Talks with Terry (Engineering in Action): Treatment Plant Commissioning

Well. Here we are again. What is it you’re going to walk me through today?

I thought I’d provide some information on commissioning, to help clear up any confusion.

Okay, but what is commissioning? Something to do with sales?

Not exactly. Simply put, commissioning is to bring something that has been newly produced into working condition. It’s also defined as the process by which equipment, facility or plant is tested to verify if it functions according to design objectives or specifications.

So…essentially making sure that people get what they paid for?

Simply put, that’s exactly right. To achieve this, after a plant is constructed there are a bunch of testing activities conducted by the manufacturer, commissioning engineers and the builders to make sure it’s been built to specification and will actually function as both the designers and client intended.

You then need to start the thing and tune it to get it working properly.

Ok…?

Because of its fairly broad definition, there are many activities included under the broad banner of ‘commissioning’.

I generally find it useful to split ‘commissioning’ into various sub-items to reduce confusion. This is useful as different parts of the commissioning process will often require different expertise.

I’m ready. Let’s break this down.

First let’s talk about mechanical/electrical commissioning.

Is that because it is up first?  

Mechanical and Electrical commissioning can be conducted either during or after the construction process, so it is good to at least talk about it first.

Mechanical commissioning is generally checks and tests for mechanical equipment to ensure that it is in working order and functions as intended, i.e. that pumps start, valves close, and tanks and pipework don’t leak. Electrical commissioning generally checks and tests electrical equipment, and involves ensuring sensors ‘talk’ to each other, that the switchboard can start and stop equipment as necessary, that drives are turning in the right direction, etc.

So who does most of the work for this?

These activities are generally conducted by suitably qualified engineers (mechanical and/or electrical) or suitably qualified tradespeople, often associated with equipment vendors or installers (specialist equipment vendors will often provide their own commissioning personnel to make sure their gear functions as intended, which can form part of equipment warranties).

Got it. What’s next?

Let’s move on to process commissioning and testing.

This generally covers testing of equipment performance against specific process requirements, and for STP’s covers Oxygen Transfer testing, Mixing tests, and Flowrate tests.

Right. STP’s. I am 100% familiar with that term.

Well, just in case, it refers to sewage or wastewater treatments plants. “WTPs” are water treatment plants.

Oh yeah….I mean, carry on.

Process commissioning and testing is conducted after M&E commissioning as the plant needs to be in an operable state. Process commissioning and testing are normally conducted by a process or chemical engineer, although mechanical engineers or tradespeople may be able to perform the work for simpler treatment systems.

For a WTP process commissioning would primarily be flow rates and mixing tests.

Okay, so you know that the mechanical/electrical parts are working and you know that the processes are doing what they’re supposed to. Are you done?

No, we still need to conduct functional testing.

What’s that? Do you have to wait until the rest of the testing is done? Oh – and why do you do it?

Functional checks and testing are conducted once the equipment performance has been confirmed via process testing. It is possible for elements of the functional tests to be carried out simultaneously to process commissioning, depending upon plant arrangement. So you don’t necessarily have to wait.

Functional testing tests the functions of the automatic control system against the control logic or philosophy. The actual execution of this testing will vary from simplistic (lifting of floats, dunking of conductivity switches into water) to much more complicated (sensor signal and fault simulation, using sensor software, contact bridging, etc) to the just plain annoying (filling of tanks to requisite levels). Generally speaking more expensive equipment will provide more options to simulate outputs for control system testing. Simpler equipment may require an electrician to bridge relays or contacts to simulate fault conditions.

Sensors are generally set-up at this point to close to their final configurations, and calibrated if necessary.

Is that it?

Not quite. The last part of a functional test is generally a test-run of the process on clean water. This can involve the use of bypasses or temporary recycle pumps to allow the return of processed water back to the head of the treatment plant.

Sounds intense. How long does that take?

A test run may last several hours or several days depending upon client requirements, and is a good way to pick up any issues with the control system or physical infrastructure.

If we’re talking about a sewage treatment plant….sorry, an “STP”, why do you use clean water?

This testing is best conducted on clean water to minimise the impact of any process failures. It’s much easier to clean up a spill of clean water than it is to clean up a spill of sewage or sludge.

Okay – and who’s in charge of this?

Functional Testing is often conducted by a process engineer, with Electrical and/or Mechanical tradespeople as backup to perform specific equipment manipulation that the process engineer is either unqualified or unable to perform.

Okay and THEN you’re done, right?

Once functional testing is complete the process is technically ‘commissioned’ as it is now ready to work as intended. For a WTP things are fine to move onto process tuning and optimisation.

Wait a minute….so STP’s have more to go before you can get to that step?

STP’s have an additional commissioning step however; biological system commissioning.

And what does that mean?

Most sewage treatment technology is based upon biological processes, the use of bacteria to degrade pollutants within the wastewater producing an effluent with reduced pollutant concentrations.

For an STP to be in ‘working condition’ the biological process needs to be started and brought online, although this can be a bit of a grey area as, to be biologically commissioned, the plant has to be ‘live’ (i.e. it has to be receiving raw sewage for treatment).

Hmmm…if it is a sewage treatment plant, does that mean it is usually connected to the actual sewer system? As in houses?

As most STP’s are generally installed as part of a ‘new’ sewage system connection to the sewage reticulation results in the ‘ramping up’ of flow over a period of time, either as existing houses are connected to the new sewage infrastructure, or as houses are occupied. Generally connection to an existing township will result in faster flow build-up than occupation of a new sub-division, and this will impact upon how the plant is started.

How do you get this up and running?

Ideally the plant is seeded with sludge from a healthy treatment plant nearby. This provides a ready-to-go (or relatively ready-to-go anyway) biomass, which kicks off process establishment and treatment relatively quickly.

What if you can’t do that?

If there is no seed sludge nearby, if you have partially toxic influent (which is likely to kill the seed biomass), or if your raw sewage ramp-up time is so slow that sufficient seed biomass would be unsustainable, it is possible to grow activated sludge from scratch. This will take a ‘sludge age’ or two (so generally somewhere between 4 and 8 weeks for an Extended Aeration style system) and will generally result in very poor treatment performance for the first 4 (or so) weeks of operation.

Does that have any on-going implications for the new plant?

The plant will also need to be set up specifically at start-up as the process will react very differently under start-up loading than it might at full load. Failure to provide appropriate start-up settings can result in the process never establishing itself, no matter how long you give it.

That seems like it will take a while. Is there a way to go faster?

It is possible to hasten the establishment of a biomass without seeding using products such as Actizyme, or providing an additional food load to increase growth rates.

It should be noted that MBR’s need to be seeded or ran specially for the initial start-up, as start-up conditions can (and will) rapidly clog the membranes the system relies upon (more on this later).

Okay, so now the STP’s are caught up with the WTP’s and ready for “tuning and optimising”?

Exactly. Once the process is operating at close to its design load (or the load it’s likely to operate at for the foreseeable future) the system needs to be tuned and optimised.

For an STP tuning is actually an ongoing process, which often starts during the latter part of biological commissioning, as the operator or engineer makes adjustments to the treatment system settings to optimise treatment or growth. The engineers or operators generally make relatively small adjustments to an STP once the process has established itself, as large changes can de-stabilise the process, which can lead to failure.

How long does all this “tuning and optimising” take?

Process changes often take a few days to have a noticeable effect, which can to an elongated tuning period. It’s best to tune with on-site tests with occasional (e.g. weekly) lab tests to make sure the on-site tests are reliable. On site testing generally targets ‘easy’ parameters such as Nitrogens and Phosphorus, although laboratory equipment for COD and TSS can also be provided. Larger plants are often provided with relatively well equipped laboratories leading to a greater range of testing data being produced during process tuning testing.

So that’s STPs. What about WTPs?

For a WTP tuning is generally a case of setting chemical dosing rates and flows. As there is no biological system to optimise the process is normally relatively rapid, as chemical dosing rates should already be roughly known based upon design and jar-testing, and process changes are generally rapidly evident, as opposed to taking a few days to have effect. Again on-site testing is generally used to tune the treatment plant. While the treated water will be tested for relevant parameters (turbidity, chlorine, metals), the operator can also assess dose rates using Jar testing (just like they would during normal operation), allowing the effect of some changes to be predicted. Again laboratory testing should be conducted to confirm site-testing (and for parameters that can’t be easily tested on site, like bacteria).

What should the outcome of all of this tuning be?

The end of the tuning period should result in a treatment plant that is hitting its’ treatment targets and running at what should be close to optimum conditions.

Hmm…this has me wondering, do you just set the plant up for what it will be doing “right now” or do you think about the long-term?

To future proof infrastructure, often new equipment is sized based upon forward projections for capacity, in our case it’s often projected population growth, normally over a 25-50 year horizon.

The need to meet significant future loads can result in treatment plants that are significantly oversized, especially at start-up.

Does that cause any issues?

For STP’s this can be problematic, as the biological process does not react well to either under or overloading. In our experience running a process below 30% capacity without special provisions will generally result in poor performance, especially with regards to TSS removal, as biomass concentrations are just not sufficient enough to promote good settlement. It is possible to provide additional food sources, stage or train a treatment plant to allow for these low loads. This can make start-up and optimisation more troublesome, and can result in multiple optimisation programs run over the course of a year or so (or even more).

For WTP’s over-sizing is not as problematic. Being chemical processes they are generally very easy to start or stop. This allows the system to be easily run as a ‘batch’, running the plant until the storage tanks are full and then simply turning it off. This will require active chlorine management in the reservoir, but this is becoming very common in water storages and wouldn’t really present much of a problem.

You mentioned something about MBR’s and Seeding?

So yeah.

Seeding is bringing in an established biomass from an alternative facility. Generally this involves the tankering of a portion of the daily waste activated sludge from the existing facility to the new facility. This sludge is pumped into the bioreactors (preferably through a screen), providing an already-settleable, biomass which can greatly improve start-up times. Ideally sludge from a ‘healthy’ system should be selected, as problems from the donor plant may be inherited by the new facility if there are any specific issues (such as filamentous bacteria bulking or septicity)

While it shortens the ‘growth’ time, as you start with a biomass, the biomass does have to undergo an acclimatisation period before it will be treating effectively, like starting a new job. As such, if the raw sewage conditions between the two plants are wildly different this acclimatisation period may be longer. If there is a toxic influent (possibly from industrial catchments) then the biomass may actually die off after seeding (often biomasses will acclimatise to toxic loads from industrial areas that will outright kill a biomass from a different treatment plant). Also, if there is insufficient load on the treatment plant (e.g. a lack of raw sewage) a portion of the biomass may also die off (until it reaches an equilibrium point).

Seeding is a useful tool for rapidly starting a treatment plant, especially if you’re starting a small plant within driving distance of a large one.

So you absolutely need to seed?

Technically, no, if you lack access to a suitable donor plant, if you have insufficient load or if you have crazy raw sewage you don’t *have* to seed your treatment plant, although this needs to be assessed on a case-by-case basis.

For an extended aeration activated sludge system based on ‘traditional’ technologies (i.e. secondary clarification) then you can get away without seeding your process. You’ll produce off-spec effluent for a few weeks and you’ll need to be careful with shock loads and over-wasting, but it is possible to nurse the plant through a build-up period. Generally there’s nothing that will be irreversibly damaged by doing so.

For MBR’s (membrane bio-reactors) however this is not the case. New activated sludge is generally ‘sticky’, producing an extracellular polysaccharide material that can promote rapid membrane fouling. An older activated sludge at a suitable concentration minimises this EPS production, making the sludge less inherently fouling. As such it’s generally a good idea to seed an MBR up to a nominal biomass concentration for operation. If an MBR can’t be seeded special consideration needs to be made to the membranes, either by ensuring a disposable ‘commissioning’ set of membranes are provided as part of the supply, or by temporarily configuring your STP as a conventional styled system (i.e. the use of a temporary clarifier or the use of the reactor tanks as an SBR system). All of these increase startup costs (and hassle) significantly.

So… do I have to get my process commissioned?

The short answer is “Yes”, the process will need to be made ready for use.

Do you need to get it fully tested, to ensure it’s been built to specification and will achieve it’s design targets?

Short answer here is “You should” as it will give you confidence as an operator, owner or client that what you’ve bought is what you thought it was, and that it will indeed do the job you bought it to do.

That said…

For a lower tech simpler system the testing and optimisation process can be shortened. These plants are simple so their setup should be able to be done relatively rapidly by someone with sufficient experience. However the site and sewage needs to be taken into account when deciding to shorten the testing and optimisation process. Complicated sites with problematic raw sewage conditions (high degree of industrial waste, highly variable flows, etc) should have longer testing and optimisation times, as these will likely result in a more successful start-up.

And if my system is more high tech?

For high-tech systems, this is different. High tech systems are generally less ‘robust’ from a process perspective, with sensitive equipment or elements that can be irreparably damaged by inappropriate operation. Proper commissioning is highly recommended for these systems as they can be rapidly damaged by inappropriate settings or even just start-up conditions that aren’t appropriately managed, and instead of just taking time to start the system will fail completely, possibly being unrecoverable.

Ok, now you’re getting confusing.

Lets simplify.

It’s always a good idea to get your plant properly commissioned. It reduces unknowns, allows for the establishment of appropriate operational procedures. It puts an expert on site which is often invaluable beyond the benefit of proper start-up, if you have an operator who can shadow them often significant amounts of on-the-job training can be conducted.

If you have a small low-tech plant you might get away without fully biologically commissioning, testing and starting the plant. Ideally you should still have an expert advise you, possibly attending site for a few days over the stabilisation period to check and alter settings, however if you’re really strapped for cash this might be able to done remotely via SCADA or a phone-hook-up.

If you have a high-tech plant like an MBR you should get it properly commissioned, tested and optimised, as the damage that can be done to the important equipment is significant.

Okay, okay, I get it. In fact, I don’t think I could absorb any more about commissioning if I wanted to! Although….I am kind of curious about all of this “testing” you’ve talked about.

No, problem. That’s plenty of info for now. And you’d like to know more about the specific tests we do when doing commissioning? That’s easy. Just stay tuned over the coming months as we’ll be going into further detail on the major ones!