Visible to the public Biblio

Filters: Author is Mallik, Nilanjan  [Clear All Filters]
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z 
D
Mallik, Nilanjan, Wali, A. S., Kuri, Narendra.  2016.  Damage Location Identification Through Neural Network Learning from Optical Fiber Signal for Structural Health Monitoring. Proceedings of the 5th International Conference on Mechatronics and Control Engineering. :157–161.

Present work deals with prediction of damage location in a composite cantilever beam using signal from optical fiber sensor coupled with a neural network with back propagation based learning mechanism. The experimental study uses glass/epoxy composite cantilever beam. Notch perpendicular to the axis of the beam and spanning throughout the width of the beam is introduced at three different locations viz. at the middle of the span, towards the free end of the beam and towards the fixed end of the beam. A plastic optical fiber of 6 cm gage length is mounted on the top surface of the beam along the axis of the beam exactly at the mid span. He-Ne laser is used as light source for the optical fiber and light emitting from other end of the fiber is converted to electrical signal through a converter. A three layer feed forward neural network architecture is adopted having one each input layer, hidden layer and output layer. Three features are extracted from the signal viz. resonance frequency, normalized amplitude and normalized area under resonance frequency. These three features act as inputs to the neural network input layer. The outputs qualitatively identify the location of the notch.