This Report has been introduced to outline the ways in which current design could be improved by increasing the complexity and cost of the unit, but ultimately to operate more effectively and efficiently. It should be noted that this will be purely conceptual and may neglect some sections of the unit. Please refer back to the full design report for the budget unit wherever possible.

The best way to come up with a solution to a problem is to clearly define the problem or in this case sections of the unit that may have been somewhat compromised to achieve a more affordable overall cost.

Bioreactor walls

It was noted that the best performing material would be #4 grade electro polished 316L SS, due to its superior tensile and torsional strength, corrosion resistance and very well defined properties and responses in Bioprocessing.

To overcome the issue of removing the ability to carry out visual inspections a viewing glass can be fitted. This should if styled correctly provide a more premium feel to unit.

O2 diffusion into fermentation broth

To improve the levels of diffused oxygen in the unit would aid more efficient respiration of the bacteria and provided an effective control system was arranged to ensure these levels were maintained the effective yields of lipase production should be increased more easily maintained. The obvious choice would be to go for a conventional ring sparger arrangement providing somewhere near 0.5 vvm of air into the unit. However, considering the ever-present constraint of space, and the pursuit of finding creative solutions, something else may be considered.

Take the original magnetic stirrer with its impeller style blades providing vertical mixing. Mould the metal with channels running through it that pass to the trailing edge of the blades. The air can be forced out at a defined pressure dependant on outlet nozzle size, to achieve enough effective trust to power the impeller. This design would only benefit such a small unit and would likely only work on such a small impeller, however mixing and air supply could be combined into a single rotating blade. Again it should be noted that a control system would be required to ensure both effective O2 diffusion and mixing, if it is at all possible.

0.22 µm filter material

The first point to note would be that, based on the superior corrosion and abrasion resistance of ceramic membranes, it would be advisable that these be selected based on the predicted, more extensive lifetime the premium product should have. These filter membranes would also be able to withstand more aggressive cleaning procedures and hence should need to be cleaned on a less regular basis as a greater level of flux should be recovered each time.

Pall Corporation suggests the following cleaning arrangement for the removal of fats, oils and proteins form their ceramic membranes;
● Solution- 0.5M NaOH with 200ppm Cl2
● Time - 30 - 60 mins
● Temperature - 50oC
● Mode of action is hydrolysis and oxidation
(From table 6.4 page 284 Cheyran 1998 source: Pall Filtron 1995):

The above cleaning operation will be used in conjunction with fluid velocities of 2 m/s of the aforementioned NaOH solution through the membranes channels.

A cleaning cycle of 45 minutes has been selected; a 30 minute cycle time for a chlorine containing solution is optimum (Cheyran 1998 pp. 279), however the mean time suggested by Pall Corporation has been selected.

Reduced filter cleaning time

One limitation of the 0.22 µm filter is that it is, by nature, perpendicular to the flow, which means it is very difficult to clean without completely stopping the flow an hence operation, which would not be ideal. However, since the filter area required is significantly less than the area of the splitter wall between the tanks there should be the possibility of housing more than one filter within the wall that can quickly replace the filter that is currently in operation without human intervention.

The simplest way of doing this would be to have a number of filters housed in a carriage, much like bullets in a revolver. After a set period of time, or once the flow rate through the membrane drops below a set value (flow meter required) then the carriage would rotate to quickly replace the existing filter with the next ‘clean’ filter. Once the unit is stopped for cleaning, the carriage with filters can be removed, cleaned and returned.

Introduce an enzyme filter

As the unit stands there is total loss of the unused nutrients that pass through the 0.22 µm filter. To improve the efficiency of the unit, by recycling the nutrients and water currently in the system, hence concentrating the enzyme product stream, a cartridge filter operating in tangential flow can used. The filter will work by allowing water and nutrient s to pass through (permeate), however the larger enzyme molecules will not be allowed to pass through (retentate). The concentrated enzyme solution will remain in the filter cartridge and will be drawn out at a specified rate using a peristaltic pump. Again a control system would be beneficial. The permeate will pass into what is the current holding tank and will be fed back round to the fermentation tank via another peristaltic pump.

Based on the lipases molecular weight of approx. 50 kDa (http://www.uniprot.org/uniprot/Q9ZG91) a molecular pore size of 10 kDa would be ideal for use in the secondary filter to achieve a sufficient degree of separation without overly affecting the flux across the membrane.

INSERT SKETCH

Enzyme filter operation

Aside from the material of the filter, its operation can be optimised to achieve the most effective filtration. By inducing greater levels of turbulence of by using back-flushing operations as outlined below, it is very easy to improve the operating life of the filters between cleaning cycles and should allow for a higher average flux across a membrane surface of the same size for a given time.

Back-flushing

A common industrial practice during tangential flow filtration is known as lumen flush, this is where the permeate flow is shut off periodically for a ‘few’ seconds, forcing flow reversal through the membrane back into the feed flow, and in doing so dislodging particles in the membrane surface. A schematic of the pressure change across the unit and the flux vs. time as derived from Cheyran 1998 (page 268), compares standard operation to lumen flush below. As is visible in Figure A-1.1-I the flux through the membrane that can be achieved is increased due to the implementation of the lumen flush, periodically during a filtration cycle. Periodic back-wash and co-current permeate flow are also options that can be used to improve the flux through the membrane, however due the need to introduce additional pumping and pipework systems to the operation of the unit they will not be used, since the membrane cartridge is already situated inside the holding tank and there will not be space for such systems.

Permeate back pressure

During lab and pilot scale testing the critical ratio between wall shear and flux will be determined to ensure the greatest possible efficiency of the filtration units; this should be in the region of 0.8-1.2 Pa/LMH as with dairy feeds (Gesan et al., 1995). This testing will also provide the information required to determine the optimum permeate backpressure for the cartridge filter. During start-up the permeate stream valve will remain closed until the desired feed pressure has been reached, then the valve will be gradually opened, until the optimum backpressure has been established. This process will be included in the methods of operation of the units, and should lead to an increased cycle time for each unit.

Turbulence

It would be worth achieving greater turbulence, even at the expense of pumping power requirements due to the potential to increase fluxes by 50-300% (Cheyran, 1998 pp. 267). Such an increase in flux across the membrane surfaces will allow for increased throughput, should this be a requirement for more industrial sized units in the future.

Introduction of a pre-filter

The inclusion of larger particles in the inlet water may cause a number of fouling issues in the subsequent filtration processes. The increased fouling will then induce reduced efficiency as well as increasing the relative rates of necessary manual cleaning in each of the filtration processes.

In addition to concerns over the filter media itself an increased operational cost will also be required. This can be attributed to the increased level of pumping power needed to cope with the increased transmembrane pressure drop, thus requiring more expensive, more powerful pumps which inherently use more power.

Wire mesh sintered together can be used due to an aperture size range made to precision down to 1 µm. These meshes present a high capital cost investment, but are easily maintained (Svarovsky, 2000) and provide excellent corrosion resistance, throughput and easy maintenance, however an aperture size of 20µm would be too large when considering normal weaves. The use of the twill Dutch weave however provides an effective particulate retention size down to 1µm. This weave also produces a less fragile membrane arrangement than a standard wire woven membrane. It lends itself well to pressure filter leaves, due to this additional strength. The wire woven media tends to cost more than standard metallic media but has a long life, stretching and shrinking are non-existent, it is resistant to blinding and is easily cleaned. As such the ease of maintenance, cleaning and extended life warrants this media to be the best membrane selection, should a pre-filter be used in this unit.

Introduce a chlorine filter

One issue not yet considered on this design, primarily in the interest of cost, is the effect of using tap water containing chlorine has on the growth of bacteria compared to using a dechlorinated water source. The introduction of a dechlorination filter acting in a similar manner as a catalytic converter in a cars exhaust system could potentially provide the benefit of increased bacterial growth rates and thus enzyme yield. Further testing will be required to determine whether this would be economically beneficial for the unit.

Exponential feed flow rate

Using an exponential feed flow rate will mean that the optimum growth rate, with respect to lipase production, can be maintained for a longer period of time, thus improving yields and maximising the effectiveness of the nutrients in the unit.

One way flow valves

As mentioned in the budget report, one potential risk to the operation of the unit is having back flow occurring if the unit overfills, leading to a potential release of recombinant bacteria being into the existing pipe systems. One means of overcoming this is to pay a bit more to have one-way flow valves to prevent such back flows.

This unit will have more sophisticated control systems including level sensors and alarms, as well as controlled flow rates into and out of the unit. This should prevent such overfilling from occurring.

A back-up safety measure could be to install a 0.22 µm filter on the inlets and outlets of the tank to ensure that even if back flow occurs, the bacteria should not be able to pass into the external environment.

Intelligent enzyme release mechanism

The budget unit has been designed to release the enzyme product stream in a continuous fashion via gravity. With the introduction of peristaltic pumps on this unit, it can be deemed that introducing a more intelligent release system could improve the effectiveness of the enzymes released into the piping, hence meaning that less will need to be produced overall – leading to a decreased unit size. The pump can be operated via a timer or by an app on the users phone to turn on the pump for say 5 seconds after the user has poured oil/molten fats down the drain.

Disposal of Bacteria

In the ideal situation additional containers should not be required for cleaning of the tank and destruction of the bacteria and its subsequent DNA held within. An internally roof mounted UVC light should be used to destroy all DNA prior to cleaning being carried out. A specific replaceable cartridge should be fitted to the base of the unit to collect the contents of the tank for safe disposal (at an external site) prior to cleaning of the filter membrane, whose waste should be safe to dispose of locally.

Safety for UVC module

The cylindrical UVC L.E.D. module when in use will be imperative to the safe running of the budget unit, however it must be considered that there is no means of determining if it is working correctly or not, since the stainless steel shield will protecting the owner from its rays. One method would be to have an external L.E.D. that lights up to alert the owner when the UVC L.E.D. has shorted since an electrical circuit will no longer be complete. Another simple method is double up the number of the UVC modules, having both run on separate circuits, to ensure if one breaks the other is still operational.