Prediction of the bond strength of FRP-to-concrete under direct tension by ACO-based ANFIS approach

• Published in: Composite structures Vol. 282; p. 115070
• Main Authors: Pei, Zuan, Wei, Yufeng
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 15.02.2022
• Subjects: Adaptive neuro-fuzzy inference system; Ant colony optimization algorithm; Fibre-reinforced polymer; Fuzzy c-means clustering method; Fibre-reinforced polymer; Fuzzy c-means clustering method; Adaptive neuro-fuzzy inference system; Ant colony optimization algorithm
• ISSN: 0263-8223; 1879-1085
• DOI: 10.1016/j.compstruct.2021.115070

•A hybrid ACO-based ANFIS model was presented to predict the bond strength between FRP and concrete.•The ACO algorithm and fuzzy c-means clustering method were employed to obtain the optimal parameters of ANFIS.•A comparison was conducted between some existing empirical models and the proposed hybrid ACO-based ANFIS model. In this study, a hybrid model integrating the ant colony optimization (ACO) algorithm and fuzzy c-means (FCM) clustering method into the adaptive neuro-fuzzy inference system (ANFIS) was proposed to predict the bond strength between fibre-reinforced polymer (FRP) sheets and concrete surface under direct tension. Eight parameters including the compressive strength of concrete, maximum aggregate size, tensile strength of FRP, thickness of FRP, elastic modulus of FRP, adhesive tensile strength, length of FRP and width of FRP are employed as the inputs, and the bond strength is used as the output variable. A comparison was conducted between some existing empirical models and the proposed hybrid ACO-based ANFIS model. The results confirmed that the developed ACO-based ANFIS model exhibits greater accuracy than the other eleven models, with higher coefficient of determination (R2 = 0.97) and Nash-Sutcliffe efficiency index (NS = 0.97), and lower root mean squared error (RMSE = 1.29 kN), mean absolute error (MAE = 0.81 kN) and mean absolute relative error (MARE = 0.053), while according to the Akaike information criterion (AIC) index, the accuracy of this model lies in its considerable complexity compared to others.