The most widely misunderstood subject in the field of energy storage is the potential for grid-based applications. They fire the imagination because the grid is so pervasive and the need is so great. They also present immense challenges to storage technology developers because the fundamental economic value per unit of grid-based energy storage is very low. While the subject of grid-based storage provides rich fodder for media reports and political posturing, the reality bears little relation to the perception. On March 9th, Lux Research published a sorely needed reality check in a new report titled “Grid Storage – Islands of Opportunity in a Sea of Failure,” which concluded that “Amongst the sea of possible scenarios, there are few combinations that offer an acceptable payback, while endless potential pitfalls exist.”
Lux analyzed the business scenario for 14 emerging energy storage technologies across 23 applications to identify the best investments for utilities, transmission operators, independent power producers and building operators in California, Germany, and China. The report was based in large part on data from a December 2010 study published by the Electric Power Research Institute, “Electricity Energy Storage Technology Options – A White Paper Primer on Applications, Costs and Benefits.” While the Lux report and the EPRI study both offer detailed insight for institutional investors that are contemplating investments in energy storage, they’re too detailed for individual investors who are mainly concerned with managing their personal portfolios.
The first thing individual investors need to understand is that while global electric power generating capacity is roughly 4,000 GW, total installed energy storage capacity is less than 128 GW, or 3.2% of generating capacity. The second thing they need to understand is that substantially all of the existing storage facilities are pumped hydro. The following graph from the EPRI report provides additional color on how much installed capacity really exists for the exciting new energy storage technologies the press is gushing over.
While EPRI’s installed capacity graph should be enough to make cautious investors pause to check their assumptions, another graph from the EPRI report is far more useful. It shows the estimated size of the potential market for 15 key energy storage applications on the horizontal axis and then shows the maximum price per kWh of storage capacity an end-user would be willing to pay on the vertical axis. The red annotations are mine.
Wholesale frequency regulation, the application that’s getting the bulk of the media attention, is shown on the left-hand side of the graph. It’s the primary target for cool storage technologies like flywheel-based systems from Beacon Power (BCOND) and lithium-ion battery based systems from Altair Nanotechnologies (ALTI), A123 Systems (AONE), Ener1 (HEV) and others. Despite the media’s excitement, the reality is wholesale frequency regulation represents less than 1% of potential demand for grid-based storage. The other 99% can only be served by cheap energy storage technologies. Less than a half of the potential market will ever be addressable by manufactured energy storage devices. The rest will remain out of reach without widespread deployment of pumped hydro, compressed air and other large-scale electro-mechanical systems.
There’s little question that the potential markets for manufactured energy storage devices in grid-based applications are big enough to support several successful companies. They’re just not as easy as the media reports would have us believe. Wholesale frequency regulation in the US is probably limited to something on the order of 400 MW, which works out to about $1.6 billion in domestic revenue potential. The bigger prize is the $16 billion of potential demand for manufactured systems that can be installed at a price point of $500 to $1,700 per kWh. Globally, those target markets are closer to $5 billion and $50 billion, respectively.
Of the electro-chemical energy storage technologies discussed in the EPRI report, conventional and advanced lead-acid batteries and flow batteries usually offered the best cost profiles for the work of transmission and distribution upgrade deferral in both fixed and transportable formats. The economics remain challenging when you include the costs of containerization, interconnect equipment and control electronics, but they are within the realm of reason. Once you get beyond short-duration frequency regulation, however, cool technologies don’t stand a chance of being competitive.
The universe of publicly traded US companies that can respond to the need for cheap grid-based energy storage is small. It includes Enersys (ENS), Exide Technologies (XIDE), and C&D Technologies (CHHPD.PK) in the established manufacturer ranks with Axion Power International (AXPW.OB) and ZBB Energy (ZBB) in the emerging company ranks. Cool technologies will probably continue to claim the lion’s share of the headlines, but cheap technologies will almost certainly claim the lion’s share of the revenues and profits. From an investor’s perspective, those are the only metrics that really matter.
Disclosure: Author is a former director of Axion Power International (AXPW.OB) and holds a substantial long position in its common stock.
you (they) are wrong about the target value for frequency regulation.
It can be $6,000/kWh or even higher as this press release clearly shows
It’s a 10MW/2.5kWh system, sold for $18 million. Total cost to INE must be close to $8,000/kWh.
So when the authors are wrong about the target value, how much emphasis should I put on their market size estimate?
I have no desire to argue about the accuracy of EPRI’s conclusions. They are what they are.
EPRI certainly has far more expertise than I could ever claim. These first round frequency regulation projects in the US and elsewhere cannot be characterized as anything other than demonstrations. The current pricing may or may not be achievable on a long-term basis. Under the circumstances, I’d be reluctant to bet against EPRI.
yes. But nevertheless I am showing you the facts, in real life. This 10MW system for frequency regulation is a first and it is absolutely a commercial sale, not a demonstration.
As far as A123 or Ener1 goes, I haven’t seen convincing evidence that their batteries can do FR commercially. Beacon obviously can but their cost may be too high.
I believe Altair Nano has a good product for the work of frequency regulation, but there is nothing beyond a single sale to indicate what its price will be. Right now everybody is doing demonstrations and getting as much revenue as they can for their installations. Over time, a price for the products will become apparent. A single sale doesn’t get us there.
The price is dependent on the region. Altair have an econometric model that shows exactly where the value is.
Apparently the value of doing FR in El Salvador is between $100 an $120 per MW per hour. At least that’s what the CEO said a while ago. Quick math shows that INE has a pay back time of roughly 2 years.
I had been wondering for a while why INE is purchasing 10MW and why they didn’t try out 2MW first. I think the answer is obvious. They will make a lot of money on this.
It’s all about finding the right locations. And Altair have (secretly) worked towards an 80MW sales pipeline in Latin America and a 180MW sales pipeline overall.
I expect the price to be roughly the same throughout Latin America and I can only guess what it would be in China for instance. The value in the US is a lot lower.
As for EPRI or LUX and their research, I think it is important to point out that companies have their secrets…
I don’t disagree that prices may vary in different parts of the world. I would only suggest that caution is in order because competition tends to decrease prices over time and with a 10-year asset like storage, demand disappears once it’s satisfied.
I’d be curious to hear John and Tom’s comments on this storage project using NGK NaS in Mexicali.
It’s a fascinating proposal. I’ll be interested to see if it happens and how it works out. The economics of a project that size strike me as tough, but I have to believe the developer knows his business.
According to the EPRI chart, it looks like NaS batteries have 9 times the market penetration of the next most successful battery technology (PbA). Why do you think that is?
NGK is my top stop pick for rebuilding the Japanese grid. See here: http://blogs.forbes.com/tomkonrad/2011/03/18/green-stock-picks-for-a-post-fukushima-world/ and also a forthcoming article tentatively scheduled to be published here Thursday night.
NaS was developed in Japan by NGK Insulators which has heavily deployed it in Japan. It’s a good battery for long discharge duration applications, but fairly expensive.
When you say NaS is fairly expensive, you mean compared to what? According to the data I compiled in 2009 (http://bit.ly/i09cPV) on a $/kWh basis, NaS is the cheapest battery except for PbA and some flow batteries (PSB & VRB.) PbA is not suitable for long discharge because it reduces cycle life, while PSB and VRB flow batteries have very low round trip efficiency (RTE). NaS has the best RTE of any battery I included in my study, which is my best guess as to why it has succeeded.
I have never seen you write about RTE (by which measure Li-ion batteries also do well.) Yet RTE becomes more and more important the more expensive the electricity is that you are using to charge the battery.
Last week I used the simple analogy of a bottle for energy and observed that the battery was thousands of times more valuable than its contents. In the case of NaS, it’s tens of thousands. The energy that goes into a battery is effectively worthless but the energy that comes out is very dear. So in my view RTE is just not that important. Cheap storage with a 50% RTE is much more economic than expensive storage with a 100% RTE.
The biggest thing NaS has going for it is years of use in grid scale applications. It’s already completed the demonstration tests that other chemistries are just starting.
Thanks, guys, for sharing your comments. Tom, I have been following you on Forbes and Twitter and look forward to the forthcoming article on NGK. Like all of us, I am trying to determine the best place to invest in storage and transmission today.
Your assumption that the price of electricity to charge storage is negligible only applies for long term storage application. Frequency regulators will be paying average wholesale prices, as will storage for T&D deferral, etc. In other words, all of the applications you’ve labeled “cheap technology targets” will be paying average wholesale prices.
For T&D deferral, we can expect the storage to cycle 1-2 x per day, and pay for electricity costing 5 cents per kwh. That means PbA with an RTE of around 72%, will buy about 21 cents of electricity/day, while NaS, with an RTE of 92%, will pay 16 cents. Over 10 years, the difference of 5 cents per day is $182.50, which is almost as much as the cost of the PbA batteries (although less if you use a time value of money calculation.) Either way, RTE is significant.
According to the numbers I’ve found, NaS batteries cost around $450/kWh, while PbA cost between $150 and $300.
If we than look at an area with expensive electricity, like Japan, the extra cost of electricity for charging should make PbA unattractive compared to NaS.
While I assume that FR is a stand-alone business that may well be performed by a service provider, T&D deferral is an in-house utility function and the installation lives are usually relatively short – a few years as opposed to decades.
Sodium sulfur is a great battery if a utility has a demand profile that needs the support for a four to six hour discharge profile. It’s overkill if the peak shaving interval is only a couple hours a day or not every day, which is typically the case with T&D deferral.
One of the biggest risks to investors in energy storage is assuming that a particular use will effectively and fully utilize the capacity of a storage system. Ultimately the engineers are going to decide what the best solution is for a particular installation and when you understand the tremendous variability, general conclusions become impossible.
NaS is the system of choice if you want to shift wind from night to day or time shift several hours of solar power. Beyond that the field is wide open.