Generation of model-based safety arguments from automatically allocated safety integrity levels

Sorokos, Ioannis

Abstract

To certify safety-critical systems, assurance arguments linking evidence of safety to appropriate requirements must be constructed. However, modern safety-critical systems feature increasing complexity and integration, which render manual approaches impractical to apply. This thesis addresses this problem by introducing a model-based method, with an exemplary application based on the aerospace domain.

Previous work has partially addressed this problem for slightly different applications, including verification-based, COTS, product-line and process-based assurance. Each of the approaches is applicable to a specialised case and does not deliver a solution applicable to a generic system in a top-down process. This thesis argues that such a solution is feasible and can be achieved based on the automatic allocation of safety requirements onto a system’s architecture. This automatic allocation is a recent development which combines model-based safety analysis and optimisation techniques. The proposed approach emphasises the use of model-based safety analysis, such as HiP-HOPS, to maximise the benefits towards the system development lifecycle.

The thesis investigates the background and earlier work regarding construction of safety arguments, safety requirements allocation and optimisation. A method for addressing the problem of optimal safety requirements allocation is first introduced, using the Tabu Search optimisation metaheuristic. The method delivers satisfactory results that are further exploited for construction of safety arguments. Using the produced requirements allocation, an instantiation algorithm is applied onto a generic safety argument pattern, which is compliant with standards, to automatically construct an argument establishing a claim that a system’s safety requirements have been met. This argument is hierarchically decomposed and shows how system and subsystem safety requirements are satisfied by architectures and analyses at low levels of decomposition. Evaluation on two abstract case studies demonstrates the feasibility and scalability of the method and indicates good performance of the algorithms proposed. Limitations and potential areas of further investigation are identified.

Citation

Sorokos, I. Generation of model-based safety arguments from automatically allocated safety integrity levels. (Thesis). University of Hull. https://hull-repository.worktribe.com/output/4222118

• Thesis Type: Thesis
• Deposit Date: Feb 14, 2020
• Publicly Available Date: Feb 23, 2023
• Keywords: Computer science
• Public URL: https://hull-repository.worktribe.com/output/4222118
• Additional Information: Department of Computer Science, The University of Hull
• Award Date: Jun 1, 2017