Modeling the Signals Collected in Cavitation Field by Stochastic and Artificial Intelligence Methods

Studies have shown that ultrasonic cavitation can give birth to electrical signals in liquids. The properties of these signals depend on the liquids’ nature, the ultrasound frequency, and the generator's functioning power. In this article, the signals collected in water in an experiment where the ultrasound generator worked at a frequency of 20 kHz and 80W are characterized by statistical methods. An autoregressive moving average (ARMA) and a generalized regression neural network (GRNN) models are built for this signal and comparisons are provided.