loss 0,50 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,50 0,2802 Customer Satisfaction and Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,5380 Time Loss 0,75 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,50 0,3115 Customer Satisfaction and Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,75 0,3993 0,5608 Time Loss 0,50 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,50 0,1712 Customer Satisfaction and Reputation 0,75 0,1109 70 4.1.4 Determining to RPN3 values RPN3 values are the numbers that help us to reprioritize the selected corrective actions. After calculating RPN3 values FMEA team determine responsible person for each corrective action completion and follow the activity. We have already calculated the RPN1 and RPN2 values. It is simple to calculate the RPN3. The eq. (4.1) shows the calculation of RPN3. 21 213 RPNRPNRPN where, 1 = 0,95 (coefficient of RPN1) 2 = 0,05 (coefficient of RPN2) Table 4.29 presents the RPN3 values of CA.1. We see that the priority of CA.1 has changed to 2. A Full list is presented in the conclusion section. Table 4.29: RPN3 value calculation of CA.1 CA. NO RPN1 value 1 RPN2 value 2 RPN3 value Reprioritization No CA.1 0,480 0,95 0,555 0,05 0,4838 3 Company should perform the corrective actions according to new prioritization that is done according to RPN3 values. After performing corrective actions a new RPN1 calculation should be done. The calculation results of new RPN1 values can be written in the same FMEA report paper. As to the last RPN1 values, re-evaluation is done and then the FMEA team decides on the necessity of corrective action. Decision is given by the evaluation of FMEA team. It is enough for the failure modes which were in the first part that failure modes have RPN values greater than 0,6. The team can determine additional corrective actions or wait for the next FMEA application. To the time that new FMEA is applied by the team, temporary precautions should be taken for these corrective actions. (4.1) 71 5. RESULTS AND CONCLUSION Risk analysis and risk management are the fundamental issues for today’s company. In recent years, the concept of occupational health and safety is growing rapidly in developed or developing countries. All workplaces which are aiming to meet the requirements of OHS should apply risk analysis methodology periodically. In this research, FMEA is chosen. Although FMEA is a great method to prevent and take corrective action to the failure modes before they occur, it has several numbers of drawbacks. The main source of these drawbacks is about the calculation of RPN. We used Fuzzy AHP and GRA methodology to eliminate drawbacks. Moreover, FMEA team members (DMs) are weighted by experienced experts to get more consistent and more realistic results. Almost all criticisms for traditional RPN calculation are eliminated but there is still some evaluation and calculation to do before performing the corrective actions. First, DMs determine the threshold value, because it is just a dream that a company get rid of from all failure modes in a short time period. Thus, DMs determine threshold interval. The value of 0,6000 is the number that separates two parts. In the first part, corrective actions have to be performed as soon as possible. In second part, corrective actions have to be performed again too, but not as short as the first part. OHS specialist determines reasonable time for each corrective action that is in the second part. Table 5.2 presents the 16 corrective actions in the first part and 34 in the second part as to RPN1 values according to RPN1 values. It is important to take the corrective actions according to the real need of company and for the workers. Five criteria are identified to consider this issue. By means of these criteria we obtained the RPN2 values. By multiplying RPN1 and RPN2 with their coefficients, we get RPN3 values. At this point, DMs determine new threshold intervals and they reprioritized the CA.s according to RPN3 values. All of the calculation results and the two prioritizations are presented in the Table 5.1. As seen in Table 5.2 the company performs 23 corrective actions in the first part and 27 in the second part according to RPN3 values. Highlighted with light grey color implies the part one 72 corrective actions and highlighted with the dark grey color implies the second part corrective actions. Table 5.1: RPN1, RPN2, RPN3 values and two prioritizations Failure Mode NO CA. NO RPN1 value Priority NO CA. NO RPN2 value RPN3 value Reprioriti zation NO 1 CA.50 0,8014 50 CA.1 0,5552 0,4838 3 2 CA.37 0,6789 37 CA.2 0,2221 0,4700 1 3 CA.36 0,6750 36 CA.3 0,1778 0,4791 2 4 CA.31 0,6490 31 CA.4 0,2913 0,5247 4 5 CA.11 0,5785 11 CA.5 0,5237 0,5373 6 6 CA.29 0,6350 29 CA.6 0,0787 0,5293 5 7 CA.38 0,6923 38 CA.7 0,7352 0,5697 11 8 CA.4 0,5373 4 CA.8 0,5652 0,5612 9 9 CA.32 0,6555 32 CA.9 0,4807 0,5674 10 10 CA.24 0,6211 24 CA.10 0,1203 0,5532 7 11 CA.3 0,4946 3 CA.11 0,4265 0,5714 13 12 CA.41 0,7390 41 CA.12 0,1213 0,5571 8 13 CA.9 0,5717 9 CA.13 0,9139 0,6043 25 14 CA.7 0,5607 7 CA.14 0,4776 0,5863 16 15 CA.2 0,4830 2 CA.15 0,1204 0,5703 12 16 CA.15 0,5940 15 CA.16 0,6574 0,6000 23 17 CA.33 0,6631 33 CA.17 0,6952 0,6029 24 18 CA.6 0,5528 6 CA.18 0,3447 0,5863 15 19 CA.48 0,7860 48 CA.19 0,6542 0,5974 18 20 CA.45 0,7616 45 CA.20 0,2814 0,5926 17 21 CA.46 0,7651 46 CA.21 0,0394 0,5834 14 22 CA.44 0,7559 44 CA.22 0,2945 0,5990 21 23 CA.25 0,6223 25 CA.23 0,5729 0,5982 20 24 CA.35 0,6677 35 CA.24 0,4290 0,5980 19 25 CA.14 0,5920 14 CA.25 0,5476 0,6183 26 26 CA.47 0,7832 47 CA.26 0,0538 0,5993 22 27 CA.16 0,5973 16 CA.27 0,3369 0,6191 27 28 CA.30 0,6357 30 CA.28 0,5666 0,6306 29 29 CA.34 0,6644 34 CA.29 0,5175 0,6291 28 30 CA.13 0,5881 13 CA.30 0,6177 0,6351 31 31 CA.12 0,5802 12 CA.31 0,3665 0,6349 30 32 CA.19 0,6072 19 CA.32 0,3077 0,6376 34 33 CA.40 0,7169 40 CA.33 0,1320 0,6365 33 34 CA.18 0,6066 18 CA.34 0,1352 0,6376 35 35 CA.21 0,6117 21 CA.35 0,0333 0,6363 32 36 CA.22 0,6150 22 CA.36 0,1033 0,6464 36 37 CA.49 0,7874 49 CA.37 0,5099 0,6705 37 38 CA.43 0,7494 43 CA.38 0,4811 0,6815 38 39 CA.27 0,6337 27 CA.39 0,6105 0,7088 40 40 CA.8 0,5609 8 CA.40 0,2007 0,6912 39 41 CA.26 0,6279 26 CA.41 0,1979 0,7119 41 42 CA.28 0,6339 28 CA.42 0,2855 0,7201 42 43 CA.17 0,6027 17 CA.43 0,4148 0,7323 44 44 CA.42 0,7427 42 CA.44 0,0907 0,7227 43 45 CA.20 0,6087 20 CA.45 0,3296 0,7404 45 46 CA.1 0,4801 1 CA.46 0,6150 0,7575 48 47 CA.23 0,6151 23 CA.47 0,1291 0,7503 46 48 CA.10 0,5764 10 CA.48 0,3489 0,7641 50 49 CA.5 0,5381 5 CA.49 0,1566 0,7555 47 50 CA.39 0,7137 39 CA.50 0,0243 0,7622 49 73 Table 5.2: CA.s’ priority changes according to the proposed methodology FM No. Corrective action No. RPN1 value RPN3 Reprioritization of CA.s 46 CA.1 0,480 0,4838 3 15 CA.2 0,483 0,4700 1 11 CA.3 0,495 0,4791 2 8 CA.4 0,537 0,5247 4 49 CA.5 0,538 0,5373 6 18 CA.6 0,553 0,5293 5 14 CA.7 0,561 0,5697 11 40 CA.8 0,561 0,5612 9 13 CA.9 0,572 0,5674 10 48 CA.10 0,576 0,5532 7 5 CA.11 0,579 0,5714 13 31 CA.12 0,580 0,5571 8 30 CA.13 0,588 0,6043 25 25 CA.14 0,592 0,5863 16 16 CA.15 0,594 0,5703 12 27 CA.16 0,597 0,6000 23 43 CA.17 0,602 0,6029 24 34 CA.18 0,606 0,5863 15 32 CA.19 0,607 0,5974 18 45 CA.20 0,609 0,5926 17 35 CA.21 0,612 0,5834 14 36 CA.22 0,615 0,5990 21 47 CA.23 0,615 0,5982 20 10 CA.24 0,621 0,5980 19 23 CA.25 0,622 0,6183 26 41 CA.26 0,628 0,5993 22 39 CA.27 0,634 0,6191 27 42 CA.28 0,634 0,6306 29 6 CA.29 0,635 0,6291 28 28 CA.30 0,636 0,6351 31 4 CA.31 0,649 0,6349 30 9 CA.32 0,655 0,6376 34 17 CA.33 0,663 0,6365 33 29 CA.34 0,664 0,6376 35 24 CA.35 0,668 0,6363 32 3 CA.36 0,675 0,6464 36 2 CA.37 0,679 0,6705 37 7 CA.38 0,692 0,6815 38 50 CA.39 0,714 0,7088 40 33 CA.40 0,717 0,6912 39 12 CA.41 0,739 0,7119 41 44 CA.42 0,743 0,7201 42 38 CA.43 0,749 0,7323 44 22 CA.44 0,756 0,7227 43 20 CA.45 0,762 0,7404 45 21 CA.46 0,765 0,7575 48 26 CA.47 0,783 0,7503 46 19 CA.48 0,786 0,7641 50 37 CA.49 0,787 0,7555 47 1 CA.50 0,801 0,7622 49 74 After all, the summarized steps of the proposed FMEA methodology are presented below: i. The FMEA team formation. ii. Determination of the importance weights for DMs (FMEA team members) by experts. iii. Determination of the importance weights for decision factors (O, S, and D) by DMs. iv. Giving scores to decision factors by DMs. v. Application of GRA by each of the DMs to obtain RPN1 values. vi. Prioritization of failure modes according to RPN1 values. vii. Determination of threshold intervals by DMs. viii. Determination of the importance weights for criteria of RPN2. ix. Giving scores to criteria from 5-point criteria scales. x. Weighted arithmetic mean calculation to obtain RPN2 values. xi. Calculation of RPN3 and reprioritization. xii. Determining second threshold value. xiii. Performing the determined corrective actions. xiv. Follow the activities and the responsible person. xv. Re-evaluate the failure modes and re-calculate the RPN3 values. xvi. FMEA team decides on the necessity of corrective action for adjusted failure modes. In the light of this study, it can be said that, the proposed FMEA methodology is useful for company. We considered company’s real needs. Furthermore, the limitations of company are also evaluated. The proposed FMEA method has the following advantages: i. In OHS concept, workers and customers safety and health are two crucial manners that the risk analysts have to give importance to. For this reason, decision factors of RPN1 calculation (O, S, and D) are given weights so that the severity gets the highest priority against other two. ii. Although the traditional FMEA has been proven as one of the most important early preventative actions in system, design, process or service, it 75 does not consider the DMs personnel data that affect their decision consistency. In this research, Buckley’s fuzzy AHP method is used to determine weights for DMs. Thanks to this, the proposed FMEA considers the experience, education level, knowledge level on risk analyzing, his/her job relevance with the current FMEA application workplace and the repetition number of risk analysis application (FMEA or another risk analysis methodology) that he/she has taken part in. iii. RPN1 is calculated by using decision factors O, S, and D. Instead of multiplication of these three, GRA is used to calculate RPN1. Thanks to GRA, high duplication problem of traditional RPN calculation is eliminated. iv. In traditional FMEA, corrective actions are performed according to on RPN values prioritization. In this research, corrective actions are performed not only based on RPN1 values but also RPN2 values. RPN2 derives from the need for reprioritization of corrective actions. RPN2 considers five additional criteria that are affecting the corrective actions rationality. These criteria are about cost, time loss, regulations, firm’s prevention policy, reputation of the firm. By adding RPN1 and RPN2 to the final RPN (RPN3) calculation, reprioritization is done based on RPN3 values. v. In this research fuzzy AHP is used for three times and GRA is used for one time to apply the proposed FMEA methodology. It seems really time consuming but it is not true in reality. Fuzzy AHP method firstly used for determining the DMs’ (FMEA team) criteria weights. Indeed, this step is not directly about the proposed FMEA application because DMs are given weights for only one time and the FMEA team can apply more FMEA application by using these weights. Secondly, fuzzy AHP is used for determining the RPN1’s decision factors O, S, and D weights. This step is also done for only one time and then calculated weights are used in other FMEA applications. Thirdly, fuzzy AHP is used for determining the RPN2’s criteria weights and this step is similar to first and the second step. While RPN1 values are calculating, GRA is used. By using GRA weights of O, S, and D are incorporated to the calculation. Besides, these methods are 76 applied in only the first application of FMEA, previously prepared criteria charts help us to do the calculations faster. Moreover, Excel templates are used for all fuzzy AHP, GRA, RPN1, RPN2, RPN3 calculations. Previously prepared templates save really enough time. vi. In contrast to traditional FMEA, proposed FMEA method gives more consistent results, considers decision makers background, evaluates corrective action’s rationality. The proposed model has also real-time applicability and it saves money, time and reputation of company. Future research can be focus on the RPN2 calculations and on its’ criteria. These criteria may change from country from country, sector to sector and company to company. Therefore, more reasonable criteria help DMs to apply FMEA more consistently. Furthermore, RPN2 calculation is partially time consuming and it gives less consistent results when it is compared with GRA. New calculation method can be developed for better results. Another point that can be developed by researchers is that, an automatic program or database can be developed for faster applicability of proposed model. Lastly, customer’s points of view are not considered during the proposed FMEA methodology. Decision factors’ scores are determined only with respect to organization’s point of view, not according to customers. This issue can be eliminated in the future research. 77 REFERENCES Books Covello, V.T. & Merkhoher, M.W., 1993. Risk Assessment Methods: Approaches for Assessing Health and Environmental Risks. New York. Plenum Press. Diberardinis, L., 1998, Handbook of occupational safety and health. Hoboken: Willey- Interscience. Özkılıç, Ö., 2005. İş sağlığı ve güvenliği, yönetim sistemleri ve risk değerlendirme metodolojileri. Ankara: Ajans-Türk Basın ve Basım A.ġ. Saaty, T. L., 1980. The analytic hierarchy process. New York: McGraw-Hill. 78 Periodicals Abdelgawad, M., Fayek, A.R. & Asce, A.M., 2010. Risk management in the construction industry using combined fuzzy FMEA and fuzzy AHP. Journal of Construction Engineering and Management.136, pp.1028–1037. Ale, B.J., 2002. Risk assessment practices in The Netherlands. Safety Science, 40(1-4), pp.105–126. Aven, T., 2010. On how to define, understand and describe risk. Reliability Engineering & System Safety, 95(6), pp.623–631. Barends, D.M. et al., 2012. Risk analysis of analytical validations by probabilistic modification of FMEA. Journal of Pharmaceutical and Biomedical Analysis, 64-65, pp.82–6. Buckley, J. J., 1985. Fuzzy hierarchical analysis. Fuzzy Sets and Systems, 17, pp 233- 247. Braglia, M., 2000. MAFMA: Multi-attribute failure mode analysis. International Journal of Quality & Reliablity Management, 17, 1017-1033. Chang, K.-H., Chang, Y.-C. & Tsai, I.-T., 2013. Enhancing FMEA assessment by integrating grey relational analysis and the decision making trial and evaluation laboratory approach. Engineering Failure Analysis, 31, pp.211–224. Chang, C.-L., Wei, C.-C., Lee, Y.-H., 1998. Failure mode and effects analysis using fuzzy method and grey theory. MCB University Press, 9, pp. 1072-1080. Chin, K. et al., 2009a. Computers & Operations Research Failure mode and effects analysis using a group-based evidential reasoning approach. , 36, pp.1768–1779. Chin, K. et al., 2009b. Failure mode and effects analysis by data envelopment analysis. Decision Support Systems, 48(1), pp.246–256. Cody, R.J., 2006. Anticipating risk for human subjects participating in clinical research: application of Failure Mode and Effects Analysis. Cancer investigation, 24(2), pp.209–14. Cohen, A.V., 1984. Overview and definition of risk: Its assessment and management. Environment International, 10(5-6), pp.359–366. 79 Deng, J., 1989. Introduction to grey system theory. Journal of Grey System, Vol.1,No. 1, pp. 1-24. Deng, J., 1982. Control problems of grey systems. Systems and Control Letter, Vol.1,No. 5, pp. 288-94. Feili, H.R. et al., 2013. Risk analysis of geothermal power plants using Failure Modes and Effects Analysis (FMEA) technique. Energy Conversion and Management, 72, pp.69–76. Ferdous, R. et al., 2009. Handling data uncertainties in event tree analysis. Process Safety and Environmental Protection, 87(5), pp.283–292. Garcia, P. a. a., Schirru, R. & Frutuoso E. Melo, P.F., 2005. A fuzzy data envelopment analysis approach for FMEA. Progress in Nuclear Energy, 46(3-4), pp.359–373. Geminiani, F.L., Smallwood, J.J. & Fee, S., 2013. A comparative analysis between contractors’ and inspectors’ perceptions of the department of labour occupational health and safety inspectorate relative to South African construction. Safety Science, 53, pp.186–192. Goldstein, S.M. et al., 2002. The service concept: the missing link in service design research? Journal of Operations Management, 20(2), pp.121–134. Helvacioglu, S. & Ozen, E., 2014. Fuzzy based failure modes and effect analysis for yacht system design. Ocean Engineering, 79, pp.131–141 Kahraman, C., Kaya, Ġ. & ġenvar, Ö., 2013. Healthcare Failure Mode and Effects Analysis Under Fuzziness. Human and Ecological Risk Assessment: An International Journal, 19(2), pp.538–552. Kahraman, C. & Çebı˙, S., 2009. A new multi-attribute decision making method: Hierarchical fuzzy axiomatic design. Expert Systems with Applications, 36(3), pp.4848–4861. Kahraman, C., Cebeci, U. & Ulukan, Z., 2003. Multi-criteria supplier selection using fuzzy AHP. Logistics Information Management, 16(6), pp.382–394. 80 Kutlu, A.C. & Ekmekçioğlu, M., 2012. Fuzzy failure modes and effects analysis by using fuzzy TOPSIS-based fuzzy AHP. Expert Systems with Applications, 39(1), pp.61–67. Lee, S.K., Mogi, G. & Kim, J.W., 2008. The competitiveness of Korea as a developer of hydrogen energy technology: The AHP approach. Energy Policy, 36(4), pp.1284– 1291. Lirer, L., Petrosino, P. & Alberico, I., 2001. Hazard assessment at volcanic fields: the Campi Flegrei case history. Journal of Volcanology and Geothermal Research, 112(1-4), pp.53–73. Liu, H.-C., You, J.-X., Fan, X.-J., & Lin, Q.-L. (2014). Failure mode and effects analysis using D numbers and grey relational projection method. Expert Systems with Applications, 41(10), 4670–4679. Liu, H.-C., Liu, L. & Liu, N., 2013. Risk evaluation approaches in failure mode and effects analysis: A literature review. Expert Systems with Applications, 40(2), pp.828–838. Liu, H.-C. et al., 2012. Risk evaluation in failure mode and effects analysis with extended VIKOR method under fuzzy environment. Expert Systems with Applications, 39(17), pp.12926–12934. Liu, H.-C. et al., 2011. Failure mode and effects analysis using fuzzy evidential reasoning approach and grey theory. Expert Systems with Applications, 38(4), pp.4403–4415. Mandal, S., & Maiti, J. (2014). Risk analysis using FMEA: Fuzzy similarity value and possibility theory based approach. Expert Systems with Applications, 41(7), pp. 3527–3537. Nieto-Morote, A. & Ruz-Vila, F., 2011. A fuzzy approach to construction project risk assessment. International Journal of Project Management, 29(2), pp.220–231. Nouri.Gharahasanlou, A. et al., 2014. Fault tree analysis of failure cause of crushing plant and mixing bed hall at Khoy cement factory in Iran. Case Studies in Engineering Failure Analysis, 2(1), pp.33–38. 81 Opricovic, S & Tzeng, G. H. 2004. Comprimise solution by MCDM methods: A comparative analysis of VIKOR and TOPSIS. European Journal of Operational Research, 156(2), pp. 445-455. Oztaysi, B., 2014. A decision model for information technology selection using AHP integrated TOPSIS-Grey: The case of content management systems. Knowledge- Based Systems. Online publication date: 1-Mar-2014. Öztürk, ġ. & Sofyalıoğlu, Ç. 2012. Application of grey relational analysis with fuzzy ahp to fmea method. Journal of Doğuş University, 13(1), pp.114–130. Pillay, A. & Wang, J., 2003. Modified failure mode and effects analysis using approximate reasoning. Reliability Engineering & System Safety, 79(1), pp.69–85. Saaty, T. L., 1986. Exploring optimization through hierarchies and ratio scales. Socio- Economic Planning Sciences, 20(6), pp. 355-360. Sant’Anna, A.P., 2012. Probabilistic priority numbers for failure modes and effects analysis. International Journal of Quality and Reliability Management, 29(3), pp.349–362. Schneider, H. & Stamatis, D.H., 1996. Failure Mode and Effect Analysis: FMEA from Theory to Execution. Technometrics, 38(1), p.80. Seçme, N.Y., Bayrakdaroğlu, A. & Kahraman, C., 2009. Fuzzy performance evaluation in Turkish Banking Sector using Analytic Hierarchy Process and TOPSIS. Expert Systems with Applications, 36(9), pp.11699–11709. Su, C.-T., Lin, H.-C., Teng, P.-W., & Yang, T. (2014). Improving the reliability of electronic paper display using FMEA and Taguchi methods: A case study. Microelectronics Reliability, 32(1), pp 324-368. Wang, Y.-M. et al., 2009. Risk evaluation in failure mode and effects analysis using fuzzy weighted geometric mean. Expert Systems with Applications, 36(2), pp.1195–1207. Xu, K. et al., 2002. Fuzzy assessment of FMEA for engine systems. Reliability Engineering & System Safety, 75(1), pp.17–29. 82 Yang, J. et al., 2011. Risk evaluation in failure mode and effects analysis of aircraft turbine rotor blades using Dempster–Shafer evidence theory under uncertainty. Engineering Failure Analysis, 18(8), pp.2084–2092. 83 Other Publications Abdelrahman, M. & Abdelgawad, M., 2011. Hybrid Decision Support System for Risk Criticality Assessment and Risk Analysis. Thesis for Ph. D. Alberta: University of Alberta: Department of Civil and Environmental Engineering. Çakır, E., 2009. Logistics Outsourcing and Selection of Third Party Logistics Service Provider (3PL) via Fuzzy AHP. Thesis for the M. A. Degree. Ġstanbu: BahçeĢehir Üniversitesi FBE. Goel, A. & Graves, R.J., 2007. Using Failure Mode Effect Analysis to Increase Electronic Systems Reliability. In 2007 30th International Spring Seminar on Electronics Technology (ISSE). IEEE, pp. 128–133. Karabulut, Ö., 2012. Türkiye’de iĢ sağlığı-güvenliği ve iĢ sağlığı ve kanunu tasarı taslağında geliĢmeler, Türkiye İşçi Sendikaları Konfederasyonu. https://www.google.com.tr/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad= rja&uact=8&ved=0CCsQFjAA&url=http%3A%2F%2Fwww.harb- is.org.tr%2Fpublic%2FUserFiles%2FOK.ppt&ei=TqZrU9vBOOah0QWKlIDIDA &usg=AFQjCNGAr03_le5Ju0zzIzq_lqRtbpJoGQ&sig2=ynE3IzDMZnXtdoyqq_y BiA. [Accessed April 14, 2014]. Reichheld, F.F., 1996. Learning from customer defections. The customer you lose hold the information you need to succeed. http://hbr.org/1996/03/learning-from- customer-defections/ar/1 [Accessed March 27, 2014]. 84 APPENDICES 85 APPENDIX A1: RPN2 Calculation of CA.2 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.2 (FM.15) DM1 (0,4269) Cost 0,00 0,1684 0,2070 0,1152 Time Loss 0,00 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,2283 Time Loss 0,00 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,2372 Time Loss 0,00 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,25 0,1109 86 APPENDIX A2: RPN2 Calculation of CA.3 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.3 (FM.11) DM1 (0,4269) Cost 0,25 0,1684 0,1504 0,1778 Time Loss 0,25 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,25 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,1664 Time Loss 0,50 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,2221 Time Loss 0,00 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,00 0,1712 Reputation 0,50 0,1109 87 APPENDIX A3: RPN2 Calculation of CA.4 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.4 (FM.8) DM1 (0,4269) Cost 0,50 0,1684 0,2112 0,2913 Time Loss 0,25 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,3154 Time Loss 0,50 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,3776 Time Loss 0,25 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,50 0,1109 88 APPENDIX A4: RPN2 Calculation of CA.5 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.5 (FM.49) DM1 (0,4269) Cost 0,75 0,1684 0,5413 0,5237 Time Loss 0,25 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,75 0,1098 0,3577 Time Loss 0,50 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,6245 Time Loss 0,25 0,0514 Obligations 0,75 0,2672 Prevention Policy 0,75 0,1712 Reputation 0,75 0,1109 89 APPENDIX A5: RPN2 Calculation of CA.6 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.6 (FM.18) DM1 (0,4269) Cost 0,0 0,1684 0,0701 0,0787 Time Loss 0,00 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,00 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,0148 Time Loss 0,00 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,1373 Time Loss 0,00 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,25 0,1109 90 APPENDIX A6: RPN2 Calculation of CA.7 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.7 (FM.14) DM1 (0,4269) Cost 0,75 0,1684 0,7688 0,7352 Time Loss 0,75 0,0783 Obligations 0,75 0,3985 Prevention Policy 0,75 0,2802 Reputation 1,00 0,0747 DM2 (0,2442) Cost 1,00 0,1098 0,7144 Time Loss 0,75 0,0924 Obligations 0,75 0,4270 Prevention Policy 0,50 0,3115 Reputation 1,00 0,0593 DM3 (0,3288) Cost 0,75 0,3993 0,7072 Time Loss 0,75 0,0514 Obligations 0,75 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,75 0,1109 91 APPENDIX A7: RPN2 Calculation of CA.8 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.8 (40) DM1 (0,4269) Cost 0,50 0,1684 0,4478 0,5652 Time Loss 0,25 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,25 0,2802 Reputation 1,00 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,6364 Time Loss 0,50 0,0924 Obligations 0,75 0,4270 Prevention Policy 0,50 0,3115 Reputation 1,00 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,6651 Time Loss 0,50 0,0514 Obligations 0,75 0,2672 Prevention Policy 0,75 0,1712 Reputation 1,00 0,1109 92 APPENDIX A8: RPN2 Calculation of CA.9 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.9 (FM13) DM1 (0,4269) Cost 0,75 0,1684 0,5608 0,4807 Time Loss 0,50 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,3302 Time Loss 0,50 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,4887 Time Loss 0,50 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,50 0,1712 Reputation 1,00 0,1109 93 APPENDIX A9: RPN2 Calculation of CA.10 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.10 (FM.48) DM1 (0,4269) Cost 0,25 0,1684 0,0990 0,1203 Time Loss 0,25 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,1721 Time Loss 0,25 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,1096 Time Loss 0,00 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,00 0,1109 94 APPENDIX A10: RPN2 Calculation of CA.11 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.11 (FM.5) DM1 (0,4269) Cost 0,00 0,1684 0,2967 0,4265 Time Loss 0,25 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,5941 Time Loss 0,50 0,0924 Obligations 0,75 0,4270 Prevention Policy 0,50 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,4707 Time Loss 0,50 0,0514 Obligations 0,50 0,2672 Prevention Policy 0,75 0,1712 Reputation 0,75 0,1109 95 APPENDIX A11: RPN2 Calculation of CA.12 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.12 (FM.31) DM1 (0,4269) Cost 0,25 0,1684 0,0617 0,1213 Time Loss 0,25 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,00 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,0654 Time Loss 0,25 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,2402 Time Loss 0,25 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,00 0,1712 Reputation 0,25 0,1109 96 APPENDIX A12: RPN2 Calculation of CA.13 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.13 (FM.30) DM1 (0,4269) Cost 0,75 0,1684 0,9384 0,9139 Time Loss 0,75 0,0783 Obligations 1,00 0,3985 Prevention Policy 1,00 0,2802 Reputation 1,00 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,9072 Time Loss 0,75 0,0924 Obligations 1,00 0,4270 Prevention Policy 1,00 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,75 0,3993 0,8873 Time Loss 0,75 0,0514 Obligations 1,00 0,2672 Prevention Policy 1,00 0,1712 Reputation 1,00 0,1109 97 APPENDIX A13: RPN2 Calculation of CA.14 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.14 (FM.25) DM1 (0,4269) Cost 0,50 0,1684 0,4004 0,4776 Time Loss 0,50 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,75 0,1098 0,5654 Time Loss 0,75 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,50 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,5129 Time Loss 0,75 0,0514 Obligations 0,50 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,50 0,1109 98 APPENDIX A14: RPN2 Calculation of CA.15 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.15 (FM.16) DM1 (0,4269) Cost 0,25 0,1684 0,1495 0,1204 Time Loss 0,00 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,0297 Time Loss 0,00 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,1502 Time Loss 0,25 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,25 0,1109 99 APPENDIX A15: RPN2 Calculation of CA.16 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.16 (FM.27) DM1 (0,4269) Cost 0,50 0,1684 0,6884 0,6574 Time Loss 0,50 0,0783 Obligations 0,75 0,3985 Prevention Policy 0,75 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,6306 Time Loss 0,75 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,75 0,3115 Reputation 1,00 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,6373 Time Loss 0,50 0,0514 Obligations 0,75 0,2672 Prevention Policy 0,75 0,1712 Reputation 0,75 0,1109 100 APPENDIX A16: RPN2 Calculation of CA.17 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.17 (FM.43) DM1 (0,4269) Cost 0,75 0,1684 0,7688 0,6952 Time Loss 0,75 0,0783 Obligations 0,75 0,3985 Prevention Policy 0,75 0,2802 Reputation 1,00 0,0747 DM2 (0,2442) Cost 0,75 0,1098 0,5506 Time Loss 0,75 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,50 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,75 0,3993 0,7072 Time Loss 0,75 0,0514 Obligations 0,75 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,75 0,1109 101 APPENDIX A17: RPN2 Calculation of CA.18 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.18 (FM.34) DM1 (0,4269) Cost 0,50 0,1684 0,3295 0,3447 Time Loss 0,25 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,3716 Time Loss 0,25 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,3445 Time Loss 0,25 0,0514 Obligations 0,50 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,50 0,1109 102 APPENDIX A18: RPN2 Calculation of CA.19 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.19 (FM.32) DM1 (0,4269) Cost 0,50 0,1684 0,4300 0,5475 Time Loss 0,50 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,5379 Time Loss 0,75 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,50 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,75 0,3993 0,7072 Time Loss 0,75 0,0514 Obligations 0,75 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,75 0,1109 103 APPENDIX A19: RPN2 Calculation of CA.20 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.20 (FM.45) DM1 (0,4269) Cost 0,00 0,1684 0,3963 0,2814 Time Loss 0,25 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,2374 Time Loss 0,25 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,1651 Time Loss 0,00 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,50 0,1109 104 APPENDIX A20: RPN2 Calculation of CA.21 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.21 (FM.35) DM1 (0,4269) Cost 0,25 0,1684 0,0421 0,0394 Time Loss 0,00 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,00 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,0506 Time Loss 0,25 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,00 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,0277 Time Loss 0,00 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,00 0,1712 Reputation 0,25 0,1109 105 APPENDIX A21: RPN2 Calculation of CA.22 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.22 (FM.36) DM1 (0,4269) Cost 0,25 0,1684 0,3488 0,2945 Time Loss 0,00 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,3568 Time Loss 0,25 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,1779 Time Loss 0,25 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,50 0,1109 106 APPENDIX A22: RPN2 Calculation of CA.23 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.23 (FM.47) DM1 (0,4269) Cost 0,00 0,1684 0,2771 0,2945 Time Loss 0,00 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,4220 Time Loss 0,25 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,50 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,1779 Time Loss 0,25 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,50 0,1109 107 APPENDIX A23: RPN2 Calculation of CA.24 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.24 (FM.10) DM1 (0,4269) Cost 0,25 0,1684 0,1604 0,1601 Time Loss 0,00 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,25 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,1284 Time Loss 0,25 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,00 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,1832 Time Loss 0,25 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,25 0,1109 108 APPENDIX A24: RPN2 Calculation of CA.25 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.25 (FM.23) DM1 (0,4269) Cost 0,25 0,1684 0,6659 0,5476 Time Loss 0,75 0,0783 Obligations 0,75 0,3985 Prevention Policy 0,75 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,5022 Time Loss 0,50 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,50 0,3115 Reputation 1,00 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,4279 Time Loss 0,50 0,0514 Obligations 0,50 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,75 0,1109 109 APPENDIX A25: RPN2 Calculation of CA.26 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.26 (FM.41) DM1 (0,4269) Cost 0,00 0,1684 0,0187 0,0538 Time Loss 0,00 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,25 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,0927 Time Loss 0,00 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,0705 Time Loss 0,00 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,50 0,1109 110 APPENDIX A26: RPN2 Calculation of CA.27 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.27 (FM.39) DM1 (0,4269) Cost 0,25 0,1684 0,3683 0,3369 Time Loss 0,25 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,3071 Time Loss 0,25 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,3183 Time Loss 0,50 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,75 0,1109 111 APPENDIX A27: RPN2 Calculation of CA.28 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.28 (FM.42) DM1 (0,4269) Cost 0,75 0,1684 0,5608 0,5666 Time Loss 0,50 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,75 0,1098 0,6870 Time Loss 0,75 0,0924 Obligations 0,75 0,4270 Prevention Policy 0,50 0,3115 Reputation 1,00 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,4849 Time Loss 0,50 0,0514 Obligations 0,50 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,75 0,1109 112 APPENDIX A28: RPN2 Calculation of CA.29 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.29 (FM.6) DM1 (0,4269) Cost 0,25 0,1684 0,4766 0,5175 Time Loss 0,50 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,7099 Time Loss 0,75 0,0924 Obligations 0,75 0,4270 Prevention Policy 0,75 0,3115 Reputation 1,00 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,4279 Time Loss 0,50 0,0514 Obligations 0,50 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,75 0,1109 113 APPENDIX A29: RPN2 Calculation of CA.30 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.30 (FM.28) DM1 (0,4269) Cost 0,25 0,1684 0,6463 0,6177 Time Loss 0,50 0,0783 Obligations 0,75 0,3985 Prevention Policy 0,75 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,6364 Time Loss 0,50 0,0924 Obligations 0,75 0,4270 Prevention Policy 0,50 0,3115 Reputation 1,00 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,5668 Time Loss 0,50 0,0514 Obligations 0,75 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,50 0,1109 114 APPENDIX A30: RPN2 Calculation of CA.31 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.31 (FM.4) DM1 (0,4269) Cost 0,75 0,1684 0,3716 0,3665 Time Loss 0,25 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,3427 Time Loss 0,25 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,50 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,3776 Time Loss 0,25 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,50 0,1109 115 APPENDIX A31: RPN2 Calculation of CA.32 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.32 (FM.9) DM1 (0,4269) Cost 0,50 0,1684 0,3108 0,3077 Time Loss 0,25 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,3427 Time Loss 0,25 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,50 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,2777 Time Loss 0,25 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,50 0,1109 116 APPENDIX A32: RPN2 Calculation of CA.33 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.33 (FM.17) DM1 (0,4269) Cost 0,25 0,1684 0,1308 0,1320 Time Loss 0,00 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,25 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,0654 Time Loss 0,25 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,1832 Time Loss 0,25 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,25 0,1109 117 APPENDIX A33: RPN2 Calculation of CA.34 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.34 (FM.29) DM1 (0,4269) Cost 0,50 0,1684 0,1216 0,1352 Time Loss 0,00 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,0571 Time Loss 0,00 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,2109 Time Loss 0,25 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,50 0,1109 118 APPENDIX A34: RPN2 Calculation of CA.35 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.35 (FM.24) DM1 (0,4269) Cost 0,00 0,1684 0,0383 0,0333 Time Loss 0,25 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,25 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,0148 Time Loss 0,00 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,0406 Time Loss 0,25 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,00 0,1712 Reputation 0,25 0,1109 119 APPENDIX A35: RPN2 Calculation of CA.36 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.36 (FM.3) DM1 (0,4269) Cost 0,25 0,1684 0,0804 0,1033 Time Loss 0,25 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,25 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,0802 Time Loss 0,25 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,1502 Time Loss 0,25 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,25 0,1109 120 APPENDIX A36: RPN2 Calculation of CA.37 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.37 (FM.2) DM1 (0,4269) Cost 0,00 0,1684 0,4345 0,5099 Time Loss 0,50 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,6046 Time Loss 0,75 0,0924 Obligations 0,75 0,4270 Prevention Policy 0,50 0,3115 Reputation 1,00 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,5375 Time Loss 0,50 0,0514 Obligations 0,75 0,2672 Prevention Policy 0,75 0,1712 Reputation 0,75 0,1109 121 APPENDIX A37: RPN2 Calculation of CA.38 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.38 (FM.7) DM1 (0,4269) Cost 0,75 0,1684 0,4416 0,4811 Time Loss 0,25 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 1,00 0,1098 0,3703 Time Loss 0,50 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,75 0,3993 0,6147 Time Loss 0,25 0,0514 Obligations 0,50 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,75 0,1109 122 APPENDIX A38: RPN2 Calculation of CA.39 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.39 (FM.50) DM1 (0,4269) Cost 0,25 0,1684 0,6463 0,6105 Time Loss 0,50 0,0783 Obligations 0,75 0,3985 Prevention Policy 0,75 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,7143 Time Loss 0,50 0,0924 Obligations 0,75 0,4270 Prevention Policy 0,75 0,3115 Reputation 1,00 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,4872 Time Loss 0,25 0,0514 Obligations 0,50 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,50 0,1109 123 APPENDIX A39: RPN2 Calculation of CA.40 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.40 (FM.23) DM1 (0,4269) Cost 0,25 0,1684 0,2305 0,2007 Time Loss 0,00 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,25 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,1202 Time Loss 0,00 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,2221 Time Loss 0,00 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,00 0,1712 Reputation 0,50 0,1109 124 APPENDIX A40: RPN2 Calculation of CA.41 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.41 (FM.12) DM1 (0,4269) Cost 0,50 0,1684 0,1216 0,1979 Time Loss 0,00 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,2544 Time Loss 0,00 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,25 0,0593 DM3 (0,3288) Cost 0,50 0,3993 0,2551 Time Loss 0,00 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,00 0,1712 Reputation 0,50 0,1109 125 APPENDIX A41: RPN2 Calculation of CA.42 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.42 (FM.44) DM1 (0,4269) Cost 0,00 0,1684 0,3262 0,2855 Time Loss 0,25 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,2291 Time Loss 0,00 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,2747 Time Loss 0,00 0,0514 Obligations 0,50 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,50 0,1109 126 APPENDIX A42: RPN2 Calculation of CA.43 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.43 (FM.38) DM1 (0,4269) Cost 0,00 0,1684 0,5651 0,4148 Time Loss 0,00 0,0783 Obligations 0,75 0,3985 Prevention Policy 0,75 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,4137 Time Loss 0,0 0,0924 Obligations 0,50 0,4270 Prevention Policy 0,50 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,2207 Time Loss 0,25 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,50 0,1109 127 APPENDIX A43: RPN2 Calculation of CA.44 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.44 (FM.22) DM1 (0,4269) Cost 0,00 0,1684 0,0383 0,0907 Time Loss 0,25 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,25 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,0297 Time Loss 0,00 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,2041 Time Loss 0,00 0,0514 Obligations 0,50 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,25 0,1109 128 APPENDIX A44: RPN2 Calculation of CA.45 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.45 (FM.20) DM1 (0,4269) Cost 0,00 0,1684 0,4150 0,3296 Time Loss 0,25 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,3301 Time Loss 0,25 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,50 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,2185 Time Loss 0,50 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,25 0,1712 Reputation 0,75 0,1109 129 APPENDIX A45: RPN2 Calculation of CA.46 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.46 (FM.21) DM1 (0,4269) Cost 0,25 0,1684 0,6463 0,6150 Time Loss 0,50 0,0783 Obligations 0,75 0,3985 Prevention Policy 0,75 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,50 0,1098 0,7226 Time Loss 0,75 0,0924 Obligations 0,75 0,4270 Prevention Policy 0,75 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,4947 Time Loss 0,50 0,0514 Obligations 0,75 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,75 0,1109 130 APPENDIX A46: RPN2 Calculation of CA.47 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.47 (FM.26) DM1 (0,4269) Cost 0,25 0,1684 0,1186 0,1291 Time Loss 0,50 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,50 0,0747 DM2 (0,2442) Cost 0,25 0,1098 0,0950 Time Loss 0,25 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,75 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,1681 Time Loss 0,25 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,00 0,1712 Reputation 0,50 0,1109 131 APPENDIX A47: RPN2 Calculation of CA.48 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.48 (FM.19) DM1 (0,4269) Cost 0,00 0,1684 0,4345 0,3489 Time Loss 0,50 0,0783 Obligations 0,50 0,3985 Prevention Policy 0,50 0,2802 Reputation 0,75 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,2374 Time Loss 0,25 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,25 0,3993 0,3205 Time Loss 0,25 0,0514 Obligations 0,25 0,2672 Prevention Policy 0,50 0,1712 Reputation 0,50 0,1109 132 APPENDIX A48: RPN2 Calculation of CA.49 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.49 (FM.37) DM1 (0,4269) Cost 0,00 0,1684 0,1884 0,1566 Time Loss 0,00 0,0783 Obligations 0,25 0,3985 Prevention Policy 0,25 0,2802 Reputation 0,25 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,2374 Time Loss 0,25 0,0924 Obligations 0,25 0,4270 Prevention Policy 0,25 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,0555 Time Loss 0,00 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,00 0,1712 Reputation 0,50 0,1109 133 APPENDIX A49: RPN2 Calculation of CA.50 Weights of DMs Criteria scores & weights RPN2 values Of each of DMs RPN2 Criteria Score Weights CA.50 (FM.1) DM1 (0,4269) Cost 0,00 0,1684 0,0187 0,0243 Time Loss 0,00 0,0783 Obligations 0,00 0,3985 Prevention Policy 0,00 0,2802 Reputation 0,25 0,0747 DM2 (0,2442) Cost 0,00 0,1098 0,0297 Time Loss 0,00 0,0924 Obligations 0,00 0,4270 Prevention Policy 0,00 0,3115 Reputation 0,50 0,0593 DM3 (0,3288) Cost 0,00 0,3993 0,0277 Time Loss 0,00 0,0514 Obligations 0,00 0,2672 Prevention Policy 0,00 0,1712 Reputation 0,25 0,1109 134 APPENDIX B1: Photos of failure modes ranked according to RPN3 values (Priority 1, Pr.1) Failure Mode 15 (FM.15) 135 APPENDIX B2: Photos of failure modes ranked according to RPN3 values (Priority 2, Pr.2) Failure Mode 11 (FM.11) 136 APPENDIX B3: Photos of failure modes ranked according to RPN3 values (Priority 3, Pr.3) Failure Mode 46 (FM.46) 137 APPENDIX B4: Photos of failure modes ranked according to RPN3 values (Priority 4, Pr.4) Failure Mode 8 (FM.8) 138 APPENDIX B5: Photos of failure modes ranked according to RPN3 values (Priority 5, Pr.5) Failure Mode 18 (FM.18) 139 APPENDIX B6: Photos of failure modes ranked according to RPN3 values (Priority 6, Pr.6) Failure Mode 49 (FM.49) 140 APPENDIX B7: Photos of failure modes ranked according to RPN3 values (Priority 7, Pr.7) Failure Mode 48 (FM.48) 141 APPENDIX B8: Photos of failure modes ranked according to RPN3 values (Priority 8, Pr.8) Failure Mode 31 (FM.31) 142 APPENDIX B9: Photos of failure modes ranked according to RPN3 values (Priority 9, Pr.9) Failure Mode 40 (FM.40) 143 APPENDIX B10: Photos of failure modes ranked according to RPN3 values (Priority 10, Pr.10) Failure Mode 13 (FM.13) 144 APPENDIX B11: Photos of failure modes ranked according to RPN3 values (Priority 11, Pr.11) Failure Mode 14 (FM.14) 145 APPENDIX B12: Photos of failure modes ranked according to RPN3 values (Priority 12, Pr.12) Failure Mode 16 (FM.16) 146 APPENDIX B13: Photos of failure modes ranked according to RPN3 values (Priority 13, Pr.13) Failure Mode 5 (FM.5) 147 APPENDIX B14: Photos of failure modes ranked according to RPN3 values (Priority 14, Pr.14) Failure Mode 35 (FM.35) 148 APPENDIX B15: Photos of failure modes ranked according to RPN3 values (Priority 15, Pr.15) Failure Mode 34 (FM.34) 149 APPENDIX B16: Photos of failure modes ranked according to RPN3 values (Priority 16, Pr.16) Failure Mode 25 (FM.25) 150 APPENDIX B17: Photos of failure modes ranked according to RPN3 values (Priority 17, Pr.17) Failure Mode 45 (FM.45) 151 APPENDIX B18: Photos of failure modes ranked according to RPN3 values (Priority 18, Pr.18) Failure Mode 30 (FM.30) 152 APPENDIX B19: Photos of failure modes ranked according to RPN3 values (Priority 19, Pr.19) Failure Mode 36 (FM.36) 153 APPENDIX B20 Photos of failure modes ranked according to RPN3 values (Priority 20, Pr.20) Failure Mode 41 (FM.41) 154 APPENDIX B21: Photos of failure modes ranked according to RPN3 values (Priority 21, Pr.21) Failure Mode 27 (FM.27) 155 APPENDIX B22: Photos of failure modes ranked according to RPN3 values (Priority 22, Pr.22) Failure Mode 32 (FM.32) 156 APPENDIX B23: Photos of failure modes ranked according to RPN3 values (Priority 23, Pr.23) Failure Mode 43 (FM.43) 157 APPENDIX B24: Photos of failure modes ranked according to RPN3 values (Priority 24, Pr.24) Failure Mode 10 (FM.10) 158 APPENDIX B25: Photos of failure modes ranked according to RPN3 values (Priority 25, Pr.25) Failure Mode 47 (FM.47) 159 APPENDIX B26: Photos of failure modes ranked according to RPN3 values (Priority 26, Pr.26) Failure Mode 23 (FM.23) 160 APPENDIX B27: Photos of failure modes ranked according to RPN3 values (Priority 27, Pr.27) Failure Mode 39 (FM.39) 161 APPENDIX B28: Photos of failure modes ranked according to RPN3 values (Priority 28, Pr.28) Failure Mode 6 (FM.6) 162 APPENDIX B29: Photos of failure modes ranked according to RPN3 values (Priority 29, Pr.29) Failure Mode 42 (FM.42) 163 APPENDIX B30: Photos of failure modes ranked according to RPN3 values (Priority 30, Pr.30) Failure Mode 4 (FM.4) 164 APPENDIX B31: Photos of failure modes ranked according to RPN3 values (Priority 31, Pr.31) Failure Mode 28 (FM.28) 165 APPENDIX B32: Photos of failure modes ranked according to RPN3 values (Priority 32, Pr.32) Failure Mode 24 (FM.24) 166 APPENDIX B33: Photos of failure modes ranked according to RPN3 values (Priority 33, Pr.33) Failure Mode 17 (FM.17) 167 APPENDIX B34: Photos of failure modes ranked according to RPN3 values (Priority 34, Pr.34) Failure Mode 9 (FM.9) 168 APPENDIX B35: Photos of failure modes ranked according to RPN3 values (Priority 35, Pr.35) Failure Mode 29 (FM.29) 169 APPENDIX B36: Photos of failure modes ranked according to RPN3 values (Priority 36, Pr.36) Failure Mode 3 (FM.3) 170 APPENDIX B37: Photos of failure modes ranked according to RPN3 values (Priority 37, Pr.37) Failure Mode 2 (FM.2) 171 APPENDIX B38: Photos of failure modes ranked according to RPN3 values (Priority 38, Pr.38) Failure Mode 7 (FM.7) 172 APPENDIX B39: Photos of failure modes ranked according to RPN3 values (Priority 39, Pr.39) Failure Mode 33 (FM.33) 173 APPENDIX B40: Photos of failure modes ranked according to RPN3 values (Priority 40, Pr.40) Failure Mode 50 (FM.50) 174 APPENDIX B41: Photos of failure modes ranked according to RPN3 values (Priority 41, Pr.41) Failure Mode 12 (FM.12) 175 APPENDIX B42: Photos of failure modes ranked according to RPN3 values (Priority 42, Pr.42) Failure Mode 44 (FM.44) 176 APPENDIX B43: Photos of failure modes ranked according to RPN3 values (Priority 43, Pr.43) Failure Mode 22 (FM.22) 177 APPENDIX B44: Photos of failure modes ranked according to RPN3 values (Priority 44, Pr44) Failure Mode 38 (FM.38) 178 APPENDIX B45: Photos of failure modes ranked according to RPN3 values (Priority 45, Pr.45) Failure Mode 20 (FM.20) 179 APPENDIX B46: Photos of failure modes ranked according to RPN3 values (Priority 46, Pr.46) Failure Mode 26 (FM.26) 180 APPENDIX B47: Photos of failure modes ranked according to RPN3 values (Priority 47, Pr.47) Failure Mode 37 (FM.37) 181 APPENDIX B48: Photos of failure modes ranked according to RPN3 values (Priority 48, Pr.48) Failure Mode 21 (FM.21) 182 APPENDIX B49: Photos of failure modes ranked according to RPN3 values (Priority 49, Pr.49) Failure Mode 1 (FM.1) 183 APPENDIX B50: Photos of failure modes ranked according to RPN3 values (Priority 50, Pr.50) Failure Mode 19 (FM.19) 184 APPENDIX C1: Risk analysis and FMEA test Question 1: According to OHS risk analysis regulation, which of the followings that are regarding with the risk control measures are wrong? a) Elimination of the hazard or sources of hazards b) Giving priority to the use of personal protective equipment c) Replacing the dangerous substance with the not dangerous or less dangerous substance. d) Combating the risks at their source Question 2: Which of the following should be considered when defining the hazards in the workplace? I. The raw materials and semi-finished products II. Procedures regarding with the residues and waste substances III. Employees' experiences and thoughts. IV. Employee education, age, gender. V. Workers' health records. VI. Consequences of workplace inspections a) I-II-V-VI b) I-III-IV c) III-IV-V-VI d) All of them Question 3: According to OHS risk analysis regulation, which of the following need re- application of risk analysis methodology? a) If the workplace is given break more than 30 days. b) After the strike made in the workplace. 185 c) After the technological changes in the workplace. d) At the begining of each year. Question 4: Which of the followings are put in order correctly according to OHS risk analysis regulation? I. To identify and analyze the risks, II. documentation, III. Deciding on the risk analysis methodology IV. Defining the hazards, V. Updating and renewal of work a) I-IV-III-II-V b) IV-I-II-III-V c) IV-I-III-II-V d) I-IV-II-III-V Question 5: Which of the following obligates the risk assessment for occupational health and safety concept? a) ISO 9001 b) IS014001 c) ISO 22000 d) OHSAS 18001 186 Question 6: I- Risk control measures II- The number of people that can be affected by hazards III- Selected risk assessment methodology IV- Potential severity of loss V- The probability of damage Which of the above is not effective in calculating the magnitude of the risk? a) I, II b) I, III c) II, V d) III, IV Question 7: Which of the following is not one of the methods of risk assessment? a) HAZOP b) FMEA c) FTA d)HACCP Question 8: Which of the following describes the FMEA risk assessment method? a) A graphical representation of a logical combination of defined adverse event’s or condition’s causes. b) It is one of the most appropriate qualitative approaches that the mechanical and electrical systems to be reviewed. c) A brainstorming approach that is comprehensive and loosely structured query using. d) Probability and severity is graded from 1 to 5 and this method is consist of the probability and severity multiplication matrix? Question 9: Which of the followings are the parameters of FMEA’s risk priority number? a) Occurrence-severity b) Occurrence- detectability 187 c) Severity- detectability d) Occurrence- severity- detectability Question 10: What is the step that the potential failure’s magnitude, detectability are evaluated in the FMEA methodology? a) The fragmentation of system or action for analyzing b) Identification of the accidents that are object to analysis c) The evaluation of potential failures that can lead to accidents d) Identification of potential error conditions for the system elements Answers : 1- b 2- d 3- c 4- c 5- d 6- b 7- d 8- b 9- d 10- c 188 CURRICULUM VITAE Name Surname : Ali Kaan PASTIRMACI Address : GMKP. Mah. Özcan Sok. Engin Baysak ĠĢ Merkezi No:11 D:9 Çerkezköy/ TEKĠRDAĞ Birthplace and Date : BĠGA/ Çanakkale, 1988 Second Language : English Primary School : Biga Ġ.Ö.O. High School : IĢıklar Askeri Lisesi Bachelor School : Turkish Military Academy (Kara Harp Okulu) Graduate School : BahçeĢehir University Institute Name : Graduate School of Natural and Applied Sciences Program : System Engineering Publications : Work Life : TSK (2010-2014) 189