Abstract
This work presents two new models, based on the use of Artificial Neural Networks (ANN) and non-local stress parameters physically related to crack initiation processes, to estimate fretting fatigue (FF) life. To train and validate these models, several FF data for different aeronautical Al alloys are gathered in the literature. To test the accuracy and generalization capability of the models, data related to two Al alloys not considered in the training phase are employed and life estimates are compared with those computed according to two well-established multiaxial fatigue models. The ANN models provide more accurate estimates than those provided by the fatigue models examined, as well as showed a good generalization capability, which suggests their potential for future industrial applications.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have