Investigating Energy Absorption Accessible by Plastic Deformation of a Seismic Damper Using Artificial Neural Network

Abstract

In this study a seismic damper of high-rise structures is analyzed parametrically using artificial neural network (ANN). The input data for ANN model was generated using experimentally validated finite element (FE) analyses. The study investigates the amount of the absorbed energy dissipated by the plastic deformation of the tubes involved in the damper. The network used in this study computes the absorbed energy of the damping system in terms of three different variables including diameter ratio, the thickness and the diameter of the outer tube. To train the network, 90% of the FE results are utilized as input, and the capability of the network is examined by the rest 10% of data. It is shown that the trained neural structure can estimate the energy dissipation with an error less than 2%. According to the results, it is observed that despite the diameter, increasing in the thickness of the outer tube improves the energy absorption measurably. The results also show that the model with the diameter ratio of 1.6, as a critical design parameter, reflects the optimum absorbed energy among all cases.