Automated Markov analysis on safety systems to evaluate the reliability

To reduce the inherent risk of industrial processes to an acceptable level, a safety instrumented system (SIS) can be added to the process. The main function of this system is to intervene when the process runs out of control. This intervention consists of taking the process back to a safe state, often a shutdown of the process. A safety instrumented system generally consists of sensors, a logic solver and actuators (like valves). To determine the (reduced) risk of the process, the safety and reliability of the safety instrumented system have to be analysed. This analysis has to comply with several standards, of which the IEC61508 standard is the most recent and, by many people also considered most relevant, standard. One of the results of a safety analysis in accordance with IEC61508, is the Safety Integrity Level of the system. As defined by this standard the SIL indicates the average probability that the safety instrumented system will not perform its safety function on demand. As proven by Rouvroye [1] Enhanced Markov Analysis is currently the most comprehensive analysis method available to calculate the Safety Integrity Level. This analysis method has a scientifically sound basis, but also has some practical drawbacks. The analysis can become quite time-consuming and complex. One way to overcome these drawbacks is an automated software tool. A prototype of such a tool has been developed and is called A-MASTER: Automated Markov Analysis on Safety systems to Evaluate the Reliability. This paper will explain several development aspects of such an automated software tool and the principles which underlie A-MASTER and its main functionality. It will also give a practical application of the tool.