Hazard Identification in Aircraft Design: A Systematic Review of ARP4761-Based Methods, STPA and MBSE-Enabled Safety Assessment

Abstract

Hazard identification remains a foundational activity in aircraft design because it shapes early safety objectives, architectural decisions, and certification strategies throughout the development lifecycle. However, the growing complexity of modern aircraft systems has challenged the sufficiency of traditional failure-centered approaches when used in isolation. This systematic literature review synthesizes the current state of knowledge on hazard identification and analysis for aircraft design, with particular emphasis on AFHA, PASA, FHA, and STPA, while also examining how these methodologies are being integrated with Model-Based Systems Engineering (MBSE). Following a PRISMA-based review protocol, the study screened records retrieved from Web of Science, Scopus, and IEEE Xplore, yielding a final corpus of 106 peer-reviewed studies for qualitative and thematic analysis. The results show that traditional ARP4761-based methods remain the dominant foundation of certification-oriented hazard identification, but that STPA has gained relevance as a complementary approach for capturing interaction hazards, software-related risks, and human-automation issues that are not naturally expressed as single-component failure conditions. The review also shows a growing movement toward hybrid methodological frameworks and MBSE-enabled safety workflows, particularly to improve traceability, front-load safety analyses, and maintain consistency between design models and safety artefacts. At the same time, the evidence reveals persistent gaps in certification alignment, empirical validation, scalability, interoperability, and the translation of systems-theoretic outputs into certification-native artefacts. Overall, the literature indicates that the future of aircraft hazard identification lies not in replacing established approaches, but in integrating traditional, systems-theoretic, and model-based methods into more coherent, traceable, and scalable safety-assessment processes.