Coupling of support vector machine and acoustic models to characterize the droplet size distribution of emulsions using ultrasonic techniques

The application of ultrasonic techniques to characterize emulsions concerning the droplet size distribution (DSD) has attracted interest in recent years, mainly due to the advantage of being a non-intrusive technique able to characterize concentrated and opaque emulsions. These characterizations occur by adjusting sound attenuation spectroscopy with the acoustic models. The most used models for this purpose are simplifying more complete models, such as ECAH, mainly due to their mathematical complexity and convergence problems. However, most of these simplified models have restrictions related to the range of applications in the wave propagation regime. Therefore, the objective of this work was to couple the acoustic models to the support vector machines (SVM), which have numerous advantages linked to their successful adaptability to nonlinear data, to acting on this limitation, thus improving the determination of the emulsion DSD. This coupling proved to be satisfactory since the SVM classified all data correctly.