RELIABILITY PAPER An alternative approach to safety integrity level determination: results from a case study Nouara Ouazraoui and Rachid Nait-Said LRPI Laboratory, Health and Occupational Safety Institute, University of Batna-2, Batna, Algeria Abstract Purpose – The purpose of this paper is to validate a fuzzy risk graph model through a case study results carried out on a safety instrumented system (SIS). Design/methodology/approach – The proposed model is based on an inference fuzzy system and deals with uncertainty data used as inputs of the conventional risk graph method. The coherence and redundancy of the developed fuzzy rules base are first verified in the case study. A new fuzzy model is suggested for a multi-criteria characterization of the avoidance possibility parameter. The fuzzy safety integrity level (SIL) is determined for two potential accident scenarios. Findings – The applicability of the proposed fuzzy model on SIS shows the importance and pertinence of the proposed fuzzy model as decision-making tools in preventing industrial hazards while taking into consideration uncertain aspects of the data used on the conventional risk graph method. The obtained results show that the use of continuous fuzzy scales solves the problem of interpreting results and provides a more flexible structure to combine risk graph parameters. Therefore, a decision is taken on the basis of precise integrity level values and protective actions in the real world are suggested. Originality/value – Fuzzy logic-based safety integrity assessment allows assessment of the SIL in a more realistic way by using the notion of the linguistic variable for representing information that is qualitative and imprecise and, therefore, ensures better decision making on risk prevention. Keywords Coherence and redundancy, Fuzzy avoidance possibility, Fuzzy risk graph, Safety instrumented system, Safety integrity level Paper type Research paper Nomenclature IEC International Electrotechnical Commission SIS safety instrumented system SIL safety integrity level SIF safety instrumented function PFD probability failure on demand RRF risk reduction factor FIS fuzzy inference system FTA fault tree analysis QRA quantitative risk assessment LOPA layers of protection analysis C consequence F occupation P avoidance possibility W demand rate Q fuzzy interval μ Q membership function of Q E(Q) distribution function of Q E*(Q) lower distribution function of Q International Journal of Quality & Reliability Management Vol. 36 No. 10, 2019 pp. 1784-1803 © Emerald Publishing Limited 0265-671X DOI 10.1108/IJQRM-02-2019-0065 Received 23 February 2019 Revised 13 May 2019 Accepted 5 June 2019 The current issue and full text archive of this journal is available on Emerald Insight at: www.emeraldinsight.com/0265-671X.htm 1784 IJQRM 36,10