Assessment of magnetoelastic resonance parameters retrieval for sensor applications

Magnetoelastic resonance sensors have been widely used for several sensing applications, as their resonance behavior is very sensitive to different external factors and they can be operated remotely. Nevertheless, under some working conditions, such as when the sensor signal is low, has considerable noise, or the medium viscosity causes damping of the signal, the accuracy in obtaining the parameters that characterize the resonance response of these sensors by direct methods can decrease. The aim of this work is to improve the performance of magnetoelastic sensors through the use of numerical fittings of the resonance curves to obtain accurately the resonance parameters used for detection. The capability of these numerical fittings to retrieve the parameters when signals present different levels of noise is evaluated, and the accuracy of this fitting method is compared with the precision of classical direct methods.