Modelling and integrating of experimental analysis for predicting the parameters of kenaf fibre-reinforced concrete beam-column joint

Abstract

To lessen the environmental impact of infrastructure projects, the construction sector has recently demonstrated a growing interest in sustainable materials. Kenaf fibre-reinforced concrete (KFRC), which has considerable mechanical qualities and biodegradability, has emerged as a possible eco-friendly substitute. The intricate interactions between material composition, geometrical factors, and load-bearing capacities make it difficult to optimise the design of structural parameters of KFRC beam-column joints. The beam column joints used in this study were designed based on ACI 318-19 shear criteria. This study suggests a novel method for precisely predicting the parameters of kenaf fibre-reinforced concrete beam-column (KFRC-BC) joints by combining machine learning modelling and experimental investigation. Experimental data were carefully documented to establish the reality, including load-displacement responses and beam-column joint parameters such as shear, stiffness, ductility, crack load, energy absorption, and ultimate load. These data were used in the modelling through GeneXproTools 5.0 and an empirical relationship with mathematical expressions has been proposed for each joint parameter. R2 statistical analysis is used to evaluate the model's efficacy. Deep learning could predict precisely concrete structure parameters. The shear spacing could be increased by 25% to 50%. Concrete strength influences all these characteristics. Kenaf fibre increased joint shear load, load at first crack, stiffness, ductility, ultimate load, and energy absorption by 4.89% to 28.5%, 10.12% to 34.1%, 6.65% to 10.74%, 14.71% to 52.06%, 10.52% to 25%, and 10% to 50.99%, respectively. These findings show that machine learning has clarified performance in the prediction aspect and proposed high accuracy of joint parameters.