Pattern-based safety engineering: Semantic enrichment of system-architecture models for semi-automated safety analysis

The growing complexity of safety-critical systems is leading to an increased complexity of safety analysis models. Automatic transformations from the system model into the safety analysis model help to reduce time and cost for safety analyses. In this paper, we describe a pattern annotation framework for development models of safety-critical systems that is used to annotate system entities of fault tolerance mechanisms. We use the annotations as a source of semantic information to automatically generate parts of the safety analysis model. The presented approach contributes to the safety analysis domain by explicitly preserving the knowledge of domain experts and engineers within the model. Our approach allows performing a semi-automated preliminary safety analysis of fault tolerance mechanisms to estimate the failure behavior of a safety-critical systems architecture model in early stages of the development. This saves effort and costs for ongoing steps of the development process and prevents setbacks in the development.