An optical fibre sensor for in situ measurement of external species in fluids based on artificial neural network pattern recognition

An optical fibre sensor system capable of detecting contaminants (e.g. part icles, inorganic or organic species) in water and other fluids is reported. In this article experimental results are presented for a single optical fi bre sensor located at a distance of 150 m from the transmitter/receiver of the system. The fibre is addressed using an optical time domain reflectomet er (OTDR) in order to achieve the spatial resolution (along the fibre lengt h) necessary for this investigation. Novel signal processing techniques involving artificial neural networks and pattern recognition have been applied to the signals arising from the sens or in order to allow cross-sensitivity effects, e.g. from fouling due to ca lcification, to be extracted from the real measurand, e.g. alcohol content.