PDA

View Full Version : Solar cells on steroids



Noclevername
2013-Sep-23, 10:35 PM
What is the maximum physically possible efficiency of a photovoltaic power satellite in Earth orbit? How far can the technology be pushed?

Swift
2013-Sep-24, 02:13 AM
The Shockley–Queisser limit (http://en.wikipedia.org/wiki/Shockley%E2%80%93Queisser_limit) is the theoretical limit of single p-n junction solar cell and is about 33.7%. This is only for a single p-n junction, and multilayer cells can theoretical go much higher, approaching 86% for "infinite" layers.

Being in Earth orbit doesn't change that, it would only take out losses from atmospheric absorption and depending on your orbit, losses from night.

The current best of actual physical cells (http://en.wikipedia.org/wiki/File:PVeff%28rev130923%29a.jpg) is in the mid 40% range for multilayer/multijunction cells.

neilzero
2013-Sep-24, 03:51 AM
Suppose we will launch the solar cells in 2020.We need to finalize the design at least a year sooner so 40 percent is optimistic for space rated photovoltaic cells. Better may be ready a few years later. Let's figure a 100 square meter panel produces 60 kilowatts on January 3 when the Earth is closest to the Sun/ almost that much the rest of the year as Earth's orbit is only slightly elliptical. The panel is in a Sun synchronous orbit approximately over sunset on Earth's surface, This will allow the energy to be sent to locations on the surface during the peak demand period when the extra electricity is most valuable.
We connect all the individual cells in series. Producing one million volts dc at one micro amp or 600,000 volts at 0.1 amps. Low current minimizes the mass of copper needed. More important we need to practice very high voltage as more powerful arrays will produce many amps. The PV panel need to face the Sun and the laser array need to face the solar receiving sites on Earth's surface at perhaps 50 different locations, So lets use a flexible power line to connect the PV panel to the laser array about one kilometer apart. We will need to shut down briefly about once per year to untangle the power cord, but damaged solar cells need to be replaced about once per year due to micro meteorites and space junk. We can shut down by rotating the PV panel until it is facing away from the Sun. I think laser diodes need about two volts, so we need 30,000 of them in series. 0.1 amps will likely be too much current, so we will need several series strings of laser diodes in parallel. Laser diodes are presently about 50% efficient. So the laser light is about 29 kilowatts. We had a small amount of copper loss in the PV array and the kilometer of flexible power cord. About 20 kilowatts will illuminate the solar array on Earth's surface, which might be 50% efficient (it does not need to be space rated) and can be steam powered instead of PV. 9 kilowatts put on the grid seems reasonable as there are some copper losses getting the 3 phase voltage exact and in phase. Several energy conversions occurred making the over all efficiency quite low, but not as low as wind turbines or roof top PV.
The lasers have the advantage of smaller illuminated area on Earth's surface, and many of the receiving sites already exist, but not optimized to the laser wave length. Also the receiving sites receive energy directly from the Sun when the laser beam is not available. The laser beam can have broad band data, which can be received over about half of Earth surface via the laser light scattered from the beam. This will work better when and if we build gigawatt solar power satellites. There are several other approaches some of which are a bit more efficient. Neil

Jens
2013-Sep-24, 03:57 AM
What is the maximum physically possible efficiency of a photovoltaic power satellite in Earth orbit? How far can the technology be pushed?

I suppose the maximum possible limit would be the 100% efficient conversion of all the solar energy hitting that area into electricity. Of course, 100% conversion is not possible, but I don't think there is any set limit on how close we can get to it. So I think the answer may be, something under that figure. I was just reading an article about plasmonics, where researchers are trying to achieve much more efficient conversion using some new technologies to make layers very thin.

Noclevername
2013-Sep-24, 04:21 AM
I suppose the maximum possible limit would be the 100% efficient conversion of all the solar energy hitting that area into electricity. Of course, 100% conversion is not possible,

Just sayin'. :)

WayneFrancis
2013-Sep-24, 05:52 AM
The issue with efficiency is multifaceted fold. One is its diminishing return each layer you put in to increase efficiency raises the cost of energy produced. Also it is a "all you eggs in one basket" type of deal.
It is fine where you have engineering considerations like the amount of space taken up by the array or something like that. I would think that, in many cases, a 1m2 solar array that is 80% efficiency would be better replaced by a 4m2 solar array that is 20% efficient.

If you are launching into space there is the weight considerations along with the mechanisms for probably unfolding the larger array. But here on Earth there is quickly a very steep curve of diminishing returns on why you would try to increase the efficiency. One of the best ideas I've heard about it work on a very low efficiency product that could be painted onto the surface of roads. Sure you might only get 1-2% but the total collection surface area makes up for that many times over.

Jens
2013-Sep-24, 07:37 AM
Just sayin'. :)

Yeah, I probably didn't express that very well. I meant to say it would be the upper limit, which you could get close to but never reach?

antoniseb
2013-Sep-24, 11:46 AM
... the theoretical limit of single p-n junction solar cell and is about 33.7%. This is only for a single p-n junction, and multilayer cells can theoretical go much higher, approaching 86% for "infinite" layers. ...
That was my first reaction too, but IIRC there is some research looking at using Carbon Nanotubes of various lengths vertically attached to a surface as a more efficient photovoltaic method. I haven't seen recent work on this, so I don't know the theoretical efficiency of this, but my memory is saying it is over 40%. So, not an answer, sorry, but it is another place to look.

profloater
2013-Sep-24, 12:27 PM
Solar cells can get knocked out by high energy particles or photons so you have to build in redundancy in the way the cells are connected in series and parallel. dead cells become high resistances. Thus the question must link design life to efficiency and even to the added cost of robot repair mechanisms or intelligent switching circuits to maximise output. On the ground it has been shown that for large powers you actually get more out of a thermal system driving a turbine and it is likely that for a large system in space this could be true too with a longer life expectancy. PV is better for small installations, portable stuff. In space large lightweight mirrors are attractive to the designer, focussing heat onto a small heat exchanger, I would expect this to be a better plan for large power generation in space.

Swift
2013-Sep-24, 01:06 PM
Solar cells can get knocked out by high energy particles or photons so you have to build in redundancy in the way the cells are connected in series and parallel. dead cells become high resistances.
Excellent point. The information I gave is for laboratory devices, and for use on Earth. A lot of semiconductive materials are sensitive to radiation damage; as are other electronic components. I know for spacecraft, the electronics are rarely the current state of the art as far as such things as memory, CPUs, or solar cells, as cost, weight, and radiation hardness can be more significant factors than efficiency. Also, as WayneFrancis points out, it can be a lot easier to make a bigger, less efficient, but more reliable array.

I found this article on spaceflightnow.com (http://www.spaceflightnow.com/news/n1105/29junosolar/) from 2011 about Juno's solar array.

profloater
2013-Sep-24, 03:37 PM
very interesting. I think it's a shame they have to explain what 450 watts is like especially in old fashioned light bulbs, why not say that's ten minutes to make a cup of tea. or just say 450 watts. I hope they are right. I reckon a mirror system with a sterling engine generator could make 450 watts too. There is also shadowing, if debris covers one cell the string goes off. And as for handling the extra power near the sun, is a switch system such a breakthrough? Great project though.

cjameshuff
2013-Sep-29, 01:44 PM
One of the best ideas I've heard about it work on a very low efficiency product that could be painted onto the surface of roads. Sure you might only get 1-2% but the total collection surface area makes up for that many times over.

You're right that efficiency isn't everything, but that's a terrible idea. You'd have to maintain that enormous collection area, which is primarily required to function as a load-bearing traction surface for vehicle traffic and is also exposed to road salt, temperature extremes, deposits of soot, oil, and tire rubber, etc. Your photovoltaics now need to provide traction at least as good as asphalt in all weather conditions. Just covering the roads with layers of conductive and solar-collecting paint would be monumentally expensive, and how long until you have to strip off the remnants of the previous layer and re-apply it? A year, a few months? What about the costs of shutting the roads down for a substantial amount of time each maintenance cycle? What are the environmental costs of this constant road painting, stripping, and re-painting going to be? And you'll likely have to constantly resurface the roads to keep them in condition suitable for painting cells onto.

You'd also have to manufacture and maintain the power collection and conversion equipment scattered all along the roads, each node only handling tens of watts of power (that is, something like a million installations per 10 megawatts), and a large number of installed nodes won't collect anything due to damaged/misapplied photovoltaic coatings, blocked sunlight, etc. These converters would need to be highly cost optimized, yet would be exposed to one of the harsher environments on Earth's surface...you're looking at a substantial failure rate there as well.

In space, for general power, the metrics of interest are more power per kilogram of launch mass and lifetime in the orbital environment...there's no shortage of sunlight or space, so conversion efficiency is relatively unimportant and thin-film photovoltaics can be useful. For large scale orbital solar power satellites, one big factor of interest is ease of manufacture...if you can make usable photovoltaics in orbit from lunar or asteroid materials, you can launch a few big payloads to set up manufacturing and then keep expanding. You obviously want a relatively simple and low-maintenance process more than efficiency for such a setup.

Here on Earth, efficiency is relatively important, as increasing efficiency directly reduces environmental impact (which is supposedly the primary reason for adopting solar power in the first place). Fabricating a large area of silicon junction is monetarily and environmentally costly, adding another layer for a multi-junction cell may have more gains than costs. Concentrated solar power can be less environmentally costly, reflector area being "greener" than semiconductor area, and high efficiency cells reduce both the relative costs of the concentrator/sun tracking hardware and the amount of heating that has to be dealt with in the photovoltaics.