Energy Destinies – Part 3: Energy Storage – Inconvenient Complications

Yves here. Satyajit Das continues his in-depth sanity-checking of green energy plans versus their ability to satisfy current, let alone expected, energy needs. Here he focuses on batteries and other energy storage mechanisms.

By Satyajit Das, a former banker and author of numerous works on derivatives and several general titles: Traders, Guns & Money: Knowns and Unknowns in the Dazzling World of Derivatives  (2006 and 2010), Extreme Money: The Masters of the Universe and the Cult of Risk (2011), A Banquet of Consequences RELOADED (2021) and Fortune’s Fool: Australia’s Choices (2022)

Abundant and cheap power is one of the foundations of modern civilisation and economies. Current changes in energy markets are perhaps the most significant for a long time. It has implications for society in the broadest sense. Energy Destinies is a multi-part series examining the role of energy, demand and supply dynamics, the shift to renewables, the transition, its relationship to emissions and possible pathways. The first and second part looked at patterns of demand and supply over time and renewable energy sources. This part looks at the need for energy storage.

Given the problems of intermittency, renewable energy sources require infrastructure for storage. For electricity, where a significant proportion of the total grid demand is supplied by renewables, external storage becomes important with more need as more such sources must be integrated.

Energy storage refers to the capture of energy produced at one time for deferred use. It involves converting energy typically from non-storable instantaneous states to a storable forms for future access. Stored energy allows supply to match demand as needed

Storage requirements can be short (covering a few hours or overnight) and long-duration (covering a period of day to months). Technologies differ in capacity and length of time of energy available. Energy storage is also differentiated by whether it is generic or specific. Batteries are useful for storing electricity and devices geared to using certain types of power. Ice tanks, used to store ice using cheaper electricity at night, can only meet peak daytime demand for cooling.

As is often not appreciated, fossil fuels, such as coal and hydro-carbons, are actually natural stores of energy from the sunlight. There are a number of potential alternative technologies but, in practice, the major forms are batteries, pumped hydro and hydrogen. Other potential alternative storage technologies, at various stage of development, include electrical or electromagnetic (capacitors and superconducting magnetic storage), mechanical (compressed air energy storage or flywheel), biological (glycogen or starch), thermal (cryogenic energy storage, liquid air energy storage or molten salt storage) or phase-change material (heat sinks using a substance which absorbs and releases sufficient energy at phase transition to provide useful heat or cooling).


Batteries, generally rechargeable, store electricity using electro-chemical reactions based on different chemistries including lead–acid,  nickel–cadmium and lithium-ion.

The key issues include:

  • Efficiency – this measures the energy retrieved relative to the amount of energy stored. The best lithium-ion batteries have efficiency approaching 90 percent in optimal conditions. Performance degrades over time. For example, if the battery is fully charged for some (short) time at an ambient temperature of 40C, its capacity (the ability to store energy) will decrease by as much as a third in a year.
  • Size and weight – batteries needed for significant energy storage are large. EVs are far heavier than traditional cars due to their large, heavy battery packs – a battery-electric Ford F-150 Lightning is 900-1,350 kilograms (2,000-3,000 pounds) heavier than an equivalent petrol- or diesel-powered model.
  • Duration – battery life duration is an issue. Typically for grid level storage, they are designed to provide a few hours of power. After a total system outage in 2018, the Australian state of South Australia installed the world’s first ‘big battery’ (Hornsdale Power Reserve), rated at over 150 Megawatts. It  can power around 50,000 homes for 3-4 hours. In fairness, the Power Reserve provides additional grid stability and system security. To keep South Australia (population 2.5 million) supplied for one half-day would require around a hundred such “world’s biggest” Tesla battery farms. Performance is also not guaranteed with the owner fined A$900,000 in 2022  after being sued by the Australian Energy Regulator failing to deliver promised capacity.
  • Battery life – Typical lithium-ion battery lives are up to 10-15 years while some other battery technologies have longer lives. On average, after 8-10 years in industrial settings, the battery capacity decreases to ‘economically disadvantageous’ levels. Degradation creates problems of disposal of lithium-ion batteries.

Pumped Hydro

The concept of pumped hydro is that excess energy (usually electrical power within a grid during times of low demand) is used to pump water from a lower to a higher reservoir. Water can be released back to a lower reservoir, body of water or waterway through a turbine, generating electricity. The technique uses the height difference between two water bodies and gravitational force. Typically, reversible turbine-generator assemblies are used as both a pump and turbine.

There are two types of pumped hydro storage:

  • Pure pumped-storage plants create two customised reservoirs dedicated to storage and generation
  • Pump-back uses existing hydroelectric plants and their reservoirs, combination pumped storage and conventional generation using natural stream-flow.

Worldwide, pumped-storage hydroelectricity is the largest-capacity form of active grid energy storage used globally. Availability is limited by terrain which requires elevation differences and ideally natural reservoirs which can be enhanced. It has low surface power density requiring large amounts of land.

There are subtler issues. Unless pure with two custom made separate reservoirs at different elevations used exclusively for power storage, these schemes are typically multi-purpose dams generating electricity and supplying water to households, agriculture and industry. If large releases are required to cover grid shortages, then any water not stored for return to the upper storage reservoir such releases into waterways may be unavailable to meet these other needs. Also once stored water is expended, no further electricity can be generated until surplus power becomes available to refill the relevant reservoir.


Excess energy, especially electricity, can be converted to a gaseous fuel such as hydrogen or, less commonly, methane. As it does not occur naturally in sufficient quantities, electricity is used to generate hydrogen through chemical processes such as electrolysis of water.

There are several types of hydrogen fuel:

  • Brown hydrogen – uses thermal coal and is cheap but highly polluting.
  • Grey hydrogen – uses natural gas via steam methane reformation without emissions capture and is the most common current form of production.
  • Blue hydrogen – similar to grey but carbon emissions are captured and stored or reused. The lack of capture availability means that it is not current extensively used.
  • Green hydrogen – uses renewable energy to electrolyse water separating the hydrogen atom from the oxygen which is currently expensive.

Unproven at scale, turquoise hydrogen uses a process called methane pyrolysis to produce hydrogen and solid carbon.

Efficiency is dependent on the energy losses involved in the hydrogen storage cycle from the electrolysis of water, liquification or compression of the hydrogen and conversion to electricity.

The interest in hydrogen derives from the potential to convert renewable energy into a zero-carbon fuel, that is, green hydrogen.

Hydrogen fuel can theoretically be used to power generation plants or heating.  It can be used in fuel cells or internal combustion engines.  Hydrogen can be used in fuel cells which are efficient, have low noise, and low maintenance requirements because of fewer moving parts. There is also potential to convert combustion engines in commercial vehicles to run on a hydrogen–diesel mix. Combustion engines using hydrogen would entail less radical change for the automotive industry, and potentially lower up-front vehicle cost compared to fully electric or fuel cell alternatives.

Hydrogen’s use as a transportation fuel is of particular interest where electric power may not be optimal, such as heavy transport, aviation and heavy industries where there is the need for greater power, longer range and quicker refuelling time.  Clean hydrogen’s is frequently presented as the ‘magic bullet’ in decarbonizing the aviation, fertilizer, long-haul trucking, maritime shipping, refining, and steel industry.

Hydrogen production currently uses fossil fuels. Scaling up green hydrogen production will require large investments to reduce production costs to make it competitive with other fuels and build infrastructure for transportation, storage and distribution. Even if sufficient green hydrogen were available at a competitive costs, there are several issues that would need to be overcome:

  • Hydrogen has a high energy content per unit mass. But at room temperature and atmospheric pressure, it has a very low energy content per unit volume compared to liquid fuels or natural gas. It must usually be compressed or liquefied by lowering its temperature to under 33 Kelvin (minus 240 Celsius). This requires high-pressure or cryogenic tanks that weigh much more than the hydrogen they can hold complicating its use in cars, trucks and airplanes.
  • Hydrogen fuel has low ignition energy, high combustion energy, and leaks easily from tanks making it hazardous. This would require careful control of the supply chain and storage.

Significant improvements in technology are needed before hydrogen fuel is a safe, viable and cost effective storage medium. Green hydrogen remains in short supply. Transport options such as pipelines are limited. Even electrolyser supply is constrained with mass-production only beginning to be ramped up. The much promoted hydrogen economy is not yet with us. 

Energy Storage Economics

The economics of energy storage is difficult to quantify as it depends on context and the type required. Different methods are not technically suited to all needs. The economics are market and location sensitive. The standalone cost is less relevant than the overall cost in the context of an energy system.

Energy storage is difficult to evaluate using traditional valuation metrics such as discounted cash flow. Some have suggested using real option analysis, which can incorporate various uncertainties and externalities (meeting intermittency, avoidance of curtailment, grid congestion avoidance, price arbitrage and carbon-free energy delivery). However, such models are highly subjective and sensitive to small changes in parameters.

Irrespective of economics, it is unlikely that currently available energy storage options are likely to allow the shift to renewables on the scale proposed. Batteries are flexible, able to respond rapidly to changes in energy demand making them suitable for fine-tuning supplies. If they have to provide energy storage for more than several hours, then their capital cost is very high. Although growth in battery demand for EVs has significantly reduced the cost, they remain expensive especially when limited life, capacity and duration are considered. Currently, batteries remain a questionable source of dispatchable power being unable to cover for variable renewable power gaps lasting for longer than a few hours. The only viable option is pumped hydro which can store energy for several hours to months, depending on storage capacity and structure.

In models with high levels of renewable power, the cost of storage can dominate the costs of the whole grid. In California, 80 percent renewable share would require 9.6 terawatts of storage but 100 percent would require 36.3 terawatts. As of 2023, the state had 5,000 megawatts of storage. While this is up 20 times since 2019 and projected to increase another 10 times to 52,000 megawatts, it is below requirements keeping in mind that one terawatt-hour is also equal to 1,000,000 megawatt-hours.  Supplying 80 percent of US demand from renewables may require a smart grid covering the whole country or battery storage capable to supply the whole system for 12 hours at cost estimated at $2.5 trillion. Others estimate the costs at much higher levels.

Building out the required battery energy storage would adversely affect the cost of power. Assuming lithium battery costs fall by two-thirds, building the level of renewable generation and storage necessary to reach California’s objective of deriving most of its power from renewables would drive up costs, based on one estimate, from $49 per megawatt-hour to as much as $1,612 at 100 percent renewables.

Relying only on renewables and energy storage may cost at least about 30-50 percent more than a comparable system that combines renewables with nuclear plants or fossil fuel plants with carbon capture and storage.

The efficiency of energy storage is bot currently optimal. Similar to Energy Return on Energy Invested (EROEI), energy stored on energy invested (ESOEI) measures the amount of energy that can be stored by a technology, divided by the amount of energy required to build that technology. The higher the ESOEI, the more efficient the storage technology.

The following table summarises the ESOEI of some common energy storage mechanisms:

Batteries have much lower ESOEI than pumped hydroelectric storage. While scientific opinion varies, without extensive pumped storage, the combination of renewables paired with existing battery technology may not be workable.

The characteristics of various energy storage system is summarised below:

Energy storage needs lowers the EROEI of renewables perhaps below the economically viable threshold.

Theory and Practice

The need for large-scale energy storage greatly complicates a renewables based energy system. It requires massive investment but also must overcome inherent inefficiencies. For battery technology, baring scientific breakthroughs that usher in revolutionary changes in its physics and chemistry, it is difficult to see the needed cost and storage efficiency improvements at least soon. Pumped hydro storage while simple is subject to other constraints.

Alongside the need for build-out of the grid and transmission capabilities, storage constraints place limits on the capacity of renewables to replace traditional fuels in modern energy systems.

In a celebrated exchange between technologists, Trygve Reenskaug states: “In theory, practice is simple.” Alexandre Boily’s response is telling: “But, is it simple to practice theory?“‘ That difference remains to be overcome in moving to a predominantly renewable driven energy system.

© 2023 Satyajit Das All Rights Reserved

A version of this piece was published in the New Indian Express.

Print Friendly, PDF & Email


  1. Grumpy Engineer

    The following paragraph had some unit-of-measure problems. I’ve added some “-hours” (highlighted in bold) in the places below where I believe the author meant to use units of energy rather than units of power.

    In models with high levels of renewable power, the cost of storage can dominate the costs of the whole grid. In California, 80 percent renewable share would require 9.6 terawatt-hours of storage but 100 percent would require 36.3 terawatt-hours. As of 2023, the state had 5,000 megawatt-hours of storage. While this is up 20 times since 2019 and projected to increase another 10 times to 52,000 megawatt-hours, it is below requirements keeping in mind that one terawatt-hour is also equal to 1,000,000 megawatt-hours.

    Why do I believe these corrections are correct? Because peak electrical for the entire continental US is about one terawatt, and California alone would certainly never require 9.6 or 36.3 terawatts of power to keep its grid up. But to get through an extended period of unfavorable weather, I can easily see California needing multiple terawatt-hours of stored energy to get through it. And given that they’re deploying less than a one gigawatt-hour of storage per year, I can see why the phrase “inconvenient complication” comes to mind.

    1. SteveD

      Agree with your correction. If we assume the US has little if any ‘spare’ industrial capacity (a good debate by itself), the interesting policy debate is “Is the shift to renewable energy important enough to *remove* capacity from industry X, Y, and Z, in order to shift production resources to producing and deploying energy storage?” I have my own short list of things we could stop doing in order to free up resources ;)

    2. Peter Nightingale

      The whole post uses power rather than energy. That makes sense, because at the same time it mentions for how long that power could be supplied. That is how much energy is stored. Both power and energy matter. A trickle from an infinite backup source will still produce a blackout.

      However that may be, I too have trouble with these numbers. The average power consumed by the U.S. is about 100 Quads per year — If you convert this idiosyncratic, U.S. unit to Standard International units you get about 3.5 terawatts. Let’s say that California represents one fifth of the U.S. economy, probably an overestimate. That’s less than one terawatt. Can somebody explain how that gets me to understand the following statement in this post” In models with high levels of renewable power, the cost of storage can dominate the costs of the whole grid. In California, 80 percent renewable share would require 9.6 terawatts of storage but 100 percent would require 36.3 terawatts.” Are the expected fluctuations really 40 times the average?

  2. BillS

    Thanks for another thought provoking article. Storage is indeed crucial if intermittent renewable sources are to be a reliable provider of electric power now and into the future. A couple of comments/questions:
    1) I would be interested in understanding how the efficiency numbers are arrived at. For example, battery charging efficiencies are stated as 80-95%. This seems reasonable for battery energy in->battery energy out, but I seriously doubt that these numbers include conversion efficiencies, i.e. the losses incurred by converting the electric energy from AC 230V (or unregulated PV output) to battery-friendly DC and back again. My experience is that this total efficiency is around 60-70% for Li batteries once all system losses are included.
    2) Is the ESOEI really so bad for Pb-acid batteries? These batteries are generally thoroughly recycled? I realize that Pb batteries are less durable, but may of the raw materials are already available from efficient recycling processes..
    3) At some point Li battery recycling needs to become much more efficient. What effect will this have on their ESOEI?
    4) Green, blue,etc. hydrogen, at this point, seems a grand scam. Electrolyzers are notoriously inefficient (unless rare metal catalysts such as platinum or iridium are employed). 75-80% is a probable optimistic level for membrane electrolyzers. Likewise for any fuel cells used for the conversion back to electricity. The electronic systems that convert the electricity to and from useful forms might be expected to have around 90% efficiency, hence total round-trip energy efficiency ends up being around 0.9*0.8*0.8*0.9=0.52 or 52%. If the H2 is compressed/cooled and/or transported, the useful energy return is even worse – perhaps 30-40%. Transporting electricity directly and using batteries seems more efficient.
    5) How might energy use timing (to correspond to when abundant renewable energy is available) and conservation play a role? Many of us who have grid-tied PV installations already do this.

    1. dave -- just dave

      A remark I’ve seen attributed to baseball great Lawrence “Yogi” Berra is:

      In theory, theory and practice are the same. In practice, they’re different.

    2. New Okie

      Thanks Bill

      I do have one question though: What exactly is the problem with hydrogen? Is it just crazy-expensive, even at scale?

      PS I’ve read recently about burning iron for energy. I gather it requires hydrogen to “recharge” iron powder so it can be burned again, but the selling point seems to be that it gets around the storage and safety issues inherent in hydrogen. Perhaps this is the future of energy storage?

      1. Jorge

        Hydrogen is a PITA to manage, v.s. other larger hydrocarbon molecules. H2 molecules are tiny, and seep out through metals, seals, plastics, etc.

        Storing energy in a gas is generally a pain; there are attempts to improve the process of turning hydrocarbon gases into ammonia, which is liquid at room temperature. Ammonia can be piped, burned, and turned into fertilizer.

  3. NoFreeWill

    “Abundant and cheap power is one of the foundations of modern civilisation and economies.”
    “Given the problems of intermittency, renewable energy sources require infrastructure for storage. ”

    Assuming current demand should continue (which assumes growth in energy demand continuing forever) and not adapt itself to reality is the problem here. There is no green transition without degrowth. In fact, many of the so-called problems of green energy are precisely their advantages, less growth means less carbon emissions which is the goal of the whole exercise in the first place. We should be adapting our energy demands to match the realities of intermittency rather than using it as a reason to dismiss the possibilities of transition. And the necessary investment in storage provides a massive economic/jobs stimulus.

    Most green growthers are indeed delusional, but the way this series has framed things so far makes it logical to conclude “well, the transition won’t work I guess we just have to keep burning carbon in insane quantities” for many readers. That’s only a bit better than the bait and switch the oil industry wants… keep underfunding green tech R&D & infrastructure and yet peddle the idea that renewables alone will fix things, then when it fails whoops I guess we have to keep building natural gas plants. While this article might poke a hole in the idea of the techno-optimists, it doesn’t provide the necessary background of alternatives to the insane logic of infinite energy growth.

    The real problem here is that degrowth essentially requires socialism, but so do any realistic hopes of transition to a sustainable economy that works for all people. And with peak/very low EROEI oil looming, reducing energy demand will happen anyways in a chaotic fashion, it’s just another reason for orderly transition to happen as fast as possible.

    An EROEI graph without a carbon intensivity and thus how fucked is our future if we keep using this power source graph is not a complete picture. Lower than 7:1 is indeed pretty bad, but again that assumes demand not adapting to supply.

    “Energy storage needs lowers the EROEI of renewables perhaps below the economically viable threshold.” there are 2 links in this sentence and both are quoted in misleading ways
    The first link says something very different than what Das implies, that EROEI for the US is 7:1, so to maintain middle class only 1st world, or rich-ish 3rd world standards of living this could be much lower, and given how bad some parts of the US already are, the disproportionate energy use of the rich, and the vast sectors of the economy which we could eliminate (luxury goods, etc.) entirely without their existence, I don’t buy this argument.
    The 2nd article he links literally contradicts his entire argument, not sure why he’s linking it in support of this article when it says right in the abstract:
    “We conclude that electrical storage is unlikely to limit the transition to renewable energy.”

  4. Synoia

    All these system need sites. Typically large ones.

    One possible large Pumped Storage (for electricity) site is the Salton Sea in the California desert, and not fit for anything. The water there is very saline, while the sea is well below sea level. The sea, in Baha California is close, but an important habitat for many endangered species.

    One can postulate, but the Environmental impact of using this for pumped storage might be technically possible, but the environmental risks so great that doing nothing appears to be the best policy, even if it results in the Salton Sea remaining a cesspit.

    I looked at this in the early years of this century, and had a good rapport on both sides of the border, jarrez and Riverside County, and and even met with Carlos Slim in Mexico City to discuss funding.

    The ecological issues were insurmountable, as I had suspected all along. No one wants what is in the Salton sea.

  5. Cine Tee

    A global electrical grid could do a lot to alleviate the need for large scale storage. One side of the Earth has sunshine when the other side needs the electricity more. Or any region experiencing unusual weather can be supplanted from the rest of the world.

    High Voltage Direct Current transmission is much more efficient and controllable than AC grids (ostensibly 100% vs 70%). China and others (ex: IEEE) have been promoting the building of a global grid for a few decades (some decent info here:

    It would also prepare the infrastructure for upcoming increases in energy use. As Africa adds widespread refrigeration, washing machines, and air conditioning, the power need will increase by orders of magnitude.

    If climate change re-distributes the areas of sunshine, wind, rainfall/drought, then the grid provides the global (and regional) ability to adjust with the climate patterns.

    Ultimately, it also prepares for new energy “killer-apps” which take energy needs to entirely new levels. Like multiple daily rocket launches for Mars colonization, or solar system industry (or use of electrical space elevators, etc.), widespread desalinization for irrigation, weather shaping, etc. Then consolidated areas of solar/wind/tidal power, or space-based solar, or centralized (or space-based) nuclear power can be distributed worldwide.

      1. Rob

        It’s science fiction masquerading as hope. Technological advances will undoubtedly prove useful in the effort to solve our linked energy/environmental needs, but planning to colonize other planets is a pure distraction from finding real solutions that work for all of humanity, not just tech billionaires.

      2. Cine Tee

        HVDC has been in production use for decades. Wikipedia lists over 200 projects for transmission, or for interconnecting incompatible AC grids.

        And the global grid is an inevitability, specifically for increasing the use and reliability of renewable energy.

        It’s not a replacement for local storage, but the particular needs for local storage will be dictated by the global grid map and its capabilities.

        And the fact that it’s also on a direct path to support future needs for scalability is not a disadvantage.

  6. flora

    Thanks for this article. My local-ish electric utility was planning to decommission one of its coal fired plants last year. However, the Texas debacle 2 years ago during the winter when the wind generators couldn’t keep up with suddenly unexpectedly freezing temps that brought down a major part of Tx power and required several midwestern electric utilities to shunt power to Tx, and at the same time induced emergency short term rolling electric blackouts in the midwest to cover the energy shortage caused by shunting power to Tx, woke up a lot of electric utilities to the true strain on the already existing power grid. And the economic hit of buying nat gas or coal on the spot market instead of prepriced longterm purchases to cover the increased power demands.

    Now with plans for more demands on the grid with electric cars and such the utility owners and managers are planning for “what if” situations. And so, some older coal fired plants that have been planned for decommission are being kept online in standby mode for a few more years. Seems like common sense even though I wish it wasn’t necessary.

  7. Piotr Berman

    EROEI table points clearly to nuclear power being by far most cost effective if we combine investments in generation (using RosAtom prices and delivery times in friendly regulatory regimes like Turkey, Egypt, Bangladesh, India and Hungary), power transmission (you can build them at reasonable distance from the consuming area) and storage (a pure nuclear solution needs to adapt to daily and seasonal cycle too, extra capacity is easiest to think about, but that adds to capital costs).

    I do not see how countries like China and India can achieve reasonable targets of consumption (by the population) and production (agriculture, industry) while lowering carbon emissions without nuclear power. Developed countries, especially US model, have a model of excess that can be pared: large homes, large vehicles, sprawl etc. (but already EU has much less of that), but what about the Global South where the carbon emission are growing because of economic growth?

    1. Odysseus

      what about the Global South where the carbon emission are growing because of economic growth?

      The global south should be using the best techologies to support that growth. It’s insane to industrialize via fossil fuels when one could electrify everything from the start.

      Massive and rising carbon emissions are straight up not a requirement for growth.

  8. tawal

    Seems like we should have learned with Covid, that office workers can work from home. We need to convert office to housing, not wipe out more indigenous cultures in our search for finite lithium.

Comments are closed.