Engineering R&D: Anaerobic collaboration
Aerobic digesters were rejected as too large and impractical for the site, and the men concluded available technology would not meet their requirements. The solution was to combine the dairy’s expertise in ultra-filtration with the consultant’s knowledge of anaerobic digesters.
Vince Taylor, now president of Daisy Brand Inc., Dallas, and John Ewing, now with Biothane LLC, call their system an anaerobic membrane bioreactor, or AnMBR. A beta version began operating at Daisy’s Garland, TX facility in 2007. A larger system now treats 50,000 gallons a day and discharges effluent with COD of less than 300 mg/L. Commercial development of AnMBR for a variety of applications is being guided by Biothane’s Graig Rosenberger, manager of the renewable energy group.
A graduate of Penn State University with a bachelor’s degree in chemical engineering and a minor in environmental engineering, Rosenberger joined Biothane in 1999 as a process engineer after working in a similar capacity with Procter & Gamble. In his current position as process group manager, he oversees process design of core anaerobic technologies and directs product development efforts by a team of engineers and technologists.
FE: Was a leap of faith required by the dairy to install an anaerobic membrane bioreactor system at its cottage cheese plant?
Rosenberger: Both John Ewing and Vince Taylor had prior anaerobic experience, and the probability of complete failure was very low. The upside was the high flux rates the system would provide, and the addition of membrane technology addressed the classic concern over washing out the biomass in the digester. Most anaerobic systems have a clarification device to retain the microbes.
Daisy Brand is an extremely progressive and innovative company. Others had dabbled with AnMBR previously, but there hadn’t been any significant installations before the Garland project. When the company decided to add cottage cheese production, the decision to aggressively pursue this technology was made. There was no risk to production, and failure analysis concluded that if COD reductions weren’t sufficient, Plan B was simply to add more membranes.<br><br>
FE: Was some prototype work done for proof of concept?
Rosenberger: After a year of discussions and another year of construction, it went directly to a beta scale digester and membrane system, a $3 million “pilot” that could process 15,000 gallons of cheese whey a day. The dairy’s engineering staff used ceramic membranes in processing and was very comfortable with tubular cross-flow membranes. They selected cross-flow polymeric and ceramic technologies from two vendors. The anaerobic digestion system included a digester large enough to hold eight days’ worth of waste stream and an equalization tank for two days’ detention. The digester design included a proprietary zonal mixing system, foam control system and reactor tank for nutrient balancing and pH management.
FE: When did the beta system start up?
Rosenberger: August 2007. Initially, the flux flow rate was only one-eighth the rate claimed by the membrane vendors. Improvements in the controls and the digester-membrane interface improved flow to about one-third of vendor claims. Much of the work that has been done since has focused on further improvements in flow rate and refinements of the process set points.
More importantly, COD reductions exceeded expectations. Effluent entered the system with an organic load exceeding 60,000 milligrams per liter [mg/L]. The best expectation was 95 percent COD removal. Instead, 99.5 percent was achieved, with discharged water at 200 mg/L and BOD of less than 10 mg/L. I was incredulous about the claim that BOD removal rates were that high until I saw the results from samples. There’s a school of thought that says that kind of reduction isn’t possible, but it’s based on old research. The beta project disproved that belief, and COD removal rates remained at 99.5 percent when the full-scale system started up. Another old paradigm is that anaerobic biomass is very sensitive, but we demonstrated that it is pretty resistant to upsets.
FE: How did you seed the digester?
Rosenberger: Sludge from a municipal treatment system was used. It provided a broad spectrum of organisms. Over a period of six to eight weeks, system efficiencies got better and better, suggesting a significant increase in the population. Because of the membranes, sludge is retained for a very long time, giving the bacteria time to do the final polishing.
Some research has been done on anaerobic bacteria, but it doesn’t approach the wealth of literature on aerobic digesters. That’s beginning to change. Researchers at Marquette University and Vanderbilt are doing detailed work on anaerobic species. We have been involved in some of the work by Dan Zitomer at Marquette in Milwaukee. He and his team have done DNA sequencing of some of the microorganisms.
FE: Why was the MBR positioned outside the digester?
Rosenberger: In an aerobic digester, submerging the MBR in the digester tank is no big deal; the tank can be subdivided, or a crane can lift the MBR out for maintenance. But submerging it in an anaerobic tank poses a host of issues, first of which is that the tank must be gas tight. Flat-sheet membranes were considered, but they must be submerged in a tank, the system must be taken off line, and the tank must be drained for cleaning and servicing. It did not pass the litmus test as a practical design.
FE: How much maintenance and operator oversight are required?
Rosenberger: Membrane fouling is the main concern, of course, but we’re able to maintain high, sustained flux rates and avoid inorganic fouling or precipitation on membrane vessels. Because of the unique reactor design and mixer, a system can run four to six weeks before being flushed. Because the membranes are all racked and in a building outside the reactor, they’re readily cleanable with a standard CIP system. Not having to clean the system daily or weekly is a huge improvement.
A specific pH range needs to be maintained, and occasionally caustic needs to be added to increase alkalinity. In the beta application, COD reductions reached 99 percent after a few months. Then it plummeted to 50 percent. Because a little caustic was good, the operator decided more would be better. It took about five days to get the excess caustic out of the system, and since then, there have been no excursions.
FE: What system refinements have you made since the full-scale Daisy AnMBR came on line?
Rosenberger: We want to increase the capacity and efficiency of the cross-flow membranes, and the reactor tank design was modified, but the changes are primarily of the cheaper, better, faster variety: more efficient pumps and piping for greater flexibility, engineering the filters on a preassembled skid and fine-tuning to handle a wider range of substrates. Besides dairy, we’re also targeting rendering and general food and beverage applications.
We’re just now breaking the paradigm that this was not possible. Until recently, people either didn’t know what AnMBR was, or they were uncomfortable with it. Now, upwards of 50 percent of the wastewater treatment proposals we’re sending out have involved anaerobic MBR.
FE: How much potential does an MBR have beyond dairy?
Rosenberger: Our only two commercial systems in operation are at dairies, but field trials with ethanol and biodiesel waste have been done, and a pilot involving chemical wastewater is being discussed. There are many applications where this can be used. If there’s a degradable substrate, it’s just a question of the economics of putting in an ultra high-efficiency system.