ETR – Eisenbahntechnische Rundschau | Ausgabe Science/2020
Prediction of Crack Growth in Railway Axles with Supervised Learning and Artificial Neural Networks
In this work, a method for predicting crack growth in railway axles with artificial neural networks is developed. The process of crack growth is simulated with randomly generated initial cracks. The crack growth can then be predicted through the neural network, which coincides with the simulated finite element analysis results. The efficiency for online diagnosis and prediction can be significantly improved with the data-driven approach.