Prediction of shear strength of reinforced concrete beams using Artificial Neural Network and evaluated by Finite Element Software

Abstract

ABSTRACT In this paper, the Artificial Neural Network (ANN) and the Adaptive Neuro-Fuzzy Inference Framework (ANFIS) are utilized to foresee the shear quality of Reinforced Concrete (RC) shafts, and the models are contrasted and American Concrete Institute (ACI) and Iranian Concrete Institute (ICI) observational codes . The ANN display, with Multi-Layer Perceptron (MLP), utilizing a Back-Propagation (BP) algorithm, is utilized to foresee the shear quality of RC pillars. Six vital parameters are chosen as info parameters counting: concrete compressive quality, longitudinal reinforcement volume, shear traverse to-profundity proportion , transverse support, compelling profundity of the bar and bar width. The ANFIS demonstrate is additionally connected to a database and results are contrasted and the ANN show and exact codes. The primary request Sugeno fuzzy is utilized in light of the fact that the resulting some portion of the Fuzzy Inference System (FIS) is direct and the parameters can be assessed by a basic minimum squares blunder technique. Correlation between the models and the observational equations demonstrates that the ANN display with the MLP/BP algorithm gives better expectation to shear quality. In addition, ANN and ANFIS models are more precise than ICI and ACI exact codes in expectation of RC bars shear quality. These study were verified by using powerful finite element program and gave suitable agreement with NNA approaches. KEYWORDS: Reinforced concrete beam, Shear strength, ANN , Adaptive neuro-fuzzy inference system, Finite Element software (Abaqus ).