Development of a design technique for the identification of fatigue initiating features

Fatigue failure is a highly complex phenomenon and any attempt to provide a non-analytical or non-empirical analysis requires the effective classification and management of the data involved. The objective of this research is to create a non-analytical technique that assists detailing of potential fatigue initiating features at design conception. This paper reveals the research methodology behind the development of a new technique, concept design for fatigue resistance (CDFR). Firstly, in a survey of UK industry, current industrial practices in designing for fatigue resistance are investigated to justify the development of a new technique. Secondly, through analysis and cataloguing of 120 fatigue failure cases trends in fatigue mechanisms, their causes and specific factors that contributed to a variety of different feature’s fatigue failure mode(s) are developed. CDFR aims to synthesise the failure process for the designer considering the most frequently occurring factors related to the feature type and failure mode as identified from the failure case database. This paper examines the development of the CDFR technique leading to the proposed utilisation and integration of three powerful and proven approaches: expert system (ES), knowledge based system (KBS) and case-based reasoning (CBR) in order to aid the process of the identification and elimination of future fatigue problems at the earliest possible stages in the development of a product.