Monday, February 13, 2012


Making the linkages between sub-systems is
important for overall resource efficiency
A fundamental principle of Cleaner Production is to optimise systems rather than individual system elements. When partitioning a system into sub-systems for analysis, it is important to draw a boundary around the broader system you are analysing to ensure that in dealing with resource efficiency problems in one part of the system, you are not introducing unintended consequences elsewhere. I will use the example of centrifugation in the brewing industry to show how resource efficiency impinges on other operational issues such as product quality and throughput, and how taking too narrow a view can result in unexpected problems. If a systems view is not taken, these problems are difficult to solve, and could actually result in a negative overall efficiency position for the system despite apparent efficiencies in an individual area.

Centrifugation is a common unit operation in industry and is used to separate materials of different densities through the application of centrifugal force. It is applied globally in the brewing industry to separate yeast from green beer, with the beer than passed on to additional clarification processes such as filtration. Disc stack centrifuges use conical discs stacked on top of each other, with the product to be clarified passing through the spaces between the discs. As the centrifuge bowl rotates at high speed, yeast is pushed against the inner surfaces of the discs and is forced downwards into solids pockets in the bowl. The clarified beer passes up the centrifuge and exits, while separated yeast accumulates in the centrifuge. This yeast is periodically removed by momentarily opening the bowl and then shutting it to prevent an excessive loss of beer.  This process (called “de-sludging”, “discharging” or “ejection”) is well illustrated by this short video
There are a number of ways to initiate a centrifuge discharge. One way is through the use of a timer, in which case the centrifuge bowl will open and shut at a pre-determined frequency. This is not really desirable, since it leads to variations in the concentration of the yeast slurry discharged and in the clarity of the centrifuged beer. The second is through the use of in-line instruments which measure the clarity of the beer passing through the centrifuge, and then prompt a discharge when this clarity level becomes unacceptable i.e. when there is too much yeast in the beer exiting the centrifuge. Turbidity meters are an example of a typical instrument that can be used in this application, and work by measuring the amount of light absorbed by the beer, which is considered to be inversely proportional to its clarity. The turbidity level in the beer exiting a centrifuge tends to increase in the periods between discharges as the bowl fills with yeast and more and more yeast is carried over into the clarified beer stream.
From a resource efficiency point of view, the objective here is to discharge the yeast from the centrifuge in as concentrated a form as possible while meeting the clarity requirements of the beer being processed. The more concentrated the yeast is, the less beer there is associated with the yeast, and beer is the valuable resource we wish to preserve in this process. It is also likely that any beer associated with waste yeast will ultimately be disposed of to the brewery effluent system, which would either impose an additional load on the brewery’s wastewater treatment plant or that of the local municipality. 
Centrifuges can typically achieve a maximum yeast concentration of 80% by mass in the centrifuge discharge, and it is common practice to dilute this yeast with water to permit pumping.  It is a simple matter to measure yeast concentration in the diluted yeast slurry (through centrifugation in a laboratory) and then to use the alcohol concentration of the supernatant, together with the alcohol concentration of the beer being processed to back-calculate the yeast concentration at the centrifuge discharge. This should generally be done at defined volume intervals over the course of a batch, along with measurements of the yeast count in the incoming beer (expressed as 10^6 cells/ml).   

The factors (outside of design) that impact on the efficiency of a centrifuge are many, and include:
  • Beer flow rate – higher flows reduce the residence time in the centrifuge, which reduces separation efficiency. Also, a centrifuge has a maximum solids (i.e. yeast) removal capacity by virtue of the capacity of the bowl and the maximum discharge frequency permitted. If the rate of solids inflow exceeds this maximum removal capacity, there will be excessive yeast carryover into the clarified beer. The solids inflow rate depends on the beer flow rate and the yeast concentration in the beer being processed. This latter factor can vary significantly over the course of a batch – see my comments about this later.
·    Inlet pressure – high inlet pressures can lead to excessive losses of product, since effectively the volume discharged increases above that desired. Yeast and clarified beer are forced out of the bowl during discharge instead of yeast only.
·    Turbidity set point – clearly, higher set points promote a thicker yeast discharge from the centrifuge. Remember however that the clarity requirements still need to be met.
·    Discharge volume – modern centrifuges are equipped with timers which control the volume discharged (obviously assuming a fixed beer inlet pressure).
·    Operating water pressure – there needs to be sufficient pressure to keep the bowl closed when required, and to close it as quickly as required.
There are also equipment-related factors that can cause the bowl to leak, though such problems can be detected by alarms (e.g. an alarm indicating low bowl speed due to the braking action of the leaking product or an alarm highlighting a reduced back-pressure) and of course, the entire control philosophy employed relies on instruments that are working correctly. A useful way to track centrifuge losses is to measure the frequency of operation of the yeast removal pump, with frequent running indicating large losses. It should be obvious from these points that the various controls available in operating a centrifuge have to be considered in concert.

So what has all this got to do with the need for a systems approach you ask? Centrifuges are typically employed as part of a much bigger process system. For example, consider the use of a centrifuge downstream of fermentation in the lager brewing process. As shown below, the centrifuge would remove yeast remaining in fermented beer, which is then transferred into the storage/maturation process. After storage, the beer is typically filtered and then filled into containers.

Maximisation of the yeast concentration in the discharge of the centrifuge would require the use of a high turbidity set-point, implying significant carryover of yeast into the clarified beer stream.  While some yeast is required for the storage/maturation process to remove residual oxygen and for continued flavour development, too much yeast can cause a range of quality problems, including:
  • Beer haze (cloudy beer and which results in excessive filter circulation, reduced throughput and increased energy and powder consumption at filtration);
  • Foam problems (since yeast contains proteolytic enzymes which can be released into the beer as the yeast autloyses – beer foam is principally comprised of high molecular weight protein materials);
  • Yeasty off-flavours in the final product, which affect drinkability.
Yeast transferred to storage along with the beer being processed will continue to sediment out, and a portion (that which settles out) is typically removed prior to filtration. However, some yeast will remain in suspension and excessive yeast carryover from the centrifuge will translate into a higher yeast count at filtration, and hence a more rapid increase in filter inlet pressure than would be the case with lower yeast counts (assuming the brewery uses constant-flowrate powder filtration). This results in increased product loss at filtration, and overall, losses could therefore be far more than the product “saved” through operating the centrifuge at a high outlet turbidity. In addition to product loss, filtration problems would increase the usage of filter powder (a non-renewable resource), increase the cleaning cycles required per unit volume of beer filtered (incurring increased cleaning chemical and energy usage) and would also increase the electrical energy required (as pressure drop across the beer filter bed increases due to increased yeast load, so does the power required to maintain the required beer flowrate). The point is that the "optimisation" of a centrifuge in isolation can be to the significant detriment of the system within which that centrifuge operates.
Taking this thinking further, this "systems approach" should also extend upstream of the process being optimised. Yeast management practices upstream of the centrifuge can have a significant impact on the resource efficiency of the centrifugation process if we consider that the amount of yeast carried over to storage/maturation can be considered a limiting factor, or constraint. The extent of this carryover depends on the turbidity setpoint employed, as discussed earlier, but also depends heavily on the amount of yeast that is in the beer being processed. Yeast flocculates and then sediments to the bottom of the fermenter once the nutrients required for yeast growth are removed from the green beer. The majority of this yeast is typically removed some time before the beer is transferred through the centrifuge.  Immediately prior to transfer, it is good practice to remove additional sedimented yeast. If this is not done, the first portion of beer being centrifuged will contain an excessive amount of yeast, and even with discharges at maximum frequency, a centrifuge will not be able to achieve the targeted turbidity level since the bowl will fill very quickly. The result could therefore be excessive yeast carryover. As a consequence, the only way that the overall yeast count at maturation could then be achieved would be to run the centrifuge at a low outlet turbidity for the remainder of the transfer, which translates into discharge at low yeast concentration, excessive ejections and therefore excessive loss of product. One control philosophy could be to use an algorithm to reduce the beer flow rate based on a combination of outlet turbidity and discharge frequency (discharge volume could also be varied depending on the capabilities of the centrifuge concerned), but it is still necessary to use good yeast management practices, since otherwise overall brewery throughput could be compromised.  This is particularly true where the transfer process is a constraint.

In conclusion, process losses at breweries are not just about minimising losses in each area, but about finding the right balance between individual areas such that the losses in the overall system are optimised. This logic extends beyond the few limited areas discussed here, and can in fact even be extended to the balancing of  losses in the manufacture of brewing raw materials such as malted barley with how those losses impact on subsequent process losses in brewing. The value of a systems approach in resource efficiency cannot be underestimated, and I will illustrate this further in future posts using examples from other industries. The implications for manufacturing organisations as regards how sustainability performance is measured and managed are fascinating and often counter-intuitive.

Copyright © 2013, Craig van Wyk, all rights reserved