Monitoring the fluidized bed granulation process based onS-statistic analysis of a pressure time series

Abstract

Pressure fluctuation measurements collected during the fluidized bed granulation of pharmaceutical granule have been analyzed using the attractor comparison technique denoted as the S-statistic. Divergence of the bed state from the reference during granulation is followed by a return to a condition statistically similar to the original state of the dry fluidized ingredients on drying. This suggests insensitivity of the S-statistic technique to the changes in particle size distribution occurring during the granulation process. Consequently, the monitoring of pressure fluctuations alone may provide an easily implemented technique for the tracking of granule moisture and process end-point determination.

Monitoring and control of coating and granulation processes in fluidized beds – A review

“…Measurements are easy to assess; they only require transmitters installed at appropriate locations, usually in the plenum chamber, as this location provides an overview of bubbles and particles movement throughout fluidization. Researches focusing on the hydrodynamic monitoring of fluidized beds and which are based on the processing of pressure fluctuation signals have been widely accepted by the international scientific community due to their great potential of application in the identification of agglomeration in the processes, such as: gasification and biomass combustion [9][10][11][12][13], coke production [14] and coating and granulation of solid particles in a fluidized bed [15][16][17][18][19][20]. The papers of Johnsson et al [4] and van Ommen et al [21] report many results on the monitoring of fluidization regimes employing various methods of analysis of pressure fluctuation signals, besides describing in detail such analysis.…”

“…Chaplin et al [16] monitored the fluid-dynamics of the granulation process in a fluidized bed applying the EARS technique (Early Agglomeration Recognition System), based on the pressure fluctuation signals on the bed and on the chaos theory. In this technique, the attractors are built based on the pressure fluctuation signals.…”

“…Nevertheless, when S is greater than 3, the initial moment of the agglomeration process or changes to the fluidization behavior can be identified. Monitoring the fluid bed granulation process, Chaplin et al [16] could observe the fluid-dynamic changes that occurred during the process granulation phase. The statistical parameter S detected the changes.…”

“…Therefore, it would be hard work to build a closed-loop control to adjust the process by only analyzing power spectra through of the identification of dominant frequencies [17,18]. The statistical processing of the pressure spectral distribution through the Gaussian spectral analysis [17,18] and processing of attractors via S statistic [16,20] has been shown to be very reliable and accurate to identify unstable behavior in the regime of fluidization. A disadvantage of using pressure fluctuation analysis is related to a rigorous field of signal analysis tools to extract the appropriate information of the system and that they are representative within a wide range of operating conditions.…”

“…Watano et al [27] 1993 Signals of the power consumption: time domain Changes in the fluidization regime were monitored by standard deviation of the signal fluctuations Park et al [30] 1999 Processing Pressure fluctuation signals: spectral analysis and chaos theory Power spectra and attractors change widely during the granulation showing the regime transitions Talu et al [15] 2001 Processing of signals stress fluctuation: spectral analysis and attractor reconstruction Spectral amplitudes and S-statistics showed the growth of particles and the alterations in the fluidization regime during the granulation Mafadi et al [23] 2003 Pressure drop across the bed Stable fluidization regime was maintained by ON-OFF control in the coating liquid flow Chaplin et al [16] 2005 Pressure fluctuation signals: attractor reconstruction S-statistic determined the fluid-dynamic regime transitions during the granulation and drying phases Moris et al [31] 2006 Pressure fluctuation signals: spectral analysis Power spectra changed drastically during the time evolution of the coating process Visnadi et al [31] 2007 Pressure fluctuation signals: spectral analysis Similar results obtained by Moris et al [31], but applied to granulation process Silva et al [35] 2009 Pressure fluctuation signals: Gaussian spectral analysis The initial moments of the defluidization phenomenon could be detected by strong oscillations of the mean central frequency below of the range stable Parise et al [17] 2011 Pressure fluctuation signals: Gaussian spectral analysis Similar results obtained by Silva et al [35], but applied to coating process Silva et al [18] 2011 Pressure fluctuation signals: Gaussian spectral analysis The fluid-dynamic conditions during the coating process was improved using PI control Nunes et al [19] 2011 Pressure fluctuation signals: spectral analysis Regime transitions could be evaluated during the coating process of a fine powder by modifications in the power spectra in a vibrofluidized bed Martín et al [20] 2011 Pressure fluctuation signals: time domain, spectral analysis, attractor reconstruction…”