NARX neural network model for strong resolution improvement in a distributed temperature sensor

Autor: Jean Carlos Cardozo da Silva, Luís C. B. Silva, Marcelo E. V. Segatto, Maria José Pontes, Jorge Leonid Aching Samatelo, Cicero Martelli, João Paulo Bazzo
Rok vydání: 2018
Předmět:
Zdroj: Applied optics. 57(20)
ISSN: 1539-4522
Popis: This paper proposes an approach to process the response of a distributed temperature sensor using a nonlinear autoregressive with external input neural network. The developed model is composed of three steps: extraction of characteristics, regression, and reconstruction of the signal. Such an approach is robust because it does not require knowledge of the characteristics of the signal; it has a reduction of data to be processed, resulting in a low processing time, besides the simultaneous improvement of spatial resolution and temperature. We obtain total correction of the temperature resolution and spatial resolution of 5 cm of the sensor.
Databáze: OpenAIRE