Document Type : Research Article
Authors
1
- Faculty of Civil Engineering, Ho Chi Minh City University of Technology (HCMUT), 268 Ly Thuong Kiet Street, District 10, Ho Chi Minh City, Vietnam. - Vietnam National University Ho Chi Minh City, Linh Trung Ward, Thu Duc City, Ho Chi Minh City, Vietnam.
2
- Faculty of Civil Engineering, Ho Chi Minh City University of Technology (HCMUT), 268 Ly Thuong Kiet Street, District 10, Ho Chi Minh City, Vietnam. - Vietnam National University Ho Chi Minh City, Linh Trung Ward, Thu Duc City, Ho Chi Minh City, Vietnam
10.24200/sci.2025.65252.9382
Abstract
The study develops a machine learning technique to predict structural deformation in corroded reinforced concrete (RC) beams, enabling a more accurate assessment of building structural health and potential long-term degradation risks. The method combines least squares support vector regression (LSSVR) with an innovative opposition sea-horse optimizer (OSH) to handle the nonlinear and multivariable aspects of deflection prediction. The OSH optimizer enhances the LSSVR's performance by fine-tuning its learning process. The research validated the predictive approach by analyzing 150 samples from deteriorating residential structures in southern Vietnam, employing cross-validation techniques to verify the model's precision and reliability. The OSH-LSSVR approach significantly surpasses traditional predictive models, including artificial neural networks, multivariate adaptive regression splines, and support vector regression. Empirical evaluation reveals exceptional performance metrics, with a root-mean-square error of 1.896 mm, mean absolute error of 1.198 mm, and a coefficient of determination of 0.891, underscoring its advanced predictive capabilities. The developed model provides civil engineers with an advanced tool for predicting RC beam deflection, opening new avenues for research in structural optimization, early warning systems, and proactive safety strategies.
Keywords
Main Subjects
)