Shear strength prediction of RC beams using adaptive neuro-fuzzy inference system

In complex engineering problems, there are some inexact conceptions, or a lot of parameters which must be considered. Soft computing is an approach that successfully applied to solve such problems. Determination of fuzzy rules for many problems has not been quite possible by an expert human. In this case, a neuro-fuzzy system which is the combination of neural network (for its ability to learn by datasets) and fuzzy system (for solving the drawback of the neural network) can be enhancing the performance of the system with several parameters or complex conditions. This paper shows the capability of a neuro-fuzzy system namely ANFIS to predicting the shear strength of reinforced concrete beams with steel stirrups. For this propose, the collection of laboratory results which was published in literatures used to train and finally test the proposed system. For this purpose, the sub-clustering approach (SC) applied for generating ANFIS. The results indicated that the considered neuro-fuzzy system was able to predict the shear strength of the RC beams which have been reinforced with steel stirrups.