Identification of the State of a Magneto-Fluidized Bed by Pressure Fluctuations

Abstract

Local instantaneous pressure signals obtained through a magneto-fluidized bed have been analyzed using both classical and advanced signal analysis methods, which can deliver the necessary information about the presence of the bubbling and turbulent flow pattern. The conventional signal processing tool such as autocorrelation and cross correlation were used as preliminary tools to analyze the data. Evaluation of the dominant bubble frequency was completed using the autocorrelation function and power spectral density function. Mutual information function was used to identify the periodicity and the predictability of the local instantaneous pressure signals. Since it does not demand any particular functional relationships between the data points, it is a better method (compared to autocorrelation function) for measuring the predictability of nonlinear systems.