A distributed temperature and strain measurement method for OPPC in distribution Internet of Things in electricity based on multilayer feedforward artificial neural network

• Published in: Journal of physics. Conference series Vol. 2187; no. 1; pp. 12068 - 12074
• Main Authors: Guan, Guofei, Song, Qingwu
• Format: Journal Article
• Language: English
• Published: Bristol IOP Publishing 01.02.2022
• Subjects: Artificial neural networks; Conductors; Internet of Things; Measurement methods; Multilayers; Neural networks; Optical fibers; Physics; Signal to noise ratio; Strain measurement
• ISSN: 1742-6588; 1742-6596
• DOI: 10.1088/1742-6596/2187/1/012068

Abstract In order to effectively measure temperature and strain along opticalphase conductor (OPPC), the multilayer feedforward artificial neural network (ANN) is applied to demodulate the temperature and strain along OPPC composite optical fiber. The basic principle and parameters of ANN for this purpose are introduced. ANN is trained by using the numerically generated Brillouin spectra with different values of signal-to-noise ratio (SNR) and Brillouin frequency shifts (BFS), and the training results are presented. The Brillouin spectra with different values of SNR, temperature and strain along the optical fiber are numerically generated. The temperature and strain along the optical fiber are demodulated by the spectrum fitting method and the ANN method. The results reveal that the multilayer feedforward artificial neural network method has similar accuracy with the spectrum fitting method. However, the computation time of the former is much less than that of the latter.