ePrints@IIScePrints@IISc Home | About | Browse | Latest Additions | Advanced Search | Contact | Help

Damage Assessment of Composite Structures- a Fuzzy Logic Integrated Neural Network Approach

Ramu, Anantha S and Johnson, VT (1995) Damage Assessment of Composite Structures- a Fuzzy Logic Integrated Neural Network Approach. In: Computers and Structures, 57 (3). pp. 491-502.

[img] PDF
DAMAGE_ASSESSMENT.pdf
Restricted to Registered users only

Download (1146Kb) | Request a copy

Abstract

Advanced composite materials are recognized as better materials for structural components and they are being increasingly used in different industries. However, these materials are prone to damage, which may occur due to uncertainty in a large number of manufacturing variables and operational conditions. Because of their primary utilization in structural application and the present trend towards damage tolerant philosophy, the non-destructive evaluation of composites exposed to an uncertain environment has become increasingly important. The present research effort is directed towards incorporating the capability of handling uncertainties within the artificial neural network strategy for making assessments of damage in composites. Fuzzy representation and its associated logic are found to be the most efficient means of treating uncertainties. A novel method of integrating the concepts of fuzzy logic with ANN has been proposed. For the ANN, capability of a multi-layer perceptron with back propagation algorithm is made use of. The present network configured can detect, classify and estimate the extent of damage from the measured vibration response of the structure. The significant feature of the newly proposed fuzzy logic integrated approach is that it requires no additional efforts at the training level, but it can be used against fuzzy input at the end user level with the integration of the vertex method module. Also a modular development of the network is proposed which makes the training fast and efficient. The potential of the above approach is illustrated with two separate case studies to show that neural network approach is a very promising tool for the damage assessment task.

Item Type: Journal Article
Additional Information: Copyright of this article belongs to Elsevier.
Department/Centre: Division of Mechanical Sciences > Civil Engineering
Date Deposited: 22 Mar 2007
Last Modified: 19 Sep 2010 04:35
URI: http://eprints.iisc.ernet.in/id/eprint/9793

Actions (login required)

View Item View Item