Neural network-based seismic design method of nonstructural components located on various floors in buildings

This paper proposes an evaluation method of floor response spectrum based on machine learning with neural networks. The proposed method transforms target response spectra into the corresponding floor response spectra with dynamic characteristics of buildings and nonstructural components. We propose a neural network with two subnetworks where the input values are the damping ratio, natural frequency of structures and spectral acceleration. Numerical examples are demonstrated for steel buildings having five or ten stories. The predicted floor response spectra on any floor have good agreement with the results from time history analysis. © 2018 Architectural Institute of Japan. All rights reserved.