基于可解释性机器学习算法的FRP筋UHPC粘结强度预测

打开文本图片集

中图分类号：TU52；TB332 文献标志码：A 文章编号：1008-0562（2026）02-0185-11

Abstract：In order to explain the bond mechanism of fiber reinforced polymer （FRP） bars and ultra-high performance concrete （UHPC），a sample dataset of 475 bond strength tests results was analyzed. Seven input variablesand one output variable （bond strength）were selected to train six machine learning algorithms.Three interpretable techniques were combined for analysis，and compared with the specifications and empirical models. The results show that the extreme gradient boosting （XGBoost） model achieves the best accuracy，with R2 of 0.881，RMSEof 3.700 andMAEof2.326.The ratio of bond length to reinforcement diameter Id，steel fiber content ρSF， the ratio of protective layer thickness to reinforcement diameter c/d，FRP bar diameter d UHPC strength fc are the key factors influencing the bond strength.The results of this study can provide a reference for the prediction of UHPC bond strength of FRP bars.

Keywords： fiber reinforced polymer （FRP） bars;ultra-high performance concrete （UHPC）；bond strength; machine learning; SHAP analysis

0 引言

纤维增强复合材料（FRP）筋与普通混凝土结合形成配FRP筋普通混凝土结构时，构件通常表现出较大的挠度和裂缝宽度，较难满足正常使用极限状态的要求。（剩余14310字）