Talk:Photovoltaics/Archive 2

From Wikipedia, the free encyclopedia

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Lengthwave

I think it may be nice to have some graphics or data about the energy corresponding to each wavelength in the energy recieved in the surface and the energy used in the photovoltaic cells. —Preceding unsigned comment added by 85.81.171.82 (talk) 19:16, 21 January 2008 (UTC)

Source data for PV Power Costs?

Anyone know what the source for this infor is/was? The fact that the decimal places are commas is a bit confusing. Also the statement "Normally photovoltaic equipment is fully functional after 30 years also." is a bit weird. Does the author mean "fully profitable"? Or paid for? -KevMoo 18:55, 6 August 2006 (UTC)

I changed the sentence to what I think it means - PV equipment should last 30 years or more, not just 20. I also changed the decimal separator to point. But this table can't stay unless there is a source cited, in case of copyvio. And the introductory paragraph still makes no sense. And I am still not sure how the table is meant to be read. Other than that, fine! ;-) Itsmejudith 20:12, 6 August 2006 (UTC)

This is my table. I created it. You read it the following way: The costs depend on the sun power you have in your country. For example in south Germany it is 1,000 kWh/year or in Spain and Sicily it is 1,800 kWh/year. California should be about 2,000 kWh/year in some regions. After that you must know how much you paid for your photovoltaic equipment and then you can see how much a kWh costs for you in cent/kWh. Here in germany we will soon produce solar modules for about 1,000 €/kW_p. The costs for customers are much higher cause the firms must grow very fast. Take a look at the german article too de:Fotovoltaik there's much more information about photovoltaic. The calculation is the following for 3,000$/kW_p, 2000 kWh/year, 4% interest and 1% cost of operation: Money costs: 3000$ * 0.04 = 120$/year, depreciation: 3,000$/20 years = 150$/year, cost of operation: 3,000$ * 0.01 = 30$/year, sum: 120$/year + 150$/year + 30$/year = 300$/year. Now devide 300$/year through the kWh/year ==> 300$/year / 2,000kWh/year = 15 cent/kWh. That's it. At the moment more as 500 million persons boost PV (most of them in europe), so 100 billion dollar isn't really much money per person. Take a look at the mobile phones, digital cameras or TFT displays. Economy grows very fast if there are several billion dollars to earn. At the moment we spent 2326 billion dollars a year for oil (75$/barrel). --212.202.193.176 10:08, 8 August 2006 (UTC)

Solar panels cost a lot more than 1,000 Euro per Kw. Here in the Middle East they are around 15,000 Euro per Kw excluding installation. Also the calculation doesn't square with the UK estimates earlier which suggest it takes 130 years to just recover the capital costs. Curry's also mentioned above quote starting prices at 12,000 Euro!--Rjstott 10:39, 8 August 2006 (UTC)

The production costs in germany are 2,000€/kW_p and on the market you can buy 93.1 kWp for 337,953€ (532 modules, 175 Wp, single crystal, 40" container, CIF Hamburg). This is 3,630€/kW_p. But 1,200€/kW_p are silicon costs cause we use chip silicon at the moment. Now the production uses solar silicon which is about 70€/kW_p. So in the near future production prices could be lowerd to 1,000€/kW_p and more due bulk production. Bulk production reduces production prices by 20% if you double the production. And we must double the production at least 10 times. Thin film moduels (amorphous silicon) are at 2,800€/kW_p at ebay: [1](scroll down to the middle of the page). If you pay 15,000 € per kW_p better ship your modules from germany.--212.202.193.176 11:37, 8 August 2006 (UTC)

Here is an auction for 7.26 kW_p [2].--212.202.193.176 14:26, 8 August 2006 (UTC)

I'm sure you will be right about panel prices some time in the future. What matters here surely is a like for like comparison of the situation now. To do that you would have to include the costs of the necessary regulators and power inverter, delivery, plus installation using actual prices including profits now. This is very topical and interesting and an article detailing practical installations with some good facts would be very useful. We're buying Kyocera panels and the prices I quoted included regulator, battery, enclosure and wiring. Delivery is a significant cost as is profits to various middlemen who add little value. Neither BP nor Shell can match the prices and all companies report significant back-order situations. For the time being price is determined by market forces similar to those driving oil prices where currently a factor of five or more can be found between extraction costs and market price.--Rjstott 03:43, 9 August 2006 (UTC)

We're still missing a source for the table, correct? Please include one. KevMoo 04:36, 12 August 2006 (UTC)

There is no source for this table. You can calculate every value as I described it above. So the source is Microsoft Excel or your pocket calculator. And yes, you have to calculate all your costs of course, not only the module price. If you are grid connected, you don't need batteries. You supply surplus power to the net and get money for it. In Germany, France, Spain and Italy this is about 50 eurocent/kWh. If you need batteries you must calculate them to the costs/kWp of course. If you are grid connected you have additional costs for the inverter. For example: Module costs: 40.000$ (10kWp), inverter: 2.000$, installation: 8.000$. So your costs are 5.000$/kWp. Now the auction ([3]) ends at 4422 Euro/kWp. It includes all, modules, inverter, cables and mountings for self-construction. So in Sicily (1800kWh/year) you can produce power for 24 eurocent/kWh. Net power costs are 21.08 eurocent/kWh. Here you can buy modules for 3,49$/Wp [4] and thin film ones for 3$/Wp.

Excel

$D$1=0.04
$E$1=20
$F$1=0.01

B5: 2400
C5: 2200
D5: 2000
E5: 1800
F5: 1600
G5: 1400
H5: 1200
I5: 1000
J5: 800

A6 : 200
A7 : 600
A8 : 1000
A9 : 1400
A10: 1800
A11: 2200
A12: 2600
A13: 3000
A14: 3400
A15: 3800
A16: 4200
A17: 4600
A18: 5000

First cell: =$A6*(($D$1+$F$1)+1/$E$1)/B$5*100

--212.202.193.176 18:00, 12 August 2006 (UTC)

Here is another calculation from diablosolar: [5]--212.202.193.176 20:49, 12 August 2006 (UTC)

I had to study the chart and read this commentary and re-study the chart before I comfortably understood what it represented. I suggest adding the following wording as introduction to it. "The label at the left end of each row shows an example price ($US) per peak Kilowatt hour (kWp) of a PV panel. The column headings indicate the actual kilowatts of production to expect from the panel. This varies by geographic region. The calculated values reflect the cost in $US cents per KwH produced, considering a 4% cost of capital, a 1% operating cost, and a 20-year lifespan of the equipment. "

comments? Leotohill 04:14, 28 August 2006 (UTC)

I think that's good and just suggest a few small changes to clarify it a little more:

The labels on the left show various total costs, per peak kilowatt (kW_p), of a PV installation. The headings across the top refer to the annual energy output expected from each installed kW_p. This varies by geographic region and according to efficiency etc. The calculated values reflect the total cost in cents per kWh produced, including a 4% cost of capital, 1% operating and maintenance cost, and depreciating the capital cost over 20 years. Normally, photovoltaic modules have 25 years warranty, but they should be fully functional even after 30-40 years.

BTW, did you also see my, and the original author's, re-calculated examples near the bottom of this page? --Nigelj 19:23, 28 August 2006 (UTC)

I've added the paragraph, more or less as above, to the article. --Nigelj 20:12, 30 August 2006 (UTC)

This still needs work--the table has no title or caption--one has to dig through the text to find a description of it. The table needs better labeling.70.109.143.5 00:44, 29 October 2007 (UTC)

Further comment on: "With lifetimes of such systems of at least 30 years" from a retired electronic engineer. I first noticed that there is no citation for this figure. While stating that capitol equipment lasts 20 years is just a rule of thumb, to state 30 years should be footnoted. There are other issues to consider here.

Silicon devices don't last forever. This applies both to the cells themselves which would suffer some degradation in performance due to annealing and passivisation failure and to the semiconductors in the inverter. Power transistors do suffer catastrophic failure. Maintenance costs would increase with time and replacement costs would become significant. MOSFETs are more reliable than Bipolar transistors and reliability continues to improve, but it is something that should be considered.

It should also be considered that a system installed today would probably be obsolete in 10 years. This would probably mean that major repairs would not be possible and/or would be not be economically justified.

In general, these are the issues faced by all early adopters of a new technology. Yet some early adoption is needed to push the technology. Tyrerj (talk) 02:26, 3 January 2008 (UTC)

I should be able to find a citation for the 30-year figure. It really applies to the PV modules, not necessarily to the whole PV system, and it is based on accelerated lifetime testing -- the modules made 30 years ago didn't really have 30 year lifetimes (though quite a few of them are still in operation). The cells themselves are well encapsulated and don't see much in the way of thermal or electrical stress, and cells removed from decades-old modules have mostly been found to work as well as they did the day they were produced. The primary degradation mechanism for PV modules is UV and thermally induced degradation of the EVA encapsulant used as a pottant to secure the solar cells in the module. Even so, mean times to failure for PV modules from published field studies are between about 500 years and 7000 years (which doesn't mean a module can be expected to last thousands of years, but that in a given year you can expect a few hundredths of a percent of modules to fail -- the failure rate will increase near the module's end of life). While nobody can claim the cells will last forever, they are very likely the longest-lasting component of a PV system (and by a longshot). Individual modules can be replaced without having to replace the entire system, and the typical PV module warranty is 20 years and guarantees no more than 1% degradation per year. I think some of this is covered in the solar cell article.

The inverter is another story. It is, in fact, the shortest-lived component of the system, though it is primarily a result of capacitor failure rather than failure of the power semiconductors (which are primarily IGBTs these days). However, the inverter can be repaired, and even if it is not economical to do so it can be replaced independently of the rest of the system at a small fraction of the cost to replace the entire system (it currently represents about 10% of the installed system cost).

Given all of this, obsolesence isn't really much of an issue. Few systems will have to replace even one module in the course of 10 years, and though many will have to replace the inverter in that time it is a relatively inexpensive repair compared to replacing the entire system. If the system is still producing enough energy to meet one's needs, there is no reason to scrap it. And if it isn't, one can always add a few modules to bump up its energy production. I'm not trying to say that there are no issues with the technology -- there most certainly are -- but I think you're picking on the wrong ones.--Squirmymcphee (talk) 03:05, 3 January 2008 (UTC)

I can offer the following:

"These attributes have led the National Renewable Energy Laboratory in Golden, Colorado to recognize CdTe’s potential for achieving the lowest production costs among current thin film technologies. (Photon International, November 2004, page 50). CdTe module costs well below $1.00/Wp have been predicted by NREL and others." http://www.firstsolar.com/company_technology.php

"PART I Item 1: Business We design and manufacture solar modules using a proprietary thin film semiconductor technology that has allowed us to reduce our average solar module manufacturing costs to among the lowest in the world. In 2007, our average manufacturing costs were $1.23 per watt, which we believe is significantly less than those of traditional crystalline silicon solar module manufacturers. " http://investor.firstsolar.com/secfiling.cfm?filingID=950153-08-367

""During the fourth quarter of 2007 we benefited from the full capacity and economies of scale of our Frankfurt/Oder plant. This combined with continued throughput and conversion efficiency gains afforded us strong operating leverage and decreased our manufacturing cost per watt by 12% year over year to $1.12 per watt in the fourth quarter of 2007, further solidifying our cost leadership position in the industry," said Michael J. Ahearn, Chief Executive Officer of First Solar." http://investor.firstsolar.com/releasedetail.cfm?ReleaseID=294090

Regards, —Preceding unsigned comment added by 189.136.158.209 (talk) 22:01, 10 April 2008 (UTC)

By the way, I don't have the data, but I would find it extremely valuable to see a visual chart of average cost of PV in dollars per watt per year. For instance, a chart showing average cost per watt from 1970 to the present. This would perhaps show a declining cost from the past to the present.

Regards,

Eric —Preceding unsigned comment added by 189.136.158.209 (talk) 22:09, 10 April 2008 (UTC)

Worldwide installed photovoltaic totals

Can anyone provide a source for the claims inthe section on Worldwide installed photovoltaic totals, specifically the one claiming total worldwide capacity at more than 5 GW and the 90% share of the three leading countries? The best sources I can find [6], [7] show lower figures (though lagging two years behind). Also, the IEA figures seem to include only figures for selected countries, so I wonder what it says for the global total. Jens Nielsen 13:10, 16 March 2007 (UTC)

Data just published by Eurobserv'er for EU countries

The current numbers are extremely misleading. First of all, they are based on cumulative production, not actual installed capacity. Secondly, they are based on the peak output values of the cells, now care to have a guess how likely it is for all cells across the globe to peak their production at once? Thirdly, the capacity factor of the cells are typically in the 17%-30% range at best, so the 12.5 GW figure is not in any way accurate. Photovoltaic never produce that amount of electricity no matter how you count.The only way you would get that output was if you put all cells in one place and measured it at its peak value. To put into perspective how bad teh figure is. According to IEA and OECD figures the production during 2005 was 840GWh giving a time-averaged output of about 0.1 GW, now it has increased since then, but not by more than 100 times. At the very least the introduction ought to specify that those are peak-values, to simply state the 12.5 GW figure is highly misleading at best, an outright lie at worst. Heck, even the source for the statement claims "12,400 megawatts, enough to power 2.4 million U.S. homes" which makes it clear they are talking about peak output and not average output. I'd say we stick to the IEA and OECD figures. They may be lagging by a year or two, but they are more reliable than the garbage that is currently used. 85.224.219.135 (talk) 11:20, 21 January 2008 (UTC)

To specify peak capacity in this manner is consistent with the way the electric power industry specifies the size of any power plant, be it PV or otherwise. The generation capacity figures for two power plants are never directly comparable without knowledge of the capacity factor, particularly if the two plants rely on different fuels. In the US, for example, natural gas is responsible for 40% of generation capacity but only 20% of actual electricity production, while for coal those figures are about 30% and 50%, respectively. I discuss this at greater length in response to another comment below. Don't get me wrong, I'm all for providing a realistic indication of the amount of energy produced by PV, but to use the generation capacity figure for that is garbage regardless of what generating technology you're talking about.--Squirmymcphee (talk) 19:22, 21 January 2008 (UTC)

Economics

Just ran into this article. Reading through this talk page, there is a good amount of discussion about the possibility of spinning off a PV economics article. As it stands, there's already an economics section in the current article (which is mostly OR). Either way, there may (or may not) be interest in referencing this tool developed by myself and an actuarial student:

Photovoltaics Economics Calculator

It's validated against PVWatts, which is a widely recognized tool for determining how much power you'll get in different parts of the country. All of the formulae used are visible in the (linked) source code, and most reference published papers on their use or show their derivations.

Anyways, I'm not going to add it to the article, as that'd be a conflict of interest on my part, as well as self publishing. However, if anyone else thinks its worth adding, feel free. It might at least help make it so the economics section has verifiable formulae and references behind it, or make for a useful "external link". -- 129.255.93.189 22:46, 1 November 2007 (UTC)

Hmm... I'm having difficulty resisting the impulse to delete the table and its explanation under Economics of PV. Basically, there are many unexplained details: Why are some cells green or red... or more confusingly yellow? I presume these are weighed against current U.S. electricity prices, but from what source? I'm not sure if this is really valuable, considering it doesn't look to include non-PV module costs (or has an overly optimistic view of what PV costs). If this is merely an exercise to show how to calculate the economics of PV, then I propose that a separate page be created that runs through the process of calculating PV economics. Anyway, I'll give it a few days and see if the original poster or anyone else wants to tweak/explain the table before I ax it, since it does look like it took quite a bit of work to create. :-) PandasCanFry (talk) 04:28, 22 January 2008 (UTC)

• All the table is, is a straight forward calculation of installed cost divided by insolation to get average kWh cost. There is nothing magic about it. The gradation of colors shows the trend from expensive to economical. Once again nothing special. Using three colors instead of two creates more of a gradient, while more colors is not necessary to see the effect. It's a pretty useful graphic and should not be deleted. The 20 years in the corner cell was confusing, so I have removed it and added a title. 199.125.109.124 (talk) 20:01, 22 January 2008 (UTC)
  • I am confused by the units of insolation. I expected watts per meter-squared, or some similar measure of capacity per unit area. What I read in the table makes no sense at all -- or am I missing something here? BTW, please register with Wikipedia, it will give us a better way to communicate with you, and it will give your contributions more weight with other editors. — Aetheling (talk) 00:30, 23 January 2008 (UTC)

It's a standard measure, but since you asked the question it clearly needs to be explained better. The units are given in the table, kWh/kWpyr, same as the far right column in the "Produced, Installed & Total Photovoltaic Peak Power Capacity" table. What you do is take the insolation number in kWh/kWpy and multiply it by your panels kWp to get your kWh/year. Couldn't be a simpler calculation. So let's say you install a 4 kWp panel and your insolation is 1000, you can expect to get 4,000 kWh each year. The insolation is a number that is averaged over a year, and is independent of the efficiency of the panel, because it is given in kWh/kWpy, and not in watts/meter-square which is also a measure of insolation but if you use that you have to determine the efficiency of your panel to find out what your output will be. Take a look at the article insolation and see what we need to change either there or here to make things less confusing. Just as an added wrinkle, if you add tracking to your panels it adds about 30% to your annual output. The easiest way to accommodate that is to add the 30% to your rated Wp but this is not universally recognized. Looking at the insolation article I see that it needs work as it confuses irradiance with insolation. 199.125.109.124 (talk) 05:01, 23 January 2008 (UTC)

In my experience, calculations on the economics of any energy system get quite complex. I'm not quite sure what is meant by "assuming a 10% capital cost." Does that means an initial loan for 90% of the system? More importantly, why these parameters?
Also, economical by what standards? Grid parity? If so, by what source? Another issue is that economical versus non-economical is very EASILY changed by toying with interest rates, depreciation calculations (straight-line versus MACRS), etc.; so without reasons for using the initial values, what useful conclusions can be drawn from this table?
I DO think that the table is valuable, but without anything to substantiate the initial conditions, it's original research at best and conjecture/wrong at worst. The other tables in this article are verifiable, but the conclusions that result from reading this table are not. Again, I think this is better suited on a separate page that describes calculating the economics of PV.
Finally, I've also never heard of the kWh/kWp as a unit of insolation, but I'm not discounting that people use this as it's much more convenient then kWh/m^2PandasCanFry (talk) 15:40, 23 January 2008 (UTC)

I'm sure the 10% capital cost is meant to refer to the effective rate of interest paid on capital, but by and large I agree with you -- at the very least, the assumptions used in the calculations must be spelled out much more explicitly. How you do the calculations properly and what assumptions you make vary considerably depending on who's paying for the system and how they can/will pay for it. The cost to a homeowner paying with an ordinary loan is higher than for a homeowner paying with a mortgage or home equity loan, and those are both much higher than the cost to a business that gets tax credits for capital investment and depreciation. Plus, I see no mention of inflation and discount rates, both of which are important when you're talking about cash flows over a period of decades. The numbers shown in the table do seem reasonably consistent with properly done calculations for a residential PV system, but I've seen a lot of improperly done calculations that nonetheless look about right.

I've never seen kWh/kWp used as a measure of insolation either. Strictly speaking it is not a measure of insolation, though it is proportional enough to insolation that it can reasonably serve as a surrogate. It expresses the energy obtained from a solar panel as a function of its rated peak output (e.g., if a panel produces 2000 kWh/kWp each year and is is rated at 100 Wp, you would expect annual energy output of 200 kWh). It is meant to level the playing field when comparing modules that have different power conversion efficiencies, but that respond differently to environmental factors like temperature and the spectral content of incident light. It is common that one module will produce 1800 kWh/kWp/year and another 1700 kWh/kWp/year under the same insolation.

Finally, the table really only shows the contribution of the solar panel itself to the cost of electricity. The table caption properly indicates this, but it seems to me that the caption and/or the article should indicate that there are other costs associated with PV systems that are not accounted for in the table so readers aren't unintentionally misled.--Squirmymcphee (talk) 18:50, 23 January 2008 (UTC)

<-Working backwards from the definition of Wp, which is a module that will generate 1 Watt with an irradiance of 1000 W/m^2 under standard conditions, and insolation in W/m^2 is the average irradiance, then insolation in kWh/kWp/year = insolation in W/m^2. You will notice that the maps which show insolation in kWh/KWp/year also include a 75% performance ratio to allow for the efficiency of the invertor and specify an optimally inclined panel. 199.125.109.54 (talk) 07:46, 1 February 2008 (UTC)

I understand perfectly well where it comes from mathematically, but there are two issues. One is that kWh/kWp does not specify insolation, but the amount of energy you could expect to generate from PV given the local insolation level -- a number that is proportional to insolation, to be sure, but it isn't insolation (unless you assume 100% efficiency). Second, the number of kWh/kWp produced by PV modules under standard conditions doesn't vary from module type to module type, but under real-world conditions it can vary by as much as 20%. These are the reasons I said kWh/kWp is "proportional enough" to insolation to make a reasonable surrogate. As for "noticing from the maps," as I said above I've never seen any such maps -- all the maps I've ever used (or seen, for that matter) specify insolation in units of kW/m^2 or kWh/m^2. Now, it's true that most people ignore the module-to-module variations in kWh/kWp and get close enough in their calculations for most purposes, especially considering annual weather variations, and that's fine, but I fail to see the advantage of adding a level of abstraction between actual insolation figures and the figures that are reported on the insolation map.--Squirmymcphee (talk) 18:12, 3 February 2008 (UTC)

It is possible that you are mistaking "your insolation" which is what you would get from a flat panel and "your output" which of course varies by 20%, as you have indicated. Insolation is simply a measure of the average irradiance. It most definitely includes weather variations (cloud cover). It does not take into account temperature variations or mounting considerations, or shading. Wp is calculated under standard conditions, at a specific temperature, and since the output of a panel is a function of temperature, "your output" will not be "your Wp" times "your insolation". Which it probably wouldn't be anyway, because it doesn't include mounting considerations anyway, even if you don't have any shading. Efficiency is not a part of the equation, because it is removed by not specifying panel area, but by specifying Wp, which is a function of efficiency and area, other than efficiency is not a constant, as mentioned. If you want "your output" you have to multiply "your insolation" times "your Wp" times "your correction factor". As I said, in the maps they arbitrarily give 0.75 as "your correction factor". 199.125.109.136 (talk) 16:44, 5 February 2008 (UTC)

Okay, I'll start off with a mea culpa of sorts -- above, where I refer to 100% efficiency, is wrong. I didn't think that statement through. In my defense, though, it is also inconsistent with everything else I've been saying.

As for the rest, I'm beginning to wonder if you're using conventional definitions for your terms. Insolation is not, as you say, "what you would get from a flat panel," it is the intensity of the sunlight that falls on the panel -- in other words, it is the energy input to the module. It can be expressed in terms of average irradiance, but for the sake of clarity I'll be explicit and say that it is most often it is expressed as instantaneous power (kW/m^2) or instantaneous power integrated over a day, month, or year (kWh/m^2). It is completely independent of the PV module, as the sun shines regardless of whether you place a module in the front of it; therefore, the kWp rating of the PV module is irrelevant and expressing insolation in kWh/kWp is not meaningful.

Output, or power and energy production, is "what you would get from a flat panel," and that is what is displayed on your map. --Squirmymcphee (talk) 19:01, 5 February 2008 (UTC)

No, the map displays insolation, multiplied by 0.75, and not. 199.125.109.38 (talk) 22:53, 5 February 2008 (UTC)

<--And not what? The (rough) energy output of the panel?--Squirmymcphee (talk) 00:36, 6 February 2008 (UTC)

It shows two scales, one is insolation multiplied by 0.75 and the other is insolation not multiplied by 0.75. The first is labeled Solar electricity kWh/kWhy, the second is labeled Global irradiation kW/m^2 per year. See Image:EU-Glob opta presentation.png 199.125.109.38 (talk) 21:17, 6 February 2008 (UTC)

So it is precisely as I have been saying it ought to be -- irradiance is shown in kW/m^2 and PV system energy production in kWh/kWp.--Squirmymcphee (talk) 23:01, 6 February 2008 (UTC)

To be sure, if insolation is 2000 kWh/m^2 and you assume your modules always operate at the STC efficiency you will calculate an energy output of 2000 kWh/kWp. The insolation and output numbers match, it's only the units that differ, but the units are inconsistent with insolation. Still, strictly speaking you can argue that this makes a perfectly serviceable and useful surrogate for insolation.

But then there's the inverter derating factor. The inverter has no impact on insolation whatsoever, unless you use it to shade your modules, so derating for it has no business in an insolation map. On top of that, your implicit assumption that the modules always operate at the STC efficiency is a false one. It's fine for a first-order level of accuracy, particularly with the derating factor you're using, but in the field you will often see low-efficiency amorphous silicon modules produce substantially more kWh/kWp than an identically mounted high-efficiency crystalline silicon module at the same location.

I think this discussion has veered off quite a bit from where it started. I now understand where your kWh/kWp number came from, which is why I entered this conversation in the first place. I stand by my original assertion that such a number is roughly proportional to insolation, but it is by definition not insolation. It does, however, appear to be a reasonable first-order estimate of the amount of energy a PV system will produce at a particular point on the map. Related to insolation, yes, but not insolation.--Squirmymcphee (talk) 19:01, 5 February 2008 (UTC)

I'm glad to hear that you understand it. The "rough proportionality" is 0.7500000000 (to 10 digits). 199.125.109.38 (talk) 22:53, 5 February 2008 (UTC)

<--Right, I misplaced my "roughly" -- the number on the map is roughly proportional to the amount of energy the system might produce. By the way, where is it that you're finding these maps expressing insolation in kWh/kWp and multiplied by an arbitrary constant? You talk like they're standard tools, but I've been working with insolation maps and data for a very long time and I've never seen one. I'm genuinely curious about who is using them and how. I suppose I can see how they might be convenient for rough sizing of PV systems according to rules of thumb, but they're quite misleading if they're actually labeled as insolation maps while displaying something other than insolation.--Squirmymcphee (talk) 00:36, 6 February 2008 (UTC)

In retrospect, I don't think the variation in kWh/kWp between modules of different types is germane to this discussion and unnecessarily complicates it, so I'm dropping it (though it would still be an important consideration in a detail PV system design). It seemed relevant at the time I brought it up, but at that point I thought you were saying something other than what you actually are. So an "insolation map" using units of kWh/kWp displays values that are proportional to insolation by some arbitrary constant, but I still maintain it would be misleading to call such a thing an insolation map. After all, that suggests that if I have zero kWp of PV then it must be dark outside.--Squirmymcphee (talk) 17:09, 6 February 2008 (UTC)

Come again? It means that if you have zero kWp installed your kWh will be zero, regardless of your insolation. 199.125.109.38 (talk) 21:17, 6 February 2008 (UTC)

Exactly my point. If irradiance is measured in kWh/kWp and you have zero kWp, then irradiance must be zero, no?--Squirmymcphee (talk) 23:01, 6 February 2008 (UTC)

No. Your kWh will also be zero and 0/0 is what is referred to as an indeterminate number - it can be anything. In this case it is your insolation. 199.125.109.136 (talk) 16:33, 8 February 2008 (UTC)

PV production and prices

Not sure where to post this message so it's going in several places. The PV articles would be greatly enhanced if we could develope a graph showing the historic prices for PV. The image below shows how the prices should be both real and inflation adjusted. As a companion to the price graph we also need a production graph.

Mrshaba 09:31, 5 November 2007 (UTC)

See Figures 3 and 4 in [8]. They're a little old, but I'm certain I've seen more recent charts in other publications by the same author -- I just don't have time for a more thorough search right now. If you have a little time you might also search for the DOE's multi-year program plan and check some of the publications at [9]. The raw data you need for such a chart, unfortunately, come almost exclusively from very expensive off-line sources (though for production data I do remember somebody spamming this talk page with a site that purported to have that awhile back...).---- Squirmymcphee (talk) 18:54, 16 November 2007 (UTC)

I know what you mean about off-line sources. I contacted Robert Margolis at NREL and asked for his data but he hasn't gotten back to me. Tear... Nobody wants to play with me. I'll keep trying. Mrshaba (talk) 02:51, 21 November 2007 (UTC)

This a a rough draft of a PV price vs PV production history. Any suggestions or requests on how to present this info. I can easily pull the production off the graph and just compare $/Watt in real and inflation adjusted prices or compare price against cummulative installed PV.

Left Y axis is a log scale $/Watt. Right Y axis is log scale Megawatts. Red boxes represent inflation adjusted $/Watt. Blue diamonds are real $/Watt. Green triangles represent production in megawatts.

I've made several other graphs from Maycock's info. The graphs and the table they are derived from are here:User:Mrshaba/Experiments#PV Graph. Check it out. Mrshaba (talk) 17:35, 10 December 2007 (UTC)

It is more conventional to plot price as a function of cumulative production on a log-log scale. I'm not saying that you have to do it that way, but it makes the relationship between price and production much more clearly than what you have here, in my opinion. Also, a slight correction to your economic terminology: costs expressed in "real" dollars are inflation-adjusted, while those expressed in "current" or "nominal" dollars are not. So in your chart, red boxes represent real $/watt and blue diamonds nominal $/watt.--Squirmymcphee (talk) 21:22, 10 December 2007 (UTC)

OK... Thanks for correcting my terminology. I was more worried about keeping prices and costs straight. When I talked to Maycock I asked whether the numbers were inflation adjusted but I might have misinterpreted his answer. The information seems correct in it's current form but I might have to redo the inflation adjustment. Does it look right to you Squirmy?

I can set up a graph of price vs production but I don't see any need to put prices on a log scale. I'd rather chop off pre-1981 data and have everything in nominal dollars or pre-1986 data and have both nominal and real dollars vs production. What do you think of this? Mrshaba (talk) 18:42, 12 December 2007 (UTC)

The inflation adjustment looks to me about like it usually does for Maycock's numbers (his numbers are a little higher than those from other sources in the mid- to late-'70s time frame). The reason both price and production are typically plotted on log scales is that when they're plotted this way you get a relatively straight line which is important in the learning curve theory of price reduction. Linear-linear and log-linear plots of the data tend to leave the impression that prices are stable, not declining, since the same decline in price now occurs over a much larger increment of cumulative production than in did in, say, 1983. But plot the data both ways and see what you think.--Squirmymcphee (talk) 20:19, 12 December 2007 (UTC)

I'm familiar with learning curves but they should plot cost against production rather than price against production. This is especially true because prices and costs are decoupled for PV. You really have to tailor your window to get the PV data to behave according to the learning curve models. Don't you agree?

Cost (presuming you mean production cost, as opposed to retail price) vs production would be ideal, but the cost data simply do not exist. Nobody has compiled them, and anybody who has tried has failed because manufacturers simply aren't willing to share that information. The price data used in most of these studies is generally what one would pay for factory-direct orders of large quantities of modules, though, so it's as close to production cost data as you're going to get.--Squirmymcphee (talk) 16:10, 24 January 2008 (UTC)

I understand and agree with your points. The thing is, I'm not sure learning curves are all that useful. The cost vs. price issue clouds the picture and there's a lot of data scatter. When I talked to Berman he said he had prices down to $20 per watt in 1972 with production at 20-30 kW. The Hanson report out of Australia corroborates this $20/watt number. This single data point dramatically throws off all the learning curves I've seen. This data point also seriously calls into question Maycock's numbers so I'm doubly stumped. I'm coming to think that this learning curve theory for PV is overly artificial. You have to tailor your window and winnow your data to get a nice price reduction path so I don't see these curves as representative of what's going on or predictive of what will occur. The upshot is - linear graphs presents the data just fine. I think the story of PV is best told by breaking things into time periods and using time-based graphs of price and production. Going forward it would seem that PV prices and costs will continue to skip along. Currently the prices are skipping off subsidies. As these go down PV should eventually start skipping off the grid prices. It will be interesting to see what happens in Italy where they have very high electricity prices, high insolation and relatively local access to suppliers and installers.

Also, according to the data Maycock gave me, prices have been rather stable for nearly 20 years. No worries though... The graphs look fine either way and I will have the tabular data available for the curious reader. I tried contacting Harmon but no luck. I'll try van Sark again one of these days. I might challenge him with the Berman data and see what he has to say. We'll see... 64.161.56.5 (talk) 01:38, 26 January 2008 (UTC)

You can't really say that data on the industry as a whole is invalid based on information about a single product. The learning curves published for the PV industry are for the PV industry as a whole, and at any given point you're going to find individual products that cost both more and less than the number reported. In other words, I don't think you can claim that a single data point for a particular product throws off the learning curve for the entire industry. It's not a very fine-grained tool, it's meant only to grossly quantify industry trends. There's always somebody ahead of the learning curve, but for anybody to really throw it off track they have to stay ahead of it (and they typically don't, for reasons that are arguably explained by the learning curve itself). There is some value to breaking the learning curve into segments -- there's a paper by a guy named Nemet, if I remember correctly, that does a very good analysis of the learning curve broken into segments over time. I'm not quite sure what you mean by "skipping off subsidies," at least not in the context of the learning curve, but if you're referring to a lack of price reductions in recent years then you are correct, at least indirectly. The prices used for learning curve data are about as close to manufacturing costs as you can reliably get, and the subsidies have increased demand to the point that (a) silicon prices are through the roof, keeping manufacturing costs artificially high, and (b) the PV market is not competitive -- many PV companies barely have sales staffs anymore. Because of those two issues, recent data points have jumped off the learning curve a bit. Some people think this represents a fundamental shift in the learning curve, but there are others who believe this is a temporary departure and that prices will align with the learning curve once again when the market becomes competitive (I'm in the latter category).

Maycock's data do show PV prices being relatively stable over the last 20 years, but only if you don't adjust for inflation. Remember, $5/watt 20 years ago translates into about $8.50/watt in 2003 dollars, and that's almost 2.5X current prices.--Squirmymcphee (talk) 22:26, 31 January 2008 (UTC)

I did find Maycock's numbers a little high when I compared them against Perlin's numbers. I asked Perlin to clear this up and he said he got his data from Dr. Berman so I wrote to Dr. Berman hoping he can give me some price and production numbers for the early 70s. What other sources of data are you comparing against Maycock's numbers. Are they on the net? Any chance you could send them me? Mrshaba

I'm speaking primary from having seen similar charts with attributions other than Maycock. The most commonly cited figures for that time period seem to come from a consulting company called Strategies Unlimited. To my knowledge, they are not available online and are very expensive to purchase -- I have only ever seen them indirectly. There's a white paper by a guy named (Christopher?) Harmon that I believe you will find online that uses these data -- if you can find the paper (which I no longer have) then you can probably ask him to share the data. There is also a guy named W. G. J. H. M. van Sark in the Department of Science, Technology, and Society at Utrecht University who will soon be publishing a paper on experience curves and cobbled together a decent data set. Hope that helps.--Squirmymcphee (talk) 21:47, 12 December 2007 (UTC)

Thanks for the leads. 71.103.233.189 (talk) 22:30, 12 December 2007 (UTC)

Robert Margolis of NREL has a few leads buried in this study: [10]. Nothing past the year 2000 unfortunately. [11] also has some data from PVMaT, which is cost data (not sure if this is public or not), and from Photon International, PiperJaffray and Prometheus Institute market analyses, all of which cost a lot to access :-( However, if someone has these lying around, well, wouldn't that be nice? —Preceding unsigned comment added by PandasCanFry (talk • contribs) 23:16, 23 January 2008 (UTC)

There is an interesting comparison of the cost of five of the renewables over the last 25 years at http://www.nrel.gov/analysis/docs/cost_curves_2002.ppt There also is an excellent comparison of cost vs. production showing the learning curve for photovoltaics at http://www1.eere.energy.gov/ba/pba/pdfs/pv_overview.pdf 199.125.109.54 (talk) 04:44, 1 February 2008 (UTC)

Disadvantages

The disadvantages section is rather small and contains several understatements. It says:

• Solar electricity can sometimes be more expensive than electricity generated by other sources.
• Solar electricity is not available at night and may be less available due to weather conditions (solar panels generated reduced amounts of power in some types of inclement weather;) therefore, a storage or complementary power system is required.

The first point says sometimes, I'm not really sure when it's actually cheaper. Does anyone have any examples? It's usually significantly more expensive then most other sources. The second point says that a storage system is required, but this does not state that such systems are also very expensive and further reduces solar's efficiency and effectiveness. Let me know what you think. Thanks!

Codingmonkey (talk) 05:15, 5 December 2007 (UTC)

The primary example of a situation where PV is cheaper than conventional electricity is in an off-grid situation where a grid extension would be required to receive grid power. The cost of a grid extension can exceed $50,000/mile and is borne by the customer requesting it, not the utility, so a PV system can easily be cheaper. You can argue that, say, a diesel generator is even cheaper, but if you're talking about an unattended commercial installation (a cell phone tower, a radio beacon, etc.) then you have to pay somebody to go refuel it periodically if you use a generator (and depending on the site's location, that may be dangerous or impossible at certain times of year).

There are also certain areas, both in the US and outside of it, where the retail price of electricity is high enough for PV to be competitive, if not cheaper. For a residential user in California, for example, each kWh beyond a certain number in a month costs well over $0.30 -- it is cheaper to generate those kWh with PV than to pay retail, even without California's generous subsidies. There are also isolated rural areas in the US -- primarily on small islands -- where PV is competitive with retail, if not cheaper. Another good example is Japan, which has high enough electric rates that PV is economically competitive in many areas.

In the end, for most of us in the industrialized world PV is currently more expensive than conventional electricity, often significantly so, but I think to say that it is sometimes cheaper is accurate without having to resort to hard-to-find special cases -- tens, if not hundreds, of millions of people experience the situations I described (though most of them are probably outside of the United States).

As for the second part of your comment, I'm not sure what you mean by storage systems reducing solar's effectiveness -- a properly designed system will be as effective as you need it to be. It does reduce overall system efficiency, particularly if you're drawing from the storage system regularly, and it drastically increases the cost of a kWh. However, to my knowledge it is quite common to choose self-generation (be it PV with storage, diesel generator, or something else) over the grid extension for grid extensions beyond a quarter mile or so.--Squirmymcphee (talk) 15:16, 5 December 2007 (UTC)

You need to include the costs of global warming in the cost of the existing electricity supply. When you add that in, solar is far cheaper everywhere. 199.125.109.134 (talk) 05:57, 6 December 2007 (UTC)

This is an argument from environmental economics and may be notable. Can you help us find a source that makes this argument specifically in relation to PV? Itsmejudith (talk) 11:53, 6 December 2007 (UTC)

It is almost impossible to put a dollar amount on the cost of mass extinction - the earth has had five mass extinctions, and we are in the middle of a sixth. The best known one of course is the extinction of the dinosaurs, but that was not the biggest of the five, and the current rate that we are losing species in this sixth one is perhaps greater than any of the others. I do not expect that we will be one of the species to go, you might be happy to learn. One of the biggest cost increases has been the insurance losses which have been increasing at a staggering rate due to the increase in damages. Once again, how do you attribute a trillion dollars in losses (mostly uninsured) to a specific level of CO2 emissions?[12] [13] [14] One person asks in a blog, "Do we have any trend data on insurance payouts for weather-related losses? It would be interesting to see if there is an upward trend." Well, duh. 199.125.109.54 (talk) 08:15, 1 February 2008 (UTC)