Sandia, Stirling Energy Systems Set New World Record of 31.25% for Solar-to-Grid Conversion Efficiency

On a perfect New Mexico winter day — with the sky almost 10 percent brighter than usual — Sandia National Laboratories and Stirling Energy Systems (SES) set a new solar-to-grid system conversion efficiency record by achieving a 31.25 percent net efficiency rate. The old 1984 record of 29.4 percent was toppled on SES’s “Serial #3” solar dish Stirling system at Sandia’s National Solar Thermal Test Facility.

The conversion efficiency is calculated by measuring the net energy delivered to the grid and dividing it by the solar energy hitting the dish mirrors. Auxiliary loads, such as water pumps, computers and tracking motors, are accounted for in the net power measurement.

Each dish unit consists of 82 mirrors formed in a dish shape to focus the light to an intense beam onto a receiver, which transmits the heat energy to a Stirling engine. The engine is a sealed system filled with hydrogen. As the gas heats and cools, its pressure rises and falls. The change in pressure drives the pistons inside the engine, producing mechanical power, which in turn drives a generator and makes electricity.

Several technical advancements to the systems led to the record-breaking solar-to-grid conversion efficiency.

The first and probably most important advancement was improved optics. The Stirling dishes are made with a low iron glass with a silver backing that make them highly reflective —focusing as much as 94 percent of the incident sunlight to the engine package, where prior efforts reflected about 91 percent. The mirror facets, patented by Sandia and Paneltec Corp. of Lafayette, Colo., are highly accurate and have minimal imperfections in shape.

Both improvements allow for the loss-control aperture to be reduced to seven inches in diameter — meaning light is highly concentrated as it enters the receiver.

Other advancements to the solar dish-engine system that helped Sandia and SES beat the energy conversion record were a new, more effective radiator that also costs less to build and a new high-efficiency generator.

While all the enhancements led to a better system, one aspect made it happen on a beautiful New Mexico winter day — the weather. The temperature, which hovered around freezing, allowed the cold portion of the engine to operate at about 23 degrees C, and the brightness means more energy was produced while most parasitic loads and losses are constant.

SES is working to commercialize the record-performing system and has signed power purchase agreements with two major Southern California utilities (Southern California Edison and San Diego Gas & Electric) for up to 1,750 megawatts (MW) of power, representing the world’s two largest solar power contracts. Collectively, these contracts require up to 70,000 solar dish engine units.

Source: Sandia

Comments

[DavidJ]

The system does appear to be designed to maximize efficiency, with cost being compromised.

[Greg Woulf]

What cost are you referring to?

The Sandia unit is just the prototype, so what you see doesn't necessarily follow, either the cost portion or the efficiency.

I think this is one of the good technologies. Better than turbines, better than PV. It doesn't seem like it has very many moving parts and none of them see very rare or expensive to manufacture.

Not being familiar with Stirling engines and this system I'm not at all sure how durable it is, or how costly it would be to manufacture.

[disdaniel]

I'm sure this is just a stock photo, but wouldn't those dishes work better if they were pointed toward the sun?
j/k

[Danzig]

The efficiency record is new, but the concept and the development plans have been chugging along for 2+ years now. A lot of people in high places have "bought in" to this technology; Southern Cal Edison and San Diego G & E signed their contracts in 2005. It boggles the mind to contemplate what they are going to do. SCE's initial contract calls for 20,000 of these dishes to produce 500 MW. 20,000? Yikes.

I wonder who will produce all those mirrors?

And I wonder what rate the electric companies have agreed to pay for the output?

[GreyFlcn]

Uhm actually, it's been in development with SES for over 20 years.

They've just been making noise about it in the last 3.

[bigTom]

I have trouble seeing this tech becoming cheap enough. These dishes look pretty heavy, and they gotta stand up to the weather for decades if they are to be a viable power tech. I thought I saw these babies on TV, they claimed $6 peak watt, and maybe potential to reach $3/watt. Steerable versus a fixed
S facing PV effciency should be roughly 50% better, but that means that a $3/watt concentrating steerable would compare to a $2/watt PV panel. The figure of merit being put out by startups with new wizzbang technology is $1/watt. CoolEarthSolar (Mylar ballon based CPV) claims they can do $.29/watt (although I suspect their system won't work in the wind).

I'd like to see real cost figures. Clearly these utilities seem to have gone for these in a big way. I can't imagine they wouldn't do due diligence for such a large scale project.

[retro tech]

A great achievement. This took 24 years to beat. Manhattan project took 3-4 years.

[GreyFlcn]

==A great achievement. This took 24 years to beat. Manhattan project took 3-4 years.==

And a generous portion of the US GDP.

[Nucbuddy]

BigTom wrote: I'd like to see real cost figures.

If we figure a December average production-density of 1 watt per square meter (10 square feet), and a land cost of $40,000/acre (~40,000 square feet), the land cost alone for December power would be $10/watt.

By the way, how would one store the electrical-energy produced?

[John Flagge]

"I have trouble seeing this tech becoming cheap enough. These dishes look pretty heavy, and they gotta stand up to the weather for decades if they are to be a viable power tech."

I would think robust structures would pay back after a few decades. I would imaging the mirrors are replaceable, or will be, and the next generation of them will almost certainly last even longer. Why, then could not a plant of this type have a lifespan of 5 decades like a conventional plant. With no fuel costs, and little maintainence per decade, I would have to think that is where these firms think these could pay out.

[Danzig]

It is certainly difficult to get one's mind (and calculator) around the magnitude of this project. If the first phase of SCE's project (500 MW, 20,000 dishes, 4 years of construction) is going to happen, it means that 100 of these babies must be built, transported, and erected PER WEEK for four years! A massive capital project, wouldn't you say? How many man-hours will be expended in that? How could the actual cost be anything close to a rational number?

[bigTom]

Nucbuddy:
Two problems with your analysis, (1) These utilities are primarily interested in peak power, i.e. the hottest part of the day during the hottest part of the year. (2) For a tracker (or an overhead system such as the CoolEarthSolar) the solar component can intercept lowangle solar radiation (it shades a considerable amount of land). If the land was only used for solar the whole land cost would count, but if the area between units (or under them for the other system), can be used for another purpose the land cost is reduced.

Again, nobody is seriously proposing these things for baseload generation, I'd prefer a healthy Nuclear component for that. The real issue for the next decade or two, is can these sorts of solutions make sense for daytime power?

John Flagge:
For reasonable interest rates (discount rates actually) infinite lifetime needs a payback period of roughly 10-13 years. Lifetimes much larger than this will have little effect on the "present value" of the system. If these things are to make financial sense, the capital costs have to be kept relatively small.

Danzig:
I wonder if the contracts actually specify the purchase of that much capability? I wouldn't be surprised if these were maximum amounts subject to contingencies (such as cosp/performance of early systems etc). If a cheaper technology came along, midbuildup, I suspect the utilities could back out (perhaps with payment of a penalty).

[Nucbuddy]

BigTom wrote: These utilities are primarily interested in peak power [...] the hottest part of the day during the hottest part of the year.

If that were the case, they could save themselves money by investing in diesel peakers.

[bigTom]

California utilities have to meet a 20% renewables target by 2020. Diesel peakers wouldn't qualify. And the rules are a bit bizarre, the older largescale hydro doesn't count. I'm in the PG&E service area, PG&E has something like 49%gas, 2% coal, and 26% nuclear. Most of the remainder is hydro, but only recent small hydro counts, they are struggling towards their 20% share.

[Kit P]

"California utilities have to meet a 20% renewables target by 2020."

This is not true. Like with EVs, California can declare vicitory and change he law.

Many of the windfarms in the PNW are contracted to Cal. If conservation works in Cal to avoid building new nukes and coal plants, that same theory should apply to all new plants.

[PD]

I notice that Sandia is about 5,000 feet above sea level. How much difference does that make for solar power generally?

[bigTom]

P.D.
Top of the atmosphere solar constant is roughly 1350watts/meter**2, the typical value that is used for the surface is 1000watts. The atmospheric scale height roughly 8km. Assuming the atmosphere is exponential at 8KM depth 5000 feet would still have
82% of the atmosphere above it. Of course many atmospheric scatterers are more concentrated lower down, so you might recover 20-25% of the lost solar, so that would give us pergaps 1080watts/meter**2.

Trying to go higher up in New Mexico and the sunshine gets stronger, but the cloudiness goes up. It probably pays to put generation at lower sunnier elevation.

[Kit P]

If you go back and read carefully the PR, you will see that the record was achieved on a very cold day. This is great if your solar system is designed to match winter peak. Say 4 am.

I always find it very interesting the information the solar industry (including DOE researchers) does not tell you. It sure looks like the solar industry likes to brag about the wrong things.

[steve]

I have been out of the country for a month. Nice to see you are still spreading your positive message there Kit. Your statement about temperature and solar is at best misleading. I would have thought as an ex power plant operator you would have mentioned that all power plants, that rely on cooling, are more efficient on cooler days. Pretty much any power plant, including your beloved coal, is going to wait until a cool day to test it’s efficiency.

In the end I don’t think efficiency is that important. What is important is the cost to build and operate these systems and the amount of electricity they produce. I live in the area they are supposed to build these systems and I have some real doubts that is will ever happen. They announced that they were going to build them but have never announced just where they are going to do it. They have also never filed with California to build them. Something they were supposed to do about 2 years ago. That Edison and San Diego have signed a contract with them does not mean much to me. Basically these contracts are signed all time for various power projects and they are far from guaranteeing they project will ever get built. Basically all they say is they will buy the power if you can you can do everything you claim but there are so many ways out of the contract that they rarely ever get built. This is the same for both conventional and renewable plants.

[Danzig]

Steve: A nice summary of the issue, and a spot-on prediction of the likely outcome of this project.

One thing we may be missing in all this is the possibility that SES has developed a promising stirling engine that could be used in other applications. Stirlings and solar power would seem to be a match made in heaven, especially for driving a pump that can lift water to a tower, thus allowing 24/7 production of electricity. This is already being done in remote areas where electricity is not available.

How about it, SES? Show us your stirling?

[Thomas Marihart]

I'd like to see how reliable the Stirling engines are now and how much/often do they need service. Poor STM didn't make it...these guys, as far as I know, are the last potential residential/farm/commercial scale stirling building block left in the US.

Their smallest unit is 25KW for ~$150K now (est) going to ~$2W or ~$50K as mentioned on invention nation recently.

I'd like to put one of these at my Dad's ranch, and we could likely service the thing on site, it seems, like one would service a regular pump engine.

I also like that the unit needs ~700 Deg C heat to run the Stirling at a nominal ~25kw, which is also the stack exhaust temperature of a certain biomass gasification system I work with...

If SES can publish their thermal mass flow vs. temp curves, that would be really helpful for others who want to interface other sources of energy (waste heat) to their product, thereby enhancing its value beyond just solar. Having a couple of engine units available for Biomass development projects would be nice too.

[mds]

Steve,
Same as Danzig. Thanks for nice summary. Sounds like an industry insider view. I'm not one. Loved the comment on efficiency. What heresy. ;-) Many seem to forget efficiency is just one factor in overall system cost per kWh produced including installation.(Balance of System, BOS) There's a good reason why CIS, CIGS, and CdTe are selling so well at half the efficiency of Si.
I've been wondering what happened to the SES developments and don't understand PPAs well. Thanks again.

[Engineer-Poet]

it means that 100 of these babies must be built, transported, and erected PER WEEK for four years!

Same steps, over and over.  Lends itself to very specialized jigs and other assistance.

These dishes are about the weight of a medium-duty truck.  Detroit could crank them out at the pace of 100 an hour.

[steve]

Engineer-Poet I agree with you that once production has started putting up this many dishes would not be a problem at all. It will take quite a lot of start up money to do so though. The problem with this is that most of these companies need outside funding which can be hard to get for a new and relatively unproven technology. My guess is if they actually build them they will start in smaller numbers and build them at a less aggressive pace than they have announced.

Yes mds I am in the industry but I don’t know much about the Sterling Dishes other than what I have read. It does seem like interesting technology but there is a big difference between a few hand built and maintained units, and having to deal with thousands of mass produced units in the field. I know the company I work for looked into them and I know a couple of people who were involved with it. My gut feeling is they may not be ready for prime quite yet and they are underestimating the amount of maintenance these units will require. The plus side is they will require less water than other forms of commercial solar but not the basically zero they first claimed. Another plus is they are small units that might work will, at high efficiency , in remote areas, as Thomas pointed out.

[Kit P]

Steve wrote,

“They announced that they were going to build them but have never announced just where they are going to do it. They have also never filed with California to build them. Something they were supposed to do about 2 years ago.”

Are you saying they have not started the environmental impact statement (EIS) process?