Title: Prediction of spectral acceleration of a light structure with a flexible secondary system using artificial neural networks

Authors: S.P. Challagulla; C. Parimi; Jagadeesh Anmala

Addresses: Department of Civil Engineering, BITS-Pilani, Hyderabad, 500078, India ' Department of Civil Engineering, BITS-Pilani, Hyderabad, 500078, India ' Department of Civil Engineering, BITS-Pilani, Hyderabad, 500078, India

Abstract: The effect of a flexible secondary system (FSS) on the design spectra is studied in this paper. An FSS affects the primary structure (PS) during ground excitation differently than a secondary system that is rigidly attached to it. A modal superposition analysis is performed on a 2DOF system (PS + FSS). The influence of various parameters such as mass ratio and length of FSS on the design spectra is addressed. A methodology to estimate the spectral acceleration of the PS with FSS is developed. A design expression is proposed using statistical nonlinear regression and artificial neural network (ANN) to estimate the spectral acceleration of PS with FSS as a function of the structural period, length of FSS, and mass ratio. ANN modelling is proved to be more efficient than nonlinear regression model. An example calculation is shown for the estimation of spectral acceleration of the PS using ANN and nonlinear regression models.

Keywords: primary structure; flexible secondary system; FSS; design spectra; seismic zones; artificial neural network; ANN; nonlinear regression; NLR.

DOI: 10.1504/IJSTRUCTE.2020.109857

International Journal of Structural Engineering, 2020 Vol.10 No.4, pp.353 - 379

Received: 03 Oct 2019
Accepted: 04 Mar 2020
Published online: 28 Sep 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article