PSI - Issue 70

N. Vignesh Kumar et al. / Procedia Structural Integrity 70 (2025) 453–460

460

Gaussian Process Regression (GPR) outperformed all other models achieving the highest prediction accuracy for both CFRP and GFRP confined concrete with R2 value greater than 0.9. Sensitivity analysis across all the 5 models identified the FRP jacket thickness tfrp and the unconfined concrete strength fco as the most influential parameters. The explicit usage of strain efficiency factor 0.31 in calculation of ffrp also contributed significantly in the strength reduction. A key enhancement of this study involved leveraging GPR’s inherent ability to assess prediction uncertainty. The analysis of 95% prediction intervals and error distributions for the test sets provided deeper insights into the model’s reliability. While the prediction intervals generally encompass the actual strength values, their width varied particularly for CFRP, indication regions of higher uncertainty that designers should consider. The error distributions revealed a slight tendency for the model to overpredict strength, through errors were predominantly clustered near zero. This explicit assessment of uncertainty significantly enhances the practical applicability of the GPR model, facilitating more risk-informed design decisions. In summary, this research article strongly supports their use of advanced ML models, particularly GPR, for FRP confinement designs, offering improved accuracy, valuable insights into parameter interactions, and crucial uncertainty quantification. Further work should focus on incorporating more detailed physical interactions between materials into models and validating these findings against fully independent external datasets to further confirm generalizability. Acknowledgement The authors would like to acknowledge the support provided by Mizoram University during the course of this research. Additionally, the authors gratefully acknowledge the financial support provided by Mizoram University through the Research Promotion Grant (SEED money) approved in December 2023 and received in March 2024. References Amin Mozumder, R., Roy, B., & Islam Laskar, A. (2016). Support Vector Regression Approach to Predict the Strength of FRP Confined Concrete. Arabian Journal for Science and Engineering, 42(3), 1129. Atf, Z., Safavi-Naini, S. A. A., Lewis, P. R., Mahjoubfar, A., Naderi, N., Savage, T. R., & Soroush, A. (2025). The challenge of uncertainty quantification of large language models in medicine. Cui, R., Yang, H., Li, J., Xiao, Y., Yao, G., & Yu, Y. (2024). Machine learning-based prediction of compressive strength in circular FRP-confined concrete columns. Frontiers in Materials, 11. Gharaei-Moghaddam, N., Arabshahi, A., & Tavakkolizadeh, M. (2023). Predictive models for the peak stress and ultimate strain of FRP confined concrete cylinders with inclined fiber orientations. Results in Engineering, 18, 101044. Hassanein, M., Y Hamed, A., A. Cashell, K., & Shao, Y. (2023). Confinement-based direct design method for fibre reinforced polymer confined CFST short columns. Thin-Walled Structures, 192, 111207. Ilyas, I., Zafar, A., Talal Afzal, M., Faisal Javed, M., Alrowais, R., Althoey, F., Mustafa Mohamed, A., Mohamed, A., & Vatin, N. (2022). Advanced Machine Learning Modeling Approach for Prediction of Compressive Strength of FRP Confined Concrete Using Multiphysics Genetic Expression Programming. Polymers, 14(9), 1789. Li, H., Yang, D., & Hu, T. (2023). Data-Driven Model for Predicting the Compressive Strengths of GFRP-Confined Reinforced Concrete Columns. Buildings, 13(5), 1309. Nain, M., Abdulazeez, M. M., & ElGawady, M. A. (2020). Behavior of high strength concrete – filled hybrid large – small rupture strains FRP tubes. Engineering Structures, 209, 110264. Napoli, A., & Realfonzo, R. (2020). Compressive strength of concrete confined with fabric reinforced cementitious matrix (FRCM): Analytical models. Composites Part C Open Access, 2, 100032. Realfonzo, R., & Napoli, A. (2011). Concrete confined by FRP systems: Confinement efficiency and design strength models. Composites Part B: Engineering, 42(4), 736. Shang, L., Isleem, H. F., Almoghayer, W. J. K., & Khishe, M. (2025). Prediction of ultimate strength and strain in FRP wrapped oval shaped concrete columns using machine learning. Scientific Reports, 15(1). Shayanfar, J., Barros, J. A. O., & Rezazadeh, M. (2021). Generalized Analysis-oriented model of FRP confined concrete circular columns. Composite Structures, 270, 114026. Valasaki, M., & G. Papakonstantinou, C. (2023). Fiber Reinforced Polymer (FRP) Confined Circular Concrete Columns: An Experimental Overview. Buildings, 13(5), 1248. Yue, C., Isleem, H. F., Qader, D. N., Mahmoudian, A., Tian, W., Jangir, P., Arpita, & Hamed, A. (2024). Utilization finite element and machine learning methods to investigation the axial compressive behavior of elliptical FRP-confined concrete columns. Structures, 70, 107681. Zarringol, M., Ishvarbhai Patel, V., & Quan Liang, Q. (2023). Artificial neural network model for strength predictions of CFST columns strengthened with CFRP. Engineering Structures, 281, 115784.