Machine Learning Models to Predict the Static Failure of Double‐Lap Shear Bolted Connections

ALMUHANNA, H., TORELLI, G. and SUSMEL, Luca (2025). Machine Learning Models to Predict the Static Failure of Double‐Lap Shear Bolted Connections. Fatigue & Fracture of Engineering Materials & Structures. [Article]

Documents
35877:968298
[thumbnail of ffe.70019.pdf]
Preview
PDF
ffe.70019.pdf - Published Version
Available under License Creative Commons Attribution.

Download (4MB) | Preview
Abstract
This study investigates the potential of machine learning models to predict the failure load and mode of double‐lap shear bolted connections. Five algorithms were evaluated: adaptive boosting, artificial neural network, decision trees, support vector machines with radial basis function kernel, and k‐nearest neighbors. A dataset comprising 221 experimental and numerical tests with varying input parameters, including different grades of stainless and carbon steel, was used to train the models. Unlike previous studies, the inclusion of diverse materials enabled the development of more generalizable models. To address data limitations, reduce biases associated with data split, and mitigate overfitting, k‐fold cross‐validation was adopted instead of the conventional 80/20 split. Results show that both regression and classification models achieved high coefficients of determination across most algorithms. Adaptive boosting delivered the most accurate failure load predictions, while artificial neural network achieved the highest accuracy in classifying failure modes. The findings highlight the potential of well‐trained machine learning models to outperform traditional codified methods in accurately predicting the structural response of bolted connections, especially when trained on diverse datasets.
More Information
Statistics

Downloads

Downloads per month over past year

View more statistics

Metrics

Altmetric Badge

Dimensions Badge

Share
Add to AnyAdd to TwitterAdd to FacebookAdd to LinkedinAdd to PinterestAdd to Email

Actions (login required)

View Item View Item