Heterogeneous measurement system based on optical fiber and ultrasonic sensors to determine ethanol concentration

This work shows an heterogeneous measurement system to determine the ethanol concentration in ethanol-water blends. The system is composed by two sensors: an optical fiber refractometric sensor based on long-period grating and a sound velocity sensor based on a pair of ultrasonic transducers operating in transmission-reception mode. An artificial neural network was used to correlate the sensor responses and determine the ethanol concentration. The results demonstrate that the heterogeneous measurement system can predict, without ambiguity for a range between 0 and 100% v/v, the ethanol concentration with maximum absolute error of 0.55% v/v and 0.14% v/v mean squared error, in validation step.