Unambiguous Signal Processing and Measuring Range Extension for Fiber Bragg Gratings Sensors Using Artificial Neural Networks— A Temperature Case

This paper describes and discusses a novel approach that uses artificial neural networks (ANN) to extend the measurement range of fiber Bragg gratings (FBGs) interrogators based on fixed narrowband filter demodulation. Interrogators with fixed spectral filters use only one edge of the filter to demodulate the signal. The proposed method uses narrowband FBG´s filters, where both edges of the entire filter bandwidth are applied to demodulate the signal. Furthermore, the approach has the possibility to concatenate n filters, obtaining a measuring range of n times filter bandwidth of one FBG filter. The great advantage of this method relies on the use of ANN to combine these signals, mitigating the ambiguities created in the both edges of each FBG filter, and generating a continuous linear output along the measuring range. Despite of the proposed demodulation method can be used for strain measurements, it is yet in the temperature sensing that the method offers better contribution. Wide temperature measuring ranges are common in the petrochemical industry, such as distillation columns (0degC to 700degC), or in power plants to measure the temperature of transformer windings (0degC to 150degC). This paper presents experimental results for two different cases of a temperature measuring application in the range between 25degC and 250degC . These approaches consider the relative superposition effect of the concatenation method to extend the measuring range. The results are then analyzed according to the proposed solution characteristics and the relative superposition of the narrowband filters.