JSTE_V3_N1_RP4 Sasikumar T. Nanda Kumar Journal on Structural Engineering 2320 - 2343 3 1 20 26 Acoustic Emission, Graphical Method, Artificial Neural Network, Composite, Flexural Testing Three point bending test on cross ply laminates leads to the activation of distinct damage mechanism, such as matrix cracking, de-lamination between the adjutant plies and the fibre breakage at various stages of loading. This study deals with the investigation of the failure of the cross ply laminates using acoustic emission (AE). Broad band AE sensors monitor the elastic waves originating from different sources of failure in coupons while loading. AE signals for different failure modes have been analysed with respect to their parameters. Assessing the health condition of the composite laminate in advance to its ultimate failure while loading is the complimentary method and much useful to the composite industry. AE data acquired until 50% of the ultimate failure load was utilised for both graphical and artificial intelligence prediction of the approximate failure load of composite specimen within the safer limit, so that the structural integrity degradation during the health assessment test is significantly reduced. Both graphical and Artificial Neural Network (ANN) method of prediction have been successfully carried out and good correlation was found between them. Feed forward back propagation network with 22 middle layer neuron was able to map the pattern of failure present in the acoustic emission data and predict the ultimate failure load well in advance. March - May 2014 Copyright © 2014 i-manager publications. All rights reserved. i-manager Publications http://www.imanagerpublications.com/Article.aspx?ArticleId=2981