Energy efficiency benchmarking in heavy industry is often discussed in broad terms, but meaningful comparison between sites has long been undermined by inconsistent system boundaries, allocation methods, and data interpretation. In a recent Applied Energy article, “A calculation method enabling energy benchmarking in the pulp and paper industry: Adopting a methodology that bridges the research–policy implementation gap,” the authors propose a standardised, transparent framework designed to finally make cross-mill energy performance comparisons both fair and actionable.
In an earlier feature, we explored why energy benchmarking has remained such a persistent challenge under the EU Industrial Emissions Directive (IED), particularly for complex, multi-product pulp and paper mills. Now, we follow up with an in-depth interview with the lead contributor behind the methodology to unpack the thinking that shaped it—from its origins in EU BAT negotiations, to the practical decisions around subprocess definition, residual heat accounting, and real-world testing with operating mills.
In this conversation with Olof Åkesson, former Senior Technical Officer at Swedish Environmental Protection Agency, we discuss what it takes to design a calculation method that can be trusted by regulators and industry alike, how policy ambitions collide with measurement realities on site, and what must happen next for the approach to gain acceptance across Europe. You can also read more here.
The following interview is presented unedited to preserve the original responses and provide a clear, practitioner-level view of how this benchmarking framework was developed, refined, and positioned for future adoption.
Acknowledgment from Olof Åkesson: Special thanks to Elin Svensson, CIT Renergy, and her colleagues, for valuable help in developing the calculation method.
What motivated your team to develop a standardised calculation method for benchmarking energy performance across pulp and paper mills, and how did you identify the key barriers to cross-site comparison?
In 2013, I participated as a representative for Sweden in the EU’s technical working group to develop BAT conclusions for pulp and paper production. I have long seen energy use in general and industrial energy use in particular as a key issue for a sustainable society. For Sweden, the pulp and paper industry is of particular importance, as it accounts for almost half of industrial energy use. My view has also been that in order to evaluate how efficient an industry’s energy use is, calculating key figures is necessary to be able to compare these with other, similar operations.
As a basis for the BAT conclusions, data was collected from European industries, regarding, for example, emissions to air and water and use of raw materials and resources, including use of water and energy. Based on this data, BAT conclusions could be formulated, in the form of intervals within which an industry using best technology should be able to stay within. However, this did not prove possible in terms of energy use. The most important reason for this was that there was no standardized method for calculating energy use. How would the system boundaries be set? How would the allocation of energy use be made for industries with multiple product manufacturing? etc. It was far too unclear what the different data represented and for that reason it was not possible to evaluate the different data against each other and not possible to draw any conclusions about which energy use corresponded to the best available technology.
With this experience, I concluded that by the next time the BAT conclusions were to be revised, a calculation method must have been developed in good time that could be accepted and used by all EU Member States.
Could you walk us through how you broke down the production process into standardised subprocesses (e.g. pulp production, purchased pulp dissolution, drying, paper production) and why those were chosen?
The idea has been to divide production into as large a process section as possible. In principle, we have started from the points where a process begins and extend as far as possible to the point where the process flow divides. Everything that happens in between is included in the key figure. For example, the first step in a kraft pulp mill is covered by the wood intake and ends where the finished pump pulp is produced and stored in the pulp towers. After the pulp towers, there are different branches: some pulp can go on to be dried and sold as market pulp, other pulp can go to the company’s own paper machines. In order to obtain a comparable process step when calculating key figures, the calculation for pulp production must therefore end at the pulp towers. Similarly, the calculation of the key figure for paper production begins at the pulp tower and ends after the finished paper comes out of the paper machines. The drying process for market pulp is calculated separately, as is the dissolution of purchased pulp. This division makes it possible to compare, for example, pulp production at integrated mills with non-integrated mills; paper production at integrated mills with non-integrated mills, etc. This in turn provides a significantly larger number of comparable processes than if only entire mills were compared.
The calculation method also means that no distinction is made between different paths that mills may have taken to reach the same product. For example, a pulp with high brightness can be produced by cooking the pulp to a low kappa number and then only a lighter bleaching process is needed. Another mill may choose to stop cooking earlier and instead perform a more intensive bleaching. In the first case, energy consumption in the cooking plant is higher, in the second case in the bleaching plant. It would be misleading to compare energy consumption in the cooking plant and bleaching plant separately. With our method, where the different process steps are included in the same key figures, this is avoided.
How does your method incorporate the recovery of residual heat (e.g. surplus heat used in district heating or greenhouses) and why is that important for fair benchmarking?
In the calculation method, the mill is credited with the secondary heat that is delivered from the mill to other users. That is, the amount of secondary heat that is delivered externally is subtracted from the mill’s own heat consumption. More specifically, it is deducted from the process step where the secondary heat is extracted, which is usually from pulp production.
In this context, it should also be mentioned that when calculating key figures for fuel use, the mill is credited in the same way with such residual products that are delivered externally and that can be used for energy extraction. Examples of this are the sale of wood chips and tall oil.
If primary heat (steam, hot water) is supplied externally, e.g. to a nearby sawmill, the fuel consumption needed to produce this steam is deducted from the fuel consumption for the mill’s own heat needs.
All of this is important to take into account in order to fairly include how the mill is integrated with the rest of society through the use of residual flows and residual energy. Energy efficiency for an industry is therefore a broader concept than just its own energy consumption.
When you tested the method with actual mills, what kinds of variations in input data or operational practices surprised you, and how did you account for those in the model?
I probably can’t answer this question fully. In the first stage, we entered fictitious data. The main purpose was to see if the calculation process worked in purely computer terms.
Test calculations with real data were mainly done by the mills themselves. We then received feedback from the mills, which was consistently that the calculation method worked and gave reasonable results. However, we also received some comments that led us to make changes. One such change was that fuel used to produce steam for a condensing turbine was excluded from the fuel key figure, since this fuel consumption has no connection to the mill’s steam requirement. Another change was that we divided the key figure for electricity consumption into two separate ones. One included all electricity consumption, the other excluded electricity for the production of heat in steam boilers. This was to ensure that the varying operation of electric boilers depending on electricity prices would not misleadingly affect the electricity key figure for the operation of the pulp and paper production itself.
Another point of view we received was that the actual measurement of steam and electricity was not always designed to be able to allocate the consumption to the different process sections for which the key figures were to be calculated. This was not something we could take into account in the method itself, but was an observandum as a source of error. In order to get the best possible calculation results, the measurement points may need to be adjusted, but when this is not possible or reasonable, assessments of the distribution must be made by the mill itself based on the knowledge it has about the plant.
From a policy and industrial perspective, how do you see this method aiding compliance with the EU Industrial Emissions Directive and motivating energy-efficiency improvements?
In my opinion, the use of key performance indicators is the only reasonable method to assess and ensure that industry meets the objectives and requirements of the IED. With a fair and transparent calculation method, it should also provide motivation for industry to take action to demonstrate its improvement in energy efficiency with calculated key performance indicators.
What challenges remain in scaling this method beyond the Swedish industry, for example, dealing with differing feedstocks, mill configurations, or data availability in other countries?
Different types of wood (softwood and hardwood of different types) are not something that affects the calculation methodology itself. Regardless of which type of wood is used, the energy flows are the same. The differences in energy consumption that the type of wood may give are something that must be taken into account when evaluating the calculations. The calculation file provides space to specify this, as well as a number of additional production conditions, as metadata.
Within the calculation methodology developed so far, there are methods for calculating key figures for all the production variants that exist in Sweden, which is 12. The vast majority of mills in Europe should be able to fit into one of these variants. However, it is possible that there are some additional combinations. For such, it will be possible to create new variants to fit these.
Looking ahead, are you planning to apply or adapt the methodology for other energy-intensive sectors, and what steps will you take to embed the method into regular industry benchmarking systems?
I have now retired from my work at the Swedish Environmental Protection Agency, and personally do not actively participate in the continued work on developing and introducing the calculation method. What I know is underway is to work on the calculation procedure itself technically. Instead of 12 separate calculation variants, the intention is to create a common calculation method for all types of pulp and paper mills where the different variants of production composition are embedded as options in the common calculation method.
The big challenge ahead is to get the calculation accepted by the industry and other EU member states. This requires that more mills use it, try it out, test it and hopefully find that the method is up to the task. An important player is CEPI, the European forest industry’s cooperation organization. If CEPI supports the calculation method, it is a big step forward.
Although the method was originally developed to be used within the framework of IED, it can of course be used in many other contexts; energy surveys, sustainability reports, permit processes, etc. Whether this will happen depends on the ambitions of both authorities and companies.
The more calculations that are made with this standardized method, the better the basis for using the results and taking into account different production conditions. It is then possible to see what significance, for example, the type of wood, the brightness of the pulp, the basis weight of the paper, etc., have on the amount of energy consumption.
Hassan graduated with a Master’s degree in Chemical Engineering from the University of Chester (UK). He currently works as a design engineering consultant for one of the largest engineering firms in the world along with being an associate member of the Institute of Chemical Engineers (IChemE).
