Winter solstice efficiency

[Mike90250]

Try to stay warm and dry Mike.

That's what the masonry heater is for. With company visiting, we got it up to a balmy 70F indoors, 45% RH, almost too dry. Masonry heaters are just so awesome, I can't believe it. makes a $4000 wood stove look like a Easy-bake-oven ! I carry wood in 5 gal buckets, clean & easy, 4 buckets a day is all it takes, 2300sf, 2 stories. Takes 2 days to heat up.

https://www.niftyhomestead.com/blog/masonry-heater/

Mine:
IMG_20151220_162550330.jpg

[peakbagger]

What are the azimuth's ? Something seems odd. The 8/12 pitch is ~ 33.7 deg. tilt and thus not that far off a 30 deg. tilt. To my experience it's not likely that any boost from snow would increase output by 1450/500 = 290 %. I've seen and estimated ~ 10-15 %. Additionally, the roof will still see some enhancement from the surroundings.

There is a confusion on terms. Both arrays face south. The pole mount is tilted 30 degrees off vertical (60 degrees off a horizontal plane). The angle you calculated for the roof is tilt off horizontal. Therefore its about a 26 degree difference between the array angles. The suns altitude on 12/21 in northern NH at noon is 22.2 degrees. On the summer solstice its 69 degrees .

[DanS26]

Just had a completely cloudless day today and did 3.65kWh/kW efficiency. This agrees pretty well with PVWatts. So I guess my system is on par with what can be expected from a properly operating system. I'll check it again later in 4-6 months when things get more exciting.

DanS26 , that 7.521 number has got to be some sort of record. where are you and what is your set up?

Dave

My array tilt is 32.5 deg.....so that means that at my latitude here in southern Indiana the first few days of April and September the sun is perpendicular to the array face. If the atmosphere is clear of moisture (ie clouds) and pollution (ie dust) then the system will produce at maximum. Wherever possible I used the golden ratio in my design and build.....it is pleasing to the eye.

I run a matched set of ground mount 36 panel arrays aligned to south using the plumb bob/solar noon method.....yes old school. I have experienced 0.7% degradation per year per panel which I had expected. Panels are 235 and 245 watt Kyocera 60 cell panels matched to two Fronius 7.5kW string inverters.

[bcroe]

I run a matched set of ground mount 36 panel arrays aligned to south using the plumb bob/solar noon method.....yes old school. I have experienced 0.7% degradation per year per panel which I had expected. Panels are 235 and 245 watt Kyocera 60 cell panels matched to two Fronius 7.5kW string inverters.

I'm running the same inverter plant here. Panels are a mixed bag, some 60 cell, some 72, but every string has 720 cells in series. The older
technology panels don't put out quite as much, but matched types are used in each string. 3 panels of unknown origin were initially swapped,
before I got every string optimized. Older panels here stay in service, still making their contribution. The clamp on ammeter on a sunny day
reveals issues; guess at some point a more precise (voltage tap) long term check should be made.

Google earth reveals my installer missed the direct south facing direction by quite a bit; easy to do when lots aren't rectangular here in the
Wild West. If Google ever is updated (from Aug 2013), it will reveal that much of the original shading is gone. A rework of the panel supports
will include quick and easy one person seasonable adjustments, and much better alignment based on my recent survey work.

I'm not sure what the term EFFICIENCY means in these parts; PV production here is tied to the (ever cloudy) weather. Even with my summer
tilt I managed to hit 67KWH this week; probably would approach 80 with a near vertical panel elevation. However I think the most impressive
thing it did this week was on Wed. With complete overcast all day, it managed 22KWH. Not huge, but at least the inverters had something to
do. The best day to date did 148 KWH through 15KW of inverters, or 9.8 sun hours. The project won't be done till it hits 10, but the real
improvements will be in partly cloudy output over all seasons. And with less labor on my part. Bruce Roe

[Sunking]

Efficiency has little to do with it if anything at all. Nature of the beast and Sun Hours.

[bcroe]

That's what the masonry heater is for. With company visiting, we got it up to a balmy 70F indoors, 45% RH, almost too dry. Masonry heaters are just so awesome, I can't believe it. makes a $4000 wood stove look like a Easy-bake-oven ! I carry wood in 5 gal buckets, clean & easy, 4 buckets a day is all it takes, 2300sf, 2 stories. Takes 2 days to heat up.

That certainly does appear to be an efficient design, for those using a lot of wood. I never see such things here. How would it compare
to the latest outdoor furnaces with the insulated water pipes? Bruce Roe

[J.P.M.]

What are the azimuth's ? Something seems odd. The 8/12 pitch is ~ 33.7 deg. tilt and thus not that far off a 30 deg. tilt. To my experience it's not likely that any boost from snow would increase output by 1450/500 = 290 %. I've seen and estimated ~ 10-15 %. Additionally, the roof will still see some enhancement from the surroundings.

There is a confusion on terms. Both arrays face south. The pole mount is tilted 30 degrees off vertical (60 degrees off a horizontal plane). The angle you calculated for the roof is tilt off horizontal. Therefore its about a 26 degree difference between the array angles. The suns altitude on 12/21 in northern NH at noon is 22.2 degrees. On the summer solstice its 69 degrees .

Understood. My bad and my apologies. I'll read more carefully.

Still, Something doesn't seem quite right on the difference in outputs if both are under approx. the same clear sky conditions. Assuming zero reflectance from the surroundings as a base case, PVWatts max. hourly output for 12/20 is ~ 681 W/D.C. kW. at a 33.7 deg. tilt and 180 az. roof mount. The max. hourly output for the same date/time for a 60 deg. pole mount, 180 deg. az. is ~ 805 w/D.C. kW. Both use 10 % system losses.

I've spent a lot of time in years past , when living in a snowy climate (Buffalo) studying enhancement techniques, particularly the (free) irradiance enhancement due to snow, and ways to estimate (calculate) performance enhancement from snow as f(array tilt, angle of incidence, atm. clearness index, snow/surface reflectance characteristics, etc.). I gotta tell ya', take it or leave it, and with all possible respect, a 1,540/500 = 290% improvement for the pole mount under the approx. same irradiance conditions based on an idea that reflection from snow has that much to do with it is a pretty long stretch based on what I think I know. Sounds like the irradiance conditions when you looked at the roof array output were a lot less than a clear sky, particularly when looking at a PVWatts hourly output est. for what is a nominally pretty good solar day.

High tilt angles and a lot of clean snow can improve P.O.A. irradiance by 20 -30% based on stuff I've actually measured (and thus, BTW, a lot less than what are, IMO, overly rosy and optimistic claims of ~ 70% enhancement by some folks - that's B.S. from what I've measured, but maybe I'm wrong). Regardless of whether my %ages are off or not, the actual performance improvement I've measured is usually about half that much as the irradiance enhancement due to the (usually) higher incidence angles of the reflected radiation vs. the beam solar radiation, as well as the higher surface reflectance of the glazing caused by those higher incidence angles, those two effects being multiplicative.

[J.P.M.]

That's what the masonry heater is for. With company visiting, we got it up to a balmy 70F indoors, 45% RH, almost too dry. Masonry heaters are just so awesome, I can't believe it. makes a $4000 wood stove look like a Easy-bake-oven ! I carry wood in 5 gal buckets, clean & easy, 4 buckets a day is all it takes, 2300sf, 2 stories. Takes 2 days to heat up.

https://www.niftyhomestead.com/blog/masonry-heater/

Mine:
IMG_20151220_162550330.jpg

Mike: A question/comment or several:

Does it take 2 days to heat the dwelling or 2 days for the assembly to heat up ? If the assembly, I bet it takes about as long to cool down, +/- some.

Any heat exchange surface inside the masonry/flue ?

Any allowance for outside combustion air ? Not trying to preach to the choir here but combine that with a tighter building envelope and interior R.H. will probably go up.

[Mike90250]

That certainly does appear to be an efficient design, for those using a lot of wood. I never see such things here. How would it compare
to the latest outdoor furnaces with the insulated water pipes? Bruce Roe

The efficiency is close to 95%. Often, the smoke exiting the chimney is so cold, it rolls down the roof, and does not rise into the air. And since they are burned full thottle, all vents open for the 90 min or so of the burn, they are as clean (and cleaner after year 2) as the best catalytic heaters..
Each time I have the flue cleaned, it's so clean, I add another year to the cycle. The sweeps can't belive how clean it burns, since there is no "try to make smoulder all night" The brickwork holds the heat for up to 48 hours. In warmer weather, we just fire it every 36 hours, to keep the chill off.

http://www.mha-net.org/

[J.P.M.]

My array tilt is 32.5 deg.....so that means that at my latitude here in southern Indiana the first few days of April and September the sun is perpendicular to the array face. If the atmosphere is clear of moisture (ie clouds) and pollution (ie dust) then the system will produce at maximum. Wherever possible I used the golden ratio in my design and build.....it is pleasing to the eye.

I run a matched set of ground mount 36 panel arrays aligned to south using the plumb bob/solar noon method.....yes old school. I have experienced 0.7% degradation per year per panel which I had expected. Panels are 235 and 245 watt Kyocera 60 cell panels matched to two Fronius 7.5kW string inverters.

FWIW, the way I understand it, those are probably the two days of the year that have the two instants of time when the beam radiation on your arrays is perpendular to the arries. That can happen at only 2 instants of time per year, not for an entire day, and, depending on array orientation, may not happen at all.

Even if oriented such that perpendicularity does happen 2X/year, that's no guarantee that such times, or days containing such times, even if identically cloud free, will be days of maximum production.

On the golden ratio: Apparently, a lot of PV panel mfgs. also like the golden ratio, or something pretty close to it given a lot of panel dimensions being ~ in the golden ratio.

FWIW, I always thought a panel shape (2 shapes actually) based on the Penrose concept of darts and kites might have some application for irregular roofs or as an easier way to accommodate oddly shaped but available surfaces. The golden ratio is contained in the relative dimensional ratios and areas of those shapes.

As an aside, and from the who gives a crap dept.: Apparently the folks who make Kleenex thought something similar (?). As Roger Penrose (or his wife as the story politely goes)discovered when perhaps sitting on the hopper and noticed (saw ?, felt? ) that the crap wrap from that company had a Penrose tile pattern. Penrose sued for patent infringement. At trial, the co. claimed it was done to reduce "bunching", whatever that means. Anyway, Penrose had patented a math concept, sued, and even though the paper pattern was different, won in court.

Question: How do you determine your annual degradation ?

[DanS26]

...........

Question: How do you determine your annual degradation ?

I had a snail trail issue show up the first year on my 235 watt panel array. So I was keen to know whether those panels would degrade faster than normal. Since I had a matched set of 245 watt panels without the defect that had the exact same tilt and orientation, I could easily compare the output of the two arrays. My panels are over 5 years old now and the output ratio of the two arrays has not changed, so I am reasonably certain that the snail trail issue has not affected production.

So here is how I determined degradation used in my analysis. Kyocera said I should expect the panels to degrade 0.71% per panel per year. But I wanted to confirm that number with my own observations. I used PVWatts to determine predicted production for each array, then I degraded the PVWatts prediction by 0.71% each year and make a comparison to monthly and annual production. Some periods are better, some worse.

So total production to date since November 2011 is 95.026MWh compared to the degraded PVWatts prediction of 95.238MWh. Slightly behind but not much. I attribute the lower production to the cooler and wetter than average weather here in the Midwest for the last two years.

[J.P.M.]

I had a snail trail issue show up the first year on my 235 watt panel array. So I was keen to know whether those panels would degrade faster than normal. Since I had a matched set of 245 watt panels without the defect that had the exact same tilt and orientation, I could easily compare the output of the two arrays. My panels are over 5 years old now and the output ratio of the two arrays has not changed, so I am reasonably certain that the snail trail issue has not affected production.

So here is how I determined degradation used in my analysis. Kyocera said I should expect the panels to degrade 0.71% per panel per year. But I wanted to confirm that number with my own observations. I used PVWatts to determine predicted production for each array, then I degraded the PVWatts prediction by 0.71% each year and make a comparison to monthly and annual production. Some periods are better, some worse.

So total production to date since November 2011 is 95.026MWh compared to the degraded PVWatts prediction of 95.238MWh. Slightly behind but not much. I attribute the lower production to the cooler and wetter than average weather here in the Midwest for the last two years.

Thank you.

Take what follows for FWIW. Or scrap it.

I believe I understand your logic, but based on what is, actually and mostly modeled data, and as the PVWatts documentation will tell you, there's probably +/- a lot more variation in what PVWatts estimates for long term average output vs. actual annual output than your claiming for your precision. I believe that makes the method you describe invalid for what you're claiming it does.

The PVWatts model mostly uses TMY data, which, when you dig down through the documentation, you'll find to be mostly modeled data itself (and with not a small amount of the solar irradiance estimates coming from good eyeball estimates of % of cloud cover in 10% increments from ground observers), with some actual, historical pyranometer data as noted in the TMY documentation, but a lot less of that data being from actual instrument measurements than most folks know or assume. See the TMY documentation for info on modeled data. That's one of the things besides weather variation that limits the PVWatts estimated long term average output to something like +/- 10% for any one year.

I'm not saying the TMY data is not representative of long term average output. It seems fit for that purpose. Problem is, most folks seem to not understand what that purpose is, and/or the model's limitations, and wind up often using it for unintended purposes, and come to what may well be incorrect conclusions based on incorrect application.

All the same to me, not my error, money or system. Just that a lot of otherwise intelligent folks make decisions based on bad assumptions borne of ignorance with sometimes less than good outcomes.

For the present purpose of estimating annual system performance degradation, what I am saying is that, to use a model that claims +/- 10% yr./yr. variation to verify an annual degradation of the order of 1% can't be done with much expectation of getting a realistic and reliable number.

As for how close PVWatts may get (or can be forced to be equal) to an actual recorded average of several or many years of actual output, as your output shows as being quite similar to a PVWatts estimate, one simple way to get any level of agreement sought is to manipulate the system loss parameter in PVWatts to force its long term estimate for any modeled system, and any actual measured average output from an actual system that's been modeled to be identically the same.

If that's done however, what you will get is probably little better than a dart throw of how the solar weather for any particular time period, such as a year, might have varied in comparison with the TMY weather (and more recently also from something called "Weather anywhere" data from an outfit started by one of my mentor's mentors), including not only insolation variations but also ambient air temp., wind vector and other variables that PVWatts uses for its estimate of long term average output.

However, still buried in that comparison along with all the rest of the variation, and without a lot more instrumentation, probably inseparable from it, is the annual degradation of the type you're looking for.

By way of some comparison:

Unfortunately, I do not have 5 years of data as you do - yet. I do however, have 1,164 continuous days of array output data gathered and logged in 5 min. intervals. I've also got roof weather data for ambient temp., wind vector, dew point and horizontal irradiance, and other data atmospheric data gathered and logged in 1 min. intervals. Over those 1,164 days, my array has produced 29,344 kWh., or an "average" of 9,202 kWh/365 day year for the 5.232 S.T.C. kW array.

If I force the PVWatts output to match the actual average annual output since startup, the system loss parameter needs to be forced to ~ 9.6 %.

But, that 9.6 % accounts for all other variables such as array fouling, which seems to vary between 0% (clean) and ~ 4-5 % depending on rain history and what I'm dong in the way of cleaning to measure that fouling, late afternoon shading of ~ 3% - 5 % or so as f(time of year), as estimated from SAM and my own software, and array degradation, not only for the first year burn in which I roughly guess/est. at ~ 2.5 % or so, but also subsequent panel degradation which I can only and very roughly SWAG at perhaps something less than the 0.4 % as listed on the panel spec sheet, and which, BTW, may or may not be a linear f(time).

If I keep a running total of the prior 365 day's system output, for 1,164-365 days = 799 days - which I've done, I find that the average of each of those 799 packets of running 365 day total array output is 9,300 kWh/yr. (and, I note, higher than the 9,202 kWh/yr. for the 1,164 day total, making the 1st year's output probably less than subsequent time periods, and perhaps counterintuitive to what might be assumed, especially considering the probable and mfg. warranty addressed 1st yr. burn in). The high for any of those 799 days of 365 prior day totals is 9,602 kWh/yr., the low, 9,033 kWh/yr., the population std. dev. is 168 kW/yr. (making the data not quite symmetric and a bit leptokurtic for those still reading or awake).

The point is, there's more variation in my output, and by some assumed similarity, yours, than can be handled or accounted for mostly, but not entirely due to weather, with that weather caused variation in annual output mostly buried/mixed in, and inseparable from, the other causes of variation, including annual cell/panel degradation.

What you and I have is the old situation of having more unknowns than independent methods to find or guess at them.

[DanS26]

I did leave out a few details about my analysis that you touched on. I used the first 12 months of panel production to true up to PVWatts....in effect using the system loss factor as you mentioned. But from that point forward I have not touched or change the setup parameters. That would make the analysis susceptible to my personal bias. It is what it is and I can accept that. I watch this data because I want to identify a misbehaving panel, string or inverter immediately.

I agree with you in the short run.....say less than 5 years data. But the longer my method correlates with the PVWatts data the more convinced I am that the 0.71% degrade factor is good approximation of my actual experience.

[J.P.M.]

I did leave out a few details about my analysis that you touched on. I used the first 12 months of panel production to true up to PVWatts....in effect using the system loss factor as you mentioned. But from that point forward I have not touched or change the setup parameters. That would make the analysis susceptible to my personal bias. It is what it is and I can accept that. I watch this data because I want to identify a misbehaving panel, string or inverter immediately.

I agree with you in the short run.....say less than 5 years data. But the longer my method correlates with the PVWatts data the more convinced I am that the 0.71% degrade factor is good approximation of my actual experience.

Understood. Thank you for the response.

I've got no problem with 0.71% number, or how you arrived at it. It may very well be dead nuts on. But the weather has changed more for you year/year than 0.71%/yr., and in both directions up and down. Also, unless you've got a way to record insolation and convert global horizontal to plane of array irradiance, any panel degradation, whatever it may be will be impossible to flag, much less measure or estimate with any reliability or precision.

But whether it's 0.71 %, or zero some years, or 1.5 % other years, part of my point, besides highlighting the limitations of PVWatts and other modeling techniques, was that, given the situation, and even with much more sophisticated equipment than you, I or most anyone may have, including some mfg. and testing outfits, and considering common residential situations, the number of variables, the uncertainties, and the small differences in degradation we're looking for, chances of fishing a reliable number out of the noise may be next to impossible, and even if possible, because of it's relative magnitude, unimportant.

A respectful suggestion: If you're interested, take a look at SAM from NREL. Sort of like PVWatts on steroids. Part of how I convinced myself that PVWatts was usually underestimating possible output, and that one way to deal with that underestimate was to fool around with the PVWatts system loss parameter to get ouputs from PVWatts closer to that of SAM, which, IMO only, is a better model.

I appreciate and share your ability to identify small changes in system parameters. Example: For the last few days, I've got a string whose voltage is running a couple of volts lower than usual. Something to keep an eye on.

[DanS26]

Well my point was that I am not particularly interested in any data point. I don't care if PVWattts is high, low or in between. Same for the SAM system.

I am not too interested in how the points on the line are calculated as long as they are calculated in a consistent manner and not subject to arbitrary adjustments as time goes on.

What I am interested in...... is the slope of the line. I want to compare the slope of the predicted production (even if it is over or under) to the slope of my actual production over a long period of time.