Component-based Decomposition of Hazard Analysis

Hazard analysis plays an important role in the assurance of systems’ safety. As systems are getting bigger and more complex, hazard analysis approaches face difficulties with the amount of memory required for the analysis. Big systems are modeled with a modularized model, like UML component model. Those difficulties with memory can be conquered if the analyst utilizes the modularization of the system model to decompose the hazard analysis. Furthermore, some of the hazard analysis results may be reused in the analyses of similar systems. This thesis presents a new approach to overcome the memory problems facing the hazard analysis of big systems by a decomposition which exploits the system modularization. This approach is implemented as a software tool, and evaluated against an already existing tool for the hazard analysis without decomposition. It also suggests an idea for the results obtained from analyzing one system to be reused when analyzing similar systems.