Prediction of particle residence times in cascading rotary dryers
The objectives of this research were to provide a better understanding of particle
motion in cascading rotary dryers. This would lead to more soundly based design
procedures. Experiments were performed, to check the validity of a proposed design
method for dryers operating in the under-loaded and design loaded conditions developed
by Matchett and Baker, on a pilot plant rig at Teesside Polytechnic using wheat and
sand in the absence of airflow. The model considers the particles to move in two
parallel phases, the airborne phase contains the material in flight and the dense phase
contains the remaining material which is caught on the flights and on the bottom of the
drum. There is continuous Interchange of material between the two phases. A
dimensionless number, dense phase velocity number 'a', has been defined which is a
measure of the axial velocity of the material in the dense phase of the drum. The 'a'
values were found to be in agreement with existing data and were found to be dependent
on material and not on dryer speed or slope.
Photographic studies of the dryer internals suggested that the assumption of a constant
0 value (measure of flight loading) In the original model was not valid and that 4)
varied with number of flights. A model was developed to predict 0 which worked
extremely well for large number of flights. The existing design model was therefore
modified to take account of the variation In 4). However, the paired t-test Indicated
that at 5% level of significance there was no difference between the original and the
modified model, even though the modified model is physically more realistic. It is,
however, recommended that the models be tested on a large number of flights and also
large equipment, because It is expected that with a large number of flights there will
be differences between the two models and the 0 model will be superior. The 'a' and am
(the am value is a modified form of the 'a' value which takes into account the variation
in flight loading) values were found to be Independent of operating conditions, flight
angle and also dryer size but were dependent on material. The 'a' and am values were
proportional to 1/number of flights. Particle motion in the dense is by bouncing,
rolling and sliding, but the high dense phase velocity numbers obtained with zero
flights (ar) suggested ii
that rolling and sliding are the important mechanisms of the dense phase motion and
may be far more important than bouncing.
A model has also been developed to study the over-loaded regime. In the over-loaded
regime It was found that dryer speed, slope, material and number of flights affected
the dense phase motion and a simple relationship between the over-loaded dense phase
velocity number (ao) and number of flights could not be developed with the limited
data. Particle motion In the over-loaded regime was found to be very complex. The ao
values could be predicted to within ± 35%. Estimates have been made of the transition
holdup, marking the change from under-loaded to over-loaded behaviour, but It was
found that the prediction of the transition holdup is also complex and could be
predicted to within ± 45%. The am values could be predicted to an accuracy of ± 10%.
Thus suggesting that the ao and the transition holdup numbers are not so reliable.
Future work has been recommended particularly in the over-loaded regime and also
on the transition region since it was found that the particle motion in these regions
was complex. It has also been suggested that the models be tested in large Industrial
units with and without air flow.