Novel modified ANFIS based fuzzy logic model for performance prediction of FRCM-to-concrete bond strength

Abstract

Construction repairs have used fiber-reinforced cement mortar (FRCM). Concrete and FRCM bond strength usually outweigh mechanical criteria. Nevertheless, testing complex bonds like the FRCM and concrete bond takes time, money, and errors. This study employed fuzzy logic (FL) based on the adaptive neuro-fuzzy inference system (ANFIS) to simplify and reliably estimate the FRCM-to-concrete bond strength (CBS) by modeling complicated and non-linear systems computationally efficiently. The models take six inputs: concrete splice length, stirrup cross-sectional area to spacing, longitudinal tension bar area to effective cross section, compressive strength, relative rib area, and minimum concrete cover. The model outputs concrete steel bar bond strength. The FLANFIS model predicts FRCM-to-CBS using the tensile testing results of the 10 specimens (5 concrete and 5 FRCM). Data instruct the model and measure its precision. This article defines the ANFIS-based FRCM-to-CBS. This research will employ 5 concrete specimens and 5 FRCM specimens, totaling 0.05 m^3 of concrete and FRCM mix. 0.25 kg of adhesive bonds the sample. Grip, control, and data gathering systems are employed with a 1 kN tensile testing equipment. This research comprises preparing concrete and FRCM specimens, bonding with adhesive, and tensile testing. The FL-ANFIS model predicts FRCM-to-CBS with a high coefficient of determination (R2) of 0.995 and a strong correlation coefficient (r) of 0.982 in training. The pattern predicted accurately with RMSE of 0.264 and MAE of 0.196. This paper shows that FL-ANFIS can predict steel bar bond strength in concrete quickly and accurately. The pattern reduces waste, design costs, and time.