Prediction of welded joint fatigue properties based on a novel hybrid SPDTRS-CS-ANN method

Fatigue life prediction of composite bolted joints based on finite element model and machine learning

“…As an advanced method in deep learning, ANN is different from traditional machine learning methods. Its robust data representation learning capability makes it extensively applicable in the field of fatigue life prediction 18,20–23 . Over the past few years, the convolutional neural network (CNN) has gained widespread utilization in hybrid neural network data prediction and data classification, owing to its exceptional feature extraction ability for complex datasets.…”

“…Its robust data representation learning capability makes it extensively applicable in the field of fatigue life prediction. 18,[20][21][22][23] Over the past few years, the convolutional neural network (CNN) has gained widespread utilization in hybrid neural network data prediction and data classification, owing to its exceptional feature extraction ability for complex datasets. Wu et al 24 proposed a CNN-BP prediction model for precipitation prediction, achieving an impressive prediction accuracy rate of 88.4%.…”

A post‐processing procedure for predicting high‐ and low‐cycle fatigue life of welded structures based on the master E–N curve

“…Besides the stress/strain parameters, researchers also developed data-driven models for predicting the fatigue life of welded/unwelded structures. [43][44][45] Post-processors are also developed to post-process the FE outputs to compute the fatigue life of structures. Nesl adek and Španiel wrote a post-processor to predict thermal/mechanical fatigue and creep life for the commercial FE software ABAQUS.…”

“…Besides the stress/strain parameters, researchers also developed data‐driven models for predicting the fatigue life of welded/unwelded structures 43–45 . Post‐processors are also developed to post‐process the FE outputs to compute the fatigue life of structures.…”

Fatigue life prediction of welded joints based on improved support vector regression model under two‐level loading

“…But considering the uncertainty and complexity of the fatigue analysis process, the mapping relationship between the given loading history and the fatigue life can not be described well by the traditional model, which leads to some limitations in practical engineering applications. In contrast, machine learning methods such as neural networks, 19 random forests, 20 and support vector machines 21 do not require any prior assumptions and are more effective in describing complex input-output mapping relationships. In recent years, machine learning methods have been used to replace traditional methods to solve the fatigue life analysis problem by more and more scholars.…”