Case study: how an industrial plant reduced its reported CO2 emissions by 8% through better fuel measurement

Case study: how an industrial plant reduced its reported CO2 emissions by 8% through better fuel measurement

Many industrial facilities will face significant challenges in the coming years in responding to National and Global efforts to reduce greenhouse gas emissions. Although Australia has been something of a laggard in this space, nevertheless it is a signatory to the Paris Agreement and we can expect a range of future emissions reductions policies to hit industrial businesses going forward.

One way to deal with your emissions challenge is to find ways to more accurately measure your major emissions sources, especially from fuel combustion. It’s not necessarily about reducing actual emissions, but improving your measurement accuracy.

In this case study, I show how I helped a clinker & cement manufacturing plant reduce its reported CO2 emissions from coal by 8%.

The plant

The plant is a clinker and cement production plant, converting limestone (Calcium carbonate, CaCO3) to clinker (Calcium Oxide, CaO). CO2 is given off during this reaction. Clinker is a precursor material used in the manufacture of cement.

It does so using a rotary kiln. This is basically a long rotating metal pipe, tilted at an angle (example image below). The rotation and angle enable the “raw meal” (limestone plus other material) to move along the kiln. A fuel, typically coal, is fired at the opposite end of the kiln. The hot gases directly contact the raw meal, converting it to clinker.

Clinker kilns require a lot of heat, most of which comes from combusting coal. This means clinker plants are heavy emitters of CO2 from combustion. They also emit more CO2 from the clinker reaction itself. Unlike many other plants, they have limited options to reduce their emissions impact.

CO2 measurement - the typical approach

Like most industrial plants, this company was reporting the CO2 from combustion of coal using the standard “Method 1” approach from the National Greenhouse and Energy Reporting (Measurement) Determination (the “Determination”).

Essentially, this calculates CO2 emissions using the formula (Determination, section 2.4):

Emissions (CO2) (tonnes) = Quantity of fuel (tonnes) x Emissions Factor (kg CO2 / Gigajoule) x Energy content (Gigajoules / tonne) / 1000

For Bituminous coal, the closest coal type to that being used at this plant, the fuel values come from the Determination (Schedule 1):

Energy content = 27.0 GJ / tonne

Emission factor (CO2 only) = 90.0 kg CO2 / GJ

If you plug these values into the 2.4 formula above, you get:

Emissions (CO2) (tonnes) = 2.43 x quantity of coal (tonnes).

This has the advantage of being simple - you simply need to measure your coal tonnage and you can estimate CO2 emissions easily.

However, this client (like many who use coal) could do better. The Emissions Factors for coal in the Determination are often conservatively high compared to emissions estimated by more accurate methods.

They wanted to see if they would get more favourable results through coal sampling and analysis. This would bring down their reported emissions, thus reducing the plant’s liabilities under any future carbon schemes.

Method 2 - more accurate CO2 emissions measurement

I helped the client to work through a process of evaluating whether more accurate emissions measurement would be worthwhile.

During the feasibility phase, we asked the client to take samples of coal on its way to the kiln from the coal conveyor belt. This was done daily for several weeks. These were sealed in plastic sample buckets and sent to an accredited lab for analysis of:

  • Carbon content (dry and ash-free basis)
  • Ash content (as received)
  • Moisture content (as received)
  • Energy content (GJ/tonne)

I then calculated new “Method 2” values using the formula from the NGER Determination (section 2.5). This allowed us to calculate varying values of the Emissions Factor and Energy Content for the coal, unlike the fixed values in the formula for method 1.

The average calculated reduction in CO2 per tonne compared to method 1 was almost 10%. In other words, it indicated a saving of reported CO2 of almost 10% could be made, simply by accurate measurement of coal properties.

Implementing more accurate measurement

The client, pleased to see the project was feasible, asked for assistance in implementing Method 2 for their plant. I stepped them through the following items to make this happen:

  1. Developed a repeatable coal sampling procedure, in accordance with the Method 2 requirements in the NGER Determination. In this example they elected to use manual sampling with a sample scoop. In other cases (especially at power stations) automated samplers are used.
  2. I led a bias testing project. NGER requires any plant using method 2 to test whether their sampling procedure introduces bias (that is, will the samples on average be different to the coal population being sampled) in accordance with AS4264.4 (Coal and Coke - Sampling - Determination of Precision and Bias). I managed the bias testing project in accordance with this standard. Basically it involved collecting lots of coal samples (in the regular way they want to use) in parallel with full across-the-belt samples of the entire coal stacked on the belt at that location. Comparing the two statistically lets you know if your sampling method is biasing your results or not.
  3. I helped the client establish a protocol for handling samples correctly (e.g. ensuring lids went on the buckets immediately to prevent samples drying out). They needed to collect daily samples for moisture and ash values, and accumulate a monthly “composite” sample (made of daily sub-samples) for carbon content and energy content.
  4. Assisted the client in making appropriate laboratory analysis arrangements. They had their own on-site lab, but it was not NATA accredited for coal analysis at the time. They made the appropriate arrangements to get this accreditation so they could meet NGER requirements for coal analysis.
  5. I developed a spreadsheet for the client to help convert lab results into final NGER CO2 results, in line with Method 2.
  6. I assisted the client in calculating measurement uncertainty. When a “higher order” method (such as method 2) is used, the Determination requires you to calculate the statistical uncertainty of the calculated CO2 emissions number. This is done by following the so-called Uncertainty Protocol.


Over the first full financial year of using method 2, the client reported CO2 from coal combustion at just over 8% below the method 1 value. Given that coal combustion represented around half the facility’s total GHG emissions, this was a significant improvement in their reported performance. The result was a little less than the feasibility numbers suggested (10% reduction) but was still a worthwhile improvement. Note that at the time, the Clean Energy Act (Carbon Tax) still applied in Australia, so this translated into a direct dollar saving for the company.

And remember, this didn’t require any operational changes - it’s just more accurate measurement of the emissions taking place.

This is one of several similar studies I have undertaken for clients. To see what percentage of CO2 emissions you could reduce, email me at for an initial estimate.

About David T Kearns

I'm an independent professional working in energy, carbon, sustainability and machine learning