A neural network approach to fatigue life prediction

• Published in: International journal of fatigue Vol. 33; no. 3; pp. 313 - 322
• Main Authors: Figueira Pujol, João Carlos, Andrade Pinto, João Mário
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 01.03.2011; Elsevier
• Subjects: Applied sciences; Distribution functions; Evolutionary computation; Exact sciences and technology; Fatigue; Fatigue (materials); Fatigue failure; Fatigue life; Lognormal distribution; Mathematical models; Mechanical properties and methods of testing. Rheology. Fracture mechanics. Tribology; Metals. Metallurgy; Neural networks; Steels; Step-stress; Fatigue; Step-stress; Neural networks; Lognormal distribution; Evolutionary computation; Fatigue life; Lifetime; Prediction; Mechanical properties; Neural network; Artificial intelligence
• ISSN: 0142-1123; 1879-3452
• DOI: 10.1016/j.ijfatigue.2010.09.003

In this work, a novel approach to fatigue life prediction under step-stress conditions is introduced, where the cumulative distribution function for the failure of components was implemented by means of a neural network. The model was fit to experimental data on the fatigue life of steel under step-stress conditions. For comparison, a standard approach based on the lognormal distribution function was also implemented and fit to the same experimental data. Both models were optimized by evolutionary computation, using a maximum likelihood estimator. The Kolmogorov–Smirnov test was applied to compare the results of the new approach to those obtained with the lognormal distribution function.