Proceedings of the
The 33rd European Safety and Reliability Conference (ESREL 2023)
3 – 8 September 2023, Southampton, UK
Identifying Test Scenarios for Simulated Safety Demonstration using STPA and CAST
1Department of Engineering Cybernetics, Norwegian University of Science and Technology, Norway.
2Risk, Security and Physical Science, Institute for Energy Technology (IFE), Norway.
3Group Research and Development, Det Norske Veritas (DNV), Norway.
ABSTRACT
Assuring safety for new technologies like a Maritime Autonomous Surface Ship (MASS) or an Uncrewed Surface Vessel (USV) is challenging due to their complexity and varying operational environments. Safety demonstrations in simulations may be used to verify operational safety, but it is impossible to test all possible scenarios. The paper proposes an approach to identify critical scenarios for scenario-based safety demonstrations based on System Theoretic Process Analysis (STPA). STPA studies the whole system including interactions between components in the hazard analysis and is, therefore, well-suited for systems like MASS or USV, involving interactions of multiple components, sub-systems, the environment, and humans. The presented approach identifies critical scenarios using STPA and generates simulation scenarios from the identified critical, as well as presumably safe, scenario spaces. In case of incidents or unexpected critical scenarios that have been uncovered during the simulated tests, a Causal Analysis using System Theory (CAST) is conducted. Thus, it is possible to improve safety in new design iterations based on the results of the evaluation. The proposed approach is demonstrated in a simplified example of a USV during remote operations.
Keywords: Safety demonstration, STPA, CAST, Scenario identification, Automation, Autonomous system, Autonomous ship, MASS, Test and verification.
