A Novel Physical Neural Network Based on Transformer Framework for Multiaxial Fatigue Life Prediction
DOI:
10.1111/ffe.14618
Publication Date:
2025-02-24T02:15:12Z
AUTHORS (7)
ABSTRACT
ABSTRACT The stability of prediction precision under complex loading paths is one the key challenges in task multiaxial fatigue life prediction. This study addresses unstable machine learning models, while further improving A novel neural network based on a transformer framework proposed to capture dependencies between data at multiple scales. Meanwhile, physical loss function with soft adjustments add constraints network. These two mechanisms assist each other accuracy and Performance validation was conducted using from nine distinct materials. Comparative analysis performed against six existing models evaluate efficacy Experimental evidence supports high predictive network, which also demonstrates robust across diverse conditions.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (47)
CITATIONS (0)
EXTERNAL LINKS
PlumX Metrics
RECOMMENDATIONS
FAIR ASSESSMENT
Coming soon ....
JUPYTER LAB
Coming soon ....