An Investigation on Residual Stress Distribution in Cold-Formed Steel Channel Sections.
Saved in:
| Title: | An Investigation on Residual Stress Distribution in Cold-Formed Steel Channel Sections. |
|---|---|
| Authors: | Panditaray, Alok Kumar, Naik, Bishal, Madhavan, Mahendrakumar Mathialagu, Muvvala, Gopinath |
| Source: | Journal of Structural Engineering; Feb2025, Vol. 152 Issue 2, p1-20, 20p |
| Subject Terms: | RESIDUAL stresses, COLD-formed steel, MACHINE learning, STRUCTURAL design, TENSILE strength, X-ray diffraction, FINITE element method |
| Abstract: | This study investigates the behavior of residual stresses (RS) in cold-formed steel (CFS) unlipped press braked channel (PBC) sections. A nondestructive X-ray diffraction (XRD) technique was employed to characterize and quantify residual stress patterns in sections with thicknesses of 1.5 and 2.5 mm. Residual stress measurements were conducted on specimens postbending through the press-braking process, encompassing 1,116 measurement points across 54 cross sections for both thicknesses. The findings revealed that maximum RS, including both tensile residual stress (TRS) and compressive residual stress (CRS), were concentrated at bent locations. However, the overall stress distribution exhibited an irregular pattern, visualized through scatter plots for each thickness across the cross sections. An extreme gradient boosting (XGBoost) machine learning (ML) model was developed to predict residual stress patterns in the CFS sections with high accuracy and precision, effectively capturing self-equilibrating stress distributions. The XGBoost model outperformed other ML models, including multilayer perceptron (MLP), decision tree, and random forest regressors. Additionally, a finite element (FE) numerical simulation using ABAQUS was conducted to predict the impact of ML predicted RS on the axial strength of press-braked channel columns of various lengths, validating its accuracy in assessing the effect of RS on the axial strength of press-braked unlipped channel columns. The results were compared with predictions from the direct strength method (DSM). DSM predictions closely aligned with finite element results, both in failure modes and ultimate load. FE simulations showed failure due to local buckling (L) in stub columns, while long columns exhibited the interaction between local and flexural buckling (L+F), consistent with DSM predictions. The largest reduction in axial load capacity was observed 16% to 17% in the 1.5 mm thickness column which is aligned well with existing research. This study highlights the significant impact of residual stress distribution on the structural performance of CFS press braked channel sections, providing valuable insights for improving their structural integrity. Practical Application: This study provides valuable insights into how residual stresses (RS) affect the strength of cold-formed steel (CFS) unlipped channel sections manufactured by press braking. Using X-ray diffraction (XRD) techniques and machine learning (ML), the research shows how RS varies across different thicknesses and locations in the section. The results indicate that RS significantly influences the axial strength of these members, especially in longer columns. Machine learning models accurately predicted RS patterns, which were then used in ABAQUS finite element simulations to assess structural performance. The simulations showed strong agreement with established design methods (direct strength method) and existing literature, confirming the reliability of the approach. This work helps engineers better understand and predict the behavior of CFS structures under load, especially when residual stresses are present. Incorporating ML-predicted RS into design can lead to more accurate assessments of strength and safety, particularly for slender columns. The findings support safer, more efficient design practices in steel construction and provide a path forward for integrating data-driven tools in structural engineering. [ABSTRACT FROM AUTHOR] |
| Copyright of Journal of Structural Engineering is the property of American Society of Civil Engineers and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Complementary Index |
Nájsť tento článok vo Web of Science Be the first to leave a comment!