A variation on the above was contained in a recent report produced for the European Commission: “in the 2050 scenarios, the large deployment of electrolyser leads to an important drop in their investment costs.”
Restating the same argument but in a different context: large-scale manufacturing of photovoltaic production equipment is important to reduce the costs……of PV panels. The reality is that the final cost of a finished PV panel has only a very small capital element due to the production equipment. That is because PV panel production is a “volume business”. The main cost is a material cost with yield a defining factor in factory profitability.
By extension, the production of hydrogen using electrolysers is a “volume business”. This statement can be proved mathematically i.e. it is objectively true, through the construction of a discounted cash flow using different capacity factors and different costs of electricity. Both Agora Energiewende and the IEA have constructed cost curves – both arriving at the same conclusion: CAPEX is a minor cost factor (circa 10% – 20%) for hydrogen from electrolysers once capacity factors of 35 – 40% are passed.
Given the above, the CAPEX of electrolysers is of passing interest compared to their efficiency and the cost of electricity.
Users of this phrase rarely include an object to complete the sentence – inefficient relative to what? There are two electrolyser technologies that sit at TRL 9 (technical readiness level 9): polymer exchange membrane (PEM) and alkali systems. Broadly speaking, each uses around 55kWh of electricity to produce 1kg of hydrogen that has an upper heating value of 40kWh. This gives an overall process efficiency of around 73%. It is forecast that technical improvements could drive system efficiencies to 80% in the next 5 years.
Electrolysers are often compared to batteries. The round-trip efficiency of a li-ion battery is about 86% – this figure is rarely criticised. The round-trip efficiency of hydrogen will depend on the end point of the trip.
Users of this phrase rarely include an object to complete the sentence – too costly relative to what? Green hydrogen is often compared in cost to hydrogen produced from fossil sources, usually using a steam methane reforming (SMR) process. This does not compare like with like since green hydrogen using, for example, electricity from wind has a carbon footprint 0.66kg of CO2 per kg of hydrogen produced. Hydrogen produced from an SMR has a footprint of 10kg CO2/kg hydrogen, i.e. a 15x larger footprint. Also overlooked is the fact that in the EU, through a political quirk, SMR systems receive state aid through free ETS allowances.
Once it is accepted that hydrogen production MUST be decarbonised (the EU uses 4 million tonnes of hydrogen for fertiliser production, this causes 40 million tonnes of CO2 emissions due to hydrogen production) then the comparison is between green hydrogen and hydrogen produced by SMRs coupled to a carbon capture and storage process. When this is done, and all externalities are included, prices are very similar (around 6eurocents/kWh). The main difference being that nobody has built an SMR+CCS process that can sequester at least 90% of CO2. By contrast, electrolyser systems have been in use at scale since the late 19th century.
Gas networks split into three types: high pressure (HP), intermediate pressure (IP) and low pressure (LP), with the final one delivering a gas supply to homes and offices and commercial premises. Much of the EU’s HP and LP pipelines are made of a steel quality that is impervious to hydrogen embrittlement. This means that most HP and IP networks can be converted to carry hydrogen through the simple expedient of changing compressor and decompressor stations. In the case of the LP network, if it is plastic (as many are, France has 150,000km of LP plastic network) it is hydrogen-ready. It is worth keeping in mind that LP networks operate at 0.2 bar or 3 psi. If your bicycle tyre is at this pressure it is, functionally, flat.
From the mid-19th through to the middle of the 20th century, large-scale gas networks in Europe carried “town gas”, produced from coal and which was 50%, by volume, hydrogen. Who can forget the explosions caused by the London gas network in 1880 that destroyed 90% of the city. The event in the previous sentence never occurred in London or indeed anywhere else. This suggests that if Europe could have a primitive gas network for 150 years carrying 50% hydrogen for the most part safely, it is likely that a modern networks will be able to carry 100% hydrogen in a similarly safe fashion.
Many reports on hydrogen deployment comment along the lines of “hydrogen, once costs are reduced, will grow in the period after 2030”. Such statements overlook the reality that some sectors such as primary steel, fertiliser and cement need green hydrogen to decarbonise – there is no alternative. Thus what is being said could be re-expressed as: we don’t see these sectors being decarbonised before 2030. This is a rather surprising view given the urgent need to decarbonise. The claim is often supported by the comment “there is not enough renewable electricity”.
Taking the final point about “not enough electricity”, and considering German electricity production in the period 1st January 2020 to 8th April 2020, there was 12TWh of “surplus” electricity due to renewables. If used to power electrolysers, it could have produced around 220,000 tonnes of hydrogen. Or, more than 5% of the EU’s hydrogen needs for ammonia production (in a year this would be around 25%).
Space heating is the largest user of natural gas in the EU with demand for residential space heat accounting for around 950TWh of gas use per year. This is a sector in urgent need of decarbonisation. (For a more detailed technical discussion, readers are directed to the blog section of this website).
There is a general consensus that the residential sector, before any move to new heating systems, needs to first implement significant energy efficiency actions on the fabric of residential buildings which will reduce the overall demand for space heat. Once this goal has been achieved, then it is possible to de-carbonise what is left of heat demand.
Conventional wisdom is that decarbonisation of the now energy-efficient building stock is to be accomplished through the use of heat pumps powered by electricity from renewable resources.
There is an automatic assumption that existing electricity distribution networks can carry this new load. This is a false assumption, for the following reason.
When electricity networks are designed, a set of rules are used to define expected loads on the network. For much of western Europe, the major expansion of electrical networks occurred in the period 1950s through to 1980. Heating systems were for the most part gas or oil-based (France is a partial exception). For the residential sector, design metrics for maximum expected demand were in the range 1 to 1.5kW per dwelling. Networks were built based on this metric which, for the most part, is semi-independent of dwelling size. Electrical distribution networks in areas where the residential sector is fossil heated see loads imposed on distribution transformers/LV networks of between 40 and 70%. Thus, electrical distribution networks where fossil heating dominates currently mostly operate well within their technical capacities. They have head-room for new loads.
The electrical demand from any heat pump is a function of: size of dwelling, thermal performance and outside temperature. It is possible to model this (undertaken by Challoch-Energy & PWR). Results suggest that heat pumps will have a continuous load when temperatures are at around zero degrees centigrade in the range 0.5 to 1kW depending on the size of the dwelling. The renovation level is assumed to be: fully insulated loft, cavity wall insulation, modern double glazing. For 250 dwellings, with a heat pump penetration of 35%, the load imposed by heat pumps will place the local distribution transformer at close to overload and, more likely, drive voltages on the network outside of their statutory limits.
Heat pumps are not the only electricity-using technology being deployed in the residential sector. Electric vehicles are starting to grow and will also impose new loads on the distribution network. These loads can be modelled. A 30% penetration of heat pumps and a 10% penetration of electric vehicles will put distribution networks very close to their design limits.
Key Point: existing electrical distribution networks were not designed to support new loads such as heat pumps and electric vehicles. This leaves two choices: substantively reinforce the European distribution network with new underground cables and substations, or find ways to embed new generation into the network.
Fuel cells that “burn” hydrogen” produce both heat and electricity. They are currently available for the residential sector from a range of European manufacturers (often the same ones that produce heat pumps). Taking one example of modelling, a 20% penetration of fuel cells, with a 30% penetration of heat pumps and 10% penetration of electric vehicles enables a given electrical distribution network to operate within its design limits.
In summary: given that Europe has to decarbonise space heat and passenger transport, it has two choices: re-build its distribution network or find a way to embed generation in the distribution network. Fuel-cells powered by hydrogen supplied by a re-purposed-for-hydrogen gas network can do that. The choice is not heat pumps or fuel-cells, it is both. Furthermore, fuel cells/electrolysers solve a related and growing problem in distribution networks: where to put electricity from PV systems on bright sunny days in the middle of the summer. Electrolysers can turn that electricity into hydrogen and push it back into the hydrogen (gas) network which has the capacity to store very large amounts of hydrogen, and thus electricity.
The EU in 2017 consumed 5,868TWh of natural gas and 3,100TWh of electricity (approx 9000TWh). The JRC in its 2018 report “Wind Potential in the EU” noted that the EU’s potential resource was: on-shore wind 11,700TWh/year, off-shore 13,800TWh/year. In total 25,500TWh/year.
Either wind resource on its own, and if fully developed, would easily meet all of Europe’s gas and electricity needs. If renewable energy from other sources such as solar, biomass or, in the future, ocean energy is included then there is more renewable resource available within the EU than could ever be consumed in any future energy scenario.
The round trip for a battery is 85%. Network losses of 10% reduce this to 75%. In the case of heat pumps, the claim is that they have a coefficient of performance (COP) of 3 such that 1 unit of electricity gives 3 units of heat. COPs for air source heat pumps (the most common and easiest to retro-fit) are related to outside temperature. As temperature declines, so does the COP. For example, at an outside temperature of -6°C the COP is less than 2.5. If boost elements are included (i.e. direct heating using electricity) then the COP falls further.
Turning water into hydrogen using renewable electricity has an efficiency (now) of 73%. Overall efficiency of a fuel cell (heat and electricity) is 98% (standard seasonal efficiency to DIN – Vitavalor PT2). This gives a round trip of 71%.
The “my system is more efficient than your system” is a sterile and academic argument. As the section on Hydrogen Application Myths shows, significant penetration of heat pumps (and electric vehicles) overloads electrical distribution networks requiring either costly and time-consuming network rebuild or the use of fuel cells.
Once the benefit of a fuel-cell/electrolyser combination is considered (for storing excess electricity from PV in the summer), it becomes clear that fuel cells have a very useful role to play in hybrid power systems of the future (hybrid = hydrogen and electricity networks). Considering one aspect of a system, “my system is more efficient than your system” misses the bigger picture.