My Energy Storage Report: Hydrogen As An Alternative To Batteries

As mentioned in the last post, my new energy storage report, The Energy Storage Conundrum, mostly deals with issues that have previously been discussed on this blog; but the Report goes into considerable further detail on some of them.

One issue where the Report contains much additional detail is the issue of hydrogen as an alternative to batteries as the medium of energy storage. For examples of previous discussion on this blog of hydrogen as the medium of storage to back up an electrical grid see, for example, “The Idiot’s Answer To Global Warming: Hydrogen” from August 12, 2021, and “Hydrogen Is Unlikely Ever To Be A Viable Solution To The Energy Storage Conundrum” from June 13, 2022.

At first blush, hydrogen may seem to offer the obvious solution to the most difficult issues of energy storage for backing up intermittent renewable generation. In particular, the seasonal patterns of generation from wind and sun require a storage solution that can receive excess power production gradually for months in a row, and then discharge the stored energy over the course of as long as a year. No existing battery technology can do anything like that, largely because most of the stored energy will simply dissipate if it is left in a battery for a year before being called upon. But if you can make hydrogen from some source, you can store it somewhere for a year or even longer without significant loss. Problem solved!

Well, there must be some problem with hydrogen, or otherwise people would already be using it extensively. And indeed, the problems with hydrogen, while different from those of battery storage, are nevertheless equivalently huge. Mostly, to produce large amounts of hydrogen without generating the very greenhouse gas emissions you are seeking to avoid, turns out to be enormously costly. And then, once you have the hydrogen, distributing it and handling it are very challenging.

Unlike, say, oxygen or nitrogen, which are ubiquitous as free gases in the atmosphere, there is almost no free hydrogen available for the taking. It is all bound up either in hydrocarbons (aka fossil fuels — coal, oil and natural gas), carbohydrates (aka plants and animals), or water. To obtain free hydrogen, it must be separated from one or another of these substances by the input of energy. The easiest and cheapest way to get free hydrogen is to separate it from the carbon in natural gas. This is commonly done by a process called “steam reformation,” which leads to the carbon from the natural gas getting emitted into the atmosphere in the form of CO2. In other words, obtaining hydrogen from natural gas by the inexpensive process of steam reformation offers no benefits in terms of carbon emissions over just burning the natural gas. So, if you insist on getting free hydrogen without carbon emissions, you are going to have to get it from water by a process of electrolysis. Hydrogen obtained from water by electrolysis is known by environmental cognoscenti as “green hydrogen,” because of the avoidance of carbon emissions. Unfortunately, the electrolysis process requires a very large input of energy.

How much is it going to cost to produce green hydrogen as the storage medium for a mainly wind/solar grid? My Report first notes that as of today there is almost no production of this green hydrogen thing:

To date, there has been almost no commercial production of green hydrogen, because electrolysis is much more expensive than steam reformation of natural gas, and is therefore uneconomic without government subsidy. The JP Morgan Asset Management 2022 Annual Energy Paper states that ‘Current green hydrogen production is negligible...’

So we don’t have any large functioning projects from which we can get figures for how expensive green hydrogen is going to be. In the absence of that, I thought to undertake an exercise to calculate how much capacity of solar panels it would take to produce 288 MW of firm power for some jurisdiction, where the panels could either provide electricity directly to the consumers or alternatively produce hydrogen by electrolysis that could be stored and then burned in a power plant to produce electricity. (The 288 MW figure was selected because GE produces a turbine for natural gas power plants with this capacity, and says that it can convert the turbine for use of hydrogen as the fuel.). Here is that exercise as written up in my Report:

Consider a jurisdiction with steady electricity demand of 288 MW. . . . The electricity needs of our jurisdiction can be fully supplied by burning natural gas in the plant. But now suppose we want to use solar panels to provide the electricity and/or hydrogen for the plant sufficient to supply the 288 MW firm throughout the year. What capacity of solar panels must we build? Here is a calculation:

• Over the course of the year, the jurisdiction will use 288 MW × 8760 hours = 2,522,880 MWh of electricity.

• We start by building 288 MW of solar panels. We will assume that the solar panels produce at a 20% capacity factor over the course of a year. (Very sunny places such as the California desert may approach a 25% capacity factor from solar panels, but cloudy places such as the Eastern US and all of Europe get far less than 20% of capacity; in the UK, typical annualised solar capacity factors are under 15%). That means that the 288 MW of solar panels will only produce 288 × 8760 × 0.2 = 504,576 MWh in a year.

• Therefore, in addition to the 288MW of solar panels directly producing electricity, we need additional solar panels to produce hydrogen to burn in the power plant sufficient to generate the remaining 2,018,304 MWh.

• At 80% efficiency in the electrolysis process, it takes 49.3 kWh of electricity to produce 1 kilogram of hydrogen. GE says that its 288 MW plant will burn 22,400 kilograms of hydrogen per hour to produce the full capacity. Therefore, it takes 49.3 × 22,400 = 1,104,320 kWh, or approximately 1,104 MWh of electricity to obtain the hydrogen to run the plant for one hour. For the 1,104 MWh of electricity input, we get back 288 MWh of electricity output from the GE plant.

• Due to the 20% capacity factor of the solar panels, we will need to run the plant for 8760 × 0.8 = 7008 hours during the year. That means that we need solar panels sufficient to produce 7008 × 1104 = 7,736,832 MWh of electricity.

• Again because of the 20% capacity factor, to generate the 7,736,832MWh of electricity using solar panels, we will need panels with capacity to produce five times that much, or 38,684,160 MWh. Dividing by 8760 hours in a year, we will need solar panels with capacity of 4,416 MW to generate the hydrogen that we need for backup.

• Plus the 288MW of solar panels that we began with. So the total capacity of solar panels we will need to provide the 288MW firm power using green hydrogen as backup is 4,704 MW.

Or in other words, to use natural gas, you just need the 288 MW plant to provide 288 MW of firm power throughout the year. But to use solar panels plus green hydrogen backup, you need the same 288MW plant to burn the hydrogen, plus more than 16 times that much, or 4,704 MW of capacity of solar panels, to provide electricity directly and to generate sufficient hydrogen for the backup.

That calculation assumed a 20% capacity factor of production from the solar panels over the course of a year. It turns out that actual solar capacity factors are more like 10-13% for Germany, 10-11% for the UK, and about 12.6% in New York. (California, with few clouds, gets capacity factors somewhat in excess of 25%.). Doing the same series of calculations using a 10% capacity factor for the solar panels, you will need something like 9,936 MW of solar panels to provide your 288 MW of firm power for the year, with the green hydrogen as your storage medium.

In other words, you will need about 35 times the capacity of solar panels as the amount of firm power that you are committed to provide. The reasons for the vast differential include: the sun doesn’t shine fully half the time; most of the time when the sun does shine it is low in the sky; places like the UK, Germany and New York are cloudy more often than not; and there are significant losses of energy both in electrolyzing the water and then again in burning the hydrogen.

Anyone and everyone should feel free to check my arithmetic here. I’m fully capable of making mistakes. However, several people have already checked this.

My Report then takes a stab at translating the enormous incremental capital cost of all these solar panels into a very rough cost comparison of trying to generate the 288 MW of firm power from solar panels and green hydrogen versus simply burning natural gas in the plant. I got cost figures for the turbine plant and the solar panels from a March 2022 report of the U.S. Energy Information Agency. Using that data:

[T]he cost of the 288MW General Electric turbine power plant [would be] around $305 million, and the cost of the 4,704 MW of solar panels [would be] around $6.25 billion.

If you needed the 9,936 MW of solar panels because you live in a cloudy area, the $6.25 billion would become about $13 billion.

My very rough calculation in the Report, with the 20% solar capacity factor assumption, is that electricity from solar panels plus green hydrogen storage would start at somewhere in the range of 5 to 10 times more expensive than electricity from just burning the natural gas. At the 10% solar capacity factor assumption, make that 10 to 20 times more expensive.

And after all of this we still haven’t gotten to the very substantial additional engineering challenges of working with the very light, explosive hydrogen gas. A few examples from the Report:

  • Making enough green hydrogen to power the world means electrolysing the ocean. Fresh water is of limited supply, and is particularly scarce in the best places for solar power, namely deserts. When you electrolyse the ocean, you electrolyse not only the water, but also the salt, which then creates large amounts of highly toxic chlorine, which must be neutralised and disposed of. Alternatively, you can desalinise the seawater prior to electrolysis, which would require yet additional input of energy. There are people working on solving these problems, but solutions are far off and could be very costly.

  • Hydrogen is only about 30% as energy dense by volume as natural gas. This means that it takes about three times the pipeline capacity to transport the same energy content of hydrogen as of natural gas. Alternatively, you can compress the hydrogen, but that would also be an additional and potentially large cost.

  • Hydrogen is much more difficult to transport and handle than natural gas. Use of the existing natural gas pipeline infrastructure for hydrogen is very problematic, because many existing gas pipelines are made of steel, and hydrogen causes steel to crack. The subsequent leaks can lead to explosions.

It’s no wonder that green hydrogen is all talk. Nobody is willing to actually try to build out a demonstration project. The so-called “hydrogen economy” is highly unlikely ever to happen.