Talks with Terry (Engineering in Action): What is BOD and De-Nitrification?

Terrence Allen, one of our friendly Engineers, is going to walk you through this topic in his own unique way.

What do “BOD and De-Nitrification” really mean and why do they matter to me?

BOD, which stands for Biochemical Oxygen Demand and Nitrogen are two of four major pollutants that we normally target during domestic sewage treatment. (The other two are Total Suspended Solids (TSS) and Total Phosphorous).

Nitrogen, measured as Total Nitrogen, generally comprises 3 main ‘forms’ in sewage, Organic Nitrogen, Ammonia and Oxidised Nitrogen. Ammonia is toxic to a variety of fish species, and excessive nitrogen pollution can cause algal blooms in waterways, severely impacting upon both visual amenity and life within the catchment.  This makes the removal of nitrogen important when a sewage treatment plant releases to a waterway.

Is Nitrogen always bad?

No, some nitrogen can often be beneficial for plant growth when treated sewage is released to land, however some removal is often needed to help make sure that any possible surface water runoff does not have the potential to cause pollution, or to ensure groundwater does not suffer nitrogen pollution.

So, how do we get rid of this pesky nitrogen?

We do so using Biological Nitrogen Removal.

So how does that work?

The “biological” part comes from the fact that we use bacteria (bugs) to remove the pollutants, by creating specific conditions that force certain bacteria to grow.

The Biological nitrogen removal process consists of 2 steps; Nitrification and De-nitrification.

Okay, so what does that mean?

Step one is called “Nitrification” and takes place in an aerobic environment (one where free oxygen is present) where the available BOD has been exhausted. In this environment our Bacteria use Ammonia to live.  This produces Nitrates and Nitrites (oxides of nitrogen, NO3 & NO2) as by products which dissolve into the water.

The nitrates produced by the nitrification reaction are then removed via the de-nitrification process.

Pardon?

I like equations. Let’s have a look at a simplified version of the reaction pathway:

Ammonia + Oxygen + Carbon Dioxide + Bacteria → More Bacteria + Nitrates + H2O

This is a (very) simple version of the overall reaction pathway; the actual pathway for the reaction is much more complex, producing nitrites and nitrous oxides as intermediate steps.  Varying conditions can cause the reaction to stall at the nitrite or nitrous oxide stages, which is bad because nitrous oxide is a very potent greenhouse gas

Okay, I get that, what about the other one?

The second step is de-nitrification.  Under low oxygen conditions certain bacteria are able to cleave or “cut away” the Oxygen molecule from the Nitrate & Nitrite Molecule and use that instead of the oxygen in air.  These bacteria then use available BOD and the Nitrate oxygen molecule to live.

Can I have the chemistry on that part too?

Absolutely. Here you go:

Nitrates + BOD + Bacteria → More Bacteria + Nitrogen Gas + Carbon Dioxide + H2O

Again this is a very simple version of the a much more complicated reaction pathway.

I notice that BOD keeps popping up – What is it?

BOD is an indicator which represents a large range of carbon based organics (usually carbohydrates and their derivatives).  Broadly speaking BOD is part of poop, and old-school sewage treatment systems generally only targeted it and Suspended Solids, not really caring about Nitrogen or Phosphorus, but I digress.

For this process BOD is necessary to achieve de-nitrification, and if your available BOD is too low it won’t work at all.  That means that there are many different types of processes which all try to create the right conditions to make the process work, generally by delivering Nitrate rich mixed liquor to BOD rich mixed liquor, or cycling conditions using timers with a consistent feed.

Can you show me how this works in the real world?

Sure. Let’s Take a look at an example using Biowin Modelling.

For this example we’ve used the Modified Ludzak Ettinger (MLE) process, which is very common for small treatment systems. It involves the use of relatively high volume (normally between 2 and 4 times the influent flow) recycle from the back-end of the aeration zone to a dedicated anoxic zone, situated before the aeration zone.

basic-mle-process

Figure 1: Basic MLE Process

Biowin modelling has been used to demonstrate the effect of an un-balanced raw sewage upon the process effluent quality.

A MLE process as per the above process diagram has been configured for a raw sewage flow of 100kL/day using a “standard” raw composition of 245mg/L of BOD and 50mg/L of TKN (a measure of the combined Organic Nitrogen and Ammonia).

I am really interested in the technical stuff. Can I see that?

The treatment plant has 75kL of total tank volume, and an average hydraulic residence time of 18hrs (excluding recycles).  The tank volumes have been split approximately 45%/55% between the anoxic and aerobic zones.  The plant has a 20 – 30 day sludge age, with a bacterial biomass concentration of approximately 4500mg/L.

The anoxic recycle has been set at 4x the influent volume (400kL/day) and the RAS recycle at 1x the influent volume (100kL/day), both returning to the pre-anoxic tank. These parameters are relatively typical for a MLE based treatment plant.

The treatment plant has been tested against a constant daily flow, using steady-state analysis.  For a full design a dynamic analysis using variable sewage flows would also be examined to assess plant performance.

So, you’re saying that the goal is Balanced Raw Sewage?

When designing a new treatment plant Manufacturers and Engineers often have to make assumptions about the raw sewage that the plant will need to treat.  Without better data the majority of treatment plants end up being designed based upon standard domestic raw sewage concentrations, with a BOD of approximately 245mg/L and a Total Kjeldahl nitrogen (TKN) concentration of approximately 50mg/L (as N).

In this example, this results in a TKN to BOD ratio of approximately 1:5.

 

standard-raw-sewage-composition-utilised-in-biowin
Figure 2: “Standard“ raw sewage Composition utilised in Biowin

Based upon the steady-state analysis, this configuration performs relatively well (the difference in influent flow can be attributed to the waste activated sludge flow).

 

final-effluent-results-balanced-sewage
Figure 3: Final Effluent Results – Balanced Sewage
influent-vs-effluent-nitrogen-concentrations
Figure 4: Influent vs Effluent Nitrogen Concentrations

The majority of the TN present in the effluent is in the form of Nitrates.  This may be further reduced by providing a post-anoxic zone or by modifying the recycle rates.

But unbalanced Raw Sewage is not at all uncommon?

It can be surprisingly common.  Treatment plants servicing itinerant work forces, construction camps (without associated accommodation blocks), schools or other similar catchments often have unbalanced raw sewages.  Due to human nature, these treatment systems often feature significant concentrations of Ammonia with lower concentrations of BOD in comparison to a domestic sewage.  At a TKN:BOD ratio of less than 1:4 the de-nitrification process will generally perform poorly, or not at all.

To simulate a BOD to TKN imbalance we’ve taken the previous model and halved the inlet BOD.

unbalanced-biowin-raw-sewage-input
Figure 5: Unbalanced Biowin raw sewage input

One might be tempted to think lower load would result in better performance.  This is often not the case, as the model predicts.

unbalanced-biowin-model-output-prediction
Figure 6: Unbalanced biowin model output prediction
unblanced-influent-vs-effluent-nitrogen-concentrations
Figure 7: Unbalanced Influent vs Effluent Nitrogen concentrations

The model predicts a worsening of the Total Nitrogen in the effluent, roughly doubling it from 7mg/L to a slightly under 14mg/L, so the process may be more lowly loaded with regards BOD, but we’re now getting relatively poor TN removal performance.

This can be even worse with increases in TN concentrations due to water saving activities.

So how do we fix it?

Unless an additional BOD load can be added to the catchment, generally the simplest method is to add a supplementary carbon source.

For small plants this is often in the form of sugar or some form of commercially available sucrose solution.  Alternatives are acetate solutions such a sodium acetate or, Methanol.  Methanol has some benefits, but represents more OH&S risk than most sites will be willing to accept.

The additional carbon loading provides a BOD source to allow the de-nitrification process to be driven to completion.

For an existing system a carbon dosing system can be retro-fitted to the pre-anoxic tank (or another suitable location).  For new systems a post-anoxic reactor may be provided, which has benefits in process control and chemical usage.

mle-system-with-methanol-dosing
Figure 8: MLE system with Methanol dosing

The model predicts that the addition of 5L/day of Methanol would return the system to producing a Total N concentration similar to that produced by the ‘balanced’ raw sewage.  The dosage as Sugar, sucrose solution or sodium acetate would be different, but can be calculated based upon the solution BOD and/or COD.

final-effluent-results-balanced-sewage
Figure 9: Effluent Results

While carbon dosing can help improve nitrogen removal, care must be exercised when designing and operating such systems, as simple carbon sources such as sugar can cause the sludge to bulk, which reduces settleability, which in turn increases TSS in the effluent, which can rapidly undo the good you’re doing by removing Nitrates by replacing it with organic nitrogen from bacterial biomass, and possibly even causing exceedances of other limits, such as TSS and BOD.

Additionally over-dosing of the carbon source can cause different problems, including increased sludge generation, increase in aeration demand or an increase in treated water BOD.

Is that all BOD dosing is used for?

Nope, additional BOD dosing can be used to boost the de-nitrification performance of an already functional plant, allowing really low levels of effluent total nitrogen to be attained, and it can be utilised to help with biological phosphorus removal as well, although this is beyond the scope of this discussion.

Uh huh?  So does this actually happen in real life to actual people who aren’t robots or just in your model?

It sure does! We’ve implemented both ad-hoc and controlled dosing systems for sugar solutions and other carbon/food sources such as dry dog biscuits.

A good example is a sizable college, operating in an unsewered area.  They operate their own sewage treatment system, but can have trouble getting effective nitrate removal.  We’ve co-ordinated with operators and developed methods for sugar and dog-food addition to both help achieve consistent nitrification and to help the plant function during the school holidays and down times.

Another example is a mining site.  The site does not include worker accommodation, and as such tends to have very high ammonia concentrations in their sewage.  We designed and installed an automatic sucrose dosing system, which included a mixer and mixing tank.  This system dosed sugar solution automatically based upon the aeration cycling in the main bioreactor.  This system achieves robust nitrogen removal using the dosing system.

This is all a bit overwhelming to me…

Don’t worry about it – that’s why we’re here. Simmonds & Bristow’s entire team of Engineers, Environmental Scientists, Operators and Trainers are here to keep your treatment plant running the way it should be. We can design and commission new plants, audit and upgrade existing plants, provide you with trained operators, trainyour staff to run the plant properly, and help you conduct the routine monitoring and scientific assessments that will keep the DEHP happy.

This is our business and we’re good at it – and we’re here to help you.