## Conclusions on Rotating Drum and Stirred Drum Bioreactors

The following conclusions can be made about the design and operation of rotat-ing-drum and stirred-drum bioreactors on the basis of the experimental work reported above:

• If a rotating-drum bioreactor is used, a decision needs to be made about the rotational rate. If a rotational rate greater than 10% of the critical speed is to be used, then it may not be essential to include baffles within the drum. However, it will require large power inputs to maintain the high rotation rate. The other option is to use quite low rotation rates but to baffle the drum in order to promote mixing.

• End-to-end mixing should be promoted by using curved baffles and inclining the drum axis. Our knowledge is not sufficient to allow detailed advice on the best way of designing curved baffles but obviously the inclination of the central axis must not be greater than the dynamic angle of repose of the solids.

• Discontinuous rotation of the drum or agitator will probably be of little benefit at large scale. Discontinuous rotation brings the added disadvantage of having to overcome inertia, both when starting and when stopping rotation.

Superficial velocity (m s" )

Fig. 8.15. Results of the residence time distribution studies of Hardin et al. (2001). (a) A descriptive model consistent with their results. (b) Effect of the superficial velocity, baffles, and fill depth on the exchange between the plug-flow and dead regions (characterized by the dimensionless variable Fmix) and the volume of the dead region (characterized by the dimensionless variable Vdead). Fmix is the volumetric exchange rate between the dead and plug-flow regions relative to the volume of the drum and the mean residence time, such that a Fmix of 1 would be equal to one volume of the drum exchanged per mean residence time. Vdead is the volume of the dead region relative to the total volume of the gas inside the drum. Key: Hollow symbols and solid lines represent an unbaffled drum. Solid symbols and dashed lines represent a baffled drum. The circles represent 26% filling, the triangles 19.5% filling and the squares 13% filling. Adapted from Hardin et al. (2001), with kind permission from John Wiley & Sons, Inc.

Superficial velocity (m s" )

Superficial velocity (m s" )

Fig. 8.15. Results of the residence time distribution studies of Hardin et al. (2001). (a) A descriptive model consistent with their results. (b) Effect of the superficial velocity, baffles, and fill depth on the exchange between the plug-flow and dead regions (characterized by the dimensionless variable Fmix) and the volume of the dead region (characterized by the dimensionless variable Vdead). Fmix is the volumetric exchange rate between the dead and plug-flow regions relative to the volume of the drum and the mean residence time, such that a Fmix of 1 would be equal to one volume of the drum exchanged per mean residence time. Vdead is the volume of the dead region relative to the total volume of the gas inside the drum. Key: Hollow symbols and solid lines represent an unbaffled drum. Solid symbols and dashed lines represent a baffled drum. The circles represent 26% filling, the triangles 19.5% filling and the squares 13% filling. Adapted from Hardin et al. (2001), with kind permission from John Wiley & Sons, Inc.

• Fractional fillings should not be more than 0.4 and may need to be less. In fact, the optimal fractional filling, that is, the filling that allows you to use as much of the drum volume as possible without compromising mixing too much, must be determined experimentally for each particular combination of substrate and microorganism.

• Our knowledge is not sufficient to allow detailed advice on the best design of mixers in the case of stirred-drum bioreactors.

0 0