Professor and chairman, Lehigh University, Packard Lab, Bethlehem, PA
Pages: 22 Published: Jan 1997
Concerns with aging infrastructure worldwide and with the life-cycle costs and management of engineered systems have placed increased emphasis on the development of methods for life prediction. To be effective as “true predictors” of future performance (i.e., to provide accurate estimates beyond the range employed in the development of supporting data), such methods must be built upon mechanistic models that capture the functional dependence on all of the key external and internal variables. The development of these methods argues strongly for multidisciplinary research that integrates mechanistic understanding with probability analysis. In this paper, a mechanistically based probability approach to life prediction (versus the more traditional statistically based parametric approach) and the processes for model development are outlined to provide a framework for discussion. The use of a coordinated, multidisciplinary approach to develop mechanistic understanding and to model material response is illustrated through examples on crack growth in a high-strength steel. The need for multidisciplinary research that broadens the perspective from testing to testing and materials is discussed.
fracture mechanics, subcritical crack growth, life prediction, materials, mechanisms, multidisciplinary approach, testing
Paper ID: STP16225S