Shear Strength Prediction of Reinforced Concrete Shear Wall Using ANN, GMDH-NN and GEP

Abstract

To provide lateral resistance in structures as well as buildings, there are some types of structural systems such as shear walls. The utilization of lateral loads occurs on a plate on the wall's vertical dimension. Conventionally, these sorts of loads are transferred to the wall collectors. There is a significant resistance between concrete shear walls and lateral seismic loading. To guarantee the building's seismic security, the shear strength of the walls has to be prognosticated by using models. This paper aims to predict shear strength by using Artificial Neural Network (ANN), Neural Network-Based Group Method of Data Handling (GMDH-NN), and Gene Expression Programming (GEP). The concrete's compressive strength, the yield strength of transverse reinforcement, the yield strength of vertical reinforcement, the axial load, the aspect ratio of the dimensions, the wall length, the thickness of the reinforced concrete shear wall, the transverse reinforcement ratio, and the vertical reinforcement ratio are the input parameters for the neural network model. And the shear strength of the reinforced concrete shear wall is considered as the target parameter of the ANN model. The results validate the capability of the models predicted by ANN, GMDH-NN, and GEP, which are suitable for use as a tool for predicting the shear strength of concrete shear walls with high accuracy.

Keywords

• artificial neural network
• genetic algorithm
• group method of data handling
• reinforced concrete shear wall
• shear strength
• regression analysis