Contribution to an improvement of quantitative evaluation model for reliability of safety-related functions

This paper presents a more accurate modeling way of safety-related function and gives new opportunities towards a quantitative evaluation of safety and availability in complex fault-tolerant architectures. The necessity of guaranteeing the availability and safety property of a process by integrating a realistic description of the studied safety function is urgent due to the quick development of fault tolerant architectures supported by fault detection procedures, auto-test and maintenance policies. Our goal is to present a high-level evaluation model, which will take into account the possible consequences of architectural choices and allow us by comparison of well-recognized safety and availability metrics to choose the best architecture in regards of requirements on performances of our safety-related function. The presented high-level evaluation model is a probabilistic evaluation model based on heterogeneous class of interacting sub-models, including Markov chains and finite automata. The choice of the best architecture is based on a compromise between availability, safety and operative cost aspects.