On Hazard Identification in the Automotive Domain
Electrical systems in vehicles are becoming more important with every new generation of cars since the functional growth in the automotive industry is mainly realized by electronics. The new functionalities can for instance be advanced safety improving systems or electronics replacing mechanical system solutions.
Many of these new automotive electronic systems are safety related and therefore need to be developed according to a safety process. This thesis presents an introduction to the development of safety related automotive systems and addresses hazard identification of these systems in an early conceptual phase.
An extended actuator-focused hazard analysis method, based on a generic failure mode approach, is proposed. Since it is the actuators that affect the systems environment, this approach is logical for an early hazard analysis when only limited information of the system implementation is available. An evaluation scheme of conceptual system solutions based on the hazard analysis is also proposed.
This work also gives an experimental assessment of the proposed method and a method from ESA, based on induction with generic low level hazards. Both methods are found to be applicable for hazard identification in automotive systems. The study also shows, with statistical significance, that the adapted FFA method is less time consuming and easier to use than the ESA method. Some improvements to the methods are also proposed.
Finally, input data to the hazard identification methods, i.e. use cases, are analyzed with regard to quality. A set of evaluation criteria, based on earlier research, is presented and applied to use cases from a project at Volvo Cars. Statistics on defect intensities and their order are provided and form the basis for proposals for improvements of industrial practice based on earlier research. Further, a comparison is made, based on the established order, between the criteria used and earlier research on guidelines and checklists for use case authoring. The results of this work are improvement propositions for these guidelines.
Quality of Use Cases