ICNPAA 2010: Mathematical Problems in
Engineering, Aerospace and Sciences

Actually, Safety Instrumented Systems (SIS) are designed to prevent and / or mitigate accidents, avoiding undesirable high potential risk scenarios and saving costs and protecting health of people, environment and equipments. It demands formal methods to assure the safety specifications. In this sense, this paper introduces a methodology for modeling diagnostic and treatment of critical faults using Bayesian networks (BN) and Petri nets (PN) respectively. This approach considers some critical faults got from hazard and operability (HAZOP) study in the equipment or system under control and define each safety instrumented function (SIF) as well determine the safety instrumented level (SIL) by each SIF. The present approach uses Bayesian networks (BN) for diagnoses and decision-making proposes, and the Petri nets (PN) for the synthesis, modeling and control proposes implemented by Safety PLC to prevent and / or mitigate accidents. A case study considering diagnostic and treatment of critical faults is presented.