Prediction of Behavior of Semi-rigid Composite Joints using Artificial Neural Networks

No Thumbnail Available
Date
2012-02
Authors
Daoud, Osama M. A.
Mohammed, Rania Salih
Journal Title
Journal ISSN
Volume Title
Publisher
UOFK
Abstract
This paper describes an artificial neural network (ANN) model developed to predict the moment-rotation response of semi-rigid beam-to-column composite joints for the full range of loading. Experimental data from 35 tests including results of tests performed by first author, cover the most frequently used types of semi-rigid beam-to-column joints with a composite metal deck floor, were used for training, testing and validating the neural network models. The data were arranged in a format represents the ultimate moment and geometric and material properties of the joints as an input and the output is the corresponding joint rotation. Aback-propagation artificial neural network model is developed and used to predict the moment-rotation response of the composite joints. Results were compared to the available experimental tests results and to a simple semi-empirical method proposed by the American Society Civil Engineers. Results indicate that the developed models can predict the nonlinear moment-rotation behavior of semi-rigid joints with a high level of accuracy.
Description
Keywords
Composite Joints, Semi-rigid, Artificial Neural Network, Rotation.
Citation