City Research Online

Abstract

Safety-critical systems such as medical devices and avionics systems are developed using systematic processes and rigorous analysis methods. This is necessary to gain strong confidence that the system is not affected by latent design problems that may lead to system failures or unintended behaviours that, ultimately, could result in damage or harm to people or the environment. Whilst different guidelines and recommended best development practices are provided in different regulatory frameworks and standards, all processes share a common initial stage, known as hazard analysis. The aim of the hazard analysis is to identify all known and foreseeable scenarios and problematic situations. It is important that the hazard analysis is as accurate and as comprehensive as possible since the entire development process builds on the hazard analysis results. Any missed scenario or overlooked problematic situation could breach the mitigation strategies designed to guarantee the safety of the system.

Several hazard analysis techniques have been introduced over the last 50 years to improve the quality of the analysis. However, a known weakness of the current generation of techniques is that they often rely on manual analysis of information recorded in textual format. For realistic,complex systems, the amount of information is usually abundant and overwhelming. Because ofthis, even the most expert analyst can accidentally overlook important aspects of the system that should have been considered to ensure the safety of the system. The research work presented in this thesis aims to provide a systematic and comprehensive way to help the expert analyst with his task.

This thesis explores the development of a novel method and supporting analysis tool for the refinement of the hazard analysis results. The method is structured into a series of stages, each of which provides feedback to the analysts to help them gain confidence in the quality of the analysis. The method also helps to identify and resolve weaknesses in the analysis, if they are present. The research builds an ontology to represent knowledge collected during the hazard analysis. Inference rules are used to reason about possible scenarios, hazards, hazard causes and their relations. Formal (i.e., mathematically-based) tools are used to mechanise the exploration of scenarios, discover relations between hazards and causes that may have been overlooked during the analysis. The effectiveness of the proposed method is evaluated using various realistic case studies from different application domain

Publication Type:	Thesis (Doctoral)
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Departments:	Doctoral Theses School of Science & Technology School of Science & Technology > Computer Science

Export

Downloads