A neuro-fuzzy approach to estimate the shear contribution of externally bonded FRP composites

The wide usage of fiber-reinforced polymer (FRP) composites as an external shear strengthening material have resulted in the development of various design guidelines. However, the accuracy of the design guidelines is a matter of concern since the shear behavior of the RC beam becomes more complex with the addition of FRP composites. On the other side, soft computing methods have been proved to be efficient enough to analyze complex systems. However, their application in structural engineering is limited. Therefore, in the current investigation, an effort has been made to evaluate the shear contribution of the EB-FRP composites with the help of adaptive neuro-fuzzy inference system (ANFIS). A total of 151 sets of data assembled from literature was utilized for the development and evaluation of the ANFIS model. The ANFIS predictions were validated against the obtained experimental results as well as with the estimations of six widely implemented design guidelines. The comparative study has shown that the ANFIS estimates are in decent agreement with that of experimental outcomes and the accurateness of the current ANFIS model is superior to the considered design guidelines. Finally, a parametric investigation was performed to explore the combined effect of various parameters as well as the impact of independent parameters.