On conceptual design reliability : an approach for identification of potential failures

Sammanfattning: This thesis is the result of research in the area of Evaluation and Decision-Making in the Design Process. It examines how to support and facilitate the evaluation activities in the design process in order to create a more solid foundation on which to base design decisions. The focus, however, is not on decision-making in general, but primarily on how the technical feasibility and reliability of a conceptual solution may be estimated with a higher degree of certainty.The current tools and methods for determine the feasibility and reliability are mostly passive, i.e. they prescribe how to proceed and do not actively contribute to the result. There is need for tools and methods that take a more active part in the process, and that may convey experience about feasibility and reliability from former designs.The objectives with this research are to acquire knowledge about how and in what ways failures occur in a product and how they may be modelled in order to increase the understanding of them; and to develop a methodology that utilises the knowledge about the failure phenomena in order to identify and avoid problems in the solutions.The research approach is mainly explorative and is based on several case studies and observations. In the case studies, the failure phenomena are examined both qualitatively and quantitatively from different perspectives. In the observations, insight and experience in the usability and shortcomings of prevalent evaluation and reliability methods are gathered.The results may be divided into two parts, one descriptive and one prescriptive part. The descriptive result consists of a phenomena model capable of explaining the appearance of failures in a system. The phenomena model is based on the results and the conclusions from the case studies and the observations, and has given rise to an enhanced system model. The main conclusion from the descriptive part is that the detection of potential failures in a system is largely a question of identifying the potentials for different physical phenomenon to occur in the system, and to determine which effects the physical phenomenon will give rise to.The prescriptive results consist mainly of an information model for failure identification, from which a failure identification process has been extracted. The fundamentals of the process are to describe the behaviour in the system by searching for effects from physical phenomena, and to determine how these effects affect the functionality. The process is to some extent verified by a comprehensive example.