Real-time Seismic Damage Detection of Concrete Shear Walls Using Artificial Neural Networks

Abstract

Concrete shear walls are widely employed in buildings as a main resistance system against lateral loads. Early identification of seismic damage to concrete shear walls is vital for deciding post-earthquake occupancy in these structures. In this article, a method based on artificial neural networks for real-time identification of seismic damage to concrete shear walls was proposed. Inter-story drifts and plastic hinge rotation of concrete walls were used as the inputs and outputs of a MLP neural network. Modal Pushover Analysis was employed to prepare well-distributed data sets for training the neural network. The proposed method was applied to a five-story concrete shear wall building. The results from the network were compared with those obtained from Nonlinear Time History Analysis. It was observed that the trained neural network successfully detected damage to concrete shear walls and accurately estimated the severity of seismic-induced damage.