Pre-installation Production Estimates vs. Reality

[J.P.M.]

Not to mention that the weather just about everywhere in the US is not following historical data very closely. There are more "extreme" incidents in the past few years to results in most of the historical weather data to no longer be accurate.

Hopefully you don't get the snow fall amount like you did last year.

An old weather saw has it that climate is what you expect, weather is what you get.

I wonder if it may be more of a perception than reality that there are more weather extremes than "before" due to the 24 hr. news cycle that also seems to be geographically everywhere "on the scene". So, for example, we perceive we're getting more tornadoes because we can now see them on TV, u-tube, storm chasers, etc. I also wonder if we're more prone to catastrophizing events rather than just sucking it up and getting on with life.

The reality is probably somewhere in between - as in, there may be more extremes, but not as many as we think, and they ain't as bad as the talking heads would have us believe.

BTW, for anyone who may care, using the above definition(s), a lot or most of the climatological data that solar estimating programs like PVWatts use (speaking of the TMY data here) probably can't be characterized as either weather or climate data. First of all, a lot of it is synthetic and not actual or recorded and second, it's not "average" data in the sense as it is usually implied in the context of "climate".

[lkstaack]

Thanks very much for the advice.

I have used PVWatts and found lots of numbers. But the estimates I got (last year) don't come in that format. Example attached.

But my main interest is hearing from members here who can compare what the contractor told them in advance, and what they're actually getting.

Thanks.

Tom in CT
[ATTACH=CONFIG]5463[/ATTACH]

The provided estimates, as well as your own estimates, can vary widely due to the variables used. For instance, did the installer include future energy price increases into the "energy savings" column? Probably did, but is it accurate? Maybe not; it will certainly figure into your return on investment.

I think any reasonably priced system installed properly will show a ROI much sooner than its 25 year life span. However, can the money spent on it work harder elsewhere? Determining the future value of money also uses variables that can be widely off the mark. If you find a way to make your money work harder, will you actually invest it there? A PV system that yields 2% is better than a 7% investment that is never made.

[sensij]

Thanks very much for the advice.

I have used PVWatts and found lots of numbers. But the estimates I got (last year) don't come in that format. Example attached.

But my main interest is hearing from members here who can compare what the contractor told them in advance, and what they're actually getting.

Thanks.

Tom in CT
[ATTACH=CONFIG]5463[/ATTACH]

With regard to the numbers you've posted... if this is for a 4 kW system, the price shown is too high. 28756 / 4000 = $7.19 W. This is double what systems in CA cost, and I doubt that prices in CT will be that much higher.

If your electric bill is $100 / mo, the year 0 energy savings column should be $1200. They used 3% increase in electric bill annually, which is slightly higher than I would use but is a defendable number.

I don't know enough about SREC's in CT or the SC Personal tax credit to comment on them at this time. It doesn't look like they are using a discount rate, so this investment is essentially being compared to stuffing cash in a mattress. While appropriate for some, you may want to consider if there are other investments you could make.

If the 0.144 entry in the first row of energy savings refers to a $ / kWh cost from the power company, this system must almost definitely be larger than 4 kW. A $1300 annual bill would be something like 9000 kWh, which you would probably need around an 8 kW array. If what was quoted is in fact 8 kW and not 4 kW, the price makes a lot more sense.

[SunEagle]

An old weather saw has it that climate is what you expect, weather is what you get.

I wonder if it may be more of a perception than reality that there are more weather extremes than "before" due to the 24 hr. news cycle that also seems to be geographically everywhere "on the scene". So, for example, we perceive we're getting more tornadoes because we can now see them on TV, u-tube, storm chasers, etc. I also wonder if we're more prone to catastrophizing events rather than just sucking it up and getting on with life.

The reality is probably somewhere in between - as in, there may be more extremes, but not as many as we think, and they ain't as bad as the talking heads would have us believe.

BTW, for anyone who may care, using the above definition(s), a lot or most of the climatological data that solar estimating programs like PVWatts use (speaking of the TMY data here) probably can't be characterized as either weather or climate data. First of all, a lot of it is synthetic and not actual or recorded and second, it's not "average" data in the sense as it is usually implied in the context of "climate".

Good point. We have become junkies to the Extreme News reports and may have developed a knee jerk reaction to catastrophic events that are happening now. Those same events may have happened in the distant past when there wasn't as good way to record the events due to not having accurate instruments or the knowledge to measure the specifics like we do now.

[SunEagle]

With regard to the numbers you've posted... if this is for a 4 kW system, the price shown is too high. 28756 / 4000 = $7.19 W. This is double what systems in CA cost, and I doubt that prices in CT will be that much higher.

If your electric bill is $100 / mo, the year 0 energy savings column should be $1200. They used 3% increase in electric bill annually, which is slightly higher than I would use but is a defendable number.

I don't know enough about SREC's in CT or the SC Personal tax credit to comment on them at this time. It doesn't look like they are using a discount rate, so this investment is essentially being compared to stuffing cash in a mattress. While appropriate for some, you may want to consider if there are other investments you could make.

If the 0.144 entry in the first row of energy savings refers to a $ / kWh cost from the power company, this system must almost definitely be larger than 4 kW. A $1300 annual bill would be something like 9000 kWh, which you would probably need around an 8 kW array. If what was quoted is in fact 8 kW and not 4 kW, the price makes a lot more sense.

That $100/mo. electric bill probably includes a basic meter fee and taxes so the OP is probably using much less than 9000 kWh a year. But I do agree with you that the $28,756 cost for a 4000 watt system is either way too high or is bigger than 4000 watts.

The sad part is that with that small $100 monthly bill and even $0.22/kWh rates the OP really doesn't use that much electricity which usually translates to a much longer payback period. To me it looks like that estimated savings document could be flawed.

[J.P.M.]

The provided estimates, as well as your own estimates, can vary widely due to the variables used. For instance, did the installer include future energy price increases into the "energy savings" column? Probably did, but is it accurate? Maybe not; it will certainly figure into your return on investment.

I think any reasonably priced system installed properly will show a ROI much sooner than its 25 year life span. However, can the money spent on it work harder elsewhere? Determining the future value of money also uses variables that can be widely off the mark. If you find a way to make your money work harder, will you actually invest it there? A PV system that yields 2% is better than a 7% investment that is never made.

FWIW: +1.
Seems like quality systems from reputable, established vendors can be had these days that make economic sense in the 7 - 10 yr. time frame or thereabouts depending on particulars.

Comparing alternative investment choices involves a bit of knowledge of the time value of money and a lot of guesswork about the future and a bit of luck.

SWAG, grossly oversimplified example: Say $10K invested in solar saves $1000/1st yr. on an elec. bill that goes up 3%/yr., vs. $10k invested in a corp. bond that pays a constant 7% for 20 yrs. and you (hopefully and probably) get the $10K back after 20 yrs. Which of those 2 investments do you GUESS will put you further ahead $$ wise 20 yrs. down the road ? Inflation, taxes, salvage (resale) value, depreciation, maintenance, etc. etc., all can play a part in the decision, as simple or as complicated as you want or think you need.

As Ikstaack quite elegantly (IMO) suggests in a few words, 2% of something is better than 7% of nothing.

The economics are important, but "further ahead" can also include life style choices that may well be as, or more important, than whether you're a few bucks ahead of the game some years down the road. Pay your money, take your choice. Just walk in with your eyes and mind open.

[pleppik]

I don't see how these extremes can be related general historical data. Bruce Roe

My thinking is that I could use historical data to figure out to what extent the insolation this year is above or below the historical average, and use that to determine how close the installer's estimate of power production was.

So for example, if the insolation in my area for October 2014 was 105% of the historical average, and my October power production was 110% of the installer's estimate, that means the installer's estimate was a little low.

But to do this estimate I need to have a source of historical insolation data for someplace nearby.

[J.P.M.]

My thinking is that I could use historical data to figure out to what extent the insolation this year is above or below the historical average, and use that to determine how close the installer's estimate of power production was.

So for example, if the insolation in my area for October 2014 was 105% of the historical average, and my October power production was 110% of the installer's estimate, that means the installer's estimate was a little low.

But to do this estimate I need to have a source of historical insolation data for someplace nearby.

Until you know and can duplicate the installer's methods and verify their logic and accuracy, and for other reasons, I don't believe what you describe will yield valid results.

[pleppik]

Until you know and can duplicate the installer's methods and verify their logic and accuracy, and for other reasons, I don't believe what you describe will yield valid results.

Perhaps I'm not understanding something... My assumption is that, by definition, an "accurate" production estimate is one where if you have a year with average insolation, the actual measured output of the array will exactly match the production estimate. There might be a lot of complexity which goes into creating the estimate, but that doesn't matter if you have production data from an average year to match against the estimate. I'm not trying to look inside the sausage factory, just decide if the bratwurst tastes good.

Of course as we all know there is never an actual average year, since weather is weather. But if we know the average insolation and the actual insolation, that should give us the data to do a first-order calculation of how high or low the production estimate was from the actual performance in the field. If you want a more precise calculation of how far off the production estimate was, you can do the calculation on a monthly, weekly, or even daily basis (the historical average insolation is available down to the half-hour from NREL). The problem is that I haven't found a place which has public insolation data available for a nearby location over the past few months.

So am I overlooking something in this approach? I actually do have monthly production estimates from my installer (and the shading factors for each month), so I should be able to do a pretty good validation of the production estimates if I can get my hands on the insolation data.

[Carl_NH]

Pleppik,

Some of us are thicker (me included) to understand the distinction between insulation vs insolation. I think we all know what insulation is.. As I understand, "Insolation" is the exposure to the sun's rays and or the amount of solar radiation reaching a given area.

Each installation "Insolation" is going to differ and provide differing output based on shading, location orientation and more however I have found it interesting with the Solaredge site I can look at installations locally to see how we are producing on a given day and array.

At the end of the day I don't think there is a good source today for insolation data other than looking at similar systems in your area.

[Mike90250]

the distinction between :
insulation Loose fluffy stuff, or Foam, to reduce heat transfer. (Roof, wall, floor insulation)
insolation Amount of sunlight shining on the earth (on your solar panels) (SOLar)
installation Placing something in position. (panels on a roof, wires into conduit)

[sensij]

Perhaps I'm not understanding something... My assumption is that, by definition, an "accurate" production estimate is one where if you have a year with average insolation, the actual measured output of the array will exactly match the production estimate. There might be a lot of complexity which goes into creating the estimate, but that doesn't matter if you have production data from an average year to match against the estimate. I'm not trying to look inside the sausage factory, just decide if the bratwurst tastes good.

Of course as we all know there is never an actual average year, since weather is weather. But if we know the average insolation and the actual insolation, that should give us the data to do a first-order calculation of how high or low the production estimate was from the actual performance in the field. If you want a more precise calculation of how far off the production estimate was, you can do the calculation on a monthly, weekly, or even daily basis (the historical average insolation is available down to the half-hour from NREL). The problem is that I haven't found a place which has public insolation data available for a nearby location over the past few months.

So am I overlooking something in this approach? I actually do have monthly production estimates from my installer (and the shading factors for each month), so I should be able to do a pretty good validation of the production estimates if I can get my hands on the insolation data.

What you are suggesting is not unreasonable, but it takes a bit more understanding of the solar resource to get to numbers that are useful enough to matter. Among other things, weather data offered by NREL in the TMY files is not actually average, but instead are selected as "typical" months based on some weighted criteria. There is extensive documentation at NREL that talks about how this works.

As time permits, I've been working through the programming necessary to take data from a local weather station and use it in the manner you are suggesting. I agree that there is value is knowing whether a particular time period is more or less sunny than the reference period used in PVWatts, for instance.

The installers estimate in this case is an unknown quantity. You could assume the installer is using PVWatts, but even there, some assumptions may not be selected as default. In other words, comparing current weather to the reference weather in PVWatts has a chance of being useful, but without knowing how the installer created the estimate, trying to compare to that is a losing battle.

One problem is that "insolation" on a particular PV array is difficult to model accurately even when a pyranometer is nearby. The meter will typically detect / report global horizontal irradiance, but the power produced by the array will be a function of how much of that irradiance is direct and how much is diffuse, among other things. You can get order of magnitude estimates by looking at nearby weather station data... you can go here and search for a weather station near you that reports solar irradiance. However, if you want numbers good enough to really determine if your array is producing what would be expected, it will take more work than that.

Careful study of the NREL documentation, and the papers it references, would be one way to learn more about it. I'll warn you, many of the equations in the PVWatts and SAM documentation contain typos, and the documents are useful as a roadmap, but not as a set of keys to the car. Another approach is to buy a reference book on residential solar. Duffie and Beckman is a starting point often recommended by J.P.M., who, from what I've seen, has demonstrated more solar knowledge in this forum than most other members combined.

Edit: I meant also to provide a link to this thread, in which my own ignorance of this subject is on display.

[bcroe]

Perhaps I'm not understanding something... My assumption is that, by definition, an "accurate" production estimate is one where if you have a year with average insolation, the actual measured output of the array will exactly match the production estimate. There might be a lot of complexity which goes into creating the estimate, but that doesn't matter if you have production data from an average year to match against the estimate.

OK, I don't see how you will get that exact match, keep us posted. With the result, what is
achieved? My own estimate of insolation is based on production, not the other way around.
Today was the first sunny day near the shortest day, and I observed today shading on the
east end lasted well into the day; cost 5 KWH lost on that end. Some day most of my panels
will probably be relocated, since I'm not willing to cut too many more trees. Bruce Roe

[J.P.M.]

Perhaps I'm not understanding something... My assumption is that, by definition, an "accurate" production estimate is one where if you have a year with average insolation, the actual measured output of the array will exactly match the production estimate. There might be a lot of complexity which goes into creating the estimate, but that doesn't matter if you have production data from an average year to match against the estimate. I'm not trying to look inside the sausage factory, just decide if the bratwurst tastes good.

Of course as we all know there is never an actual average year, since weather is weather. But if we know the average insolation and the actual insolation, that should give us the data to do a first-order calculation of how high or low the production estimate was from the actual performance in the field. If you want a more precise calculation of how far off the production estimate was, you can do the calculation on a monthly, weekly, or even daily basis (the historical average insolation is available down to the half-hour from NREL). The problem is that I haven't found a place which has public insolation data available for a nearby location over the past few months.

So am I overlooking something in this approach? I actually do have monthly production estimates from my installer (and the shading factors for each month), so I should be able to do a pretty good validation of the production estimates if I can get my hands on the insolation data.

Pleppik: First off - Welcome back. Where you been ?

I think there may be a lot you're not understanding, at least the not in the same way I think I understand it, starting with how meteorological data for solar simulations is generated and used for simulations. But, rather than get into a protracted, and for this venue inappropriate discussion of solar resource assessment, which will probably get me into a pissing match with you or someone else, and also get me accused of long winded and condescending explanations, which for a lot of reasons I'd rather avoid just now, I'll try this:

1.) To get an accurate production estimate of your existing system, you will need to be able to measure what goes in the front end, and what comes out the back end. You will also need that estimate to gauge the accuracy of any estimating schemes, vendor or otherwise. You can do this on an hourly, or yearly basis or anything in between. You will need some instrumentation (cost ~ $1K). With that instrumentation you can do your own estimates of system efficiency if you choose and, depending on the time resolution of the questioned software, estimate its accuracy.

2.) Looks like in your post you first define what an accurate production estimate is and then in the next paragraph state that what is needed for an accurate production estimate is impossible to get. I'm not sure where I can go with that.

3.) You will need to know how the vendor's production estimate was done before you can use the same algorithms/software and methods to duplicate the method under other (actual) conditions for purposes of verification.

4.) Alternately, and probably faster and cheaper, get and learn how to use SAM and compare that output to what the vendor gave you. Chances are the vendor used PVWatts, SAM or some other canned software anyway. If so, find out what the vendor used, verify the input and do the same with the same software for an estimate of yearly output. Any year actual output is probably ~ +/- about 10% compared to the software estimate due to natural variation in the weather year/year.

Take what you want of the above, leave the rest.

Happy New Year.

J.P.M.

[pleppik]

Pleppik: First off - Welcome back. Where you been ?

Thanks! Life got in the way...

I think there may be a lot you're not understanding

Probably true.

2.) Looks like in your post you first define what an accurate production estimate is and then in the next paragraph state that what is needed for an accurate production estimate is impossible to get. I'm not sure where I can go with that.

I was trying to come up with a way to compare the estimate to real-world production data to determine how accurate the estimate was. So the production estimate is an input, not an output.

3.) You will need to know how the vendor's production estimate was done before you can use the same algorithms/software and methods to duplicate the method under other (actual) conditions for purposes of verification.

This is the part I'm not understanding. If I'm taking the estimate as an input and trying to compare it to real-world data, why do I need to know the details of how the estimate was produced? It either is close to the actual production or it isn't.

Of course, I realize that with weather being weather it's not reasonable to expect five months' of production data to be within 20% of the original estimate. That's why I was thinking about how to use insolation data (which I don't have anyway) to correct for this and come up with an estimate of the accuracy of the estimate.

4.) Alternately, and probably faster and cheaper, get and learn how to use SAM and compare that output to what the vendor gave you. Chances are the vendor used PVWatts, SAM or some other canned software anyway. If so, find out what the vendor used, verify the input and do the same with the same software for an estimate of yearly output. Any year actual output is probably ~ +/- about 10% compared to the software estimate due to natural variation in the weather year/year.

Before signing the contract, I actually did plug in the installer's parameters and shading factors into PVWatts to verify that I got the same results. And they were within a percent or so.

But all that tells me is that the two different tools agree with each other, not whether they agree with reality. For example, if inputs like the efficiency of the inverters were over/underestimated, the tools would agree with each other but not match reality.

After only five months of non-average weather, at this point the estimate could have been wrong by 20% in either direction and I wouldn't know. I could of course just wait five years and average my annual production over that time to get a pretty solid answer as to whether the production estimate was accurate. But I'm not that patient.