Classification of Temperature-Strain Effects on Apodized Fiber Bragg Grating Sensor using Artificial Neural Network

An apodized Fiber Bragg Grating (FBG) sensor is introduced to analyze the importance of side lobe reduction in the reflection spectrum through apodization method for the estimation of measurands. The neighbouring FBG spectrums may intersect with one another as a result of the effects of measurands on account of the appearance of large-scale side lobes in the quasi-distributed sensing network. The Bragg wavelength varies relying upon the parameters could impact the period and the index modulation of the FBG. To estimate the parameters efficiently, it is highly desirable to differentiate even if the alterations on the Bragg wavelength are following by the consequences of the temperature or to the strain as both the parameters can influence the Bragg wavelength. The Artificial Neural Network (ANN) is introduced with different optimizers (like-Adamax, Adagrad, Adam and RMSProp) to classify the impacts of the temperature as well as the strain on an individual estimation of Bragg wavelength variation to exceed the obstacles of existent approaches. It has been reported that, the accuracy of the ANN with Adam optimizer is greater than the other optimizers for the discrimination analysis of the temperature as well as the strain.