A NOVEL FRAMEWORK FOR IMPROVING FMEA APPLICATION IN COMPLEX V-MODEL DEVELOPMENT

Abstract

Management of design risk has persisted as a topic for as long as formal design processes have existed. In 1949, based on the lessons from World War II development efforts, the U.S. Department of Defense published the first standard for Failure Modes, Effects, and Criticality Analysis (FMECA), which later shortened in common use to Failure Modes and Effects Analysis (FMEA). Over the following 70 years, this approach became arguably the most ubiquitous tool in the design risk toolkit, with standards and application across virtually every aspect of product and process design. FMEA has also become an active topic for academic research, in part because of persistent problems with its application, effectiveness, and usability. Counterintuitively, the bulk of research into the topic has done little to stem the continued reporting of these problems. In fact, the last 35 years have seen continued publication of peer-reviewed articles documenting recurring challenges in FMEA application. At roughly the same time that these challenges became more visible, the same Department of Defense published standards that contributed to the origins of the systems engineering V-model as a means of managing the growing complexity of modern weapon systems. In short order, this model migrated into software development and then into broader product development, where it has become similarly influential. This dissertation argues that the persistence of FMEA application challenges is not coincidental with the growth of the V-model as a dominant development approach. Rather, many recurring problems in FMEA can be understood as symptoms of a structural mismatch between a traditionally bottom-up risk analysis method and a top-down staged development model. This dissertation addresses that argument through three stages of research. First, a literature review is performed to identify recurring criticisms of FMEA and examine commonality in these issues over time. Second, a structural analysis is performed comparing the FMEA methodology, as defined by internationally accepted standards and common practice, against the information structure and staged logic of V-model development. This analysis identifies systemic mismatches between the two methods, including proliferation of failure modes, incompatibility between FMEA information needs and the timing of information availability in the V-model, and difficulty aligning FMEA decomposition with systems engineering decomposition. Third, a novel framework is developed to address these problems while preserving the core value of FMEA as a design risk management tool. The proposed framework improves FMEA application in complex V-model development through four related elements. First, a New, Unique, and Difficult (NUD) design risk filter is used to identify the areas of a system architecture that warrant focused risk analysis. Second, requirements-based definitions of failure modes and causes are used to align FMEA more directly with systems engineering requirements and architecture. Third, mitigations may be expressed as new or revised requirements and recursively reviewed to determine whether they introduce additional design risk. Fourth, Critical-to-Quality (CtQ) designations are reframed as a mechanism for tracking mitigations and requirements that cross architectural or engineering discipline boundaries. A key portion of this framework, the NUD design risk filter, is validated in an industrial setting through application to a complex technology development project. The validation trial demonstrates that the method can be taught and deployed with limited training and facilitation, and that users perceive value in its ability to identify key risk areas in a system architecture. The results provide initial evidence that targeted design risk filtering can help address FMEA proliferation and improve the practical deployment of FMEA in complex V-model environments. As FMEA continues to be a fundamental expectation of sound design risk practice, the proposed framework does not seek to replace FMEA. Instead, it improves how FMEA is scoped, timed, decomposed, and connected to systems engineering development activities. From a research perspective, the framework opens opportunities for continued improvement in FMEA through integration with Model-Based Systems Engineering (MBSE), requirements management tools, and improved measures of design maturity and design risk management throughout the V-model process.