John Petersen
Last week I appeared as a luncheon speaker at EESAT 2009, a biennial international technical conference sponsored by the DOE, Sandia National Laboratories and the Electricity Storage Association that focuses on storage technologies for utility applications. The conference included dozens of high-level technical presentations from storage technology developers and was far and away the best-organized event I’ve ever attended. The only notable absence was a large contingent of buyers, which left some participants wondering whether they were preaching to the choir. Nevertheless, I was encouraged by rapid growth in the number and size of utility-scale demonstration projects and the growing body of proof that storage will be a critical enabling technology for the smart grid. I left Seattle more convinced than ever that the opportunities in grid-based energy storage are huge, but that successful investing will require study, patience, diligence and a firm grasp of economics.
The theme of my presentation was that some developers of energy storage devices are destined to follow in the footsteps of Arkwright, Fulton, Vanderbilt, Carnegie, Rockefeller, Ford, Moore, Gates, and Brin, and become the next generation of industrial legends for one simple reason: we’re entering an era where 500 million people in North America and Western Europe can no longer lay claim to the lion’s share of global resources because the other 6 billion inhabitants of our planet know for the first time that there’s more to life than mere subsistence. While each of them may only want a small piece of the pie, the law of large numbers will give rise to explosive increases in global demand for everything and the only way to avoid armed conflict or catastrophic environmental damage is to minimize waste in all its forms, beginning with energy.
On the cautionary side I returned often to the unpleasant reality that most grid-connected storage applications won’t pay under current economic conditions because the spread between the cost of storage and the value of storage remains narrow. That cost-benefit equation is changing rapidly as energy costs rise and renewables are added, but as long as waste is cheaper than storage, waste will prevail. The following graph comes from a November 2004 presentation by John Broyes of Sandia National Laboratories that provided an overview of the DOE’s Energy Storage Systems Program. The chart focused on the California utility market and showed the clear inverse relationship between the installed cost of energy storage systems and total demand for those systems. It merits more than a passing glance from investors who want to know where the business is (see p. 11 of the presentation for an expanded version).
While the graph contains a wealth of information on the wide variety of potential uses for storage in the utility market, the most important lesson for energy storage investors is price sensitivity. When total installed costs for energy storage systems are $1,000 per kW or higher, demand for storage is almost insignificant. As installed costs fall into the $600 per kW range, the number of cost-effective utility applications soars. I’ve been told that an updated version of the graph is in the works and will be released shortly. You can bet that I’ll be among the first to write about it.
There were several EESAT presentations that focused on important but expensive frequency regulation technologies that are priced beyond the high-range of the graph. Over the last year, demonstration systems from Beacon Power (BCON), Altair Nanotechnologies (ALTI) and A123 Systems (AONE) have shown a remarkable ability to respond to regulation signals in microseconds and provide up and down regulation at speeds that traditional systems can’t even begin to match. Based on estimates from the PJM Interconnection, one of the independent system operators that manage the U.S. grid, national demand for frequency regulation installations is on the order of 6,000 MW and could be much higher if flywheel and battery systems prove capable of handling longer duration load ramping intervals. The ongoing tests are not conclusive because the new systems have not been in service long enough to establish their useful lives, but the preliminary results are promising.
There were also several EESAT presentations that dealt with more mundane energy storage applications that were priced in the mid-range of the graph. Those projects ranged from the use of flow batteries at cellular telephone installations in Africa to a recently completed 12-year demonstration where Exide Technologies (XIDE) used lead-acid batteries to effectively eliminate the need for diesel fueled backup power on a remote island where the primary power source was renewable. Yet another presentation showed how computer analysis of satellite maps was being used to identify new locations in Ireland for pumped hydro, a technology that generally falls in the low-range of the graph but is commonly believed to have limited potential because most of the desirable locations are already developed.
Overall, the most important takeaways from EESAT were that from a utility perspective:
- Storage is the economic equivalent of a dispatchable generating asset;
- Installed cost and reliability will be the primary drivers of decisions to implement storage solutions;
- Maintenance and cycle life will be secondary decision drivers;
- An optimal smart grid configuration will need storage equal to at least 5% of peak system load; and
- As renewables become prevalent, storage will become increasingly critical to grid stability.
In Energy Storage on the Smart Grid Will Be 99.45% Cheap and 0.55% Cool, I explained that the required annual storage build in the State of California was estimated at 500 MW per year for the next decade. Of this total, 50 MW would need to be fast storage in the form of flywheels and Li-ion batteries and the 450 MW balance would be 4 to 6 hour storage in the form of pumped hydro, compressed air, flow batteries and advanced lead acid batteries. When the California numbers are scaled up to a national level, they translate to billions in new annual demand for as far as the eye can see. When you add in billions in new demand for transportation, it’s clear that the sector isn’t even close to ready for the near-term demands. To compound the problem, essential raw material supply chains aren’t ready either.
In preparation for my EESAT presentation, I spent a good deal of time analyzing how the energy storage industry of today is different from the industry that existed a few years ago. My most important conclusion was that energy storage devices are rapidly evolving from minor components in high-value durable goods to stand-alone end user products. As a result, the cost of energy storage is rocketing from less than 5% of product cost in the case of portable elec
tronics to more than 50% of product cost in the case of an EV like the Tesla roadster. When you get into the utility arena, the storage devices are the products and represent 100% of the product costs. Since consumers generally have higher payback expectations and shorter investment horizons than utilities, I believe consumer price sensitivity will be very high notwithstanding the current flood of optimistic stories, speeches and reports from the mainstream media, politicians and environmental activists.
While some of the stock market valuations in the energy storage sector reflect the emerging reality that energy storage is and will remain a highly price sensitive product, others do not. As a result, we have a weird market dynamic where Enersys (ENS), the world’s largest manufacturer, marketer and distributor of industrial batteries, trades at a 50% discount to a newcomer like A123 Systems (AONE); and Exide Technologies (XIDE), the world’s second largest manufacturer of OEM automotive batteries, trades at a 28% discount to a newcomer like Ener1 (HEV). While the valuation disparities might be justified if either of the newcomers had a technology that would displace the established leaders or significantly erode their revenues or margins, that outcome can’t be expected in the foreseeable future because the newcomers are focused on far more expensive products for markets that don’t even exist yet.
Over the last fifteen months I’ve written 92 blog entries that focus exclusively on the energy storage sector; the established and emerging energy storage technologies; and the principal competitors in the industry. My recurring simple hypothesis has been that cheap energy storage will beat cool energy storage in the market and that companies that manufacture objectively cheap products will experience far more rapid and sustained stock price growth than companies that are developing objectively expensive products. Over that time, my personal trading account that includes Active Power (ACPW), Enersys (ENS), Exide Technologies (XIDE), ZBB Energy (ZBB) and Great Western Minerals Group (GWMGF.PK) has gained over 300%. Nevertheless, I think I’ve finally reached a point where I’ve said most things that can be said. Accordingly I plan to slack off a bit and write in response to current events instead of trying to maintain a regular schedule.
Over the next decade, I believe that every energy storage company that brings a product to market will have more business than it can handle. Nevertheless, I believe that companies that have attained lofty market valuations based on ambitious plans to develop exotic products are likely to trade flat or decline in price while the companies that have less ambitious goals and less expensive products have substantial upside potential.
My favorite short-term holding is ZBB Energy (ZBB) because its ZESS 50 and ZESS 500 flow battery systems are market ready and carry an attractive mid-range price while its market capitalization of $15.3 million is but a small fraction of the peer group average. My favorite mid- to long-term holding is Axion Power International (AXPW.OB) because its first generation PbC batteries are in production and have been delivered to select end users for testing, the PbC battery promises a cheap solution for a wide variety of mundane energy storage applications and Axion’s market capitalization of roughly $80 million is well below the peer group average.
The only thing that will prove me right or wrong is time.
DISCLOSURE: Author is a former director of Axion Power International (AXPW.OB) and has a substantial long position in its stock. He also has small long positions in Active Power (ACPW), Enersys (ENS), Exide Technologies (XIDE), ZBB Energy (ZBB) and Great Western Minerals Group (GWMGF.PK).
What would a company be worth if it could produce a reliable battery storage unit that costs $100 per kw/h and it could be easily mass produced? The weight would be much lighter than LiOn battery of similar capacity.
It could be worth a great deal, but the numbers you’ve cited are so far below industry averages that I would be very cautious about the too good to be true rule.
Mr. Peterson,
As a Beacon Power Shareholder, your articles are very well read and respected in web surfing on these subjects.
My gut in finance and math and doing years of layman research around these storage technologies tells me there is some rigorous financial analysis that still needs to be done within the cyclic technologies versus the duration technologies.
The niches and strengths of these technologies can be measured in a variety of ways and compared (roughly) with each other as well as in their separate niches; however in my opinion there are hidden costs and hidden savings when also making comparisons, are there not?
However, the point I bring forth is that over a 15 – 20 year time framework, with all speculated future costs and benefits, many tables that epri and the industry in whole reference, in my opinion do not properly reflect the “true” long term costs and benefits comparisons.
As a Beacon Power shareholder and laymen study of these technologies over the years, I have sent several emails to epri themselves questioning the metrics in which these measurements and cost tables are derived from, their response seems to indicate that indeed there may be more than meets the eye. I believe their intentions are to separate and drill deeper into the numbers.
We know Beacon Power “plans” to build 20MW facilities that will sit underground and hum along with very little maintenance over a 15 year – 20 year life, more software triggered than human. Let’s assume they achieve what they set out to accomplish.
How does one measure a lifespan and “costs” during 2 decades?
What long term “costs of ownership” need to be taken into account for while making these LONG RUN comparisons?
When Comparing Batteries to Beacon Power Flywheels only for instance (within second responders):
On epris cost table published in February 2009 page 29 alludes to these assumptions.
In the footnote section itself, epri references {not including battery replacement costs and I would assume hazardous disposal costs}. If over a 15 year period batteries need to be replaced and disposed of 3 times (5 year life) do we then make the assumption that Batteries cost 3 times more in this table versus flywheels?
In fact these footnotes almost seem to say, “This whole thing is a wag long term”.
It is also not only in the storing of power, what about moving of excess power? We know that at times renewable grids may “overproduce” power that essentially gets wasted at times. We also know that flywheels may offer thousands more cycles at a higher round trip efficiency than batteries can. This cyclic advantage may allow Flywheels to store “excess” power from one renewable grid that may be overproducing power and allow flywheels to move that excess power into an adjoining grid that is need of power within minutes. This event can repeat itself many times in a day or twice an hour. It would appear that these cyclic events may cause the battery systems to wear down sooner. This moving or wheeling excess power with the cyclic advantage that flywheels have could be much more effective to “bump” up the efficiencies of these wind and solar farms at these times making their total output increase slightly.
Another cost saving which can include battery systems as well as flywheels is in the reduced costs of transmission line expenses planned for these new projects. While chatting with an important caiso member, he alluded to me that with proper Frequency balancing that transmission line expenses possibly could be brought down 20 percent or more. They essentially would not need to build overcapacity in the transmission lines to handle the swings in these new projects.
If true, how does one take these savings into play while doing these analysis?
We also have hydro plants that seem to be quite effective on these cost tables. What are the long term maintenance costs as well as long term costs that need to be considered over a 15 year cycle?
As you well know, also hydro, a great cheap duration technology also came under some environmental attack this summer in the BPA region as an unforeseen event caused havoc during a two day event. Also, I also believe a paper put out last year by some knowledge fellows alludes to the facts that flywheel systems may actually compliment hydro very well creating long term efficiencies for the hydro plant and what they refer to as the “hydro curve”.
How does one measure statically how much money these efficiencies may save a hydro plant over a 15 year stretch? Does one properly take these savings into account????
The scalability of sprinkling sprinkle flywheel plants or battery installations plants also gives some measure of cost savings in my opinion as one can’t drop a hydro plant anywhere.
What are these costs to the grid system that may be mitigated, with fast acting technology as a circuit breaker before these duration technologies “kick in”? Many papers discuss this as you well know.
How do we measure carbon footprint of these technologies. Do we figure any carbon trading type credits into this future analysis, if true, as you know KEMA has a report that bodes quite well for flywheel technology?
Cost analysis sometimes is not what meets the eye in a static view; it could also be how much you are going to save short term and long term on the grid architecture itself as well as taking an “out of the box” look at efficiencies created.
The analysis is a complex one that at times seems to be comparing apples to oranges?
I certainly would have included links to the epri table, the hydro event this summer and the paper discussing the flywheel/hydro curve from 2008, as well as the KEMA reference, but do not know what your policies are in regards to providing links, however I am sure you are aware of them.
I would be interested in your thoughts.
Edward, I don’t normally post lengthy comments but yours brought out a critical distinction that many readers miss – the difference between energy devices and power devices.
Beacon’s flywheels and the fast response battery systems are great at storing and releasing power quickly, but they’re designed for 5 second to 15 minute cycles and are terrible at storing large amounts of energy. For example, the nameplate energy rating on Beacon’s 20 MW facility will be roughly 5 MWh.
When it comes to durability, the only thing anyone has to work with are expectations. Beacon is claiming a 20 year useful life which is not yet proven and the cycle life claims from various battery system developers are comparably aggressive. The jury is still out on both sets of claims.
I like the idea behind flywheel systems because the laws of physics usually leave more room for improvement than the laws of chemistry. But an efficient smart grid will require a combination of energy storage systems for the seconds, minutes and hours time ranges identified in the Broyes presentation and it would be a grave mistake for anyone to assume that a single solution is more important or valuable than all the others.
In energy storage like most other sectors, investment diversification is critical.
Mr. Peterson,
I appreciate your response and your thoughts.
I also usually do not post lengthy posts, but also agree that many readers miss the devil in the details, thus I felt compelled to.
Agreed, that many niche areas will be deployed in the future smart grid, as you and I both point out in our statements. I also agree that Flywheels niche will be the short term, “circuit breaker” type technology.
Both you and I can also point to the fact that many around the “j boyes circle” point out that “fast acting technology” and the “ramping” capabilities of these new technologies can actually be compared to double traditional generators, i.e. 20MW of fast acting = close to 40MW traditional. These details also are rarely brought up. We can quote, hawkins, gyuk, boyes and many others in this area as you are well aware of.
I was merely trying to point out that the cost tables in my opinion are rather static in their comparisons, and do not take into the account these niche properties that both you and I discuss.
It’s the cyclic advantage as well over the long run versus batteries that I was also making a case of not only storage but energy arbitrage and the long run efficiencies that this technology brings to a variety of other already existing technologies as well as transmission infrastructure costs.
When you refer to the nameplate rating of 5MWh, is this purely derived from using 20MW for 15 minutes (.25 hour)?
If flywheels make it into commercial use and can store for 15 minutes and release in 15 minutes, I am always intrigued why when quoted this way, the second half hour is always ignored. Why would it not have a 10MWh rating?
Also, since Beacon has now had flywheels attached to the grid for a year now in the NEISO pilot program; do you think it is now possible to start doing proper long term analysis on this type of technology?
Your response in regards to the fact that some batteries are making those long term life claims and cycle advantages was new to me and I guess I will have to do more research in that area.
edward, I got the nameplate ratings from a fairly recent Beacon presentation that says its Gen 4 unit has a capacity of 25 kWh / 100 kW. See page 6:
http://www.usea.org/Publications/Documents/Beacon_Power_Presentation_USEA_21Oct08_handouts.pdf
As far as I know, the Broyes graph only takes the primary value stream into account and does not necessarily account for secondary values which are typically very location specific. It’s not a perfect presentation, but it does provide a simplified overview of a very complex subject area. The more detailed data points will almost certainly be discussed in negotiations between sellers and buyers, but I don’t think I’ll ever be able to take things beyond the overview from 50,000 feet level.
That graph is garbage!!! Seriously. You will tarnish your image by republishing it.
I’m familiar with the source data used to construct that graph. The source paper lists possible benefits of storage. Then, for each benefit, it lists the value and market potential.
The people who made this graph are misrepresenting the underlying data, unfortunately. They simply added up the different value streams multiplied by their market potential to construct the pictured demand curve.
However, the key mistake is that many of the benefits are mutually exclusive – you cannot be doing all of these things at once. Furthermore, many of these benefits overlap – you may have two names for slightly different applications of the same benefit.
I have worked at a utility doing cost benefit analysis of storage – I guarantee you no one is interested in storage that costs $1,000/kWh.
Some further comments:
I know the graph comes from Sandia, which usually provides wonderful data, but in that 2004 presentation, someone screwed up.
Their data is great. I have no issue with the original paper. The only problem is the incorrect way it’s presented here.
I highly recommend looking at the source data to check what I’m saying.
Or even spend some time thinking about how a single battery could do all of those things at once.
Or spend some time thinking about how ancillary services, renewables firming, and power quality are similar and overlapping, etc.
John,
I’d be interested in hearing anything you have to say about the non-battery, non-flywheel storage technologies like CAES and PHES, after hearing the EESAT presentations. Especially CAES, since you say a bit about pumped hydro in the above.
Ted, my biggest problem with the graph is age rather than construction methodology. It may not be perfect, but it’s far from garbage and certainly better than nothing. One of the big problems facing storage investors is the PR and hype about huge markets for massively expensive battery and flywheel systems. They have their uses and in certain respects represent low hanging fruit in the small world of frequency regulation, but the real market will be very price sensitive. If I can get that message clear for now we can worry about more accurate price and demand data when it’s released.
Identifying and capturing each of the discrete value streams from grid-based storage will be a challenge, but it will also be the key to most buying decisions.
Tom, CAES is wonderful but terribly constrained by geology, which is why there are only two working CAES systems in the world. A number of companies are looking at the potential for above-ground CAES but the technology is still pretty new. I would love to see a pure play emerge in CAES (or even a significant play) but for now I don’t see any way to invest in the technology so I don’t talk much about it.