First learn and read about how transformers work. They are bi-directional by nature, law of physics my friend. They are called step-down transformers by the power company (cause that is the way power normally goes), but they just as easily step-up cause the electrons don't care what the label says. Its done all the time with standard utility transformers! You're arguing it as if this is a theory. It done all the time, passe, old news.
And of course this hinders centralized control - if a power line needs to be worked on, how do you guarantee that no current is flowing in it?
All grid-tie inverters on the market autosense the power lines going down or short-circuit to pass the NEC. This is called anti-islanding.
convert the pole transformers to be dual and rig it so that if you generate excess power, your metering runs backwards
Nonsense, nothing new required. Hook PV panels on one side of a standard meter, and it spins backwards feeding power onto the grid. PV owners do this all the time even without the utilities co's knowledge.
slightest mismanagement can cause blackouts and other system failures
The whole point of distributed generation is a highly redundant system of local power sources produces a more reliable power grid. Old news, not theory.
Look. I'm not interested in a argument for its sake. Every time you pull some new reason out of the air that solar can't work I've proved you wrong and uninformed. If you want to learn more, or have informed opinions you want to share, then fine - but I am starting to feel this is going nowhere. (have you learned anything about solar in this process?)
you can't expect to transfer 100 volts cross country without serious inefficiencies
It just occured to me you don't know how transformers work. They step up one direction, and step down the other. Even if I generate only 240 volts AC with PV using a small residential inverter it goes through my transformer and steps up to medium voltage distribution lines (4-20kV), it can go from there through a distribution station transformer to 230kV+ for long distance transmission power lines. Transformers are TWO WAY. There is no low voltage transmission requirement here.
PV is no different in this regard from any other power source.
Like I said this is being done all the time, there are over a 200,000 small residential systems (ave. 10kW peak) grid tied in the world right now, and working fine. (40,000 new grid tied systems in just 2003). The growth in the grid tied market is 60% per year, meaning another 64,000 systems went up in 2004.
FYI, here's a small primer on how large scale generation works: first there is a generation source, producing low (5-20 kV) voltage electricity, which is then 'stepped up' by a transformer to above 100 kV over alternating current. This is then sent through power wires to a substation which then 'steps down' the current to residential levels (there may be multiple transformers up and down).
You answered your own question here. Solar can work with the grid like *every technology* works with the grid. Inverters can output 240V or 3 phase multi-kilovolt output (such as this 20kV unit). Works just like any source, it is stepped up or down by transformers. No difference here.
The big difference is PV is distributed . This means a far more efficient, redundant, and secure grid. But it also mean less grid stress, because more power is generated locally. For example, normally my power comes from my roof (distance 20 ft). Sometimes some of my power comes from my neighbor (dis. 500 ft). Occasionally some my power comes from the shopping mall (distance 3 miles). When the insolation is low some of my power is imported from 2 states away (600 miles). Say the weighted average distance my power travels is 1 mile (down the same wires it would have before). Now compare that to the centralized infrastructure we currently use which 90% of the time its traveling 600 miles! Transmission efficiency is improved and grid utilization is reduced.
For the 'grid' is a misnomer - its NOT a two-way street
It is, in fact, more efficient as a two way street. This is very foundation of concept of distributed generation which has been successful at reducing grid stress and $ for a couple decades already (mostly NG turbines), and which PV is a good example of. Centralized power is everything that is wrong with the grid today. If you want to learn about DG read: here, here, here, here, or here.
And that the EROEI is quite different when you consider solar taking the major power role.
Huh? Just because you want it to? So you can support your argument? Not only is there no evidence for this, but it defies all the fundamental tenets of mass production and the benefits of scaled industries. If anything EROEI will rise. (because of improvements in technology, manufacturing process, installation efficiency, density of systems will reduce maintenance costs, etc, etc).
We've already shown it to have a better EROEI than fossil fuels, even when favoring the fossil fuels with less stringent EROEI calculations (i.e. not counting embodied energy of equipment). If PV has a EROEI of 15 (minimum), and since the fuel/sun is a free resource, its energy output can replicate itself by x^15. Hardly a problem
Hence we'd need to reinvent ourselves in a big way in order to use solar energy in load-based systems.
Not at all. You're still underinformed about PV technology. PV itself is low voltage, however its converted to AC using an inverter (same thing for fuel cell BTW) and transmitted over the grid just like normal electricity. Done all the time, every day, all over the world. 93-98% efficient (I know, we haven't accounted for that yet in our land calcs). PV very nicely integrates into the current electricity structure - no change needed The EROI numbers given already account for the inverter energy inputs, and its conversion efficiency. The EROI stands as it is.
I didn't say Odum's notions of transformities is useless. Its quite interesting from a perspective of environmental capital. But it disregards time and space. It took a lot of solar and geological work to produce oil or coal thus giving it more "eMergy". While there is a lot of environmental capital invested in oil and coal (by the earth and sun), that isn't very useful measure of energy viability, and has nothing to do with EROI. Outside the study of biosystems natural ecological capital, it's really not useful at all.
If for example, it takes 10 sun units to make 1 coal unit over 100 million years, that doesn't imply coal has any real measure of "betterness". Cleveland's point is the important questions from a energy policy/economic perspective are:
What is the economic ROI?
What is the energy ROI?
What is the environmental impact?
What is the quality of energy source? (how useable is it)
What is its realibility?
etc.
People are "still bothering" with oil because it is a reasonable storage medium, not nessesarily a energy source (similar to hydrogen). More importantly, people are using oil because there still is cheap middle eastern oil (which has an OK EROI, 3 times US oil).
Odum's work is interesting but not useful for EROI comparisons (even if updated with current solar data), because it is looking at the social use of resources as energy aggregation in terms of solar energy units. So coal, IS by nature further up the energy aggregate food chain than the solar energy that made it in the first place. And from a economic, resource availability, EROI, and environmental impact of use view point, So what?
Cutler Cleveland is his review of Net Energy Analysis Methods (A very good overview) says this:
"It is important to differentiate between two aspects of Odum's contribution. The first is his development of a biophysically-based, systems-oriented model of the relationship between society and the environment. Here Odum's (1971; Odum and Odum, 1976) early contributions helped lay the foundation for the biophysical analysis of energy and material flows, an area of research that forms part of the intellectual backbone of ecological economics...
The second aspect of Odum's work, which we are concerned with here, is a specific empirical issue: the identification, measurement, and aggregation of energy inputs to the economy. Emergy (with an "m") analysis is a pure cost-of-production approach that measures the quality of a particular type of energy by its transformity. Transformity is the amount of one type of energy required to produce a heat equivalent of another type of energy. To account for the difference in quality of thermal equivalents among different energies, all energy costs are measured in solar emjoules (SEJ), the quantity of solar energy used to produce another type of energy. Fuels with higher transformities require larger amounts of sunlight for their production and therefore are more economically useful (Odum, 1988)...
This approach raises a fundamental question about the appropriateness of transformities to reflect energy quality: Is the usefulness of a fuel as an input to production related to its transformity? Probably not...Thus, while Odum's method provides a useful framework for highlighting he important role the environment plays in generating energy and material resources, it is of dubious value in comparing and aggregating energy flows in economic applications...
In addition to this conceptual issue, there are computational problems with emergy analysis that make transformities incomplete indicators of energy quality. The calculation and application of transformities are time, location, and technology specific, yet Odum and his colleagues mix the temporal, spatial, and technical scales of their analysis in ways that are poorly defined. First, Odum presents the transformities as constants, but based on the method used to calculate them (Odum and Odum, 1983), the transformities are clearly dynamic because they are based on the first law efficiency of technologies such as power plants, coal liquefaction, and oil refineries. The efficiency of those technologies have changed dramatically over time. Second, the emergy calculations also contain an ad hoc mixture of spatial scales. The basis for the calculation of the transformities is the thermal efficiency of a wood-fired power plant in Brazil, but the efficiency of power plants vary throughout the world (Smil,1991) as do all the other energy conversion technologies used in the emergy calculations. Similarly, energy/output data from the New Zealand economy are mixed with the Brazil power plant data to calculate the transformities, which are then applied to many other economies throughout the world (Odum and Odum, 1983; Odum et al., 1987; Odum and Arding, 1990; Huang and Odum, 1991). Third, the values of the transformities are highly sensitive to technological assumptions made by Odum and Odum (1983). They calculate the relative quality of oil, gas, and coal based in part on the fact that the first law thermal efficiency of converting natural gas in boilers is 20 percent more efficient than the conversion of coal. However, the relative thermal ef
Yes, we are comparing apples or pears. But not the way you think. I've been reading more EROI papers, and it comes down to this: Solar, wind, hydro, and nuclear EROI numbers are largely inclusive of their externalities, conversion efficiencies, and construction embodied energy. Whereas typically fossil fuels EROIs are based on a simple energy in-to-thermal energy out at the well head. So in comparison, fossil fuels EROIs are very optimistic.
Cutler Cleveland (Director of the Center for Energy and Environmental Studies at Boston University) appears to be one of the leading energy analysts these days, his work is quite broad. In "Net Energy from the extraction of oil and gas in the united states", Int. Journal of Energy 30 (2005), he shows that US Oil production has a EROI of 11 for energy in/thermal energy out. And gasoline is 30-50% of this value (ie 3.3-4.5 EROI). Now that doesn't include conversion efficiency in a car or power turbine, nor does it include the embodied energy of the extraction equipment or ICE/power plant to burn it.
If you make calculations just for conversion efficiency (33% ave) of US oil converted to electricity/mechanical power has EROI of 3.6, and gasoline in a car is less than 1 (meaning that the energy to do mechanical work in a car is being subsidized by by electricity (coal) to run the extraction equipment). And still we haven't considered the embodied energy in the extraction equipment or the ICE. (now of course middle eastern oil is 3 times better than this) That is very poor EROI! And coal isn't looking much better. Both of these resources EROIs have dropped by at least a order of magnitude over the last 100 years as extraction becomes more difficult. The future of fossil fuels by EROI analysis looks bad.
As for Alsema, he does review the added embodied energy of infrastructural components in section 4.5 of the paper. For complete balance of systems analysis (inc. frames, structures, concrete, maintenance, etc) the best technologies (thin films and ribbon Si) have EPBP of 1.2-2 (15-25 EROI @ 30y, 25-41 EROI @ 50y). scSi is around 3.3 years (9 EROI @ 30y, 15 EROI @ 50y). Analysis shows that PV energy is manufacture side heavy, with little continuing energy inputs as would be expected from a solid state, fixed, and essentially maintenance free device (how often do you maintain your current roof shingles?). However, even his latest numbers are out of date as he notes getting information from manufacturers is difficult because EROI calcs involve knowing trade and financial secretes so it takes a long time to get agreements in place. Also his calcs don't use the best efficiency panels on the market, which underestimates EROI, if that was the criteria on which we made purchases. Also multijunction concentrators, should be significantly better since 1) they use less material per peak watt and 2) they have a higher efficiency. References: Here,Here, Here, Here
What's the end result? EROI calculations beyond first or second order become quite tricky and controversial. But we can show that solar in a detailed "second order" or more EROI estimate looks very favorable compared to even a "first order" estimate of oil, NG, or coal.
Array DC efficiency (ranges from 2% to 14%)...'30% efficient cell' becomes 14% DC efficiency.
WRONG! Read before saying complete NONSENSE. This graph ISN'T Entechs multijunction concentrators, but a number of different OLD mid 1980s PV technologies in test at the PVUSA site since the 80's. READ to the end of the doc PLEASE! (and read the graph titles for goodness sakes) Or read the other resources I gave you. The rest of your argument thus follows as pure nonsense (feeling silly?)
if we need to cover the equivalent of texas/alaska with solar cells...
Obviously this follows as nonsense also. Our numbers stand at 5% of Texas (13,491 Mi^2)with 30% concentrators. Land use is not a factor . We have 30,000 Mi^2 of parking lots alone in the US add that to 9,400 Mi^2 of building roofs and we have 3 times the space in just these two unutilized surfaces. Even if we used new land, it would be very favorable compared to the current energy infrastructure.
You can squint your eyes, stomp your feet, and don that helmet believing whatever you want, but reality is waiting when you want to join us.
...do the same studies as does Odum, I'd be interested. But by a cursory glance at his website, he does sound like he's interested in a fairly narrow bit of the energy production spectrum
Odum's "ENVIRONMENTAL ACCOUNTING" is out of date (1995), if it took him 2-3 years to write, his numbers have come from older published works, which themselves old by publication, would make the data from the mid-80s. Also, he gets only 9 periodical references in the last 15 years on Compendex & Nexus Lexus Environmental. Vaclav Smil gets only 3. BOTH HAVE ZERO PUBLICATIONS on photovoltaics or solar energy, let alone PV EROEI. They do not appear to be experts in this area AT ALL. They themselves have never done an indepth study of solar EROEI, or it would be published (they are siteing others work that is very out of date, read references please!).
His 'worse case' scenarios are always current technology - and even his base cases are a *lot* more optimistic than his worse case scenarios.
The PV market is expanding at 45% per year, the technology at least as fast. If you publish a study, and you see what is being done, and phased out, versus and what is being done on pilot plants and will hit commericalization in 1-2 years, which would you pick as your base case? If you pick the former you study is out of date by publication time. Remember that study was published in '96, the data was coming from 94-95, we are now in 2004. The decade has been good to PV. Read his more current stuff if you are interested.
Alsema (Professor in the Dept. of science technology and society at Utrecht University, Naterlands) has 13 peer reviewed papers published on PV all relating in some way to EROEI or environmental impact. In his most recent publication (2004) in Refocus he says:
"Recent studies give the impression of photovoltaics having considerable environmental impact. Looking closer at the data however, it is clear that these studies are based on photovoltaic systems of the late eighties, with only minor recalculations. Since the photovoltaic market has increased rapidly, a lot of progress has been made regarding the environmental profile of photovoltaics." He goes on to show, for example how current production ribbon silicon panels have a payback period of 1.2
years (and that is silicon, not even thin flims like CIS).
I favor Odum's numbers (from 'Environmental Accounting'), which include...
Read Alsemas numbers, he does too. He breaks it down for you though. Thin films have a EPBP of ~.5 (~60 EROEI @30 years) for frameless panels such as these roof shingles. With poles, mounts, concrete, yada, yada the EPBP is 1-2 years (or a 15-30 year EROEI).
If you don't like Alsema cause he blows your outdated arguement read Kato. Kato (prof. Japan agriculture university) has 25 publication all dealing with photovoltaics. He has several publications dealing with EROEI.
Or read this (note this is dated and the 2 and 1 modules are already on the market).
We have done a thorough job of calculating this from the principal numbers. An order of magnitude more accurate than using some rough, back of the envelope 30W/m^2 number you gleaned from a book. I don't necessarily have a problem with this author. But he is either spinning the facts, or you are misunderstanding them. I don't know what to tell you, you've already worked the numbers and seen for yourself.
Now, there is only so much radiation per m^2, so it takes approximately 150% as much area to do tracking collectors. So the average *real* efficiency for these things is going to be 40% * 4.3 / 6.6 = 26%. At least before we consider the next part.
No the real efficiency IS their efficiency! As for land utilization, that is something else. Still, you are misunderstanding your calcs (you divided the wrong direction). If 4.3 = 1 (base case). 6.6/4.3 = 1.53 OR 53% more with 50% more land. As I said a wash. Also we have already been calculating based on the land area for tilted (39 degrees) fixed panels having 28% extra land use due to shadowing, not trackers (so your point is moot, even if wrong).
DIRECT BEAM SOLAR RADIATION FOR CONCENTRATING COLLECTORS
Good thinking, but you are uninformed. These numbers are for IMAGING OPTICS. Many concentrators are using NON-IMAGING optics, Which is what Sharp is using in their 28% efficiency concentrator, which I've given you link for. These take in direct and global insolation. These don't even require tracking.
we have invested in the ICE...
I started with land use for electricity only. You wanted to talk about total energy, so we have. No matter if you use SOLAR OR NUCLEAR (or wind) you won't be using a ICE. And frankly while the ICE will be around for a long time, the industry is already moving into electric hybrid autos, because when you can get 50+% versus 20%, guess who'll rule the market in 30 years, when the EROEI for oil is below 3? There are other good transisional options as well, but that's not the topic.
I've seen everything from.75 to an EROEI of 2.5 for them
Frankly, either you pulled these numbers out of the air, or are using 30 year old numbers for stand-alone systems that include lots of lead-acid battery storage (which have horrible embodied energy). Yes storage, I know, we will get to that once you've admited the land and EROEI arguments don't hold water, we can't move forward until you understand what we already have done (Binders off, math reality set in). I've shown you a detailed peer-reviewed government comissioned paper showing 15-30+ MINIMUM EROEI(30 year). A EROEI of 15 (25 @ 50 years) assumes ground mounted with steel structure and concrete footings (but recycled Al or polymer composite mounts would bring it to 30). Building mounts using frameless roof integration take you up to 50+ @ 30 years (83 @ 50 years).
If you really want to understand this issue, read the journal article thoroughly, read his references, read his other studies, read "Energy pay-back time and life-cycle CO2 emission of residential PV power system with silicon PV module", Progress in Photovoltaics v6 1998 by K. Kato. And "Energy pay-back time and CO2 emissions of PV systems" Progress in Photovoltaics v8 2000 by K. Alsema.
You will find different EROEI numbers around for PV. The newer the study, the better they are (unlike most energy sources, PV's EROEI is going up not down). The industry has focused more on EROEI in the last 5-7 years as manufacturing has become more efficient.
Alsema is a leading expert on this field on study (you'll see his name on many in-depth studies) and shows the energy pay back period to be:
Multicrystal Si: 0.8 years (EROEI 37.5 @30y, 62.5 @yr)
CIS: 0.4 years (EROEI 75 @30y, 125 @50y)
CdTe: 0.6 years (EROEI 50 @ 30y, 83 @ 50y)
Crystal Si: 3.3 years (EROEI 9 @30y, 15 @50y)
I should note that these studies are again becoming 5-10 years old again and don't reflect the improvements in efficiency. CIS for example in this study was modeled at 12% efficiency, but the best efficiency (2003) is now 19.2%. Same with crystal Si, efficiency have edged up about 3-4%. I wasn't able to find info on concentrators, but because of low materials-to-power ratio I expect they will be at least as good.
I calculated the numbers for both 30 years and 50 years. Comparisons are usually done on a 30 year basis since this is the build life of other power plants and is a typical load period. However it is a bogus, made up number for easy comparison. Many PV manufacturers are guarantying their panels for 25 years, with no or little power degradation from the specs. No reason to artificially cut their life short.
The numbers I used were the base case, not best. Of course its possible to package them with material having more embodied energy (as in the case of the silicon numbers which has a significant amount of aluminum in the frame). Now these don't include infrastructural embodied energy such as inverters, mounting systems etc, which would decrease the EROEI to about 30 @ 30 years with current techniques. But neither do the EROEIs for traditional fuels contain the externalities of generation plant embodied energy. Other energy sources are: Coal: 9 EROEI Oil (middle east): 10-30 EROEI Oil (US): 3 EROEI Light water Nuclear: 4 EROEI current (12 with improvements)
Ethanol: Likely zero EROEI, maybe negative
Now you might quibble with some of the numbers, however I think it shows that PV is at worst good, and at best really good.
I think we've shown it doesn't matter if we use 10%, 20%, 30% or 40% efficiency. The land consumption is not a factor. First, its small, and second we can synergistically utilize other surfaces with no or little other space needed. Even at 17% efficiency panels, the US could generate near all its ENERGY (triple its electricity) with the US rooftop space as we've shown. Since not all of that has solar access, we'll throw in some parking-lots. Heck, we could generate 1/7th of our electricity needs with just the land from the Hanford nuclear superfund site (570 mi^2).
Now take into consideration that the spectrapower cells Entech is using are now up to 37.3% (2004) efficiency, which will increase module efficiency to 33.5% from their 2001 announcement (which is in line with a claims of the VC I spoke with).
So at 30% efficiency (using published value) we need to increase our land base values by 33%. So All US ENERGY Needs from 13,491 Mi^2 or 5% of TEXAS (including shading at an average of 1800 kwh/m^2/year).
Thanks for calling that one, I'll update my database of facts. I haven't been reading the solar journals very closely over the last 4-5 years as the company I am working for is developing storage technologies, so I put most my time that technology and market trends therein (which we will get to).
I accidentally switched the values for Transportation and Electricity in the post.
Transportation = 35.30 Quads
Electricity = 25.65 Quads
Though the total was posted right 50.3 Quads.
Trackers are a red herring
It is true that trackers are a wash. You get more energy, but you shadow more ground (however, only in flat 2D arrays). Of course, I didn't use them in my calcs so it doesn't play in this discussion.
These parts, whilst efficient, take an energy penalty of their own
100 times more costly in terms of real estate than our current scheme!
WRONG. The US uses 24% of the worlds energy (97.7 Quads US, 404 Quads World). So if the US uses 3.5% of Texas, the world would uses 14.6% of Texas (100,800 km^2) or LESS THAN HALF the current infrastructure. Not that it matters since we've shown it can be done with no new space utilization.
20-60 W/m^2 of energy for solar - which is the reference figure that I've seen everywhere except for you.
Everywhere? Not NREL, not DOE, not NASA, not the EU PV program, not PV manufacturers. Where, Smil? I think at this point his credibility is crumbling. I've given you many primary source material references. Read them! Learn how to use the data sets. Make sure you know what you are using before calculating. Is it ground plane, flat plate at angle, concentrator, tracking, plane normal radiation, diffuse radiation, gobal radiation, average, worst case, monthly, or amortized?
The only way 20-60W/m^2 makes any sense is if it is amortizing the light hours over the 24 hour day and multipled by 50% efficiency. If that is the case you can't multiply by 40% efficiency agian and then multiply by only 6 hours of insolation per day! AGAIN look at the data sets, read the instructions, and see for yourself. Or check out a PV design handbook.
Yes I know we haven't addressed use, efficiency, storage, and seasonality implications in our land use numbers, and they will rise some. Ready?
Hmmm. I take that post as agreement. Obviously Smil is massaging the truth (and you are starting to look silly defending this guy, who in 7 posts is proved wrong again and again - time to read a broader group of authors?). I'm looking forward to getting past your FUD, and discussing your other questions, but first we must get pass the nonsense.
I avoided anything that was tracking, took all the fixed rates, added them up together, averaged them,and multiplied by 365.
As did I, only i used NRELs annual statistics not your uninformed calculations of their raw data (no offence). Since we seem to go over this again and again I will spell it out for you. Though at this point I think you are just being obstinate. From the rredec database:
Now look at that last column, "Year". Look at the row "Lat Average". 4.9kWh/m^2/day. Got it? If not, want to see a map of the same data? I don't know what you want, whack you over the head with a dozen sources? Here, here, here
So 7% of texas land mass to produce ALL of our energy use, only 0.8% for our electricity need. Using fixed panels (not even adjusting the angle seasonally), including shading. I didn't use the best location, but an average location. This doesn't translate by any stretch of the imagination into all of Texas. More importantly I showed we don't to use any new space at all.
You also are naively using 3 * 10 ^ 13 kW as our total energy source that would need to be replaced - the difference is that 90+% of that energy is in a form that we can directly use - natural gas for heating and gasoline for burning, coal for making steel, etc.
WRONG! JUST THE OPPOSITE. We've compared solar for PRODUCTION EFFICIENCY OUTPUT to US GROSS ENERGY CONSUMPTION (my mistake really). So if you want to do a REAL comparison, we need to calculate a conversion efficiency of current energy sources based on end-use (for oil, coal, Gas).
Transportation = 25.65 Quads @ 20% ave conv efficiency
Heat = 23.09 Quads @ 90% ave conv efficiency
Electricity = 35.30 Quads @ 33% ave conv efficiency
Nuclear+Renewable Electricity = 13.99 Quads @ "100%" efficiency (numbers ARE net)
-----------
50.3 Quads Net energy produced. NOW WE ARE AT ONLY 3.5% OF TEXAS.
Ready to discuss storage, transmission, grids, seasonality, etc Yet? I think you've lost this part of the argument.
I forgot to respond to the 40% concentrators question. Mature Si Technology (currently the market leader) is generating 17-18% panel efficiency based on 21+% efficiency cells. This alone is far more efficiency than required to provide the US electricity needs OR energy needs without using any new land (by using roofs, parking lots, electric distribution land, train track sidings [great for wind breaks in the mid-west], highway midians or any of hundreds of unused or synergistic lands).
Since nulcear proponents are keen of saying "but nuclear stuff would be great if only we used this as yet uncommericalized techonology, that we could build the infrastructure in ONLY 5-15 YEARS...", I thought it would be fair to use PV technology that will be on the market in 3 MONTHS!
Now the cells used in these concentrators are nothing new, they've been used in space applications for alomost 2 decades and are well tested. They have become a bit more efficient in recent years. With concentrators (@500 suns) they are even more efficient, 39% real-world efficiency for PRODUCTION cells as of 2001 (TECSTAR and Spectrolab). 42% Eff are on in the works (ie. it hasn't been maxed out yet). But they are expensive, much more so than other PV technologies. So arrives the concentrator idea. Use cheap tubes of plastic fesnel lenses to focus onto a line of PV cells. At 500 suns, 500 times less cell area. Voila, very efficient super cheap panels!
Sharp will begin selling similar product in 2005 (three months), with a total panel efficiency of 36.5% at a production cost of less than $1/Wp. Entech also has a similar technology that will retail for around $2/Wp next year. I talked with a VC recently, who owns a large share in an energy company that has contracted Entech's complete production capacity for building for-profit solar electric plants in the western states.
Even still, its just one competitor in the market. As the land calcs have shown, land is not the issue, price is. It is widely held in the industry that anything above 8-10% efficient is land efficient enough to be market viable. So if I can produce solar cells for $0.50/Wpeak @ 10% I've got a winner. This is the stratigy of thin-film PV manufacturers (aSi, ribbon-Si, CIS, dye-TiO, CdTe, organic cells, etc). For instance UniSolar makes a amorphous Si solar panel that looks/works just like a roll of 3-tab asphalt shingles that you nail down just like normal (stuff is literally bulletproof, nail right through it). Only 10% efficient, but who cares? A roof-full would still produce 3-4 times the homes electricity needs, and you don't have to pay for a separate roofing, lowering your overhead! Konarka is making flexable polymer solar cells at 8% efficiency but a simple dirt cheap polymer manufacturing process. They are in early production, but the technology has a 30% efficiency theorectical limit, looking at the history of Si cells, they'll likely acheive 15-20% in the next 5-10 years. Other technologies are more mature: Chalcogenide (18%), CIS (13% efficient), Ribbon Si (12-14%), cast Si (12-14%), Si microsphere (12%), etc.
So 40% concentrators are looking really competitive, but there are market niches for every technology. Any references I've already given, else google them.
Link works fine from here. Try:
http://rredc.nrel.gov/solar/old_data/nsrdb/redbook/sum2
You've realized that my numbers are in fact right, yes?
-Double dipping-
These are quick numbers to show you the ballpark of land-consumption, we assumed a 2-D array, we didn't count shadowing NOR did we count that the panels are angled around 39 degrees thus taking only 77% of the land space we counted.
1. Even if you used fixed panels in one giant array at an angle equal to latitude, ~ 39 deg angle the land used becomes only 28% greater than the panel area. For an average insolation location like Kansas City having 1800 kWh/m^2/day x 40% efficiency = 720 kWh/m^2/year. 2.88E13 kWh Total US energy/720 = 15444 mi^2 panel area x 128% to account for shadowing = 19768 mi^2 OR a 140 mile square. Now we are up to a WHOPPING 7% OF THE TEXAS LAND MASS with fixed no moving parts panels. WooHoo. Trackers would of course do much better.
2. This is a visualization exercise. In fact, all the US power would not be generated in one place. Shadowing is only a problem in 2D arrays, however we can arrange them however we want. 2D arrays on buildings do not have shadowing due to the slope of the roof. 1D arrays don't have this problem. Though not possible everywhere, they are in some places such as the 1500 mi^2 of idle land sitting below large electric transmission lines, or highway midians, etc. You see the great thing about solar is we literally can use almost any under utilized space for power production (roofs, parking lots, superfund sites, etc).
-An Example-
Near my house there is a typical suburban shopping plaza. Contains 5 big box stores (homedepot, target, walmart, etc) all 150-200,000 ft^2 plus a dozen 50,000 ft^2 stores. That's 1.6 million ft^2. With 40% efficient fixed panels at an average of 2000 kWh/m^2/year insolation that's enough energy for 30,000 homes! Put panels over the parking lot (as is being done in California - dual purpose shade and energy) which is 150% as big, and together we have enough energy to power 72,000 homes FROM ONE SHOPPING PLAZA with no extra land used and no trackers.
-Alabama-
First use the REAL 30 year average numbers not a cherry picked number (how did you come up with that? Your number is far less than the worst year out of 30 years for a panel FLAT on the ground: 1553 kWh/m^2/year). The average annual insolation on a flat panel in Birmingham is 1606 kWh/m^2/year at 0 degrees (flat on the ground), 1788 for a panel at latitude (this case 33 degrees), 2263 kWh/m^2/year at latitude with 1 axis tracker, and 2336 kWh/m^2/year for a 2-axis tracker.
I'm still interested in answering your other questions, but I want to make sure you really understand that land use for solar is small (at most any efficiency 10-40% with or without trackers), smaller than many other energy resources currently in use (e.g. coal), AND in fact we don't need to use new land at all (which is pretty much unique to solar. And perhaps arguably wind, wave, and geothermal). If you understand that point now, and are done nitpicking technical details of infinitesimal returns, I'm happy to move on.
Sorry if I don't believe you. First, 964kWh/m^2/year means that you are extracting 110 W/m^2, when in Kansas the total amount there is about 140 W/m^2 - which means you are getting 80% efficiency.
Before we can move on the rest of your argument, you need to use REAL numbers for insolation. I've given you the links to the definitive government resources, but you keep on using the same bogus numbers. Come on, you are a smart guy!
Dude, read the references I gave you before giving uninformed responses. You can't make an argument by pulling your own numbers out of the air.
Size
* Wisconsin has a average solar gain of 4.5 kWh/m^2/day on a fixed flat solar panel at an angle equal to latitude based on 30 year statistics.
* Typical PV panels has 17% whole module efficiency (Sharp, BP, etc), so each the solar panel generates.765 kWh/m^2/day
* Average house has 2000 ft^2 of roof area (185.8 M^2). Thus average house roof in Wisconsin generates 142 kWh/day.
* The average US household consumption is 28 kWh/day OR 1/5 the roof area! And 30% more if you use trackers. Twice that with multijuction Concentrators.
Cost
Current PV is being produced for $1/Wp. But due to huge international demand, and lack of maturity in the market retail price is about $3.50/Wp. Even with current prices the $/kWh over the 30 year warranty and load life is $.07/kWh in wisconsin ($3.50/4.5 x 365 x 30). You are getting mixed up between capital costs and electricity cost. Further, as volumes increase the market will mature like all commodities products to be around a 30% profit margin, or $1.30/Wp. And that without counting new technologies which are already headed out the door.
Snow
Wisconsin gets more snow than Snowmass or Aspen where there are TONS of PV systems. WRONG. Many systems are very remote with no maintenance at 10,000 feet such as the Tenth Mountain ski huts. Snow not a problem.
Land mass
Do the calculation yourself. To Replace ALL US electricity a 46 mile square. 0.8% of the land in Texas.
Storage
The solar cycle follows the electricity peak use cycle. Connected to the grid, solar could replace ~40% of our power without substantially needing any storage. Beyond that there are many ways to store energy (Hydrogen, hydro, air flywheel, superconductors, capacitors, vanadium flux batteries, etc).
byproducts
PV is not the IC industry. Many PV technologies (CIS, aSi, Ribbon Si, Dye-TiO) have essentially no pollutants.
1) residential use is 3% per month, meaning 36% annual.
As noted
2) the mistake that you made in overestimating square footage is far more endemic than I think you realize. For the *apartments* are also counted double (or triple, or quadruple) by just taking the average and multiplying it by the number of units.
Good thinking, but not the case. Apartment buildings above 2 stories (5 or more units) only make up 6.5% of total units. My numbers if anything underestimate the sqft, because I didn't factor the larger single family housing sqft which averages 2527 sqft and make up 88% of units. So just considering single family units we have 94.2 million units with 75% floor sqft to roof (50% single story, 25% double story w/single story garage). With 6% slope addition = 1.9E11 or 8% MORE while ignoring all apartment buildings. Adding the apartment buildings back in: 2-4 unit buildings (6% of total units) we'll say they are all 2 story (1393 sqft * 6.4E6 units *.5) and we get another 4.5E9 sqft. 5 and up units (6.5% of total) lets be generous and say 20% sqft->roof (847 sqft * 6.95E6 units *.2) and we get another 1.2E9 sqft. So TOTAL Residential SQFT = 1.95E11 SQFT (sure enough apartments don't add much). Add the commercial space and we get 2.6E11 sqft - a bit better than previously.
Even the EIA doesn't see solar power rising any time soon - out to 2025, solar is.1-.5%....
Of course they are not too sanguine about nuclear either, but then again I think that they are massively overestimating how much oil is left in the ground..
True enough. The EIA tracks very accurate numbers for what is, and what has been, but they bad with the future. Which works OK for consumption models (sort of, they have been horribly wrong on that to - in the 70s they assumed exponential growth, when in fact efficiency made up for it), but they do not consider technological changes, cost reductions, geopolitics, etc (for instance a decade ago their wind power forecasts were FAR too low). They are primarily a current energy information outlet for congress.
Dude, get a grip. You are so busy trying to prove a point you read from this one book (which is either wrong, or trying to prove a point by comparing some future, yet as unbuilt unproven nuclear reactor with 1970s solar technology), that it continues to fog your argument. My numbers are accurate, knowledgeable of the field (as it is my field), and well thought out. Every time I prove my point, you try to come up with something else, which I again prove wrong. Please, thoroughly read some books and periodicals on renewable energy, energy systems, and energy policy. Hopefully this exercise has been useful and you are learning something (I commend you for taking the time)
* Roofspace is not as you stated. From the census data and DOE data: The average housing unit size is 2066 sq ft. There are 107 million units. 50% of houses are 1 story (roof=sq footage). The other 50% are 2 story or more (Census), which I estimated roof space is half of living space (this averages in the added garage roof space of some with the loss of roof space to 3 or more levels). The result was increased by 6% for the added area of the average roof slope. Commercial was 67 billion sqft. If you want to do a more detailed analysis, It would be great, send it to me. The average single family unit is 2527 sqft which I didn't use in these calcs, and are 88% of total housing units, so these numbers are likely underestimating roof space by around 15% or so. However the outcome will not likely be more than +/-10%.
* Of course not all roofs will be usable. The point is to get perspective on the land area needed. Even if 1/3 of roofs are usable, then problem solved. If you don't use roofs, the land area required is still VERY small. THE LAND AREA IS NOT THE SIZE OF TEXAS! With 17% panels on trackers the land area is a 46 mile square - 22% smaller than Dugway Proving grounds OR 0.8% THE SIZE OF TEXAS. With multijuction concentrators, its less than half of that. The problem here is you've been programmed to believe it should take a huge amount of space, BUT IT JUST DOESN'T. Clear yet?
*OK Say you want to replace ALL the US energy with solar(oil, coal, Natural Gas, wood, etc). How much land would it take? The US uses 98.3 Quads a year, or 2.88E13 kWh. Using 40% efficient multijunction concentrators ($1/Wp!) on trackers in average location (Kansas City) you get 964 kWh/m^2/year. LAND REQUIRED: a 100 mile square. OR 4% the size of Texas! VERY SMALL! 1/10th the 290,000 km^2 number you cite (Reference for this number please).
*Obviously once you see the real numbers - infrastructure isn't a problem. In fact a distributed PV system, some on roofs, some in local grids, some in large arrays would reduce distribution and transmission infrastructure substantially
Perhaps next year you might want to tour the variety of solar homes in wisconsin and see for yourself that they really do work and talk to owners who have been living in them for years. Friday, September 30th and Saturday, October 1st 2005.
If he means by this average usable power per day including darkness, then hes not far off. (1000 W/m^2 X 6 hours of peak sun)/24 hours per day = 250W/m^2 per 24 hours. BUT this is misleading, watts are peak POWER measurements not ENERGY. The energy stays the same in either case: 1000 W/m^2 X 6 hours = 6000Wh OR 250W/m^2 X 24 hours = 6000Wh. See?
In reality, the solar flux density (same as power density) varies between 250 and 2500 kilowatt hours per meter squared per year (kWh/m^2/year).
Tucson = 2372 (3285)
San Francisco = 1971 (2591)
Kansas City = 1788 (2409)
Seattle = 1350 (1788)
Denver = 1825 (2701)
Columbus = 1533 (1971)
Boston = 1679 (2153)
Buffalo = 1496 (1934)
Anchorage = 1095 (1460)
To get the annual energy produced per m^2 multiply by the PV panel efficiency. For example, at 17% efficiency in Denver 310 kWh/m^2/year on a fixed panel, 459 kWh/m^2/year on a tracking panel. Multiply this by the number of square meters of roof in the US, and you get 6.99E12 kWh/year for Denver fixed panels, and 1E13 kWh/year for Denver tracker (not all places are the same as Denver, but they aren't all that different either - its an example;). See? Here we get 200%-300% more power than needed, like I said.
I'm sorry my friend but you sound like you have a SERIOUS agenda. (silicon with solar having a greater energy production per pound than nuclear fuel? Yeah right.
The quote (which is true, Lovins is very accurate with his numbers) is not meant to be more than it is: an interesting comparison between solar and the current and real state of nuclear power in the US (light water reactors are capable of using only a fraction of their fuel before they are spent). Of course breeders or other designs could produce 100-1000 times as much energy per pound, but they have their tradeoffs too (which is why we chose not to use them).
And of course, that doesn't even touch the fact that the major problem that we are facing is not going to be electricity shortage, but energy carrier shortage.
Exactly. Solar power is available everywhere (did you notice the deviation between the alaska and arizona is only 2-1?). Solar IS the ultamate distributed power source. If most of the power is generated locally, they carrier requirement of transmission is HUGELY reduced, and overall costs come way down.
Tell you what - would you agree to having the government both subsidize the development of solar technologies AND next generation nuclear ones, and see which one wins?
Please learn about the subject before you respond. Energy is my area of expertise, am I'm always appalled by how engineers and geeks can tell you the latest in computer technology to the day, but are 30 years out of date (or just completely misinformed) when it comes to renewable energy.
averages 170 W/m^2 when it reaches the ground.
Yikes! Here the first problem with your calculations! Solar insolation is 1300 W/m^2 outside the atmosphere, 1000 W/m^2 on the ground in peak sun conditions. NOT 170! (look it up yourself you'll find tens of thousands of refs on Google)
expect to use intermittently depending on weather and time of year.
insolation FOR A FIXED panel at an angle equal to latitude provides an average of 6 hours of peak sun per day in the average US location. (of course the solar insolation is changing based on time of day. However this is how it is specified in the industry: pre-integrated to an equal number of peak hours). That equals 2190 kWh/m^2/year. Some locations a little more, some a little less. With trackers this goes up 25-50%. See the National Renewable energy laboratory insolation database and mapservers for more data.
inefficiency of incorrect angles in capturing the energy
Already considered see above numbers are already based on tilted fixed panels. Trackers of course improve the angle and thus the energy, but I'm giving a simple case, not best case.
Spacing is accounted for, 17% is total edge to edge module efficiency not cell efficiency. Maintenance costs, essentially are none (solid state revolution man) no moving parts, no dusting, no snow removal required (the benefits of dusting/cleaning has been proven to be of small benefit. less than 4%). Storage is an issue. There are many storage technologies and they do cost money (some solar technologies, not PV, are self storing such as Solar 2's phase change salt storage). However, energy profile on the grid tracks the solar cycle closely. 40%-60% of our energy could be replaced without substantial storage added to the system. (another 20-30% could come from wind, as the Dutch have shown, and the base load could be largely provided with geothermal, biomass, and wave. Thought I do think storage is an important piece of the puzzle.)
I don't know where you got your 'roof space' figure (2.43e11) but it seems high
The number you show is ENERGY consumption NOT ELECTRICITY consumption, and its a little too high (I guess the are spooks not energy experts). From the Department of Energy, Energy Information Administration total energy consumption is 2.88E13 kWh. The total US ELECRICITY consumption is 3.4E12 kWh - which is what we are talking about.
Don't get me wrong, I really *want* to believe that solar is our best bet.
Today is your lucky day. The numbers are very much right (as you can now see). And we didn't have to even invoke any extra land consumption OR higher efficiency cells OR Dye-sensitized solar cells which can be used as windows on high rise buildings, etc. PV is amazing stuff with incredible potential, 40% annual market growth, prices are nearing $1/peak watt (33
Its not my opinion or a model, it's 30 years of solar insolation statistics gathered by the Department Of Energy (please look at the link I gave before responding next time).
Snow isn't a problem, as I've already noted elsewhere. In fact its a bessing. In winter months PV panels are angled steeply in northern latitudes. Snow slides right off the panels. The snow increases the energy production because it effectivly reflects the sun from the ground/roof onto the panels. Winter can be the best solar producing month in many locales (like the rocky mountains which get a whole lot more than 3 feet of snow).
Solar still makes energy even on cloudy days. Solar energy is remarkably constant throughout most of the world/US.
For a flat panel, the deviation from the best southern Nevada site to the worst northern Washington state site is only 2-to-1! The rest of the country is surprisingly small deviation within this range. See rredc.nrel.gov/solar/ [nrel.gov]
Wisconsin gets an average daily insolation of 4-5 kWh/m^2 verses 6.5 kWh/m^2 for best locale in Arizona for a fixed panel. So Wisconsin is still 70% of the best solar location. not too much difference.
Slashdot Mirror
Here are some links on how the grid and customer transformers work.
You can use your existing power transformer so long as you don't exceed its power rating. Here is one utilities regulations.
All grid-tie inverters on the market autosense the power lines going down or short-circuit to pass the NEC. This is called anti-islanding. Nonsense, nothing new required. Hook PV panels on one side of a standard meter, and it spins backwards feeding power onto the grid. PV owners do this all the time even without the utilities co's knowledge. The whole point of distributed generation is a highly redundant system of local power sources produces a more reliable power grid. Old news, not theory.Look. I'm not interested in a argument for its sake. Every time you pull some new reason out of the air that solar can't work I've proved you wrong and uninformed. If you want to learn more, or have informed opinions you want to share, then fine - but I am starting to feel this is going nowhere. (have you learned anything about solar in this process?)
PV is no different in this regard from any other power source.
Like I said this is being done all the time, there are over a 200,000 small residential systems (ave. 10kW peak) grid tied in the world right now, and working fine. (40,000 new grid tied systems in just 2003). The growth in the grid tied market is 60% per year, meaning another 64,000 systems went up in 2004.
The big difference is PV is distributed . This means a far more efficient, redundant, and secure grid. But it also mean less grid stress, because more power is generated locally. For example, normally my power comes from my roof (distance 20 ft). Sometimes some of my power comes from my neighbor (dis. 500 ft). Occasionally some my power comes from the shopping mall (distance 3 miles). When the insolation is low some of my power is imported from 2 states away (600 miles). Say the weighted average distance my power travels is 1 mile (down the same wires it would have before). Now compare that to the centralized infrastructure we currently use which 90% of the time its traveling 600 miles! Transmission efficiency is improved and grid utilization is reduced.
It is, in fact, more efficient as a two way street. This is very foundation of concept of distributed generation which has been successful at reducing grid stress and $ for a couple decades already (mostly NG turbines), and which PV is a good example of. Centralized power is everything that is wrong with the grid today. If you want to learn about DG read: here, here, here, here, or here. Huh? Just because you want it to? So you can support your argument? Not only is there no evidence for this, but it defies all the fundamental tenets of mass production and the benefits of scaled industries. If anything EROEI will rise. (because of improvements in technology, manufacturing process, installation efficiency, density of systems will reduce maintenance costs, etc, etc).We've already shown it to have a better EROEI than fossil fuels, even when favoring the fossil fuels with less stringent EROEI calculations (i.e. not counting embodied energy of equipment). If PV has a EROEI of 15 (minimum), and since the fuel/sun is a free resource, its energy output can replicate itself by x^15. Hardly a problem
I didn't say Odum's notions of transformities is useless. Its quite interesting from a perspective of environmental capital. But it disregards time and space. It took a lot of solar and geological work to produce oil or coal thus giving it more "eMergy". While there is a lot of environmental capital invested in oil and coal (by the earth and sun), that isn't very useful measure of energy viability, and has nothing to do with EROI. Outside the study of biosystems natural ecological capital, it's really not useful at all.
If for example, it takes 10 sun units to make 1 coal unit over 100 million years, that doesn't imply coal has any real measure of "betterness". Cleveland's point is the important questions from a energy policy/economic perspective are:
What is the economic ROI?
What is the energy ROI?
What is the environmental impact?
What is the quality of energy source? (how useable is it)
What is its realibility?
etc.
People are "still bothering" with oil because it is a reasonable storage medium, not nessesarily a energy source (similar to hydrogen). More importantly, people are using oil because there still is cheap middle eastern oil (which has an OK EROI, 3 times US oil).
Cutler Cleveland is his review of Net Energy Analysis Methods (A very good overview) says this:
"It is important to differentiate between two aspects of Odum's contribution. The first is his development of a biophysically-based, systems-oriented model of the relationship between society and the environment. Here Odum's (1971; Odum and Odum, 1976) early contributions helped lay the foundation for the biophysical analysis of energy and material flows, an area of research that forms part of the intellectual backbone of ecological economics...
The second aspect of Odum's work, which we are concerned with here, is a specific empirical issue: the identification, measurement, and aggregation of energy inputs to the economy. Emergy (with an "m") analysis is a pure cost-of-production approach that measures the quality of a particular type of energy by its transformity. Transformity is the amount of one type of energy required to produce a heat equivalent of another type of energy. To account for the difference in quality of thermal equivalents among different energies, all energy costs are measured in solar emjoules (SEJ), the quantity of solar energy used to produce another type of energy. Fuels with higher transformities require larger amounts of sunlight for their production and therefore are more economically useful (Odum, 1988)...
This approach raises a fundamental question about the appropriateness of transformities to reflect energy quality: Is the usefulness of a fuel as an input to production related to its transformity? Probably not...Thus, while Odum's method provides a useful framework for highlighting he important role the environment plays in generating energy and material resources, it is of dubious value in comparing and aggregating energy flows in economic applications...
In addition to this conceptual issue, there are computational problems with emergy analysis that make transformities incomplete indicators of energy quality. The calculation and application of transformities are time, location, and technology specific, yet Odum and his colleagues mix the temporal, spatial, and technical scales of their analysis in ways that are poorly defined. First, Odum presents the transformities as constants, but based on the method used to calculate them (Odum and Odum, 1983), the transformities are clearly dynamic because they are based on the first law efficiency of technologies such as power plants, coal liquefaction, and oil refineries. The efficiency of those technologies have changed dramatically over time. Second, the emergy calculations also contain an ad hoc mixture of spatial scales. The basis for the calculation of the transformities is the thermal efficiency of a wood-fired power plant in Brazil, but the efficiency of power plants vary throughout the world (Smil,1991) as do all the other energy conversion technologies used in the emergy calculations. Similarly, energy/output data from the New Zealand economy are mixed with the Brazil power plant data to calculate the transformities, which are then applied to many other economies throughout the world (Odum and Odum, 1983; Odum et al., 1987; Odum and Arding, 1990; Huang and Odum, 1991). Third, the values of the transformities are highly sensitive to technological assumptions made by Odum and Odum (1983). They calculate the relative quality of oil, gas, and coal based in part on the fact that the first law thermal efficiency of converting natural gas in boilers is 20 percent more efficient than the conversion of coal. However, the relative thermal ef
Cutler Cleveland (Director of the Center for Energy and Environmental Studies at Boston University) appears to be one of the leading energy analysts these days, his work is quite broad. In "Net Energy from the extraction of oil and gas in the united states", Int. Journal of Energy 30 (2005), he shows that US Oil production has a EROI of 11 for energy in/thermal energy out. And gasoline is 30-50% of this value (ie 3.3-4.5 EROI). Now that doesn't include conversion efficiency in a car or power turbine, nor does it include the embodied energy of the extraction equipment or ICE/power plant to burn it.
If you make calculations just for conversion efficiency (33% ave) of US oil converted to electricity/mechanical power has EROI of 3.6, and gasoline in a car is less than 1 (meaning that the energy to do mechanical work in a car is being subsidized by by electricity (coal) to run the extraction equipment). And still we haven't considered the embodied energy in the extraction equipment or the ICE. (now of course middle eastern oil is 3 times better than this) That is very poor EROI! And coal isn't looking much better. Both of these resources EROIs have dropped by at least a order of magnitude over the last 100 years as extraction becomes more difficult. The future of fossil fuels by EROI analysis looks bad.
As for Alsema, he does review the added embodied energy of infrastructural components in section 4.5 of the paper. For complete balance of systems analysis (inc. frames, structures, concrete, maintenance, etc) the best technologies (thin films and ribbon Si) have EPBP of 1.2-2 (15-25 EROI @ 30y, 25-41 EROI @ 50y). scSi is around 3.3 years (9 EROI @ 30y, 15 EROI @ 50y). Analysis shows that PV energy is manufacture side heavy, with little continuing energy inputs as would be expected from a solid state, fixed, and essentially maintenance free device (how often do you maintain your current roof shingles?). However, even his latest numbers are out of date as he notes getting information from manufacturers is difficult because EROI calcs involve knowing trade and financial secretes so it takes a long time to get agreements in place. Also his calcs don't use the best efficiency panels on the market, which underestimates EROI, if that was the criteria on which we made purchases. Also multijunction concentrators, should be significantly better since 1) they use less material per peak watt and 2) they have a higher efficiency. References: Here,Here, Here, Here
What's the end result? EROI calculations beyond first or second order become quite tricky and controversial. But we can show that solar in a detailed "second order" or more EROI estimate looks very favorable compared to even a "first order" estimate of oil, NG, or coal.
You can squint your eyes, stomp your feet, and don that helmet believing whatever you want, but reality is waiting when you want to join us.
Alsema (Professor in the Dept. of science technology and society at Utrecht University, Naterlands) has 13 peer reviewed papers published on PV all relating in some way to EROEI or environmental impact. In his most recent publication (2004) in Refocus he says:
Read Alsemas numbers, he does too. He breaks it down for you though. Thin films have a EPBP of ~.5 (~60 EROEI @30 years) for frameless panels such as these roof shingles. With poles, mounts, concrete, yada, yada the EPBP is 1-2 years (or a 15-30 year EROEI)."Recent studies give the impression of photovoltaics having considerable environmental impact. Looking closer at the data however, it is clear that these studies are based on photovoltaic systems of the late eighties, with only minor recalculations. Since the photovoltaic market has increased rapidly, a lot of progress has been made regarding the environmental profile of photovoltaics." He goes on to show, for example how current production ribbon silicon panels have a payback period of 1.2 years (and that is silicon, not even thin flims like CIS).
If you don't like Alsema cause he blows your outdated arguement read Kato. Kato (prof. Japan agriculture university) has 25 publication all dealing with photovoltaics. He has several publications dealing with EROEI.
Or read this (note this is dated and the 2 and 1 modules are already on the market).
I've shown you a detailed peer-reviewed government comissioned paper showing 15-30+ MINIMUM EROEI(30 year). A EROEI of 15 (25 @ 50 years) assumes ground mounted with steel structure and concrete footings (but recycled Al or polymer composite mounts would bring it to 30). Building mounts using frameless roof integration take you up to 50+ @ 30 years (83 @ 50 years).
If you really want to understand this issue, read the journal article thoroughly, read his references, read his other studies, read "Energy pay-back time and life-cycle CO2 emission of residential PV power system with silicon PV module", Progress in Photovoltaics v6 1998 by K. Kato. And "Energy pay-back time and CO2 emissions of PV systems" Progress in Photovoltaics v8 2000 by K. Alsema.
That gives it a EROEI of 60 over 30 years.
Very good
Alsema is a leading expert on this field on study (you'll see his name on many in-depth studies) and shows the energy pay back period to be:
Multicrystal Si: 0.8 years (EROEI 37.5 @30y, 62.5 @yr)
CIS: 0.4 years (EROEI 75 @30y, 125 @50y)
CdTe: 0.6 years (EROEI 50 @ 30y, 83 @ 50y)
Crystal Si: 3.3 years (EROEI 9 @30y, 15 @50y)
I should note that these studies are again becoming 5-10 years old again and don't reflect the improvements in efficiency. CIS for example in this study was modeled at 12% efficiency, but the best efficiency (2003) is now 19.2%. Same with crystal Si, efficiency have edged up about 3-4%. I wasn't able to find info on concentrators, but because of low materials-to-power ratio I expect they will be at least as good.
I calculated the numbers for both 30 years and 50 years. Comparisons are usually done on a 30 year basis since this is the build life of other power plants and is a typical load period. However it is a bogus, made up number for easy comparison. Many PV manufacturers are guarantying their panels for 25 years, with no or little power degradation from the specs. No reason to artificially cut their life short.
The numbers I used were the base case, not best. Of course its possible to package them with material having more embodied energy (as in the case of the silicon numbers which has a significant amount of aluminum in the frame). Now these don't include infrastructural embodied energy such as inverters, mounting systems etc, which would decrease the EROEI to about 30 @ 30 years with current techniques. But neither do the EROEIs for traditional fuels contain the externalities of generation plant embodied energy. Other energy sources are:
Coal: 9 EROEI
Oil (middle east): 10-30 EROEI
Oil (US): 3 EROEI
Light water Nuclear: 4 EROEI current (12 with improvements)
Ethanol: Likely zero EROEI, maybe negative
Now you might quibble with some of the numbers, however I think it shows that PV is at worst good, and at best really good.
The land consumption is not a factor. First, its small, and second we can synergistically utilize other surfaces with no or little other space needed. Even at 17% efficiency panels, the US could generate near all its ENERGY (triple its electricity) with the US rooftop space as we've shown. Since not all of that has solar access, we'll throw in some parking-lots. Heck, we could generate 1/7th of our electricity needs with just the land from the Hanford nuclear superfund site (570 mi^2).
Indeed, I took the cell numbers as functionally equivalent to module efficiency (ie mirrors can be 98+% efficient). But reading the literature, It's clear that cheap is the goal (cheap focusing elements). In fact, the production price for multijunction concentrators being discussed is 12-50 cents/Wp. WOW! $0.12/Wp for 30 years is $0.0015/kWh! (of course this doesn't include BOS, but even with, its amazing)
Commercial Efficiencies:
Entech - 30% net concentrator efficiency, 33% cells (2001)
Sharp - 28% net concentrator efficiency (FYI-uses non imaging optics)
Sharp - 17.4% MODULE efficiency (not cell)
Sunpower - 16.5% MODULE efficiency (21.5% cells)
Now take into consideration that the spectrapower cells Entech is using are now up to 37.3% (2004) efficiency, which will increase module efficiency to 33.5% from their 2001 announcement (which is in line with a claims of the VC I spoke with).
So at 30% efficiency (using published value) we need to increase our land base values by 33%. So All US ENERGY Needs from 13,491 Mi^2 or 5% of TEXAS (including shading at an average of 1800 kwh/m^2/year).
Thanks for calling that one, I'll update my database of facts. I haven't been reading the solar journals very closely over the last 4-5 years as the company I am working for is developing storage technologies, so I put most my time that technology and market trends therein (which we will get to).
Transportation = 35.30 Quads
Electricity = 25.65 Quads
Though the total was posted right 50.3 Quads. It is true that trackers are a wash. You get more energy, but you shadow more ground (however, only in flat 2D arrays). Of course, I didn't use them in my calcs so it doesn't play in this discussion. Not really. Passive trackers use no extra electricity at all, while active trackers use only 3/100th of a % of the power. FYI these are power numbers not energy WRONG. The US uses 24% of the worlds energy (97.7 Quads US, 404 Quads World). So if the US uses 3.5% of Texas, the world would uses 14.6% of Texas (100,800 km^2) or LESS THAN HALF the current infrastructure. Not that it matters since we've shown it can be done with no new space utilization. Everywhere? Not NREL, not DOE, not NASA, not the EU PV program, not PV manufacturers. Where, Smil? I think at this point his credibility is crumbling. I've given you many primary source material references. Read them! Learn how to use the data sets. Make sure you know what you are using before calculating. Is it ground plane, flat plate at angle, concentrator, tracking, plane normal radiation, diffuse radiation, gobal radiation, average, worst case, monthly, or amortized?
The only way 20-60W/m^2 makes any sense is if it is amortizing the light hours over the 24 hour day and multipled by 50% efficiency. If that is the case you can't multiply by 40% efficiency agian and then multiply by only 6 hours of insolation per day! AGAIN look at the data sets, read the instructions, and see for yourself. Or check out a PV design handbook.
Yes I know we haven't addressed use, efficiency, storage, and seasonality implications in our land use numbers, and they will rise some. Ready?
As did I, only i used NRELs annual statistics not your uninformed calculations of their raw data (no offence). Since we seem to go over this again and again I will spell it out for you. Though at this point I think you are just being obstinate. From the rredec database:
"City: ","KANSAS CITY ", "Sep","Oct","Nov","Dec","Year"
"SOLAR RADIATION FOR FLAT-PLATE COLLECTORS FACING SOUTH AT A FIXED-TILT (kWh/m2/day)
"Tilt(deg)"," ","Jan","Feb","Mar","Apr","May","Jun","Jul","Aug"
"Lat ","Average", 3.8, 4.3, 4.8, 5.4, 5.6, 5.8, 6.0, 5.9, 5.4, 5.0, 3.8, 3.3, 4.9
" ","Minimum", 2.7, 3.3, 3.5, 4.2, 4.6, 4.9, 5.2, 4.8, 3.5, 3.8, 2.7, 2.5, 4.5
" ","Maximum", 4.8, 5.4, 5.7, 6.4, 6.4, 6.6, 6.8, 6.6, 6.8, 6.6, 4.9, 4.3, 5.5
Now look at that last column, "Year". Look at the row "Lat Average". 4.9kWh/m^2/day. Got it? If not, want to see a map of the same data? I don't know what you want, whack you over the head with a dozen sources? Here, here, here
So 7% of texas land mass to produce ALL of our energy use, only 0.8% for our electricity need. Using fixed panels (not even adjusting the angle seasonally), including shading. I didn't use the best location, but an average location. This doesn't translate by any stretch of the imagination into all of Texas. More importantly I showed we don't to use any new space at all.
WRONG! JUST THE OPPOSITE. We've compared solar for PRODUCTION EFFICIENCY OUTPUT to US GROSS ENERGY CONSUMPTION (my mistake really). So if you want to do a REAL comparison, we need to calculate a conversion efficiency of current energy sources based on end-use (for oil, coal, Gas).Transportation = 25.65 Quads @ 20% ave conv efficiency
Heat = 23.09 Quads @ 90% ave conv efficiency
Electricity = 35.30 Quads @ 33% ave conv efficiency
Nuclear+Renewable Electricity = 13.99 Quads @ "100%" efficiency (numbers ARE net)
-----------
50.3 Quads Net energy produced. NOW WE ARE AT ONLY 3.5% OF TEXAS.
Ready to discuss storage, transmission, grids, seasonality, etc Yet? I think you've lost this part of the argument.
Since nulcear proponents are keen of saying "but nuclear stuff would be great if only we used this as yet uncommericalized techonology, that we could build the infrastructure in ONLY 5-15 YEARS...", I thought it would be fair to use PV technology that will be on the market in 3 MONTHS!
Now the cells used in these concentrators are nothing new, they've been used in space applications for alomost 2 decades and are well tested. They have become a bit more efficient in recent years. With concentrators (@500 suns) they are even more efficient, 39% real-world efficiency for PRODUCTION cells as of 2001 (TECSTAR and Spectrolab). 42% Eff are on in the works (ie. it hasn't been maxed out yet). But they are expensive, much more so than other PV technologies. So arrives the concentrator idea. Use cheap tubes of plastic fesnel lenses to focus onto a line of PV cells. At 500 suns, 500 times less cell area. Voila, very efficient super cheap panels!
Sharp will begin selling similar product in 2005 (three months), with a total panel efficiency of 36.5% at a production cost of less than $1/Wp. Entech also has a similar technology that will retail for around $2/Wp next year. I talked with a VC recently, who owns a large share in an energy company that has contracted Entech's complete production capacity for building for-profit solar electric plants in the western states.
Even still, its just one competitor in the market. As the land calcs have shown, land is not the issue, price is. It is widely held in the industry that anything above 8-10% efficient is land efficient enough to be market viable. So if I can produce solar cells for $0.50/Wpeak @ 10% I've got a winner. This is the stratigy of thin-film PV manufacturers (aSi, ribbon-Si, CIS, dye-TiO, CdTe, organic cells, etc). For instance UniSolar makes a amorphous Si solar panel that looks/works just like a roll of 3-tab asphalt shingles that you nail down just like normal (stuff is literally bulletproof, nail right through it). Only 10% efficient, but who cares? A roof-full would still produce 3-4 times the homes electricity needs, and you don't have to pay for a separate roofing, lowering your overhead! Konarka is making flexable polymer solar cells at 8% efficiency but a simple dirt cheap polymer manufacturing process. They are in early production, but the technology has a 30% efficiency theorectical limit, looking at the history of Si cells, they'll likely acheive 15-20% in the next 5-10 years. Other technologies are more mature: Chalcogenide (18%), CIS (13% efficient), Ribbon Si (12-14%), cast Si (12-14%), Si microsphere (12%), etc.
So 40% concentrators are looking really competitive, but there are market niches for every technology. Any references I've already given, else google them.
You've realized that my numbers are in fact right, yes?
-Double dipping-
These are quick numbers to show you the ballpark of land-consumption, we assumed a 2-D array, we didn't count shadowing NOR did we count that the panels are angled around 39 degrees thus taking only 77% of the land space we counted.
1. Even if you used fixed panels in one giant array at an angle equal to latitude, ~ 39 deg angle the land used becomes only 28% greater than the panel area. For an average insolation location like Kansas City having 1800 kWh/m^2/day x 40% efficiency = 720 kWh/m^2/year. 2.88E13 kWh Total US energy/720 = 15444 mi^2 panel area x 128% to account for shadowing = 19768 mi^2 OR a 140 mile square. Now we are up to a WHOPPING 7% OF THE TEXAS LAND MASS with fixed no moving parts panels. WooHoo. Trackers would of course do much better.
2. This is a visualization exercise. In fact, all the US power would not be generated in one place. Shadowing is only a problem in 2D arrays, however we can arrange them however we want. 2D arrays on buildings do not have shadowing due to the slope of the roof. 1D arrays don't have this problem. Though not possible everywhere, they are in some places such as the 1500 mi^2 of idle land sitting below large electric transmission lines, or highway midians, etc. You see the great thing about solar is we literally can use almost any under utilized space for power production (roofs, parking lots, superfund sites, etc).
-An Example- Near my house there is a typical suburban shopping plaza. Contains 5 big box stores (homedepot, target, walmart, etc) all 150-200,000 ft^2 plus a dozen 50,000 ft^2 stores. That's 1.6 million ft^2. With 40% efficient fixed panels at an average of 2000 kWh/m^2/year insolation that's enough energy for 30,000 homes! Put panels over the parking lot (as is being done in California - dual purpose shade and energy) which is 150% as big, and together we have enough energy to power 72,000 homes FROM ONE SHOPPING PLAZA with no extra land used and no trackers.
-Alabama-
First use the REAL 30 year average numbers not a cherry picked number (how did you come up with that? Your number is far less than the worst year out of 30 years for a panel FLAT on the ground: 1553 kWh/m^2/year). The average annual insolation on a flat panel in Birmingham is 1606 kWh/m^2/year at 0 degrees (flat on the ground), 1788 for a panel at latitude (this case 33 degrees), 2263 kWh/m^2/year at latitude with 1 axis tracker, and 2336 kWh/m^2/year for a 2-axis tracker.
I'm still interested in answering your other questions, but I want to make sure you really understand that land use for solar is small (at most any efficiency 10-40% with or without trackers), smaller than many other energy resources currently in use (e.g. coal), AND in fact we don't need to use new land at all (which is pretty much unique to solar. And perhaps arguably wind, wave, and geothermal). If you understand that point now, and are done nitpicking technical details of infinitesimal returns, I'm happy to move on.
Before we can move on the rest of your argument, you need to use REAL numbers for insolation. I've given you the links to the definitive government resources, but you keep on using the same bogus numbers. Come on, you are a smart guy!
So here it is:
Kansas City = 6.6 kWh/m^2/day average for a 2-axis tracker.
x 365 days/year
= 2409 kWh/m^2/year
x 40% concentrator module efficiency @ 500 suns (NREL, Entech, Sharp)
= 963 kWh/m^2/year. GOT IT?
US Energy consumption = 2.88E13 kWh/year
/ 963 kWh/m^2/year
= 3E10 m^2 OR 11544 mi^2 (a 107 mile square)
/ Texas 267,277 square miles .043 OR 4.3%
of the Texas land area.
---------
=
Which part of this isn't clear?
Even if it were a third that efficiency, the land area is minuscule.
Size .765 kWh/m^2/day
* Wisconsin has a average solar gain of 4.5 kWh/m^2/day on a fixed flat solar panel at an angle equal to latitude based on 30 year statistics.
* Typical PV panels has 17% whole module efficiency (Sharp, BP, etc), so each the solar panel generates
* Average house has 2000 ft^2 of roof area (185.8 M^2). Thus average house roof in Wisconsin generates 142 kWh/day.
* The average US household consumption is 28 kWh/day OR 1/5 the roof area! And 30% more if you use trackers. Twice that with multijuction Concentrators.
Cost
Current PV is being produced for $1/Wp. But due to huge international demand, and lack of maturity in the market retail price is about $3.50/Wp. Even with current prices the $/kWh over the 30 year warranty and load life is $.07/kWh in wisconsin ($3.50/4.5 x 365 x 30). You are getting mixed up between capital costs and electricity cost. Further, as volumes increase the market will mature like all commodities products to be around a 30% profit margin, or $1.30/Wp. And that without counting new technologies which are already headed out the door.
Snow
Wisconsin gets more snow than Snowmass or Aspen where there are TONS of PV systems. WRONG. Many systems are very remote with no maintenance at 10,000 feet such as the Tenth Mountain ski huts. Snow not a problem.
Land mass Do the calculation yourself. To Replace ALL US electricity a 46 mile square. 0.8% of the land in Texas.
Storage The solar cycle follows the electricity peak use cycle. Connected to the grid, solar could replace ~40% of our power without substantially needing any storage. Beyond that there are many ways to store energy (Hydrogen, hydro, air flywheel, superconductors, capacitors, vanadium flux batteries, etc).
byproducts
PV is not the IC industry. Many PV technologies (CIS, aSi, Ribbon Si, Dye-TiO) have essentially no pollutants.
As noted
Good thinking, but not the case. Apartment buildings above 2 stories (5 or more units) only make up 6.5% of total units. My numbers if anything underestimate the sqft, because I didn't factor the larger single family housing sqft which averages 2527 sqft and make up 88% of units. So just considering single family units we have 94.2 million units with 75% floor sqft to roof (50% single story, 25% double story w/single story garage). With 6% slope addition = 1.9E11 or 8% MORE while ignoring all apartment buildings.Adding the apartment buildings back in: 2-4 unit buildings (6% of total units) we'll say they are all 2 story (1393 sqft * 6.4E6 units *
True enough. The EIA tracks very accurate numbers for what is, and what has been, but they bad with the future. Which works OK for consumption models (sort of, they have been horribly wrong on that to - in the 70s they assumed exponential growth, when in fact efficiency made up for it), but they do not consider technological changes, cost reductions, geopolitics, etc (for instance a decade ago their wind power forecasts were FAR too low). They are primarily a current energy information outlet for congress.
FACTS:
* Residential electricity consumption is 35% of the total (not 3% as you state)
* Roofspace is not as you stated. From the census data and DOE data: The average housing unit size is 2066 sq ft. There are 107 million units. 50% of houses are 1 story (roof=sq footage). The other 50% are 2 story or more (Census), which I estimated roof space is half of living space (this averages in the added garage roof space of some with the loss of roof space to 3 or more levels). The result was increased by 6% for the added area of the average roof slope. Commercial was 67 billion sqft. If you want to do a more detailed analysis, It would be great, send it to me. The average single family unit is 2527 sqft which I didn't use in these calcs, and are 88% of total housing units, so these numbers are likely underestimating roof space by around 15% or so. However the outcome will not likely be more than +/-10%.
* Of course not all roofs will be usable. The point is to get perspective on the land area needed. Even if 1/3 of roofs are usable, then problem solved. If you don't use roofs, the land area required is still VERY small. THE LAND AREA IS NOT THE SIZE OF TEXAS! With 17% panels on trackers the land area is a 46 mile square - 22% smaller than Dugway Proving grounds OR 0.8% THE SIZE OF TEXAS. With multijuction concentrators, its less than half of that. The problem here is you've been programmed to believe it should take a huge amount of space, BUT IT JUST DOESN'T. Clear yet?
*OK Say you want to replace ALL the US energy with solar(oil, coal, Natural Gas, wood, etc). How much land would it take? The US uses 98.3 Quads a year, or 2.88E13 kWh. Using 40% efficient multijunction concentrators ($1/Wp!) on trackers in average location (Kansas City) you get 964 kWh/m^2/year. LAND REQUIRED: a 100 mile square. OR 4% the size of Texas! VERY SMALL! 1/10th the 290,000 km^2 number you cite (Reference for this number please).
*Obviously once you see the real numbers - infrastructure isn't a problem. In fact a distributed PV system, some on roofs, some in local grids, some in large arrays would reduce distribution and transmission infrastructure substantially
It would be very instructive for you.
If he means by this average usable power per day including darkness, then hes not far off. (1000 W/m^2 X 6 hours of peak sun)/24 hours per day = 250W/m^2 per 24 hours. BUT this is misleading, watts are peak POWER measurements not ENERGY. The energy stays the same in either case: 1000 W/m^2 X 6 hours = 6000Wh OR 250W/m^2 X 24 hours = 6000Wh. See?
These numbers are right on, similar to what I've been showing you (1000 W/m^2 x 6hour peak/day x 365 day =2190 Wh/m^2/day). Except 250 is way too low(even Barrow, AK 375 miles north of the arctic circle gets 912 Wh/m^2/year, and 1314 with a tracker). Here a selection numbers right from the 30 year average weather history statistics for flat plate solar panels taking into consideration incident of insolation (first number is a fixed panel at an angle equal to degrees latitude of the city, second number is a panel on a tracker - numbers in Wh/m^2/year):
Tucson = 2372 (3285)
San Francisco = 1971 (2591)
Kansas City = 1788 (2409)
Seattle = 1350 (1788)
Denver = 1825 (2701)
Columbus = 1533 (1971)
Boston = 1679 (2153)
Buffalo = 1496 (1934)
Anchorage = 1095 (1460)
To get the annual energy produced per m^2 multiply by the PV panel efficiency. For example, at 17% efficiency in Denver 310 kWh/m^2/year on a fixed panel, 459 kWh/m^2/year on a tracking panel. Multiply this by the number of square meters of roof in the US, and you get 6.99E12 kWh/year for Denver fixed panels, and 1E13 kWh/year for Denver tracker (not all places are the same as Denver, but they aren't all that different either - its an example ;). See? Here we get 200%-300% more power than needed, like I said.
The quote (which is true, Lovins is very accurate with his numbers) is not meant to be more than it is: an interesting comparison between solar and the current and real state of nuclear power in the US (light water reactors are capable of using only a fraction of their fuel before they are spent). Of course breeders or other designs could produce 100-1000 times as much energy per pound, but they have their tradeoffs too (which is why we chose not to use them).
Exactly. Solar power is available everywhere (did you notice the deviation between the alaska and arizona is only 2-1?). Solar IS the ultamate distributed power source. If most of the power is generated locally, they carrier requirement of transmission is HUGELY reduced, and overall costs come way down.
The problem is
Yikes! Here the first problem with your calculations! Solar insolation is 1300 W/m^2 outside the atmosphere, 1000 W/m^2 on the ground in peak sun conditions. NOT 170! (look it up yourself you'll find tens of thousands of refs on Google)
insolation FOR A FIXED panel at an angle equal to latitude provides an average of 6 hours of peak sun per day in the average US location. (of course the solar insolation is changing based on time of day. However this is how it is specified in the industry: pre-integrated to an equal number of peak hours). That equals 2190 kWh/m^2/year. Some locations a little more, some a little less. With trackers this goes up 25-50%. See the National Renewable energy laboratory insolation database and mapservers for more data.
Already considered see above numbers are already based on tilted fixed panels. Trackers of course improve the angle and thus the energy, but I'm giving a simple case, not best case.
Spacing is accounted for, 17% is total edge to edge module efficiency not cell efficiency. Maintenance costs, essentially are none (solid state revolution man) no moving parts, no dusting, no snow removal required (the benefits of dusting/cleaning has been proven to be of small benefit. less than 4%). Storage is an issue. There are many storage technologies and they do cost money (some solar technologies, not PV, are self storing such as Solar 2's phase change salt storage). However, energy profile on the grid tracks the solar cycle closely. 40%-60% of our energy could be replaced without substantial storage added to the system. (another 20-30% could come from wind, as the Dutch have shown, and the base load could be largely provided with geothermal, biomass, and wave. Thought I do think storage is an important piece of the puzzle.)
From the 2000 census data for households and the DOE for commercial buildings
The number you show is ENERGY consumption NOT ELECTRICITY consumption, and its a little too high (I guess the are spooks not energy experts). From the Department of Energy, Energy Information Administration total energy consumption is 2.88E13 kWh. The total US ELECRICITY consumption is 3.4E12 kWh - which is what we are talking about.
Today is your lucky day. The numbers are very much right (as you can now see). And we didn't have to even invoke any extra land consumption OR higher efficiency cells OR Dye-sensitized solar cells which can be used as windows on high rise buildings, etc. PV is amazing stuff with incredible potential, 40% annual market growth, prices are nearing $1/peak watt (33
Snow isn't a problem, as I've already noted elsewhere. In fact its a bessing. In winter months PV panels are angled steeply in northern latitudes. Snow slides right off the panels. The snow increases the energy production because it effectivly reflects the sun from the ground/roof onto the panels. Winter can be the best solar producing month in many locales (like the rocky mountains which get a whole lot more than 3 feet of snow).
Solar still makes energy even on cloudy days. Solar energy is remarkably constant throughout most of the world/US.
For a flat panel, the deviation from the best southern Nevada site to the worst northern Washington state site is only 2-to-1! The rest of the country is surprisingly small deviation within this range. See rredc.nrel.gov/solar/ [nrel.gov]
Wisconsin gets an average daily insolation of 4-5 kWh/m^2 verses 6.5 kWh/m^2 for best locale in Arizona for a fixed panel. So Wisconsin is still 70% of the best solar location. not too much difference.