Uncertainties in LCOE of EGS geothermal - an open question

Contents

LCOE – the love/hate energy metric;?? LCOE variability;?? LCOE & geothermal EGS;?? EGS LCOE Insights from Western US;?? EGS LCOE insights from Iceland and France;?? NREL 2024 Update;?? External Lazard (2023) LCOE estimates;?? Where are the hoped for improvements coming from?;?? Where would these improvements put EGS competitively in 2035?;? Sense checking against older EGS LCOE projections; Summary; References

LCOE – the love/hate energy metric

LCOE is an imperfect metric - as every metric always is.? This we know.? It is the levelised cost of energy, designed to compare the life cycle costs of different types of energy delivery, per MWh delivered.

It deals with the costs of a success case plant, so does not really account for the chance of success & failure.? I.e. for geothermal the upfront costs required to verify a resource before final investment decision (FID) are greater than surface-based options, and not accounted for in LCOE where the FID is a no.??

LCOE has unavoidable subjectivities about where the life cycle analysis begins and stops, and whether various things like environmental impacts (e.g. emissions etc.) are included financially in some manner.?

LCOE makes no direct account of the different attractions of "flavours" of energy, like constraints on land footprint, 24/7 supply reliability, absolute size and scalability, etc.?

LCOE may be selective in how it captures costs of various operational risks, impacts or accident, and related monitoring.? For geothermal EGS an example is an induced seismic event.

LCOE looks at the metric per type of project and does not consider the wider energy needs of a mix of sources and their constraints and ceilings in total possible supply. Something may be the cheapest LCOE, but it’s no panacea if ultimately only enough is available, for whatever reason, to supply 10% of what is required at a holistic national level.

Yet, it is a useful term.? So it is good to understand what it is for geothermal EGS.?

That is, an Enhanced Geothermal System where fractures are either created or enhanced for permeability through artificially engineered stimulations.? At one end of that spectrum is “fracking” which artificially creates and props open at high pressure tensile fractures with chemical “proppants”.? At the other end, more common in EGS geothermal contexts is a less highly pressured shear fracture stimulation of existing fractures – a “tickling”.?

Induced seismicity risks are much less with the latter, and a function of the background seismicity vulnerabilities that are already present in an areas, but which might be prompted prematurely & “released”. ?Monitoring in any such projects these days is typically fierce and operations moderated and/or halted at any untoward signs of an unwanted size seismicity.

LCOE variability

It goes with the territory, from the above, that LCOE estimates vary a lot.? It is also a fairly frequent habit to provide projections of future LCOE which are even more subjective as they involve assumptions about future costs and prices and productions that no-one can really know for sure.

Without suggesting it tells us any answers, one useful reality check is always to go back in time and see what LCOE reductions were forecast historically and assess whether they actually came about.? Probably needless to say, we often have a propensity to be aggressively positive in forecasting production and cost and price figures for the things we like.? This is good to be aware of.? It is always of course helpful too, to understand how much variation in current LCOE costs is occurring.? Without choosing or needing to know which is right and which is wrong, it gives a feel for uncertainty.?

LCOE & geothermal EGS

Figure 1 gives a feel for some of the things that either aren’t fully captured for LCOE with EGS geothermal or are strong modifiers on input assumptions.??

In Figure 1a we see how the amount of investment required to prove up a resource is a much larger component of any subsurface project like a geothermal one, than it is for a surface one, where the energy source is easily measured.? In the subsurface we have to spend a lot of money to prove up a resource before we truly know we have a feasible project and a final investment decision (FID) can take place.? In contrast, for a wind, or solar, project, it is relatively easy to ascertain what the resource is prior to FID.? Now in a success case, these costs are captured in any LCOE estimate, but importantly what is not captured in an LCOE estimate, is the risk that the FID decision is no-go.? In that case the red part of Figure 1a is an unredeemed loss.? LCOE only looks at the costs involved in active plants, it doesn’t (normally) incorporate the risk of failure for those that never get to that point.? Some authors like Park et al. (2021) are trying to do that better.

In Figure 1b we see that well depth and cost model is an important component, and that well cost is a nonlinear function of depth, though it is possible to approximate a constant value of cost per m with an error bar that is not too wild (~$2.7M per km +/- 25%).? We also see that there are possible costs related to any damage of induced seismicity.? Now admittedly a lot of lessons have been learnt since the largest EGS related earthquake (Mw 5.5) to date at Pohang, Korea, in 2017 ((Woo et al., 2019), but the cost of just under a billion dollars, is to be noted.?? So such things can be captured in LCOE as an insurance cost, or if such insurance is not available then a risked cost – noting the risk of such an event is very small, but smaller costs of around a few million dollars are not impossible associated with smaller magnitude Mw 4 events, depending of course on the proximity of the project to vulnerable infrastructures and people, and the local background seismic risk.

In Figure 1c we see the good news story being recognised in the United States - of the good improvement in consecutive well drilling costs that is possible in EGS projects, with the Fervo Nevada example revealing cost behaviours nearly 70% cheaper relative to the first well - occurring by the fifth well in a project.? How representative this is, of course depends on a much larger database of long-lived commercial EGS projects than we currently have, but it is encouraging.? EGS projects are now likely in the ~ 60-75 bracket globally, given that 64 were accessible for study in 2020 (Pollack et al., 2020)? - see Figure 2.? How many remain in commercial or pilot R&D operation is a harder question to answer, but it is likely in the ball-park 25-35% -i.e. 20-30 (Pollack et al., 2020).? I stand open to correction or assistance if anyone has more up to date figures.

It remains then to ask the question what geothermal EGS LCOE information do we have currently and from the past, how variable is it, and how successful have past predictions of change been?

EGS LCOE insights from Western US

Anyone who knows anything about EGS will recognise that a lot of interesting stuff is happening in the western US at the moment, and the US DOE (Department of Energy) and the National Renewable Energy Laboratory (www.nrel.gov ) have long been active in promoting and analysing the technology.? NREL in particular has been very active in the techno-economic analysis of geothermal applications and indeed all renewable energy options.? Their annual technology baseline (NREL, 2024a) is a particularly useful reference, based for geothermal on the assumptions and analysis of the GeoVision: Harnessing the Heat Beneath our Feet study (US DOE Department of Energy Geothermal Technologies Office, 2019) and associated summaries (Mccabe et al., 2019; Susan Hamm et al., 2019).?

We can see some of this analysis in Figure 3.? We note an assessment of mature existing/conventional geothermal technologies as having an LCOE of about $70/MWh in 2022 and similarly now, with a projection to about $55 +/- $10/MWh by 2050.??? Helpfully, there is also on the site (NREL, 2024a), a raft of comparisons available for other energy sources, but perhaps the most interesting comparison is against utility scale solar PV with battery storage – which is similarly estimated at $80/MWh in 2022, down to $70/MWh in 2024, and projected to reach about $50 +/- $15/MWh by 2050.?

We can compare to their estimates of “nascent” geothermal technology, predominantly EGS which has a much wider range, but is pinned as ~ $95-100/MWh in 2024, down from ~ $170/MWh in 2022, and projected to range between $40/MWh and $140/MWh, i.e. $90 +/- $50MWh by 2050.? Without commenting on the likelihood or otherwise, their lower estimates include a very aggressive decline in LCOE between 2023 and 2035 from?$113/MWh to $42/MWh – a decline of 63% in that period.

Also in Figure 3, we can see a statement issued by the US DOE in 2022 (US DOE (Department of Energy), 2022) which rather cryptically in terms of the NREL LOCE estimates, talks of a 90% reduction in geothermal EGS LCOE to $45/MWh by 2035.? So it shares a similar view of the best-case scenario LCOE end-game, but the implied 2022 value for EGS LCOE is $450/MWh, which is considerably harsher than the NREL view – nearly three times more expensive.? Is the latter a bit of political licence being deployed for dramatic effect?? Or is it real?? Not sure, any insight into reasons behind the scale of this difference would be appreciated, but this is discussed further in a later section. ??

EGS LCOE insights from Iceland and France

Another very useful set of analyses is provided for individual examples from Iceland and France (Cook et al., 2022; Sigurjónsson et al., 2021).? While these are specific individual examples, note that any figures are future estimates when relating to production, costs and pricing, and these projects have not been producing for any extended length of time.? A fuller description of the two studies is provided in Figure 4 (Sigurjónsson et al., 2021).?

Of particular importance to note is that the Iceland example is an “add on” to an existing facility without any new power plant construction - and given Iceland’s unique volcanic heat resource - is relatively shallow at 2.5 km.? The Iceland example is then a relatively unusual outlier from this cost analysis perspective.? The Vendenheim project from the Upper Rhine Graben in Alsace France targeted fractured granites at around 4.5 km so is perhaps more representative of the typical fare happening today.?

Both projects were part of a paired “H2020-DEEPEGS” study, and more detailed information, particularly on the more representative Vendenheim example is informative, including associated induced seismicity events (Cabral, 2018; Gaucher et al., 2021; Peter-Borie et al., 2019; Sanjuan et al., 2019).

For our purposes though, it is the comparative techno-economic analysis that is of most interest, and Figure 5 details some of the cost assumptions for both examples (Cook et al., 2022).? Including a production plant, the Vendenheim project is estimated at roughly 100 million Euros for delivery of a 40-80 MW power output (Sigurjónsson et al., 2021).

In Figure 6, we see the very different cost-breakdown of the two projects (Cook et al., 2022) and the more representative Vendenheim one is notable in a number of features.? Perhaps most notably is that when production plant costs are included, then only 31% of the total costs are related to drilling, i.e. 69% of costs are non-drilling costs.? Any improvement in drilling costs for EGS projects therefore needs to be taken in this context.?

While it is perhaps a little older, an interesting analysis of CAPEX costs for different types of geothermal projects is also provided in Figure 7 (Sigfússon & Uihlein, 2015).? Importantly, a breakdown for ORC binary plant hydrothermal is compared against ORC binary plant EGS (see Figure 17).? The most notable difference is the greater proportion of costs for exploration, drilling and stimulation, which together constitute 52% for the EGS breakdown and 34% for the conventional hydrothermal one.? This is slightly at odds with Figure 6, but less so if we note that the latter also includes OPEX costs, and there may be a difference in where stimulation costs are recorded.? In other words it seems likely that a large part of the 18% difference in exploration, drilling, and stimulation costs between the EGS and conventional is related to the stimulation costs. ??Also of note is a near doubling of insurance costs in Figure 7.

Cutting to the chase though, the projected LCOE, based on an assumption of success and the output assumptions of Figure 4 (5-30 MW Iceland, 40-80 MW France), as calculated by (Sigfússon & Uihlein, 2015) is of the order €16-18/MWh for the Iceland example, and €45-55/MWh for France, as shown in Figure 8, for a variety of assumed discount rates as used for life cycle time discounting in NPV calculation (2%, 5%, 8%).? The Icelandic example again reflects the lack of any need for a plant, and the ability to take advantage of existing infrastructures, so the French example is more representative.?

These French estimates compare with the best-case 2050 scenarios of NREL (NREL, 2024b, 2024a) as shown in Figure 3 and Figure 9.? Again a key point not to lose sight of is that LCOE takes no account of project failure chance – it assumes success, so when comparing with options for which a failure chance is much less, that should be remembered in assessing competitiveness.? Success in this context really relates to delivery of temperatures and flow rates within the expected ranges/probability distribution functions.

NREL 2024 Update

However, quite why the LCOE estimates of (Cook et al., 2022) are quite so good for EGS is not very clear without a more detailed investigation.? A 2024 update for the NREL estimations has also been published online (NREL, 2024b) and includes a greater resolution of the difference between mature hydrothermal flash, and mature hydrothermal binary.? These results are shown in Figure 9.? Hydrothermal flash plants are typically reserved for large volcanic-associated high enthalpy geothermal fields of a large scale – or for EGS.? As such the number of new hydrothermal flash plants likely is more limited, and the more typical “new” conventional geothermal plant today is a smaller scale medium enthalpy ORC binary plant.?

Figure 9 shows a table with the estimated LCOE’s for three selected years – 2024, 2035, and 2050, with the advanced (best), moderate, and conservative estimates of LCOE for each of hydrothermal flash, hydrothermal binary, EGS flash, and utility scale solar plus battery storage.? The NREL estimates indicate that EGS best case scenarios are anticipated to be between 60-100% of the LCOE of a hydrothermal binary plant, but note this assumes any EGS plant is flash and not binary (see Figure 17).? The typical assumed ratio of a binary plant LCOE relative to a flash plant is 135% to 150%, so any EGS shift to binary would have a significant impact on LCOE.?

Comparing to solar with storage, the NREL estimates of Figure 9 conventional binary (again not the lower LCOE flash) has an LCOE ratio (i.e. worsening if > 100%) of 110% to 185% relative to solar depending on scenario, but in general worsening with time – presumably factoring in a cheapening of both solar and battery storage. Meanwhile for EGS relative to solar, the figures range from about 100% to 145%, but again this assumes flash plants.?

In all the comparisons, only the like scenarios are compared, i.e. conservative/conservative, not conservative/advanced.? This is probably a fair assumption for all the inter-geothermal comparisons, but it is possible the solar geothermal comparisons might be more subject to different scenario permutations, so for example, the worst case EGS would be ~225% of the best case solar scenario in 2035 and ~250%? in 2050.? One additional point worth noting from Figure 9 is that of all the worst-case LCOE scenarios in the table, the EGS (with flash plant) one is the worst.? EGS with binary would by implication be worse again.

External Lazard (2023) LCOE estimates

How do these renewable energy focused estimates compare with a wider external estimate of LCOE for 2023 (Lazard, 2023)?? These estimates are broadly compatible with NREL’s 2024 unsubsidised conventional geothermal estimates, and Lazard only assume conventional geothermal, as shown in Figure 10 , with a value of $61-$102/MWh.? Although wind & solar with storage have a range that extends cheaper, with the storage included, the averages are not very different. The same is true of gas combined cycle.? Nuclear, it is suggested, is more expensive but it does offer a somewhat different “flavour” at a potentially very large GW scale. That is arguably relevant because it potentially provides a security "backup" option if cheaper LCOE low carbon options can't deliver all of the required amount.

Where are the hoped for improvements coming from?

Having observed some quite aggressively positive “best case scenario” EGS costs from the US DOE, NREL and other sources, it is obvious to ask what is driving those.? This topic is covered in some detail in a number of key references and reports (Augustine et al., 2021; Dupriest & Noynaert, 2024; El-Sadi et al., 2024; US DOE et al., 2024; US DOE Department of Energy Geothermal Technologies Office, 2019), but is reviewed accessibly in the GeoVision and Pathways to Commercial Uplift documents form the USDOE (US DOE et al., 2024; US DOE Department of Energy Geothermal Technologies Office, 2019).

As Figure 11 shows, there are a number of key assumptions in the 2019 US DOE GeoVision document, and these include:

1.????? Exploration costs a half to a quarter that of hydrothermal.? Presumably this is most related to the reduced need to detect naturally porous reservoir.

2.????? A one third reduction in exploration drilling costs.

3.????? Significant reductions in appraisal well costs.

4.????? A 50% increase in appraisal well success rate.

5.????? A 2/3 reduction in the number of appraisal wells required.

6.????? References to the drilling cost well cases of previous work (Lowry et al., 2017), where the base case is approximately 2.5 to 4 times the cost of the “ideal” case for depths of 3km and more, and about twice for shallower depths, which means a base plus 50% as referred to for EGS appraisal well is about 3.5 to 6 times more costly than the “ideal plus 0%" envisaged.? This is not a plus or minus of some fraction in going from the base case to the ideal – it is huge improvement – and shown in more detail in Figure 12.

7.????? An ability to improve the manufactured (i.e. stimulated) flow rates for EGS wells by a factor of 2-3.

8.????? An ability to improve the well productivity (well flow rate as a function of well pressure drawdown) by a factor of ten for EGS projects. In simple terms, a measure of how much more you get as a function of how hard you suck.

What should be apparent is that all of these improvements are major tasks.? Not impossible, but the best cases on eight very fundamental aspects.? That should temper our expectations of the best cases being delivered on all fronts.? Such a chain would be improbable, but on the positive side, there are lots of variables where we can look to for some improvement, and where improvements indeed look to be happening. Notably one thing the US DOE (2019) study did not incorporate is any consideration of supercritical geothermal, in an EGS or other context (see inset Figure 11). The technology was considered too immature for formal LCOE analysis at that point.

In the recent “Pathways to Commercial liftoff: Next-Generation Geothermal power” document (US DOE et al., 2024), this staircase to improvement is elaborated on more, and in part (not wholly) explains their “90%” reduction in EGS LCOE proposed in a 2022 “Energy Earthshot” as per Figure 3.? This staircase is shown in Figure 13.? What is evident, is that their “starting gun” for EGS LCOE is a base level in 2021, where just under $28K/kW is envisaged, or roughly ~￡340/MWh.? The target reduction to something like $45/MWh (87%) involves the following observations from 2021 to 2023:

• Drilling cost, well stimulation, and exploration well number improvements, suggesting nearly 50% reduction is possible, implying an LCOE of ~ $175/MWh.

Subsequent improvements envisaged from here on in, to take this to ~ $45/MWh, in a decade, include:

1.????? Improved data collection and analysis.

2.????? Lowering further still the number of exploration wells required.

3.????? Reducing further the average drilling cost.

4.????? Increased ability to create larger wells capable of larger flow rates.

5.????? Improvements in cementing techniques and casing costs.

6.????? Improved ability to deal at the power plant with the higher fluid volumes delivered by item 4 above (an operations and maintenance issue – scale, corrosion. etc.)

7.????? Models of centralised development hubs +/- modular “spokes”.

1-7 explicitly take to about $60-70/MWh (best case), and a further quite significant “increment” of 25-30% reduction – likely ongoing improvements on all the above (?), is envisaged as an item 8 to take to $45/MWh, with the goal to do so by 2035.

This is another aggressively positive list of goals.? That’s not a criticism, such targets are fine, but they need to be treated as goals and not “likelihoods”.?

?It should be emphasised that improvements are likely based heavily on a relatively small number of projects – notably FORGE in UTAH and Fervo in Nevada ((El-Sadi et al., 2024; Norbeck et al., 2023; Simmons et al., 2021), and that an LCOE calculation necessarily involves assumptions about full life cycle production, price obtained, and costs – and none of these projects have reached that full life cycle yet.?

That’s not to dismiss many reasons to be cheerful, it is just to highlight a simple fact that an LCOE figure is not demonstrated definitively until project end of life & decommissioning, when all the operational issues, breakthroughs, costs and rewards are fully appreciated and documented. That is recognised in Figure 13 in describing this near-halving as being the result of “FOAK” or “first of a kind”.? ?Great news, but good to note that repeated long-term verification and replication at a global level is the pre-requisite for any real “game-changing”.

Where would these improvements put EGS competitively in 2035?

The question of where such improvements might place geothermal EGS competitively, in a decade’s time in 2035, is also addressed by the US DOE document (US DOE et al., 2024), and shown in Figure 14.? What we observe is that the best case EGS outcomes place it in a competitive position, and the worst-case outcomes don’t.? Onto that we need to add the uncertainty around the chance of success case at the initiation of any individual EGS project, for any time in the future.?

Any geothermal project is a very bespoke beast. As we see in Figure 11, the chance of success assumed for any individual production well of 50% now, is assumed to improve to 75% in the future (US DOE Department of Energy Geothermal Technologies Office, 2019).? Maybe it will even go better than that – conventional geothermal production wells are often around 80% success chance these days.? Yet to do geothermal is to recognise that there is a level of well-scale in-situ rock variability that inherently holds some level of irreducible pre-drill risk.? That is to say there are some things remote geophysics will just never see in a deep opaque subsurface until we go there with a drill bit.? This particular technical risk is not present for surface energy sources and is not captured by LCOE figures.?

This issue is discussed more by (Park et al., 2021), who describe two mathematical treatments of a “risk adjusted LCOE” (LCOE RAD) and a “certainty equivalent LCOE” (LCOE CE).? Their approach is to treat both parameters as a function of the probability distribution of CAPEX and OPEX standard deviation.? Arguably that is more a case of capturing uncertainty than outright failure risk, but whatever the case, the assertion holds that one thing traditional LCOE does not capture is the cost of project-busting failures. They might not be very common, but it is good to realise it.

A key point to note is that nobody is hiding the uncertainty – Figure 14 is explicit in this regard.? The scenarios between $45/MWh and $145/MWh ($95+/-50/MWh) may not be strictly uniform (equiprobable) in distribution, but this is a real described range, and the high end is not attractive, especially when the strongest competition is renewables plus storage, which is routinely being predicted at around $45-85/MWh by 2035, I.e. $65+/-20, with an [allegedly] much smaller level of uncertainty.? Utility-scale solar is the most common direct competitor, particularly in the regions of North America, the Middle East and North Africa, the circum-Mediterranean, Australia, and South Asia, where EGS is not infrequently proposed. ?

As already stated, LCOE is far from being the only important metric, and land footprint considerations are key too, amongst other things.? Critical mineral equations are less of an issue at this point but may become more so, but this situation is very dynamic and also technology led, especially on a battery storage and turbine-type front.? An example is sodium batteries as opposed to lithium – especially in static large-scale storage (Lu et al., 2021; Niu, 2024), along with other developments – see Figure 15. ??

Solar PV itself is most dependent on copper and aluminium like any electric deployment, while Geothermal depends heavily on high temperature steel alloys and requirements for nickel, chromium, molybdenum, and titanium in particular.? Quite a lot of distracting nonsense is spoken about the various critical mineral requirements of different clean energy options, so it is worthwhile acquainting ourselves more with those issues, and the dynamic, constantly evolving nature of those requirements as technologies advance – often rapidly and unpredictably.

A useful introduction to the subject from the International Energy Agency was updated in 2022 (IEA, 2022) and selected elements are shown in Figure 15.? It is important to realise, that while improvements and workarounds are equally possible in a geothermal arena, the critical mineral and mineral implications of any large scale up of geothermal deployment, are not negligible, just as this is equally true of any other renewable energy option.? Different, but present and not absent.

Sense checking against older EGS LCOE projections

As referenced early on in this article, it is useful when we consider present day EGS LCOE projections, to examine how successfully or otherwise historic LCOE projections have been.?? That is not to suggest current day ones are incorrect, it is just to get some sense of the uncertainties and variations that exist. ?There are of course real developments, not available to past authors, that change things, nevertheless, the range of past LCOE predictions is a helpful qualifier.

Figure 16 shows the discussions of two such studies (Limberger et al., 2014; S. M. Lu, 2018).? Lu (2018) envisages ( or at least suggests as possible) a 77% reduction in EGS LCOE from around $235/MWh in 2011 to $55/MWh in 2030.?? The latter is not wildly off what other studies are suggesting for 2035 (Figure 9) as best-case scenarios (NREL, 2024b).? Limberger et al. (2014) operating from a European perspective, are somewhat more conservative, envisaging €100/MWh by 2050, from a base of €215/MWh in 2020.? Usefully Limberger et al. (2014) also offers a sensitivity treatment of LCOE, with the four relatively similarly-impacting biggest hitting parameters to affect LCOE being temperature, flow rate, stimulation cost, and well cost, with variations in the range +/- 20-40% recognised for each parameter (Figure 16).

Summary

What can we say then??? Clearly LCOE is a bit of a subjective and mixed bag metric, by no means the only important metric, and needs to be treated with care - yet provides a useful comparison of life cycle costs & competitiveness of a success-case power generation facility type.?

What is also clear is that exciting improvements are happening in EGS geothermal technologies, with improvements in many aspects of such developments, but notably drilling.?

These improvements yield encouragement, that in the best case scenarios, LCOE improvements for deep geothermal might be LCOE competitive with other renewables such as utility scale solar with battery storage, in as little as a decade.? Onto this comparison of a single metric needs to be considered other elements – such as land footprints, and geological risk, that LCOE does not capture, but it is interesting that such competitiveness seems - at least in theory - within reach.

However, we also need to take note that the range of LCOE uncertainty is much greater for geothermal EGS LCOE than for its prime competitors, and the moderate and conservative LCOE estimates for EGS are significantly less competitive on that LCOE basis alone.? That is not to say that there can’t be other aspects which contribute to overall strategic competitiveness in those scenarios, but a competitiveness on a standalone LCOE basis is far from assured.?

Fundamental to this, is the understanding that a LCOE estimate is a projection and not an actual until a full life cycle, including decommissioning, of a facility, has been completed – with a full understanding of life cycle costs, rewards, successes and failures. Geothermal projects are inherently bespoke to the unique geology of each location, and important time dependent factors such as temperature, geochemistry, pressure and flow rate, can critically affect projects - for better and worse than expectation. Importantly, in subsurface fluid flow, sudden changes in permeability connectivities are not unheard of.

The sophistication of geological modelling is improving fantastically year on year, but all models are always at the mercy of the scenarios envisaged, and this historically is always the trickiest thing to fully capture in a subsurface setting. We are excellent at modelling the scenarios we imagine, we are not always so good at capturing all the potential scenarios.

It is also important to recognise that dependence on “first of a kind” (FOAK) project information, while encouraging, is to be treated appropriately.? The confidence in any LCOE projections for the future is a function of how many data points are involved.? That is to say the replicability of success.?The FOAK data points are undeniable traliblazers we sincerely hope can be routinely replicated, and reasons to be cheerful on that front exist - but we don't know they can with any certainty until they actually are.

On this basis we can take encouragement that a $40-55/MWh best-case scenario of geothermal EGS LCOE by 2050 seems to be a relatively recurrent theme amongst those involved in projecting such things from active projects and global reviews.? However, we can also take from the literature another relative consensus, that this is a best case, and that the range of outcomes varies from that lower bound of ~ $40/MWh to something like $140/MWh ($90 +/- 50/MWh) by 2035 to 2050.?

It is also worth noting that estimates generally seem to base EGS application on an assumption of flash power plant type costs, whereas any binary plant deployment would likely - taking NREL analysis as read - add to the LCOE by a factor of a third to a half. For those less acquainted with these terms, Figure 17 shows some of the differences. ORC binary plants involve heat exchange to a second working fluid to allow vapour generation at lower temperatures for turbine function and power generation, whereas flash power plants are a simpler operation.

The higher figures of estimated EGS LCOE would in general be far more problematic from a competitiveness perspective, given rapid advances in other renewable technologies.? Notwithstanding particular requirements of particular customers for particular “flavours” of energy supply, going beyond just LCOE.

Moreover, the probability distribution within that $40-$140/MWh range remains highly uncertain, and the data points informing the character of that distribution are few.? So we continue to watch with interest on a worthy new energy player for consideration, but we also watch with caution.

References

Augustine, C., Fisher, S., Ho, J., Warren, I., & Witter, E. (2021). Enhanced Geothermal Shot Analysis for the Geothermal Technologies Office. www.nrel.gov/publications .

Cabral, E. (2018). Thermoeconomic analysis of EGS/Deep Geothermal Resources in the region of Alsace, France.

Cook, D., Sigurjónsson, H., Davíesdóttir, B., & Bogason, S. (2022). An environmental life cycle cost assessment of the costs of deep enhanced geothermal systems-the case studies of Reykjanes, Iceland and.

Dupriest, F. E., & Noynaert, S. F. (2024). Continued Advances in Performance in Geothermal Operations at FORGE Through Limiter-Redesign Drilling Practices. SPE - International Association of Drilling Contractors Drilling Conference Proceedings, 2024-March. https://doi.org/10.2118/217725-MS

El-Sadi, K., Gierke, B., & Howard, E. (2024). Review Of Drilling Performance In A Horizontal EGS Development.

Gaucher, E., Jolivet, M., & Peter-Borie3, M. (2021). Preliminary Results of the Seismicity Induced During the Stimulation of the Vendenheim Enhanced Geothermal System (France).

IEA. (2022). The Role of Critical World Energy Outlook Special Report Minerals in Clean Energy Transitions. www.iea.org/t&c/

Lazard. (2023). LCOE Plus 2023.

Limberger, J., Calcagno, P., Manzella, A., Trumpy, E., Boxem, T., Pluymaekers, M. P. D., & Van Wees, J. D. (2014). Assessing the prospective resource base for enhanced geothermal systems in Europe. Geothermal Energy Science, 2(1), 55–71. https://doi.org/10.5194/gtes-2-55-2014

Lowry, T. S., Foris, A., Finger, J. T., Pye, S., & Blankenship, D. A. (2017). Implications of Drilling Technology Improvements on the Availability of Exploitable EGS Resources.

Lu, S. M. (2018). A global review of enhanced geothermal system (EGS). In Renewable and Sustainable Energy Reviews (Vol. 81, pp. 2902–2921). Elsevier Ltd. https://doi.org/10.1016/j.rser.2017.06.097

Lu, W., Wang, Z., & Zhong, S. (2021). Sodium-ion battery technology: Advanced anodes, cathodes and electrolytes. In Journal of Physics: Conference Series (Vol. 2109, Issue 1). IOP Publishing Ltd. https://doi.org/10.1088/1742-6596/2109/1/012004

Mccabe, K., Beckers, K., Young, K. R., & Blair, N. (2019). GeoVision Analysis Supporting Task Force Report: Thermal Applications Quantifying Technical, Economic, and Market Potential of Geothermal District Heating Systems in the United States. www.nrel.gov/publications .

Niu, Z. (2024). Research Progress on Cathode Materials for sodium-ion batteries. In Highlights in Science, Engineering and Technology GTREE (Vol. 2023). https://wap.cnki.net/qikan-SDQE202312042.html

Norbeck, J., Latimer, T., Gradl, C., Agarwal, S., Dadi, S., Eddy, E., Fercho, S., Lang, C., McConville, E., Titov, A., Voller, K., & Woitt, M. (2023). A Review of Drilling, Completion, and Stimulation of a Horizontal Geothermal Well System in North-Central Nevada.

NREL. (2024a). NREL Annual Technology Baseline 2023. https://atb.nrel.gov/electricity/2023/technologies

NREL. (2024b). NREL Annual Technology Baseline 2024. https://atb.nrel.gov/electricity/2024/technologies

Park, S., Langat, A., Lee, K., & Yoon, Y. (2021). Measuring the impact of risk on LCOE (levelized cost of energy) in geothermal technology. Geothermal Energy, 9(1). https://doi.org/10.1186/s40517-021-00209-9

Peter-Borie, M., Loschetter, A., Maury, J., Peter-borie, M., Loschetter, A., Maury, J., & team, D. (2019). Assessment of Geothermal Well Productivity Improvement Technologies: an Overview from the DEEPEGS Project. https://brgm.hal.science/hal-02320214

Pollack, A., Horne, R., & Mukerji, T. (2020). What Are the Challenges in Developing Enhanced Geothermal Systems (EGS)? Observations from 64 EGS Sites.

Sanjuan, B., Negrel, G., Le Lous, M., Poulmarch, E., Gal, F., Damy, P.-C., & Négrel, G. (2019). Main geochemical characteristics of the deep geothermal brine at Vendenheim (Alsace, France) with constraints on temperature and fluid circulation. https://hal.science/hal-02335810

Sigfússon, B., & Uihlein, A. (2015). 2015 JRC Geothermal Energy Status Report. https://doi.org/10.2790/959587

Sigurjónsson, Cook, D., Davíesdóttir, B., & Bogason, S. (2021). A life-cycle analysis of deep enhanced geothermal systems-the case studies.

Simmons, S. F., Kirby, S., Allis, R., Bartley, J., Miller, J., Hardwick, C., Jones, C., Wannamaker, P., Podgorney, R., & Moore, J. (2021). The Current Geoscientific Understanding of the Utah FORGE Site.

Susan Hamm, P. G., Augustine, C., Tasca, C., & Winick, J. (2019). An Overview of the U.S. Department of Energy’s GeoVision Report: Preprint. https://www.nrel.gov/docs/fy20osti/74378.pdf

US DOE, Blankenship, D., Porse, S., Getler, C., O’Connor, M., & Kamaludeen, M. (2024). Pathways to Commercial Liftoff: Next-Generation Geothermal Power.

US DOE (Department of Energy). (2022). DOE launches New Energy Earthshot to Slash the Cost of Geothermal Power. www.energy.gov/articles/doe-launches-new-energy-earthshot-slash-cost-geothermal-power

US DOE Department of Energy Geothermal Technologies Office. (2019). GeoVision: Harnessing the Heat Beneath Our Feet. https://www.osti.gov/scitechonlineordering:https://www.ntis.gov/help/ordermethods.aspx

Woo, J. U., Kim, M., Sheen, D. H., Kang, T. S., Rhie, J., Grigoli, F., Ellsworth, W. L., & Giardini, D. (2019). An In-Depth Seismological Analysis Revealing a Causal Link Between the 2017 MW 5.5 Pohang Earthquake and EGS Project. Journal of Geophysical Research: Solid Earth, 124(12), 13060–13078. https://doi.org/10.1029/2019JB018368

?