Shang Li, Isleem Haytham F, Almoghayer Walaa J K, Khishe Mohammad
School of Civil and Hydraulic Engineering, Xichang University, Xichang, 615000, China.
Department of Computer Science, University of York, York, YO10 5DD, UK.
Sci Rep. 2025 Mar 28;15(1):10724. doi: 10.1038/s41598-025-95272-8.
The accurate prediction of the strength enhancement ratio ([Formula: see text]) and strain enhancement ratio (ε/ε) in FRP-wrapped elliptical concrete columns is crucial for optimizing structural performance. This study employs machine learning (ML) techniques to enhance prediction accuracy and reliability. A dataset of 181 samples, derived from experimental studies and finite element modeling, was utilized, with a 70:30 train-test split (127 training samples and 54 testing samples). Four ML models: Decision Tree (DT), Adaptive Boosting (ADB), Stochastic Gradient Boosting (SGB), and Extreme Gradient Boosting (XGB) were trained and optimized using Bayesian Optimization to refine their hyperparameters and improve performance.Results demonstrate that SGB achieved the best performance for predicting [Formula: see text], with an R of 0.850, the lowest RMSE (0.190), and the highest generalization capability, making it the most reliable model for strength enhancement predictions. For strain enhancement prediction (ε/ε), XGB outperformed other models, achieving an R of 0.779 with the lowest RMSE (2.162), indicating a better balance between accuracy, generalization, and minimal overfitting. DT and ADB exhibited lower predictive performance, with higher residual errors and lower generalization capacity. Furthermore, Shapley Additive exPlanations analysis identified the FRP thickness-elastic modulus product (t × E) and concrete compressive strength ([Formula: see text]) as the most influential features impacting both enhancement ratios. To facilitate real-world applications, an interactive graphical user interface was developed, enabling engineers to input ten structural parameters and obtain real-time predictions.
准确预测纤维增强塑料（FRP）包裹椭圆混凝土柱的强度增强比（[公式：见原文]）和应变增强比（ε/ε）对于优化结构性能至关重要。本研究采用机器学习（ML）技术提高预测的准确性和可靠性。利用了一个由181个样本组成的数据集，该数据集来自实验研究和有限元建模，采用70:30的训练-测试分割（127个训练样本和54个测试样本）。使用贝叶斯优化对四个ML模型：决策树（DT）、自适应提升（ADB）、随机梯度提升（SGB）和极端梯度提升（XGB）进行训练和优化，以细化其超参数并提高性能。结果表明，SGB在预测[公式：见原文]方面表现最佳，R为0.850，均方根误差（RMSE）最低（0.190），泛化能力最强，使其成为强度增强预测最可靠的模型。对于应变增强预测（ε/ε），XGB优于其他模型，R为0.779，RMSE最低（2.162），表明在准确性、泛化能力和最小过拟合之间取得了更好的平衡。DT和ADB的预测性能较低，具有较高的残差误差和较低的泛化能力。此外，夏普利加法解释（Shapley Additive exPlanations）分析确定，FRP厚度-弹性模量乘积（t×E）和混凝土抗压强度（[公式：见原文]）是影响两种增强比的最具影响力的特征。为便于实际应用，开发了一个交互式图形用户界面，使工程师能够输入十个结构参数并获得实时预测结果。
