In the design process of a safety related system, three main steps are necessary: (i) Hazard and risk analysis, which amongst others results in a number of safety functions that have to be implemented, (ii) conceptual design which specifies a design that will implement the safety functions, (iii) evaluation of the safety functions where the Safety Integrity Level (SIL), for example, needs to be quantified (ref. functional safety standard “IEC 26262”). This last step requires considering reliability information of the safety-instrumented system.
Analyzing the current applied approaches that industries follow to quantify the Safety Integrity Level, some issues are observed. First of all, the additional safety information is not structured in the same way as in the design concept. As a consequence, it is difficult to relate the safety information to the different components in the design concept. When design iterations are required, it is hard to know which safety information is affected.
A second issue is the lack of re-use. Although some tools, like FTA (Fault Tree Analysis), offer the ability to specify failure propagation patterns, it was found that such patterns were not used in the failure specifications of the different safety functions.
At Flanders Make, these issues have been investigated in the framework of the industrial project Experimental Validation for Safety Integrity Level (VAL4SIL) resulting in the development of a new model-based safety methodology to deal with the aforementioned issues.