Premium
Hydrodynamic study of biogranules obtained from an anaerobic hybrid reactor
Author(s) -
Saravanan V.,
Sreekrishnan T.R.
Publication year - 2005
Publication title -
biotechnology and bioengineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.136
H-Index - 189
eISSN - 1097-0290
pISSN - 0006-3592
DOI - 10.1002/bit.20567
Subject(s) - fluidized bed , settling , reynolds number , mechanics , porosity , superficial velocity , granule (geology) , drag , particle size , logarithm , particle size distribution , mathematics , materials science , thermodynamics , physics , chemistry , flow (mathematics) , composite material , mathematical analysis , turbulence
The bed expansion characteristics of a fluidized bed containing bacterial granules have been studied. These biogranules were obtained from an anaerobic hybrid reactor, which uses biogranules (without carrier particle) in fluidized condition. The settling velocity study of biogranules has shown that the drag coefficient of biogranule is greater than that of the rigid particle at the same Reynolds number. A new correlation based on this finding has been developed. The bed expansion study has demonstrated that a linear relationship exists between the natural logarithm of bed porosity and the natural logarithm of upflow superficial liquid velocity for the bed containing either a particular fraction of biogranule size or biogranules with wide size distribution. For a fluidized bed having a particular granule size, the bed porosity, and liquid superficial velocity could be related by the classic equation suggested by Richardson and Zaki (1954). The characteristic parameter of this correlation, the slope of the line n , has been related with Reynolds number. The intercept of the line gave a smaller value than the unhindered settling velocity of the particle. For fluidized bed having wide size distribution, the characteristic parameter n could not be related to Reynolds number. But the correlation suggested for single biogranule size has been found to predict n value with an average error of 2.3%. © 2005 Wiley Periodicals, Inc.