Prediction of obsolete FBG sensor using ANN for efficient and robust operation of SHM systems

摘要

Increased use of FRP composites for critical load bearing components and structures in recent years has raised the alarm for urgent need of a comprehensive health mentoring system to alert users about integrity and the health condition of advanced composite structures. A few decades of research and development work on structural health monitoring systems using Fibre Bragg Grating (FBG) sensors have come to an accelerated phase at the moment to address these demands in advanced composite industries. However, there are many unresolved problems with identification of damage status of composite structures using FBG spectra and many engineering challenges for implementation of such FBG based SHM system in real life situations. This paper details a research work that was conducted to address one of the critical problems of FBG network, the procedures for immediate rehabilitation of FBG sensor networks due to obsolete/broken sensors. In this study an artificial neural network (ANN) was developed and successfully deployed to virtually simulate the broken/obsolete sensors in a FBG sensor network. It has been found that the prediction of ANN network was within 0.1% error levels.