Power mirrors: ESolar's solar thermal test bed in Lancaster, CA, is set to start producing power for the grid later this summer. The field's 24,000 mirrors can produce five megawatts of electricity by reflecting solar radiation to tower-based water boilers that drive turbines.
eSolar
ESolar expects to start up its large solar thermal plant soon.
The hitch with solar power has always been its sky-high cost: the sun may be free, but the materials and the equipment needed to convert rays into electricity certainly aren't. That's why entrepreneurs have long been searching for a way to create a solar company with an economic model that resembles a software startup--selling a sophisticated computer program that drives cheap, commodity hardware.
Bill Gross, founder of the startup incubator Idealab, based in Pasadena, CA, believes that he's got it: an enterprise designed to kick off what he calls a "disruptive revolution" in carbon-free energy. A serial entrepreneur who has launched more than 30 tech companies, Gross is CEO of eSolar, a Pasadena-based solar thermal venture that will go live with a five-megawatt test bed of its utility-scale technology on the grid later this summer.
But that's only a tiny fraction of what's to come. The privately held eSolar and its power plant operating partner, NRG Energy, have announced agreements with three electric utilities to install 500 megawatts of thermal solar capacity over the next few years.
To put that into perspective, that is more than the current 450 megawatts of solar thermal capacity that's online in the United States today, says Daniel Englander, a solar energy analyst with GTM Research, in Cambridge, MA. And it's a significant fraction of the total of 1.5 gigawatts of photovoltaic solar capacity currently installed nationwide.
The agreements--with Pacific Gas and Electric, Southern California Edison, and El Paso Electric--are aimed at producing power beginning in 2011. While the utilities aren't disclosing the target price for the electricity in the contracts, Gross claims that he can deliver it at about 10 cents per kilowatt-hour--less than the typical rate for wind power or even a natural gas plant.
ESolar's thermal power technology has only a few basic components. Giant fields of tabletop-sized glass panels track and reflect the sun. The beams shoot at towers where water is boiled to make steam that can drive a traditional turbine. Meanwhile, special software that costs $100 million to develop runs on a bank of Dell servers. The software coordinates with cheap video cameras that continually monitor the angle of the panels as the sun rises and sets.
The company thinks adding solar-thermal technology to its natural-gas plants will make them run more cheaply and efficiently
Federal funds could help 15 gigawatts of solar projects move forward.
A simpler design could reduce the cost of solar power generated by concentrating sunlight on Stirling engines.
$/kW; ¢/kWh; on a 24/365 basis
For apples to apples comparison.
$/kW; ¢/kWh; on a 24/365 basis
This is all great, but did they solve any of the traditional problems with solar? Like the night? Or dust settling on the mirrors? Cloudy days?
What we need is a electrical energy storage unit of some type (EESTOR?) to store the energy during times when the sunlight fails.
A hundred mil does sound like a lot for software that tracks the sun, but perhaps it does a bit more like align the mirrors optimally or something.
HJ :O)
You can use the wind power while the sun is not shining or one of the grid energy storage methods, wich are becoming more and more popular, or use small biomass plants burning wood cut by sustainable forestry.
Besides the problem "the sun suddenly was obscured by clouds, a.k.a the wind stopped blowing" is no different than the problem "every household turned its oven after coming back from work". What I mean that the fluctuations in the demand of the electricity are no different than the fluctuations in the production of electricity. And we have managed to cope with that at least 40 years ago.
The next generation concentrated solar plants will store energy by heating oil which is held in an insulated tank. This compensates for short term cloud shadows.
Also, storage allows for shifting of power to the late afternoon peak need.
Skyfuel, in which I have invested as a VC, has developed a solar collector that uses a plasitc film in place of mirrors. The reduction in weight reduces the cost of the steel supporting
structure.
The initial installation will go online this fall in California.
I know that several companies offer flywheel storage for UPS systems at a mid-sized scale. Does anyone know of any plans for large scale grid-levelling flywheel systems?
My believe is the same as the articles, cost will dictate the winner. Ok, Excuse the grammar error this was typed on a small screen and my eyes are straining. Although their solution is cheap, it is nowhere as cheap as it could be. A bank of servers, bull junk. Just put a self contained aiming control system on each mirror, and figure out how to make them very cheap! Also I am not convinced that the mirror and large collector concept will win, I was thinking scaled solar collector parks may be the winner. Why not also uses a small mirror array to power a larger solar panel?
On the business side, the local governments funding these plants is a unwise solution, because as we can see the state's budgets are in bad shape. Following the traditional business model of building your own plants then selling the electricity to recoup the investment will prevail.
In my opinion their is plenty of room for improvement on the technology and business model will be exactly the same!
Also I think there will be many winners because once a large solar plant opens up, (5000MW) there will be room for additional solar providers until the local electricity demands are met. Then as always local electrical provider has a monopoly over the supply side of the equation. And they have plenty of time to make their operations more efficient.
Dr. Brian Glassman
Ph.D Innovation in Innovation Management
Purdue University
electricity costs about 5-6c to make with NG; wind costs about 10c, proven. all experiments to make concentrating solar electricity have wound up costing a lot more than advertised b/c of service costs. i look forward to someone getting this right, but i will not buy the hype until i see some certified accounting
You are right, right now wind is cheaper, but this may be an economy of scale issue. Once solar power plants get built they will become cheaper as well. That's what I hope at least.. ;)
And one big advanatge: Electrical energy can hardly be stored, which means that you will always need large capacities of backup power plants for wind power to run in case the wind just isn't blowing, and the wind conditions can change rather quickly, putting stress on these power plants.
Solar thermal plants, by contrast, are already being built with thermal storage. Example: The Andasol power plants in Spain, which can run for 7.5 hours after sunset. These plants thus produce a predictable and controllable output and can thus be integrated in a grid much better.
Manufacturing in the United States is in trouble. That's bad news not just for the country's economy but for the future of innovation.
On Twitter
Become a Fan on Facebook
Subscribe to the Feed
ppinheiro76
1 Comment
Tracking control
I don't understand the question of the need of having a "special software that costs $100 million to develop runs on a bank of Dell servers. The software coordinates with cheap video cameras that continually monitor the angle of the panels as the sun rises and sets."
The tracking of celestial bodies (and in this case, the alignment of a mirror to always point the sun's reflection towards a fixed point) is a very easy problem, as the orbit and rotation of the Earth are stable. There are really cheap telescopes that do star-tracking with minimum software/hardware. Deploying a solution based on such hardware (even if modified to cope with larger more powerful servos) would be much more economical.
Reply
z0rr0
98 Comments
Re: Tracking control
My sentiments exactly - granted, it involves aiming 24,000 panels, but still, the geometry is pretty predictable, I would think.
Call me jaundiced, but I wonder what Congressional plus up is funding this magnificient piece of software, or are gullible investors more impressed by the smoke than by the mirrors?
Reply
briang1621
170 Comments
Re: Tracking control
"smoke and mirrors!" Hillarious
100 Million for software is rediculous the story must be wrong! That 100 million must be for the design and testing of everything.
Reply
erbium
331 Comments
Cheaper tracking control
if it really cost $100 million, I'm surprised they didn't hire a bunch of illegal aliens, as companies are apt to do in the SW, to simply walk around the field as needed and adjust the mirrors.
Reply
carlhage
84 Comments
Re: Tracking control
The "$100M" software is to track the mirror focus, not the unreflected sun. If you have imprecise mirrors and/or tracker-mounts, then an open-loop control system where the computer moves the mirror motor to some specific angle won't be precise enough to maintain a calibrated partial-degree angle. (The weight of the mirror might flex the mount, for example.) The solution is trivial-- put a video camera (or a set of 3 or 4) at the tower, and the cameras "see" the field of mirrors. When a mirror is in focus, it will be bright, otherwise not. With one camera, the mirror could move slightly, or with multiple ones, the brighter ones to one side or another indicate a shift is required. Of course, you need something to keep the cameras from melting, e.g. a partial reflector into a pinhole, or an actively cooled pinhole cover (no lens required!).
Once aligned, relative movement of the tracker would be fairly predictable, so a single camera calibration of a mirror might only need a few seconds of back-and-forth movement every few minutes to optimize. Also, the calibration might be pretty similar day-to-day, so the computed tracker motor position should be pretty close.
I don't know if this is really what they do-- it's just what I thought of many years ago when I read the mirror mounts cost $10K each (or whatever it was).
I don't understand why it takes $100M for the software-- they paid too much (or stretched the truth). Seems like $100K is more reasonable. The Keck telescope cost $100M including all the mirrors, actuators, _and_ software. It also uses active focusing to align the mirrors as the telescope moves.
Reply
briang1621
170 Comments
Re: Tracking control
See already one person on this forum solved a problem that cost $100Million for around $100K. This is a testament to the wisdom of the larger community. See how open innovation work, you have a problem and come up with a dumb $100Million solution, then you put the idea out there, and some random guy gives you the solution you were looking for! Maybe I should email "esolar" this Smart solution but they probably will not like it because it is not branded with Dell! Go dell sales guys!
Dr. Brian Glassman
Ph.D in Innovation Management
from Purdue University
Reply