By Jim Fiske
Editor’s Note: This is the first in a series of guest columns on innovative energy technologies, discussing current trends and new solutions. This article looks at the pros and cons of today’s grid-scale storage efforts and proposes that storage tech offering geographical flexibility as well as cost- and space savings will be the future of the industry.
Over the last several years electricity storage has grown at a furious pace, reaching an installation rate of 6 gigawatts per year in 2017, according to the Energy Storage Association. But the build-out has barely begun. Recent research results from the Carnegie Institution for Science at Stanford suggest that achieving 80% renewable energy in the next 20-30 years could require vastly more storage than currently contemplated—up to 12 hours of energy for the entire grid. Approaching 100% renewables will require additional storage with much longer duration (days to months), and efforts to find such storage have already begun. But first things first. Our primary need for the next two or three decades will be daily storage.
Some have suggested that the plummeting cost of Li-ion batteries make this a solved problem. Let’s put that in perspective. In many regions of the world the cost of wind and solar has dropped below 3¢/kWh and is rapidly approaching 2¢, while Lazard estimates the 2018 Levelized Cost of Storage (LCOS) for Li-ion batteries at nearly 27¢/kWh. Recent contracts in Hawaii put the cost of Li-ion storage somewhat lower at roughly 16¢/kWh, but this is still pretty dismal. Flow batteries have similar costs. Obviously this is far too high to provide the necessary bulk storage capacity to support a major expansion of renewable energy, and further expected declines in price are unlikely to be enough.
What qualifies a technology as a general solution to the bulk storage problem? There are no firm rules, and ultimately the marketplace will decide, but we can suggest some likely requirements:
- A low enough LCOS to allow renewables to undercut fossil-fueled generation, both in developed and developing countries, for all 24 hours of an average day;
- High Round-Trip Efficiency (RTE); (Some manufacturers advertise DC-to-DC efficiency, or even thermal efficiency, which gives an incomplete view. Only AC-to-AC efficiency tells the full story.)
- Long lifetime;
- High local content, particularly in developing countries;
- No barriers to widespread and terawatt-hour scale deployment.
Of all available technologies only Pumped Storage Hydropower (PSH), the “Gold Standard” of bulk storage, has come close to meeting these requirements. As a result it provides over 96% (more than 150 gigawatts) of the world’s stationary storage. But its land area and topology requirements and its environmental impact constrain further growth.
What are the alternatives? Established companies and newcomers are exploring almost every conceivable method to replace pumped hydro using chemical, thermal and mechanical storage techniques of all kinds. Here are some of the top candidates:
Aquion developed a relatively low-cost battery using a saltwater electrolyte, a manganese oxide cathode and a carbon-based anode. It raised nearly $200 million, intending to sell to wind and solar energy suppliers, then went bankrupt before it could scale production high enough to bring costs low enough to compete with Li-ion. After restructuring it intends to continue, but its difficulty competing with Li-ion suggests it is unlikely to reach bulk storage cost levels.
Ambri is developing a battery with liquid magnesium and antimony electrodes separated by a molten salt electrolyte that operates at 500°C. Developed at MIT, the technology’s claimed advantages include deep discharge, long battery lifetime, and capital requirements of only $30 to $40 million to build a factory to make batteries at the gigawatt scale—very low compared to Li-ion. They originally planned to ramp up production and begin sales some time ago, but re-grouped in reaction to falling Li-ion prices. They now plan on commercial release in the next two years. Again, their difficulty competing with Li-ion does not bode well for bulk storage applications.
Ionic Materials developed a liquid crystal polymer that will allow solid-state Li-ion batteries and rechargeable alkaline batteries. This could both increase performance and decrease costs. They plan to manufacture their polymer and supply it to battery companies, but no details on actual battery performance improvements or cost reductions have been released.
Form Energy has begun a quest to develop long duration electrochemical storage, i.e. weeks to months, allowing dispatchable renewable energy. Although they have released no details about technology, the nature of the application suggests it will be some form of flow battery where storage duration can be increased simply by building more or larger containers and adding cheap electrolytes. They have chosen a formidable task. Where diurnal storage can provide revenue on a daily basis, long duration storage might go weeks between paydays, making it difficult to achieve an acceptable return on investment.
To date none of the advanced battery technologies have shown they can reduce costs far enough to provide viable bulk storage. Some may also find it difficult to provide an acceptable RTE in hot climates (cooling will use a sizable chunk of energy), to incorporate substantial local content, and/or to deploy at large enough scale to make a significant dent in the bulk storage problem.
Compressed Air Energy Storage (CAES)
Two commercial plants have been built, the most recent over 25 years ago. Both have apparently functioned well but require natural gas to heat compressed air before expansion, depend on cheap caverns in salt structures that are not available in most regions, and provide a low RTE (29-54%). Neither plant design has been duplicated, and implementing improved CAES has turned out to be much more difficult and expensive than expected. Over the last couple of decades this category has been most noteworthy for its casualties (e.g. Iowa CAES, Seneca CAES, Norton CAES, General Compression, SustainX, and Lightsail). Other organizations are still attempting to deal with the constraints of thermodynamics and overcome the CAES curse using new technical approaches.
The ADELE project, led by RWE in Germany, began with the intention of implementing adiabatic CAES (where compression heat is removed and stored for later use), but after many delays and revisions appears to be reverting to something closer to classic CAES, retaining reliance on salt caverns and gas heating. This indicates the difficulty of implementing adiabatic CAES and will limit its market penetration even if it manages to be otherwise successful.
Storelectric offers a version of adiabatic CAES that eliminates the need for natural gas and claims a round-trip efficiency of 62% or more (they don’t specify if this is AC-to-AC). It still uses salt caverns, however, which again will limit market penetration.
Hydrostor provides another version of adiabatic CAES using thermal storage, deep underground caverns mined from rock, and a water reservoir at ground level to maintain near-constant pressure. Its construction cost and claimed RTE of 60% will make it difficult to achieve a low LCOS.
Many organizations have attempted to convert electricity to heat, which can be stored inexpensively, and use that heat to generate electricity on demand. Accomplishing this cost effectively and achieving an acceptable RTE has proven difficult.
Isentropic, located at Newcastle University, used a heat pump to create hot gas (500°C), which was used to heat one container of gravel, and cold gas (-150°C), which was used to cool another container of gravel. To regenerate electricity the heat pump was run in reverse as a heat engine, drawing on the stored hot and cold materials. They claimed RTE over 70%, but the company apparently dissolved in 2017.
Malta is also developing a heat pump/engine, but uses molten salt storage on the hot side and water on the cold side. With the resources of Google X behind them perhaps they have a better chance of success. Achievable RTE and LCOS have not yet been revealed.
Highview has gone in the opposite direction, storing energy in the form of cryogenic liquid air. To regenerate electricity it uses ambient air or waste heat from industrial processes to evaporate and expand the liquid air through a turbine. Liquid air is a dense storage medium that does not need a high pressure tank, so substantial storage capacity is feasible. However, this won’t be a general solution due to its low RTE of ~55%. It could be a good niche solution, particularly in conjunction with waste heat from industrial processes that can raise its RTE above 70%.
ARES (Advanced Rail Energy Storage) uses the same underlying principle as PSH, i.e. it elevates mass to store gravitational potential energy, but in place of pump/turbines, water and pipes it uses electric motors, concrete blocks and rail cars. This falls in the category of “too simple to fail” technologies, and could satisfy some niche requirements. Its need for substantial land with large elevation changes means it is not a general solution to the bulk storage problem.
Quidnet uses fracking to create a reservoir of fractured rock deep underground. To store energy water is pumped into the reservoir under high pressure, elastically expanding the fractures and lifting the land above. High pressure water is then released through a turbine to generate power on demand. While conceptually simple, success of this approach depends on Quidnet’s ability to overcome numerous complications that can impact performance and cost, e.g. reservoir leakage, hydraulic friction, the relatively low efficiency of small hydro pumps and turbines, and variable pressure that further reduces pump/turbine performance, among other issues. Success also depends on availability of suitable rock formations. This looks like a long-shot for bulk storage.
Gravity Power Plants
All the companies described above are trying to find a replacement for PSH. On the other hand Gravity Power LLC, a startup company in California, is modifying PSH using another technology that’s “too simple to fail”. Like PSH the Gravity Power Plant (GPP) elevates mass to store energy, but instead of pumping water from a low reservoir to a high reservoir it pumps water to hydraulically elevate a massive piston in a water-filled shaft (see figure). On demand the piston descends, forcing water back through the pump-turbine to spin a generator and produce power.
Net energy stored is “mgh”, i.e. piston mass m (minus the mass of water displaced) multiplied by the acceleration of gravity g and the height the piston is lifted h. Both m and h can be large, providing enormous storage on very little land.
Some of the parameters for practical GPPs are:
Shaft diameter: 20 – 100 meters
Shaft depth: 500 – 1000 meters
Piston height: 250 – 500 meters
Storage mass: 0.5 – 9 million tonnes
Piston seals: Sliding hydraulic-type
Turbine pressure: 375–750 meters of head
Power output: 50 – 1600 megawatts
Storage capacity: 200 – 6400 megawatt-hours
GPPs remove the need for reservoirs, providing an alternative with similar scale to PSH, but with lower cost and better operating characteristics. They require no new science, technological breakthroughs or factories. They use standard hydropower components and conventional shaft construction techniques, making it practical for civil construction companies to build bulk storage in countless places where previously it was considered impossible, including cities, and resulting in a large fraction of costs coming from local labor and materials. Like PHS, plant lifetime can easily exceed 40 years. GPPs need no make-up water, are silent, and are barely visible above ground. They can ramp from zero to full power in seconds, useful when compensating for variable wind and solar power. Cost per kilowatt-hour decreases and efficiency increases as plant size increases, with RTE (AC-to-AC) reaching 84% in large plants. Gravity Power claims detailed cost modeling shows the potential for an LCOS below 2¢/kWh. Combined with solar PV this could produce 24-hour power at a cost below 4¢/kWh, cheaper than almost any other source. GPPs have not yet been demonstrated at scale, so all of these claims are speculative, to some degree. If they come to fruition, GPPs may come close to meeting all of the requirements for a general solution to bulk storage.
Battery storage does have advantages; it can be installed quickly, its input and output power can vary over a wide range, and small installations can be nearly as cost effective as large ones. Storage customers must decide if those advantages are worth paying high prices. In some cases the answer will be yes. For large plants and large scale grid conversion to renewable energy, however, bulk storage alternatives such as Gravity Power Plants could be hard to beat. The precipitous drop in the cost of solar and wind energy has inspired some governments to consider goals of 50, 80 or even 100% renewable energy. Cheap, ubiquitous electricity storage might make those goals achievable. If those governments are truly serious about their goals they should incentivize the development and demonstration of any and all bulk energy storage technologies that might meet the requirements for a general bulk storage solution. The short-term costs would be a tiny fraction of the long-term benefits.
Jim Fiske is founder and CTO at Gravity Power LLC. He leads the team developing the new Gravity Power Plant (GPP), a revolutionary grid-scale electricity storage system. Jim has spent over 30 years in technology R&D and commercialization and has founded three venture-funded companies and led R&D efforts in fields ranging from high speed digital electronics to electro-mechanical energy storage systems. He received his Electrical Engineering & Computer Science degree from the Massachusetts Institute of Technology in 1978. He can be reached at firstname.lastname@example.org.