Interpreting complex data from a dual element multipoint optical fibre sensor system for use in process water systems

D. King, W. B. Lyons, C. Flanagan, E. Lewis

Research output: Contribution to conferencePaperpeer-review

Abstract

A dual element multipoint optical fibre sensor system capable of detecting ethanol in water supplies is reported. A U-bend configuration is used for each sensor element in order to maximize sensitivity and the sensor system is interrogated using Optical Time Domain Reflectometry (OTDR). It has been proposed to apply Artificial Neural Network Pattern Recognition Techniques to the optical fibre sensor system to accurately classify each sensor element test condition. Novel signal processing techniques utilising Fourier Transform methods are applied to the resulting OTDR data, with the aim of reducing the computational resources of the ANN required to accurately classify the sensor system data.

Original languageEnglish
Pages975-980
Number of pages6
Publication statusPublished - 2003
EventSmart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Complex Systems and Artificial Life - Proceedings of the Artificial Neural Networks in Engineering Conference - St. Louis, MO., United States
Duration: 2 Nov 20035 Nov 2003

Conference

ConferenceSmart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Complex Systems and Artificial Life - Proceedings of the Artificial Neural Networks in Engineering Conference
Country/TerritoryUnited States
CitySt. Louis, MO.
Period2/11/035/11/03

Keywords

  • ANN
  • Backpropagation Algorithm
  • Discrete Signals Fourier Transform
  • Fast Fourier Transform
  • Feedforward Neural Networks
  • Multi-Layer Perceptron
  • Optimisation
  • Pattern Recognition
  • Sensor Recognition
  • Signal Processing
  • Supervised Training

Fingerprint

Dive into the research topics of 'Interpreting complex data from a dual element multipoint optical fibre sensor system for use in process water systems'. Together they form a unique fingerprint.

Cite this