The aim of this study is to develop a modeling procedure based on using artificial neural networks (ANNs) for predicting the frequency-dependent behavior of a microwave split-ring resonator (SRR) used for the dielectric characterization of liquid samples. The SRR device was designed and fabricated using the inkjet printing technology and, then, calibrated by means of water/ethanol mixtures with varying concentrations. By observing the variations in the forward transmission coefficient (i.e., S21) of the studied microwave device, a frequency shift of the resonant frequency and variations in the magnitude of S21 were recorded, which were related to the ethanol volume fraction. Using this calibration data, an ANN-based model is developed, which takes the ethanol volume fraction as input feature and, then, predicts the SRR sensor resonant parameters. The accuracy of the ANN-based model is reported and discussed.
Artificial Neural Network Modeling of Microwave Sensors for Dielectric Liquids Characterization
Gugliandolo, Giovanni
Primo
;De Marchis, Cristiano;Battaglia, Filippo;Latino, Mariangela;Campobello, Giuseppe;Crupi, GiovanniPenultimo
;Donato, NicolaUltimo
2023-01-01
Abstract
The aim of this study is to develop a modeling procedure based on using artificial neural networks (ANNs) for predicting the frequency-dependent behavior of a microwave split-ring resonator (SRR) used for the dielectric characterization of liquid samples. The SRR device was designed and fabricated using the inkjet printing technology and, then, calibrated by means of water/ethanol mixtures with varying concentrations. By observing the variations in the forward transmission coefficient (i.e., S21) of the studied microwave device, a frequency shift of the resonant frequency and variations in the magnitude of S21 were recorded, which were related to the ethanol volume fraction. Using this calibration data, an ANN-based model is developed, which takes the ethanol volume fraction as input feature and, then, predicts the SRR sensor resonant parameters. The accuracy of the ANN-based model is reported and discussed.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.