Mixing Characteristics of a Reactor

 The data you obtained for this section was some of the best I have ever seen form a class. However there were some problems with the analysis. Too many of you took times for tmix which did not reflect your results. When we are dealing with an asymptotic curve for a system reaching a new equilibrium the time to reach  a strictly steady value can verge on the ¥. So we use a characteristic time for the system to substantially achieve its new equilibrium e.g. 80%. So we do not wait until our curve becomes a straight line, we take a point at which it has substantially reached the new position ± say 10 - 20% of the change. This is far more useful because it gives us a measure of the major differences in rate which are going to have a significant impact on the output from our system. If you are running a reactor you worry about the 90% of the output not the last 10% when you are trying to get it right. So many of the tmix values were far too large and this led to unreasonable values for Bo.
In the presentation of the data many of you used rather squashed graphs with small y axis and long x axes, this made it further difficult to assess the tmix accurately. You have to do this from the graph, not by looking at a table of figures, where it is very difficult to assess the relationship between time and rate of change. It would also have been better to plot all 3 curves for each condition on one graph so that you could immediately and directly compare differences in mixing behaviour. Some of you also used peculiar x-axis scales that made it difficult to view the data correctly
Only a few reports included the results of gas-hold up and even fewer discussed this data. A pity since this data complements the observations with KLa, and both relate to bubble  conditions.

Viscosity
 All the curves were a little strange, in that after starting with a  high viscosity at low shear and then decreasing, a rise in viscosity was observed at the highest shear rates. The most probable explanation for this behaviour is that at high rates the sedimented culture on the bottom of the vessel was stirred up again. The viscometer disc is unlikely to cause shear damage even at the highest rates used as several suggested. A trend which will give us a negative viscosity is strange, and it is not very meaningful to have a straight line with a negative slope!

Flow properties
 In the stirring experiment many of you apparently only obtained very low Re numbers indicative of laminar flow. However when you did the expt.  you could also remember that the agitation could be so vigorous that the liquid would splash out of the container. The reason for this discrepancy is that when using dimensionless numbers if the units don’t cancel out they become meaningless, so you have to be careful to make sure that the correct units are used. The measurements of viscosity that you made gave a result in mPa.s (=cP) so that the value you used should be x 10-3. Putting this on the bottom of the equation for Re will make the resultant values safely in the turbulent region!

Mass Transfer Of Gases
 The results here were excellent and you had no difficulty deriving useful values for KLa, pity that this was not usually followed up by a similar discussion. This experiment allowed you to examine the effects of changes in surface tension (increased by NaCl, decreased by detergent) on oxygen transfer, which are due to the changes in bubble dimensions caused by surface tension.
 

Simulation
The simple questions were all answered very well. However A lot of you were still confused by the dilution rate in question 2. Even through the biomass concentration drops the reactor does still produce more biomass because it’s volume has increased. However the volume efficiency of the reactor will be less.

The final question was the most important part of this report and the discussions were generally disappointing with few useful attempts to properly explain what was influencing the shape of the curves observed.

The calculation of the yields raised a slight problem in that I gave you the calculation with two terms inadvertently transposed so that the resultant sign was negative. However the difficulty of achieving a negative yield did present itself to some of you who converted the yield to a positive value.

Stirring rate
For the first 2 expts. with stirring rate and gassing rate a very similar curve should have been obtained for both product & yield with a rapid increase at first and then a levelling off  at high rates. Or stirring rate as it increases we get better mixing & aeration and therefore better productivity, however the production is still limited by supply of substrate and mixing a culture with nothing to eat cannot make it produce any better, so the curve will flatten out.

Gassing rate
There were some strange explanations for the drop in productivity at high gassing rates, which involved suggested that there was an increase in ethanol production. This is normally a consequence of anaerobic metabolism! Similarly to the stirring curve there will no longer be any benefit from high gassing rates when the cultures are already receiving sufficient oxygen to fully utilise the available substrate.

Substrate rate
 As for the other curves increases in substrate will give an increase in production until another factor, in this case, oxygen, becomes limiting. The difference here is that the yield curve will give the opposite curve since we are following the yield of product per unit substrate. At high substrate addition rates the culture will only produce a small amount of extra product for the addition of a large quantity of substrate, so the yield will be very low. This does not mean though that the culture is producing less or dying off, it simply means that it is not using the substrate that you are adding.
 
 

Dr. L. Ramsden
Botany 22/1/001