# Realistic Costs of Renewables plus Hydrogen Storage

### How the costs outlined in the Royal Society Report would change by using more realistic assumptions

# Introduction

A significant portion of the criticisms I made in the article about the Royal Society Large Scale Storage report were about the cost and efficiency assumptions made in the original report. I believe many of their assumptions were unrealistic and created the impression that the overall system costs of the proposals would be cheap. Indeed, the report made the claim that electricity provided by wind and solar supported by large scale hydrogen storage would compare favourably to alternatives. Since then, I have been working on a model to estimate more realistic costs of an electricity system solely reliant upon wind and solar renewables plus hydrogen storage.

This article examines the cost assumptions in the original RS report in more detail and makes alternative assumptions to show how the system costs will vary with more realistic assumptions.

# Methodology

The starting point is the assumption that the new system can be modelled as follows. First consider a “lump” of renewables with a Levelised Cost of Energy (LCOE) of ~£44.64/MWh using a 5% discount rate. The approach to calculating the LCOE follows the same approach as the Government in its various generation cost reports. 5% is the baseline discount rate used in the RS report. This is slightly different to the discount rate used by the Government. They use 5% for large scale solar, 5.2% for onshore wind and 6.3% for offshore wind in their 2020 generation cost report. For simplicity, the 5% used in the RS report has been used as a blanket figure. The £44.64/MWh result is close to the starting point of the high-end estimates of £45/MWh in Figure 24 of the RS report, shown below as Figure A.

These renewables are configured to deliver 5.7TWh per year or 1% of the proposed system. Scaling the system to the full size makes no difference to the cost per MWh, except if more expensive renewables such as floating offshore wind are required to meet demand. The rationale behind the RS report was that renewables on their own cannot match supply and demand and so storage is required. In the RS report, their storage was modelled as a single hydrogen store. However, they also raised the possibility of a dual store solution using a fast-cycle advanced compressed air system (ACAES) together with a slow-cycle hydrogen store. For the purposes of this financial model, the single hydrogen store has been used.

In financial terms, this store can be modelled as an additional lump of capex and annual operating cost (opex), together with a 30% uplift in generation capacity to 7.4TWh for the same net output of 5.7TWh as per the RS report. The uplift in generation capacity is required to cover the round-trip losses from producing the hydrogen, storing it and then using it to generate electricity again.

Using the baseline costs of the storage system in the RS report and adding the capex and opex using a 5% discount rate gives a result of £77.48/MWh which is close to the central green dot in Figure A above for the base renewables cost of £45/MWh. This establishes the validity of the modelling approach.

# Sensitivity Analysis

We can now vary the cost assumptions in the RS report and determine the impact on the levelised cost of energy. The sensitivity analysis has been carried out in two ways. The first is to calculate the LCOE based on today’s cost of the components. This is summarised in Figure B1 below. The second approach is based upon average costs over the life of the project if the cost reductions envisaged by 2050 in the RS report are realised. The results are summarised in Figure B2 below. Both sets of results are explained in more detail in the following sections.

# Adjusting the LCOE of Renewables

The first item to adjust is the starting point of the cost of renewables. The RS report starting assumption of £45/MWh compares to the actual results of ~£70/MWh for onshore wind and £63/MWh for solar (both in 2023 money) awarded in AR5. Currently, we pay ~£177/MWh for offshore wind CfDs and £110/MWh for onshore. The average FiT price which is mostly solar was £196/MWh in 2022, with current solar CfDs around £106/MWh in 2023 money. Of course, no offshore wind was awarded in AR5.

Plainly, the weighted average base generation cost of wind and solar is far higher than estimated in this report. By using the costs of onshore wind and solar in AR5 and estimating £100/MWh for offshore wind (in-line with the recent AR6 announcement), the weighted average generation cost doubles from their estimate to ~£90/MWh. The £100/MWh in 2023 money for offshore wind is also in line with the recent statement by Tom Glover, chairman of RWE, calling for a 70% increase in offshore wind prices.

Adjusting the discount rate up to 11%, increasing capex and reducing average load factor increases the cost of our lump of renewables above to £89.05, close to the assumed weighted average of £90/MWh indicated above. The 11% discount rate for 2023 was calculated by averaging the premium over the average of 10Y and 30Y gilt yields in the Government 2016 and 2020 generation cost reports (see Figure C).

Gilt yields have fallen somewhat since that analysis was conducted, so this rate might be considered a little too high. However, the H21 report used as a reference in the RS report decides in Figure 3.53 that:

“It is not considered practical to source low carbon hydrogen produced from offshore wind and electrolyser technology…The technical, project execution and commercial risks and commercial risks are simply too high.”

It could therefore be argued that investors would demand even higher returns before investing in such immature and high-risk technology.

# Applying the increased discount rate to Renewables plus Storage

Adding the same RS storage capex and opex as above to the adjusted renewables cost above, using an overall 11% discount rate gives a system cost of £143.68/MWh as shown by Item D in Figures B1 and B2 above. This is much larger than the highest system cost of ~£92/MWh in Figure A above. This is because the RS report used a high-end cost of capital of 10% and did not consider that rising interest rates (and so the discount rate) would also impact the underlying cost of renewable generation.

# Applying Realistic Electrolyser and Generator Costs

## Electrolyser Costs

Now to examine the cost assumptions made in the RS report about electrolysers to produce hydrogen and the hydrogen-fuelled four-stroke engines assumed to be used for back up electricity generation.

The RS report used £333/kW as the average cost of electrolysers. This compares to current costs of £880-£1,440/kW (using a $1.25/£ exchange rate) according to the IEA quoted in the RS report and 2050 costs of £160-720/kW. It is unclear whether these costs represent a base cost for the electrolyser or a fully installed cost.

The H21 report used as a source for other items in the RS report found the lowest current base costs for PEM electrolysers to be £1,600-£1,900/kW. The H21 report also found the current installed costs to be in the range £2,000-£3,000/kW. Applying the same improvement factors as the IEA would give an installed cost range of £364-£1,500/kW in 2050. Clearly the average cost of electrolysers from now until 2050 is going to be much higher than the £333/kW assumed in the RS report. Moreover, today’s costs are considerably larger than those assumed in the RS report.

Arriving at a suitable average figure is a matter of judgement. The logic applied here is first that fully installed costs are the most appropriate starting point. Second, the IEA figures are likely for the US, so greater weight has been given to the costs calculated for the UK by the H21 project. Thus, the starting point for 2018 is £2,000-£3,000/kW and the end point for 2050 is £364-1,500/kW. We could then apply various averaging techniques to determine the average cost for 2050 as shown in Figure D.

I decided to calculate the geometric mean of the geometric mean of current (2018) costs and the geometric mean of anticipated 2050 costs (£1,345/kW) and rounded it up to £1,350/kW as the average from now to 2050.Applying a similar approach to the IEA figures would yield a result of ~£650/kW. For today’s cost, £2,500/kW has been used from the H21 report and £1,160 using IEA figures.

## Generator Costs

The cost of the four-stroke engines used to generate electricity in the RS report were assumed to be £315/kW. The RS report also cited two references for similar (non-hydrogen) reciprocating engine power plants. These were the Goodman Energy Centre in Kansas and the IPP3 power plant in Jordan. The Goodman plant was built in 2008 and the IPP3 plant in 2014. The costs per kW of the IPP3 plant in Jordan was £1,067/kW and the Goodman plant cost £653/kW, both much larger than assumed in the RS report as can be seen in Figure E.

I calculated an approximate, round figure average estimate of £850/kW in today’s terms, ignoring the inflation since 2008 and 2014.

## Sensitivity to Higher Electrolyser and Generator Costs

Combining the new estimates for the average cost electrolysers and generators gives a levelised costs of £180.16/MWh and £206.14/MWh as shown by Item E in Figures B1 and B2 above. Some might argue that using the H21 electrolyser cost estimate is too harsh and the impact of using an average cost derived from IEA estimates will be discussed below.

# Cost of Storage Caverns

The RS report uses £727/MWh_{e} as the cost of the storage caverns. However, this is adjusted for the output efficiency of the generators, so the cost of the gross storage is £727*0.55 = £400/MWh. However, it is not clear how this figure was derived.

The H21 report used as a source (p392) estimates a cost of £1,991m for 8,052GWh of gross storage which equates to £247/MWh. At first glance it looks like the RS report has been conservative in its £400/MWh estimate. However, the H21 report also adds a further £3,427m of capex for the hydrogen transmission system. They assume quite an extensive network across the North of England. Adding in this transmission network gives a total capex of £5,418m resulting in £673/MWh of gross storage including the associated pipework and other equipment to transport the hydrogen from the electrolysers to the storage caverns and then on to the generators.

The location of the caverns will be constrained by geography; it is doubtful that there will be space available in the same geography for the electrolysers and generators. Moreover, the location of the electrolysers will be constrained by the availability of a reliable supply of fresh water. It is unlikely that the ideal locations for electrolysers and storage caverns will coincide. Indeed, the gross area taken up by the electrolysers, caverns and generators will be around 144km^{2} as shown in Figure F below.

Therefore, some kind of transmission network will be required. It is unclear whether the £400/MWh assumed in the RS report allows for enough of a transmission system. For the purposes of this analysis, I have not included a sensitivity factor for increased unit costs of the storage caverns, except for in the inflation adjustment discussed below. Nor have I estimated a cost for a reservoir to provide a secure supply of water to the electrolysers.

The other issue that could impact the size of storage caverns is the scale of inter-annual weather variations. Theoretically, more demand for electricity during cold years that are also less windy, like 2010 will lead to a need for more capacity in the system. This may push up the required storage capacity to 180TWh or more. Some combination of extra renewable generation capacity, more electrolysers and/or a bigger store would be required. Whatever the optimum configuration, this will lead to more capex to deliver the same average output and so push up the levelised cost of energy. This sensitivity factor has not been included in this analysis.

# Impact of Electrolyser and Generator Efficiency

The RS report assumed an average 74% efficiency for the electrolysers it proposes. This is at the top end of the 67-74% the IEA believes will be achievable by 2050. Current efficiency is in the range 55-60% according to the IEA.

The RS report assumes 55% efficiency for the hydrogen-fuelled generators it proposes. The Goodman Energy project it uses as a reference has an efficiency of 44.6%. Another project using hydrogen engines in locomotives expected a peak efficiency of 46%. Both are clearly well below 55%.

For the purposes of this sensitivity analysis, 55% has been used as the current electrolyser efficiency and 45% for the hydrogen generators. Figures of 65% and 48% respectively have been used for as the averages from now until 2050. The capex for the electrolysers has simply been scaled for the electrolysers by the ratio of the different efficiencies. For the time being I have ignored that more renewable generation capacity will likely be required to make the hydrogen at a lower electrolyser efficiency. Similarly, the capex of the storage caverns and the hydrogen generators have been scaled by the ratio of the different efficiencies because more hydrogen and more generators will be required to generate the same amount of electricity. I have ignored the potential compounding effect of potentially even more electrolyser and renewable generation capacity being required.

This gives the result shown by Item F in Figures B1 and B2 above, with the levelised cost rising to a cost today of £233.18 with an average from now to 2050 of £189.33/MWh.

# Impact of Inflation

All the cost estimates used in the RS report use sources that calculated their costs before the recent inflationary surge and normalisation of interest rates. The most recent sources are from 2018. To get a realistic view of levelised cost on today’s terms we need to add an inflationary uplift to the electrolysers, storage caverns and generators. The recent inflationary surge and rise in interest rates for the renewable generation is already included in Item C in Figure B above.

Recently, wind farm developers have complained of cost increases of 40% or more for manufacturing turbines. For the purposes of this analysis, I assumed a blanket 30% uplift in the capex values for electrolysers, storage caverns and hydrogen-fuelled generators to bring them into line with today’s prices.

This results in today’s levelised cost rising to £268.40/MWh and the average over the life of the project rising to £211.40/MWh as shown by Item G in Figures B1 and B2. As mentioned above, if the IEA-derived estimates for electrolyser costs are used instead, this inflation-adjusted costs become £215.47/MWh and £188.01/MWh as indicated by Item G* in Figures B1 and B2.

# Asset Life

All the above calculations have been made assuming an asset life of 30 years. This is generous for the renewable technologies, where 20-25 years would be more appropriate for the wind turbines and solar panels. This would add to the capex requirements over a 30-year lifespan.

The current expected life of an electrolyser is 10-20 years with the stack inside having a life of less than 20,000 hours. Clearly, the electrolysers would need to be refurbished and replaced during the thirty-year life, further adding to capex requirements.

The salt caverns themselves may well last longer than 30 years, but it is likely that the wells used to access them will need some form of refurbishment during their lives and perhaps re-casing.

The impact of the shorter asset lives has not been accounted for in this analysis, but it is safe to say that costs would increase further if they were.

# Impact of Higher Electricity Demand

The original RS report assumed a very low estimate of final electricity demand at 570TWh. As discussed here, the lead author of the RS report was also a named author of another report suggesting 1,500TWh was a reasonable estimate of final electricity demand in 2050. This second report suggested that the capacity for fixed offshore wind is limited, and as final electricity demand rises, we would have to start to rely upon floating offshore wind. The trouble is, in the recent announcement of the AR6 subsidy levels next year, in today’s money floating offshore wind will cost £241/MWh. Adding floating offshore wind into the energy mix will therefore push up the starting point of the average cost of renewable generation considerably.

# Conclusions

The Royal Society set out to demonstrate that the cost of energy using renewables plus hydrogen storage was reasonable. Their range of overall costs ranged from a mid-point of around £58/MWh assuming a 5% discount rate with baseline renewables costing £30.20/MWh to a mid-point around £85/MWh using a 10% discount rate with £45/MWh as the cost of renewables.

The highest mid-point figure for the whole renewables plus hydrogen system is about the same as the blended average cost of renewables indicated by the results of the recent renewables auction, assuming offshore wind around £100/MWh.

The unrealistic starting point for the cost of renewables together with low-ball estimates of the discount rate and the cost of the hydrogen system compounded to create the impression we could add hydrogen storage to renewables, and it wouldn’t cost any more than renewables today.

When realistic cost, efficiency and discount rates are used, the levelised cost of energy from renewables supported by hydrogen storage increases very dramatically, probably to well over £250/MWh using today’s technology. Cost might fall in the future, but even then, the average cost would be over £200/MWh in today’s terms. The extra sensitivity factors ignored in this analysis such as extra storage requirements, hydrogen transport system, shorter asset lives and the introduction of floating offshore wind to meet realistic demand estimates would push these costs up even further. This is too high to even contemplate when UK electricity prices are already amongst the most expensive in the developed world. The whole idea of an electricity system run on wind and solar renewables plus hydrogen should be dropped.

*If you have enjoyed this article, please share it with your family, friends and colleagues and sign up to receive more content.*

I’m wondering what happens when you generate the green hydrogen in one place and transport it to another.

For example, here in Atlantic Canada, we have at least two organizations vying to plant offshore and onshore wind turbines all over the place and then use the power generated to make hydrogen, which they hope to sell overseas. Beyond the costs you’ve calculated, it seems to me that adding in transport by ship (likely burning fossil fuels!) would likely make the whole enterprise even less financially viable.

The other bit of stupidity, in our case, is that over half of our electricity in Nova Scotia comes from coal, so they should have a market to sell wind power here, without going through all the nonsense of making hydrogen.

Sadly, our provincial and federal governments seem quite supportive of the whole hydrogen project, even to the extent of subsidies, which of course may be the reason why the business people wish to proceed.

One does not have to look far to discover that the energy to produce and isolate 1 kg of hydrogen exceeds the amount of energy that 1 kg of hydrogen contains.

Probably best to start there.