Reporting of "final" energy for electrolysis and ammonia synthesis

[pkyle]

Apologies for missing the calls where this may have been addressed, but has there been discussion of the appropriate reporting conventions for onsite electrolytic production and use of hydrogen? Energy commodity inputs and outputs involved with electrolytic ammonia synthesis seem especially difficult to pin down. As an example, the data table below indicates the flows of mass and energy involved with producing green ammonia from 1 EJ of electricity, assuming 60% electrolysis efficiency, 100% ammonia synthesis efficiency, 120 GJ/t for hydrogen and 18.8 GJ/t for ammonia:

Commodity         Mass (Mt)     Energy (EJ)
Electricity       NA            1
Hydrogen          5             0.6
Ammonia           28.3          0.53

There are several complexities that bear on the categorization of this set of transformations.

1. How the ammonia is used determines whether these transformations are appropriately categorized as "secondary energy" or "final energy". If the ammonia is used for agricultural purposes, then the production would be classified as "final energy consumption" whereas if it is used for energy purposes then its production should be classified as secondary energy production. The same logic applies to hydrogen as well.
2. Both the electricity and the hydrogen may be purchased or produced onsite. In fact the models might even represent green ammonia production without explicit tracking of either electricity generation or hydrogen production (i.e., all of the transformations could just be rolled into an estimated cost). If electricity is produced (and consumed) onsite, whether implicitly or explicitly in the models, should this be reported as electricity generation? Should it be reported as electricity consumption? The same questions apply to hydrogen. If e.g. we want to be able to estimate economy-wide capacity of solar panels, wind turbines, and/or electrolyzers in a scenario, we'd need to be tracking these quantities of both electricity generation and hydrogen production, even if both are produced (and consumed) onsite, in support of producing something else (hydrogen or ammonia / other chemicals). But there's also potential for inconsistency across all fuels in the energy system; the convention of reporting all within-facility transformations is quite different from standard reporting conventions. For example we don't currently report steam methane reforming as an energy transformation; that's just reported as final energy consumption of gas.

As far as I know, to this point the energy statistics world hasn't yet adapted to the use of hydrogen and/or ammonia as energy commodities, and given the complexities involved I expect them to put it off for at least another few years. The conventions that they have used for electricity, of tracking all gross flows (i.e., reporting all within-facility production and use), work reasonably well when there's only one energy commodity involved. But with three, that can all be used to produce one another, the existing conventions simply won't work (the IEA Energy Balances' already nebulous "own use" category would become a mess of double counting), and it's hard to foresee what will be adopted as the standard accounting convention.

[christophbertram]

This topic was not discussed in detail at today's call. But I think issues 20, 21, and 22, are somewhat related, so we should make sure to have one place to discuss these. I think the subgroup on Primary and Secondary energy will be the place for this. I think the accounting in our models should be somewhat easier than the one in the real world, given that the models do not differentiate between different sites, and the question of onsite vs. imported electricity/hydrogen (or at least in my view they should not: of course one could build in technologies that just produce ammonia from money, but I think most models will not do this, as this would also mean you cannot judge the resource availability anymore). Your example of SMR is also useful: I think it actually would be good to represent hydrogen production from natural gas in the historic time series of our models as much as possible, as this shows a subset of natural gas demand that is actually easier to supply directly by hydrogen.

So the hydrogen - electricity example should apply: And in my understanding we aim to represent total electricity generation, even if some of it used for hydrogen production, some of which in turn will later in that year be re-transformed to electricity. But from a grid balance (and analogously hydrogen supply-demand balance) perspective, it still makes sense to add all electricity generation (and hydrogen generation) up, even if one could call that partial double-counting.

Within electricity, knowing the amount of losses from storage in any technology that does not involve a dedicated energy carrier like hydrogen, transmission and curtailment is important, so creating an extra category like "Quasi-secondary Energy|Electricity|Storage loss/transmission loss/curtailment" or similar could be the solution for issues 20-22 (The "quasi" would indicate that this contains aggregate variables that annual models might track in reduced form, and which help to compare them to results of hourly models)

But perhaps the colleagues from the ECEMF study have already a another solution for this? @Renato-Rodrigues

[pkyle]

Thanks @christophbertram ! Just to briefly follow up on one point, I'm not aware of any datasets that estimate the production of hydrogen by nation, economic sector, and technology, even for a single historical year. If such a dataset exists, please let me know and we can see about using it for historical calibration.