Impact of Vapor Bypasses on Separation Performance of Rotating Packed Beds in Distillation

The impact of different sealing types and their wearing
on the
separation performance in distillation processes in rotating packed
beds (RPB) has been investigated, which potentially causes bypass
flows inside the unit. In the literature, a maximum separation performance
is observed, which can have several possible explanations. Next to
the theory that with increasing rotational speed a lower residence
time counteracts the mass transfer intensification, one theory is
that there are undetected bypasses from the casing of the unit to
the eye of the rotor. Besides the effect of the maximum, the latter
would indicate the fact that bypasses generally induce lower separation
performance in RPBs. However, with a suitable sealing concept, higher
separation performances could be achieved in the units. Therefore,
the influence of bypass flows on the separation performance was investigated
by installing different sealings in the RPB, which has not yet been
discussed. For this purpose, two sealing types, a labyrinth and a
floating ring sealing, were tested in two different RPBs of similar
design. The results showed that the labyrinth sealing was less effective
than the floating ring sealing in reducing the bypass mass flow. For
the floating ring sealing, the separation performance dropped with
increasing operation time due to abrasion and wearing. The results
also showed that, in general, with increasing rotational speed, the
tightness of the sealings is reduced, leading to the formation of
bypass streams that reduce the separation performance of the plant.
The study highlights the need for greater awareness of bypasses in
RPBs especially concerning the separation performance and emphasizes
the demand to develop more efficient and effective sealing systems
that are ideally nonabrasive, require no additional fluids, and have
higher tightness. These findings could call for re-evaluation of previous
results, packing evaluations, and derived models.