What actually comes out the stack?

Chapter 6 followed the part of the fuel that does not burn. Some leaves through the residue system, some may remain as deposits, and some can travel onward with the combustion gases.

That leads to the obvious next question:

What actually comes out of the stack?

The Glan Devon air-quality assessment tries to answer that using emissions data from a Xetrov test conducted somewhere else. According to Vipac, the source data came from a client-supplied document called the “Xetrov DRAX FBA Test Report.” The underlying report is not included in the public DA material I have reviewed. What we have is Vipac's summary: the test burned polyurethane dust, the Xetrov was operating at about 70 per cent capacity, and there was no emissions-control equipment at the test site. Some of those measurements were then scaled by 1.42 for use in modelling the proposed Glan Devon plant.

There are obvious questions about using polyurethane dust to represent agricultural HDPE and timber, and about scaling a 70 per cent test to full load.

But I think those are actually the second problem.

There is a more basic one.

An emissions measurement does not belong simply to “a Xetrov”. It belongs to the complete physical system through which the gas travelled before somebody measured it.

And once I look at the problem that way, the relevance of the polyurethane test becomes much harder to establish.

Start by making the comparison as generous as possible

Imagine removing the fuel difference completely.

Suppose the proposed Glan Devon Xetrov were fed exactly the same polyurethane dust as the machine used for the original test.

Would I expect the Glan Devon stack measurements to be the same?

Not necessarily.

For that comparison to be convincing, much more than the fuel would have to be equivalent.

The two machines would need to burn that fuel under comparable conditions: similar load, feed rate, air supply, oxygen levels, combustion temperature, residence time and control settings. The relevant internal Xetrov configuration would also have to be the same, or there would need to be evidence showing that any design changes did not materially affect the emissions.

But even that would not be enough.

Once combustion has finished, the gases still have to travel from the Vortex to the sampling point. The two plants would therefore also need a sufficiently comparable downstream journey: similar heat recovery, similar cooling rates, similar gas residence times, similar ducting and surfaces, similar pollution controls, and a sampling point that measured the gas at a comparable stage of that process.

Only then would I start to feel comfortable treating one set of measurements as representative of the other.

That is the part of the comparison that has been easy to miss.

Because for some pollutants, what leaves the combustion chamber is not necessarily what eventually leaves the stack.

At Glan Devon, the gas is deliberately being cooled

This becomes particularly striking at Glan Devon because recovering heat is not a side benefit of the Xetrov.

It is the reason the machine is there.

The process information uses a nominal fuel input of about 1 tonne per hour at 25 MJ/kg, giving roughly 25 GJ/h of thermal input, and shows around 20 GJ/h of recovered heat being transferred into the process used to dry the biosolids.

That is a large amount of energy being deliberately removed from the hot combustion system.

So even if the gas leaving the Glan Devon Vortex were chemically identical to the gas leaving the polyurethane test machine, Glan Devon then does something very important to it:

it takes heat out of it to dry biosolids.

As energy is removed, the gas cools.

How quickly it cools, how long it remains within particular temperature ranges, what surfaces it encounters and what happens to particles during that journey can all matter to the final stack gas.

The available Glan Devon material does not yet provide a final supplier-integrated heat-recovery and dryer configuration that allows that complete thermal journey to be reconstructed. The DA uses a recovered-heat/hot-oil concept, while the Andritz material attached to the application is a non-binding Caboolture proposal based on a different steam duty. The final heat-transfer, off-gas and control interface remains unresolved in the evidence reviewed.

For ordinary heat calculations that is already an integration issue.

For emissions, it becomes something more serious.

Because the chemistry does not necessarily stop when the flame goes out.

SARA's dioxin concern shows why this matters

SARA identified dioxins and furans as a particular problem.

The reason is slightly counter-intuitive.

Very high combustion temperature can help destroy organic pollutants. But dioxins and furans can form or reform later, as the gases cool under suitable conditions.

SARA specifically identified the roughly 450–200°C range and said the Glan Devon material had not demonstrated the required rapid cooling through that region to below about 200°C.

That gives us a very simple thought experiment.

Suppose two Xetrov units burn exactly the same polyurethane dust under exactly the same combustion conditions and produce identical gas at the outlet of the combustion chamber.

One plant cools that gas rapidly through the temperature range associated with dioxin re-formation.

The other removes heat differently and the gas spends longer passing through that range.

The final stack chemistry need not be the same even though the combustion was the same.

That is the key point.

And at Glan Devon, heat is deliberately being extracted from the combustion system because the plant needs that energy for drying.

So the downstream cooling path is not some minor detail that can be added after the air assessment.

It is part of the emissions process.

Which raises a very awkward question about the polyurethane test

I now want to know something quite basic about the original polyurethane-dust measurements.

Where were they taken?

The underlying DRAX/FBA report is not public, so I cannot reconstruct the tested system from the material available to me.

I do not know from the public DA documents whether the gas was sampled close to the Vortex, after substantial heat recovery, after some other cooling stage, or at a final stack sampling point.

I do not know the gas-temperature history between combustion and sampling.

I do not know how long the gas had spent cooling before it was measured.

I do not know whether that downstream arrangement resembled the heat-recovery system proposed for Glan Devon.

Those details could radically change the meaning of some of the test results.

If a pollutant was measured while the gas was still very hot, the result may tell us something useful about the combustion stage while telling us much less about what would exist after subsequent cooling.

If the measurement was taken at the end of a complete stack train, it tells us considerably more about that particular test plant — but I would still need to show that its cooling and gas-treatment pathway was representative of Glan Devon.

At present, I cannot establish either proposition.

This is why I no longer think it is enough to say that the polyurethane data came from “a Xetrov test”.

For emissions, I need to know which Xetrov system was tested and where in that system the emissions became a measurement.

And Glan Devon's downstream system is still unfinished

There is another missing part of the comparison.

The final Glan Devon air-pollution-control system has not yet been specified.

The February air assessment says the abatement package had not been determined and models the Xetrov source without abatement. Later material says an emissions-abatement package and continuous emissions monitoring will be installed, but does not identify the final components or demonstrate their performance. SARA specifically objected to leaving selection of the pollution-control system until after later stack testing.

Again, that does not mean suitable pollution controls cannot be designed.

The problem is representativeness.

The final gas reaching the Glan Devon stack will be the product of a sequence something like:

Xetrov combustion → heat recovery and cooling → pollution control → stack

Yet the final heat-recovery pathway is unresolved and the final pollution-control system is unresolved.

Meanwhile, the physical pathway associated with the polyurethane test is not available publicly because the underlying test report is missing.

So I cannot place the two systems side by side and demonstrate that the emissions measurements from one are representative of the other.

That would remain true even if Glan Devon burned the exact same polyurethane dust.

But it will not.

Then the comparison gets harder again

Only after that do the more obvious differences return.

Glan Devon proposes fuels including agricultural HDPE and timber-derived materials rather than polyurethane dust. SARA expressly questioned whether the polyurethane test represented those fuels and identified gaps involving metals, HCl and HF, PAHs, dioxins and furans and other pollutant pathways.

The test was also conducted at about 70 per cent load.

For the Glan Devon modelling, selected values were multiplied by 1.42 to estimate a nominal full-capacity case.

That calculation simply asks what the measured number would become if it increased directly in proportion to load.

It does not demonstrate that the machine actually behaves that way at full load.

But by this stage, the 1.42 multiplier almost looks like the smaller problem.

The transfer is not merely:

70 per cent → 100 per cent.

It is:

one fuel → other fuels

through

one operating condition → another operating condition

inside

a tested Xetrov configuration → an incompletely fixed Glan Devon configuration

followed by

an unknown test cooling and sampling pathway → the Glan Devon heat-recovery and cooling pathway

and then

no test-site APC → a future Glan Devon APC system that has not yet been selected.

Multiplying by 1.42 deals with one number inside that much larger chain.

It does not establish the chain itself.

Even the machine may not be equivalent

There is one more uncertainty sitting behind the comparison.

I cannot establish from the public material exactly which Xetrov configuration produced the polyurethane test data.

Since then, Xetrov has filed its 2024 patent dealing with practical changes to airflow, particle return, residue handling and other internal features. Separate Australian trial material also refers to a Xetrov Vortex V4, although the DA does not clearly establish that V4 as the final Glan Devon configuration.

I cannot say that the polyurethane test was performed on an earlier machine.

I also cannot say that Glan Devon will incorporate every later patented feature.

The problem is precisely that I cannot establish configuration equivalence.

And if airflow and particle movement are changed inside the combustor, it is reasonable to require evidence before assuming that earlier emissions measurements remain representative.

The public record does not provide that bridge.

So what does the air model actually demonstrate?

This is where I think the entire Chapter 7 argument becomes much simpler.

The dispersion model takes an emissions source and asks what happens to it after it leaves the proposed Glan Devon stack.

That is useful.

But the most important unresolved question occurs before the atmospheric model starts:

what source will the completed Glan Devon plant actually put into that stack?

The current source terms are transferred from limited Xetrov testing.

For the polyurethane test, the public record does not establish the complete gas-cooling and sampling pathway that produced those measurements.

For Glan Devon, the final heat-recovery, cooling and pollution-control pathway has not yet been demonstrated.

For some pollutants, particularly dioxins and furans, those missing stages are themselves part of the process that determines the final emission. SARA's concern about cooling through the 450–200°C range makes that explicit.

So the atmospheric model can perform exactly the calculation it was asked to perform and still leave the central industrial question unanswered.

It can tell us:

If the Glan Devon stack emits these assumed pollutant rates, where will those pollutants go?

It cannot tell us:

Will this particular Xetrov, burning these fuels, while surrendering this much heat to this dryer, cooling through this gas pathway and passing through this future pollution-control system actually produce those stack emissions?

That second question has to be answered before the first one becomes fully representative of the proposed plant.

The startling part is the dryer

For me, this is now the central weakness in the Glan Devon emissions evidence.

The same feature being promoted as one of the project's principal benefits — using the Xetrov's heat to dry biosolids — is also one of the reasons emissions data from another Xetrov cannot simply be treated as portable.

The heat has to go somewhere.

At Glan Devon, a large part of it is deliberately being removed for the drying process.

That means the combustion gas must pass through a Glan Devon-specific thermal journey before reaching the atmosphere.

And SARA has already identified one part of that journey as environmentally important.

So even the most generous possible comparison — same polyurethane fuel in both machines — would not establish equal stack performance unless the combustion conditions, machine configuration, heat recovery, cooling history, gas pathway, pollution controls and sampling point were also shown to be sufficiently equivalent.

They have not been.

Glan Devon then makes the comparison harder by changing the fuel, changing the operating load, using an incompletely defined downstream system and potentially using a different generation of the Xetrov itself.

None of this proves that the eventual plant will exceed emissions limits.

It establishes something more precise:

the public record does not demonstrate that the emissions source used in the Glan Devon air model represents what the completed Glan Devon stack will actually emit.

That is not fundamentally a CALPUFF problem.

It is a problem with the source that CALPUFF has been asked to disperse.

And it leads directly into the next chapter, because the missing emissions pathway is not separate from the rest of the plant.

It is created by the way the Xetrov, heat-recovery system, dryer, pollution controls, fans and stack are joined together.

At Glan Devon, those pieces have to become one machine.

Next Chapter