Title: Failure of infilled frames – a study using artificial neural network

Authors: K.M. Mini, K. Subramanian

Addresses: Civil Engineering Department, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, Tamil Nadu 641 105, India. ' Civil Engineering Department, Coimbatore Institute of Technology, Coimbatore, Tamil Nadu, India

Abstract: A neural network model to determine the failure load and drift of infilled frames under lateral loading is developed in the present paper. The backpropagation neural network is used to evaluate the failure criteria on the infilled frames using the analytically generated data. Training of the network is done by considering the aspect ratio, number of bays, area of column, area of beam, grade of concrete, grade of steel used for the construction and a non-dimensional parameter λh as the input parameters. To validate the efficacy of the model, an experimental investigation was carried out and the results are compared with that obtained using the ANN model. The experimentation is carried out under the same conditions used for the generation of the analytical data. The agreement was found to be good.

Keywords: evolutionary computation; infilled frames; artificial neural networks; ANN; failure load; equivalent strut method; backpropagation; training; testing; lateral loading; design engineering.

DOI: 10.1504/IJDE.2010.032823

International Journal of Design Engineering, 2010 Vol.3 No.1, pp.77 - 96

Published online: 24 Apr 2010 *

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