Evaluating the bond strength of FRP-to-concrete composite joints using metaheuristic-optimized least-squares support vector regression

Junfei Zhang, Yuhang Wang

Research output: Contribution to journalArticlepeer-review

11 Citations (Scopus)

Abstract

The reinforced concrete (RC) infrastructure can be retrofitted by adhesively bonding fiber-reinforced polymers (FRPs) to the tension face. In the FRP-to-concrete bonding system, the debonding of the FRP plate from the member is the most common failure type. Predicting the bond strength of FRP-to-concrete joints using traditional predictive models is far from being satisfactory because of the highly nonlinear relationships between the bond strength and a large number of influencing variables. To address this issue, this study proposes a metaheuristic-optimized least-squares support vector regression (LSSVR) model to predict the bond strength of FRP-to-concrete joints. The hyperparameters of the LSSVR model are tuned using a recently proposed beetle antennae search (BAS) algorithm. In addition, the Levy flight is incorporated in the BAS algorithm to improve its searching efficiency. The proposed model is then trained on a dataset collected from internationally published literature. To understand the importance of each input variable on the bond strength, the variable importance is calculated using the random forest algorithm. The results show that the proposed LBAS-LSSVR model has comparatively high prediction accuracy, as indicated by a high correlation coefficient (0.983) and low root mean square error (1.99 MPa) on the test set. Width of FRP is the most sensitive variable to the bond strength. The proposed model can be extended to solve other regression problems in structural engineering.

Original languageEnglish
Pages (from-to)3621-3635
Number of pages15
JournalNeural Computing and Applications
Volume33
Issue number8
Early online date30 Jul 2020
DOIs
Publication statusPublished - Apr 2021

Fingerprint

Dive into the research topics of 'Evaluating the bond strength of FRP-to-concrete composite joints using metaheuristic-optimized least-squares support vector regression'. Together they form a unique fingerprint.

Cite this