CA23109 - Fatigue Benchmark Repository (FABER)

Fatigue damage is the key factor in 80-90% of in-service failures of structural components. In any design, it is essential to be able to estimate the potential fatigue damage safely and expertly. Commercial fatigue solvers are program tools dedicated to allowing even a not very expert user to perform a straightforward fatigue life prediction. However, the fatigue damage process is very complex. It involves a large number and a large range of variables. The computation process is based mostly on empirical experience. Currently-used fatigue solvers are based on experimental evidence acquired during first two thirds of the 20th century. Although very many experiments have been conducted in the meantime, the computational basis has remained fixed, and little effort has been dedicated to a redefinition. The lack of interest of academia in the topic, the tendency of fatigue solver developers to implement computational strategies without understanding the problem adequately, the use by inexperienced and inexpert users, and final warranty denial for the results of fatigue solvers have led to a critical mass of problems, which can be marked as a loss of responsibility. Our project focuses on preparing a database of experimental fatigue data, which will be easily accessible for creating benchmark sets. Users from academia and from engineering sectors will be able to adopt the data quickly for testing various prediction hypotheses and various computational tools. An open-source fatigue software will be prepared. Only such joint action can restore a responsible attitude for the computational results presented by fatigue solvers.