Low carbon hydrogen will be vital for meeting our legally binding commitment to achieving net zero by 2050, with the potential to help decarbonise vital UK industry sectors and provide flexible energy across heat, power, and transport. Hydrogen production costs are a fundamental part of energy market analysis, and a good understanding of these costs is important when analysing and designing policy to make progress towards net zero. A recent report produced by the Department for Business, Energy, and Industrial Strategy (BEIS), presents estimates of the costs and technical specifications for different production technologies. REGlobal presents an extract of this report..
How levelised costs are calculated
The levelised cost of hydrogen (LCOH) is the discounted lifetime cost of building and operating a production asset, expressed as a cost per energy unit of hydrogen produced (£/MWh). It covers all relevant costs faced by the producer, including capital, operating, fuel and financing costs. The levelised cost of a hydrogen production technology is the ratio of the total costs of a generic/illustrative plant to the total amount of hydrogen expected to be produced over the plant’s lifetime. Both are expressed in net present value terms. This means that future costs and outputs are discounted, when compared to costs and outputs today. Technologies’ financing cost is applied as the discount rate (. This means it is not possible to express financing cost as a £/MWh component of the cost directly. Levelised cost estimates do not consider revenue streams available to producers (for example from sale of hydrogen). The main intention of a levelised cost metric is to provide a simple “rule of thumb” comparison between different types of hydrogen production technologies. However, the simplicity of this metric means some relevant issues are not considered.
Levelised cost estimates can be reported for different milestones associated with a project including the project start, the financial close and the online/commissioning year. In this publication, we report levelised cost estimates by online/commissioning year. As the evidence currently assumes a uniform build time of three years across technologies and excludes pre-development timings (due to lack of evidence), assuming project start of financial close would not introduce differences between technologies. In reality there is likely to be variation in pre development and construction timings across technologies and we are inviting stakeholder views.
Technologies
This section explains the technologies covered in this report. Explanations are based on the Hydrogen supply chain: evidence base that Element Energy put together for BEIS in late 2018.
CCUS-enabled methane reformation
Steam methane reformer with carbon capture, usage, and storage: A Steam Methane Reformer (SMR) is a mature production process in which an external heat source provides high-temperature steam for the reforming reaction that produces hydrogen and CO2 from a gas source, such as methane. Any excess steam can be used to generate power, which is sufficient to meet the power demand of the overall plant. This report assumes that SMR plants achieve a conversion efficiency of 74% (HHV).
There are two main sources of CO2 emissions; one source is the CO2 that is produced alongside hydrogen in the reforming reaction, with the other being the CO2 produced by the external heat source that provides the high-temperature steam for the reaction. The process of capturing CO2 is far simpler for the former, with the capture of CO2 from fuel combustion being relatively expensive, as it needs to be separated from nitrogen. Overall, 90% of all CO2 is assumed to be captured.
Due to the design of the plant, SMRs cannot be readily turned down and it takes a number of days to turn on or off. Therefore, an SMR acts very much as a baseload producer, i.e. a producer that is operating at constant (usually high, up to 95%) load factors year-round.
The technology varies in scale but is most likely to be deployed in the 100s of MW scale. This report covers a 300MW and a 1000MW illustrative example for 2020 to 2050 online years to show the impact of economies of scale, however other sizes are possible. Whilst 2020 has been included for comparison purposes, SMR with CCUS can only come forward once a CO2 transport and storage (T&S) infrastructure is in place. The technology’s technical life is assumed to be 40 years.
Autothermal reformer with carbon capture and storage: An Autothermal Reformer (ATR), unlike an SMR, is ‘self-heating’ (autothermal) as it partially combusts some of the natural gas feed to generate heat for the endothermic reforming reaction. To achieve partial combustion, oxygen is often used instead of air, which requires stripping out nitrogen. Whilst this requires an additional Air Separation Unit (ASU) and creates significant additional power demand that needs to be imported (a proportion of which could be offset by electricity generated from the heat of the hydrogen CO2 mix coming out of the reformer), it avoids the need for expensive post combustion separation of CO2 from nitrogen (required in an SMR). The autothermal process allows ATRs to achieve higher conversion efficiencies and, due to only a single CO2 stream, higher CO2 capture rates than SMRs. This report assumes a conversion efficiency of 84% (HHV) and a CO2 capture rate of 95%.
Due to the high-temperature thermal processes, an ATR’s output cannot be readily turned on and off. Whilst, unlike SMR, it has some ability to ramp up and down it remains far slower at doing so than an electrolyser. Therefore, an ATR acts very much as a baseload producer, i.e. a producer that is operating at constant (usually high, up to 95%) load factors year-round.
The technology is most likely to be deployed in the 100s of MW scale. This report covers a 300MW and a 1000MW illustrative example for 2020 to 2050 online years to show the impact of economies of scale, however other sizes are possible. Whilst 2020 has been included for comparison purposes, ATR with CCUS can only come forward once a CO2 T&S infrastructure is in place. The technology’s technical life is assumed to be 40 years.
Autothermal Reformer with Gas Heated Reformer with carbon capture, usage, and storage: A Gas Heated Reformer (GHR) could be added to a methane reformer, typically to an ATR, improving the overall conversion efficiency of the plant. This report assumes a conversion efficiency of 86% (HHV). This is based on simulation estimates undertaken by Jacobs in 2018 and should be considered aspirational as a future case, given that the GHR design is less proven at scale. This is achieved as the hot gases coming off the ATR are used to heat a mixture of natural gas and steam as it enters the GHR, which is partially reformed to produce a mixture of syngas (CO, CO2 & H2), unreacted methane and steam. The gases then pass to the ATR which completes the conversion to syngas and the gas from the ATR is then used to heat the methane and steam coming into the GHR in a continuous process. The GHR process means that more power needs to be imported as the heat of the hydrogen CO2 mix coming out of the reformer cannot be used to generate electricity.
Like a standard ATR, the high-temperature thermal processes mean an ATR+GHR cannot be readily turned on and off but has some ability to ramp up and down, however far slower than an electrolyser. Like ATR, ATR+GHR has a single CO2 stream, from the reformer, therefore high CO2 capture rates of 96% can be achieved.
The technology is most likely to be deployed in the 100s of MW scale. This report covers a 300MW and 1000MW example for 2025 (earliest technical availability in Element Energy dataset) to 2050 online years to show the impact of economies of scale, however other sizes are possible. The technology’s technical life is assumed to be 40 years.
Electrolysis
Electrolysis is the process of using electricity to split water into hydrogen and oxygen. There are plans for electrolysis units or plants to be built in various sizes into the 100s of MW scale, however, these larger projects plans are made up of a series of smaller modules or stacks. Currently, stack sizes are typically up to 5 MW in size.
Alkaline electrolysis is the most mature form of electrolysis with around 90 years of operational experience. In alkaline electrolysis the reaction that separates the water into hydrogen and oxygen occurs between two electrodes in a solution composed of water and liquid electrolyte. Alkaline’s electrical conversion efficiency is assumed to increase from 77% in 2020 to 82% for a plant coming online in 2050. Alkaline’s ability to ramp up and down is quicker than that of gas reforming plants. However, when compared to other electrolyser technologies, such as Proton Exchange Membrane (PEM), it is slower at responding to a fluctuating power supply, so it might be more difficult and costly to pair them with renewable energy sources efficiently.
The levelised costs presented in this report assume different operating modes for electrolysis, including grid-connected, dedicated electricity sources and curtailed electricity. Each configuration assumes different load factors. For our calculations, we have assumed a 30-year lifetime for Alkaline electrolysis, with a plant size of 10MW (made up of smaller stacks) for online years from 2020 to 2050. Proton Exchange Membrane (PEM) electrolysis splits water by using an ionically conductive solid polymer and is assumed to achieve electrical conversion efficiencies of 72% in 2020 up to 82% for a plant coming online in 2050. PEM electrolysis offers rapid dispatchability and turns down to follow energy output, for example from renewables. Therefore, it is ideal for pairing with, for example, dedicated wind farms for low carbon hydrogen production or the provision of rapid response to the grid.
Like for Alkaline, the report shows different operating modes, including grid-connected, dedicated electricity sources, and curtailed electricity. For our calculations we have assumed a 30-year lifetime for PEM electrolysis, with a plant size of 10MW (made up of smaller stacks) for online years from 2020 to 2050.
Whilst Alkaline and PEM electrolysis are both low-temperature electrolysis, Solid Oxide Electrolysis (SOE) uses high-temperature electrolysis (~500 degrees centigrade). Although SOE is not yet widely available commercially, it is included in this report due to the potential of the technology at a large scale, once mature. One advantage of SOE is that the higher temperatures render electrolysis more efficient. The report assumes a conversion efficiency of 74% in 2020 up to 86% for plants coming online in 2050. If, in addition, the temperature can be created through waste heat, electrical efficiencies over 100% can be achieved. Note, that in this report, for simplicity, we assume that waste heat can be accessed at £0/MWh. This is an optimistic scenario and SOE would be more expensive if the price for waste heat was non-zero. One possible future application is to pair SOE with future nuclear power sources, where it can benefit from both high-temperature heat and electricity from the same source. However, due to the high temperatures required, SOE is less suitable for cycling but is expected to be capable of a cycle in less than one day. For our calculations we have assumed a mature SOE technology, with a 30-year lifetime and a size of 10MW (made up of smaller stacks) for online years from 2020 to 2050. Whilst 2020 has been included for comparison purposes, SOE is not yet available at MW scale.
CCUS-enabled biomass gasification
Gasification is a mature technology that heats a solid feedstock, such as coal or biomass, in a reduced concentration atmosphere (to avoid combustion) comprising air, oxygen, or steam to produce a synthetic gas (syngas). Hydrogen is then separated out of the syngas. Whilst various different feedstocks can be gasified, this report focuses on biomass. Biomass feedstocks can include almost any organic material including purpose-grown woody crops (e.g. short rotation beech), purpose-grown herbaceous crops (e.g. grasses), agricultural waste, commercial waste, and dry sewage waste. For simplicity, our analysis assumes that plants consume biomass wood pellets. In general, information on gasification technologies with CCUS is sparse and this report includes one example of biomass gasification with CCUS (also referred to as Bio-Energy with Carbon Capture and Storage (BECCS)). BEIS is currently undertaking further work on gasification technologies to improve the evidence base. This will include getting a better understanding of the costs faced by First-Of-A-Kind (FOAK) projects. The Element Energy data underlying this report does not make a distinction between FOAK and Nth-Of-A-Kind (NOAK) projects. This report therefore only covers 2030 to 2050 online years. The technology is most likely to be deployed in the 100s of MW scale. This report covers a 59MW and a 473MW illustrative example to highlight economies of scale, however other sizes are possible. The technology’s technical life is assumed to be 30 years.
Other technologies
The above does not represent an exhaustive list of hydrogen production technologies. There are other technologies, including newer technologies such as Natural Gas Partial Oxidation (POX) or Compact Hydrogen Generation (CHG) using Sorption Enhanced Reforming (SER). These technologies integrate previously independent process steps and can achieve superior efficiency. These more novel technologies are not covered in this report, but BEIS will be considering these further in the future. Additionally, it is important to note that some of the technologies covered in the report by E4tech and the Ludwig-Bölkow-Systemtechnik (LBST) on behalf of BEIS on a low carbon hydrogen standard6 have not been considered for LCOH at this stage, but future work will explore these further. They include CCUS-enabled gasification that uses residual mixed waste as a feedstock, gas reforming with CCUS that uses biogas formed from food waste, and the Chlor-alkali process, which is an electrolytic process that produces H2 as a by-product from water and salt.
Levelised costs
This section summarises the analysis of the levelised cost of hydrogen for a selection of core technology configurations. For CCUS-enabled methane reformation we are presenting levelised costs for the three core technologies at a 300MW and a 1000MW illustrative scale, respectively. This is to reflect economies of scale. For electrolysis we are presenting levelised costs for the three core technologies using four different types of electricity sources and associated running patterns, respectively. For CCUS-enabled biomass gasification we are presenting levelised costs for one core technology at a 59MW and 473MW scale to reflect economies of scale. All values presented in this section are in 2020 real prices.
CCUS-enabled methane reformation
The chart shows that LCOH for CCUS-enabled reformers are overall fairly constant over time as two effects take place. On the one hand, there is global technological learning, which reduces CAPEX over time, whilst on the other hand fuel and carbon costs increase. The chart also shows the effect of economies of scale, with 1000MW sites being less costly on a £/MWh basis than 300MW sites. As noted in Section 4, on a LCOH basis, when running at baseload operation, the impact of economies of scale is relatively small. The chart uses retail gas and electricity prices; if a “wholesale price plus” (proxied through the LRVC) or wholesale prices are used, LCOH are around £10/MWh lower.
Electrolysis
Chart 6.2 shows levelised costs for online dates from 2020 to 2050 for electrolysis technologies that are grid-connected with baseload operation at a 98% load factor for Alkaline and PEM and a 90% load factor for SOE, connected to dedicated offshore wind at a 51% load factor for 2025 online dates up to a 63% load factor for 2050 online dates, and curtailment-using at a 25% load factor. When grid-connected we are showing the impact of either facing the industrial retail price or a “wholesale price plus” (proxied through LRVC).
There is obviously a much larger range of combinations that could be modelled. For example, different dedicated renewable sources, mixed renewable sources or coupled with electricity storage, overplanting of renewables, or use of other low carbon generation, such as dedicated nuclear. This report does not explore all of these possibilities but notes that they will require further exploration going forward. We have shown a dedicated offshore connection as one potential example
The chart shows that LCOH are highest when using grid electricity, for which our price series include either all (retail) or a reduced amount of policy costs (LRVC). If electrolysis plants do not face these costs, grid-connected operation would become more attractive. Dedicated offshore connection reduces LCOH, but also increases the fixed cost elements of the LCOH due to lower load factors. Using curtailed electricity does not (under our assumptions) incur electricity costs but sees significantly higher fixed costs per MWh due to lower load factors (25%). Technologies that have higher upfront costs (such as SOE) see the largest impacts.
CCUS-enabled biomass gasification
Chart 6.3 shows levelised costs for online dates from 2030 to 2050 for CCUS-enabled biomass gasification. Whilst other gasification routes are possible, such as coal gasification or waste gasification, these have not been considered in this report. To show the economies of scale, this report covers both a 59MW and a 473MW illustrative plant size. Further detail can be found in the annex, published alongside this report.The chart shows that LCOH for biomass gasification with CCUS are fairly constant over time, reducing slightly due to CAPEX learning. There are also significant economies of scale when moving to a larger site. As noted in Section 5, for simplicity, this report values negative carbon emissions at the same carbon price as carbon emissions for the purposes of the analysis. This assumption allows production costs to be partially and eventually more than offset by the value associated with negative emissions. This is demonstrated by the red “net LCOH” dots. As this policy area evolves, we will update our illustrative assessments.
Overarching Conclusions
Chart 6.4 summarises and compares the findings from 6.1, 6.2 and 6.3. To keep the chart manageable, it focuses on one electrolyser technology (PEM), one CCUS-enabled methane reformation technology (ATR+GHR 300MW with CCUS) and a 59MW biomass gasification with CCUS plant, but all data is available in the annex, published alongside this report.
The chart shows that currently CCUS-enabled methane reformation technologies are the lowest cost hydrogen production technology. However, over time and depending on fuel price assumptions, different electrolysis configurations are coming down in costs and in some cases become cost competitive with CCUS-enabled methane reformation technologies. It is not possible to determine precise LCOH ‘switch points’ between different technologies as these vary depending on the different assumptions made.
For example, PEM using only curtailed electricity could become cost competitive from 2025 onwards. Importantly, curtailment by 2030 is likely to be limited, with the actual amount dependent on build out in the power sector, and electrolysers would have to compete with other flexible technologies and solutions (such as Demand Side Response, storage, interconnection), which may mean less electricity available for electrolysers or increasing prices for curtailed electricity (i.e., not £0/MWh, which is assumed in the estimates in this report). The limited amount of curtailment means that, even if electrolysers are cost competitive (through use of curtailed electricity), hydrogen production would be at low volumes, at least in the short term. Biomass gasification with CCUS is relatively high cost, but the value associated with the negative emissions assumed for this analysis results in rapidly declining and even negative costs to 2050. As this policy area evolves, we will update our illustrative assessments.
Sensitivities
Levelised cost estimates are highly sensitive to the underlying data and assumptions used. Within this, different technologies are sensitive to different input assumptions.
Fuel and Electricity price sensitivity
Chart 7.1 shows the impact on central LCOH (shown for reference on the vertical axis) of varying fuel (natural gas for CCUS-enabled methane reformation and biomass for CCUS-enabled gasification) and grid electricity prices, relevant to all technologies except for SMRs and gasification for a 2030 online year. 2030 is chosen as all technologies are assumed to be available by that year. Other years for technologies where these are available are covered in the annex, published alongside this report. Whilst the chart focuses on retail prices (except for biomass, where prices refer simply to the cost of wood pellets), the annex shows additional sensitivity results when a “wholesale price plus” (proxied by LRVC) is assumed.
The chart also only focuses on small plants (300MW for methane reformation with CCUS and 59MW for biomass gasification with CCUS) as the difference to large sites (due to similar efficiencies) was insignificant. The annex captures the larger units as well.
For plants coming online in 2030 the chart shows that by moving from central to low/high fuel and electricity prices, LCOH vary by up to £13/MWh for CCUS-enabled methane reformation technologies. The impact reduces for more efficient technologies, like ATR+GHR with CCUS. The LCOH of electrolysis technologies varies by up to £22/MWh. The impact is larger for electrolysis due to generally lower efficiencies, with the exception of SOE, which has higher efficiencies than other forms of electrolysis and shows similar impacts as for CCUS-enabled technologies. Biomass prices are assumed to have a larger upward than downward risk, therefore LCOH increase by up to £15/MWh if high biomass prices are used but only reduce by £5/MWh if low biomass prices are used.
Technology cost and efficiency sensitivity
The Element Energy evidence base provides us with high and low technology costs (CAPEX and OPEX) and efficiencies for electrolysis technologies, which are significantly more uncertain than CCUS-enabled methane reformation and gasification technologies. Chart 7.2 shows the impact on central LCOH (shown for reference on the vertical axis) of varying technology costs and efficiencies.
It shows that LCOH could vary substantially with higher or lower assumptions. The right-hand side of the chart shows the impact of a high-cost scenario coupled with lower efficiency, whereas the left-hand side shows the impact on LCOH of a low-cost scenario along with higher efficiency. The impact of high/low fixed cost (CAPEX and fixed OPEX) on LCOH increase with lower load factors, as we move from grid-connected electrolysis to only using curtailed electricity. The size of the impact of high/low efficiencies is strongly linked to the electricity price assumed. The highest price (industrial retail price) results in the biggest swing, whilst efficiencies do not have an impact on LCOH if the electricity price is £0/MWh (assumed for curtailed electricity). As Alkaline is the most mature technology, there is less uncertainty and LCOH are impacted the least out of all three technologies. The newest form of electrolysis, SOE, faces the largest uncertainty and as such LCOH are impacted the most, with the exception of efficiencies, which vary less than for other technologies and therefore results in lower LCOH variation.
Load factor sensitivity
Whilst all assumed load factors are uncertain and will depend on the system the technologies will operate in, the shape of demand and the amount of hydrogen storage, we are focusing the sensitivity testing on the assumed 25% load factor for electrolysis using curtailed electricity. If those plants assumed to run at a baseload maximum load factor (equal to availability) had lower load factors LCOH would increase. Chart 7.4 shows the impact of a 10-percentage point higher/lower load factor (i.e. 15% and 35%). As curtailed electricity is assumed to be £0/MWh, the change only affects the fixed CAPEX and OPEX elements of the LCOH. The larger the fixed costs, the more a change in load factors affects the LCOH.
Overarching conclusions
The use of different sensitivities has shown that LCOH estimates are subject to large uncertainties and has highlighted the importance of considering the individual circumstances of technologies and scenarios. Certain technologies may be heavily impacted by a scenario change, whereas other may not.
The above report has been sourced from United Kingdom’s Department for Business, Energy, and Industrial Strategy (BEIS) and can be accessed by clicking here.
