A neural network based model is developed for the structural analysis of masonry infilled steel frames, which can account for the non-linearities in the material properties and structural behaviour. Using the data available from the analytical methods, an ANN model with input parameters consisting of dimension of frame, size of infill, properties of steel and infill was developed. It was found to be acceptable in predicting the failure modes of infilled frames and corresponding failure load subject to limitations in the training data and the predicted results are tested using the available experimental results. The study shows the importance of validating the ANN models in simulating structural behaviour especially when the data are limited. The ANN model was also compared with the available experimental results and was found to perform well.
K. Subramanian, Dr. Mini K. M., and S. Florence, J. Kelvina, “Neural Network Based Modeling of Infilled Steel Frames”, International Journal of Structural Engineering and Mechanics, vol. 21, no. 5, pp. 495-506, 2005.