Answers are grounded in this paper and cite their sources.
Deep learning predicts fatigue life in GH4169 alloy joints, indicating enhanced accuracy with strain data.
The low cycle fatigue (LCF) life for the GH4169 inertia friction welding joints is predicted through deep learning approaches. Experiments are conducted to obtain three types of LCF datasets for convolutional neural network (CNN) establishment, including deformation images, average strain, and combined average strain and damage fraction. The role of dataset types and sizes on the CNN accuracy in LCF life prediction is discussed, utilizing 1222–2397 deformation images and 1175–2585 numerical values (including average strain and damage fractions). Results show the dataset of strain and damage fraction brings higher , especially under a larger dataset (0.9560). This data‐physics integrated approach innovatively utilizes peak strain features as the CNN model input, and it demonstrates that higher accuracy of LCF life prediction for nickel‐based materials can be achieved after introducing a physics model. The method is promising for non‐destructive supervision of in‐service equipment and structures.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: