Standard clear clear in San Diego county

[J.P.M.]

This is a bit parochial, but for those in San Diego county who may want to know what their array might put out on what's close to a standard clear day around here for a time of the year when array output is high due to more perpendicular incidence angles of solar radiation, today, 04/29/2017 was about as standard a day in terms of what might be expected for a clear day's irradiance as is likely to happen.

Using something called the HDKR model of clear sky irradiance, today's modeled day long global horizontal insolation at my location in zip 92026 was ~ 7,794 W/m^2 for the total day. The Davis instrument that records horizontal insolation about 4 ft. north of my array recorded 7,735 W/m^2 for the day, or ~ 7,735/7,794 = .992 of a "standard" clear atmosphere's insolation for this date. Using another common clear sky model - the "Bird' model (after Richard E. Bird) - the modeled clear sky daylong horizontal insolation comes up 7,825 W/m^2. The numbers for both models are corrected for elevation above sea level and barometric pressure. The Bird approx. is also corrected for estimated precipitable water vapor in the atmos. as f(dew point) which may be one reason why it gave a bit higher daylong total, this being a rather dry day w/ dew points of ~ 20 F.- 30 F. at my location, about 20 miles inland and ~ 1,400 ft. above sea level. The Davis numbers are corrected for sensor temp. The Davis sensor is 9 months into it's 2 yr. calibration cycle.

This time of year in this area is one of two times/yr. when a lot of arrays have their minimum solar incidence angles (the other time being early/mid Aug.). If an array is mostly south facing and tilted something like 20 deg. as is somewhat common around here, the angle of incidence of beam radiation will probably be within ~ 5 or so deg., +/- some of zero around solar noon. Solar noon is about 12:45 or so P.D.T. this time of year in San Diego.

So, if it's sunny and clear (and no clouds) around 12:45 P.D.T for the next few days as expected, and your array is mostly south facing at ~ 20 deg. tilt (as many around here are), check your array's instantaneous output at about 12:15 - 1:15 P.D.T.or so. and you'll get a decent if still ballpark approx. of you array's max. instantaneous output, particularly if the weather is on the cool side and/or it's a bit windy. I've found it a nice number to have around.

Take what you want of the above. Scrap the rest.

[CharlieEscCA]

I've been peaking at about 6.65 kW the last several days. This is with two panels getting a little bit of shade (shown in layout). Still very happy with my first 5 days.

Based on production vs my TOU consumption, I've opted to switch to the DR-SES tariff (takes affect at next bill). This will lock in the current TOU periods for five years. Some future owner of my house will thank me -- it looks like we are buying a much larger vacant property in VC; build in 2018 or 2019. Wasn't planned when I signed the solar contract March 1, but stuff happens.

[J.P.M.]

I've been peaking at about 6.65 kW the last several days. This is with two panels getting a little bit of shade (shown in layout). Still very happy with my first 5 days.

Based on production vs my TOU consumption, I've opted to switch to the DR-SES tariff (takes affect at next bill). This will lock in the current TOU periods for five years. Some future owner of my house will thank me -- it looks like we are buying a much larger vacant property in VC; build in 2018 or 2019. Wasn't planned when I signed the solar contract March 1, but stuff happens.

FWIW, under current DR-SES tariff, and NEM 2.0, your array ought to offset something like ~~ $4,020/yr. or so of the electric bill associated with that array, LESS any production loss from shading, and with about 6% - 8% annual system production variation due to weather, and maybe about 0.5% annual loss due to system deterioration.

If the shading is more predominant in the P.M., expect a greater revenue hit (i.e. lower annual system production and total annual revenue) because that shading will be during peak revenue (or billing) time.

So, whatever bills the occupants incur, the PV system will eliminate probably something less than ~ $4K/yr. of it, less any shading penalty, +/- about 6-8 %/yr. or so, depending on weather. .

At current T.O.U. -SES tariff rates, the average $/kWh rate of your system's production before any shading allowance is about $0.274/kWh, again, before any shading penalty. Afternoon shading will tend to decrease that average.

That total annual offset, of course, will also change as tariff rates change. That, and the rest of the variables mentioned above and others, make all this a SWAG, but, it's maybe an educated number.

[sensij]

It is a shame the 29th was so hot. Last year was cloudier, with a peak temp on my rooftop just uphill form the array of 73.2 deg F. This year it hit 87.8 deg F, on the same sensor. Total insolation measured (uncorrected) was 7515 W/m^2 for the day, adjusting for the palm tree shade that swipes across the sensor in the afternoon would give 7711 W/m^2 shade free.

Max array output on the 29th was 2577 W (0.826% of DC rating). It was higher on the 20th, at 2630 W. That was another fairly clear day but cooler, peak temp of 81 deg F. 2% more power from 3% less irradiance suggests the penalty for the high temp was about 4.8%. Published temperature coefficient of power for these panels is -0.43%/deg C, which would translate to a 5.3% power hit under these conditions. I'm happy to see that kind of agreement for this coarse level of analysis.

April 20th to 21st was actually a major shift in conditions, with the peak temp jumping from 81 deg to 93.9 deg F. Array power took a beating, as expected. There was some relief when the breeze picked up and shifted to be from the south (which is probably the most effective direction for cooling my array); the peak temp was at 1:25 pm, but rooftop temp dropped 10 deg F in 30 min after that, with array power boosting from 2396 W to 2501 W even as irradiance was falling.

(All power reported for 2017 is from the inverter, I don't have a revenue grade meter on this setup anymore. My estimate is up to 1% error for clear conditions)

[CharlieEscCA]

FWIW, under current DR-SES tariff, and NEM 2.0, your array ought to offset something like ~~ $4,020/yr. or so of the electric bill associated with that array, LESS any production loss from shading, and with about 6% - 8% annual system production variation due to weather, and maybe about 0.5% annual loss due to system deterioration.

If the shading is more predominant in the P.M., expect a greater revenue hit (i.e. lower annual system production and total annual revenue) because that shading will be during peak revenue (or billing) time.

So, whatever bills the occupants incur, the PV system will eliminate probably something less than ~ $4K/yr. of it, less any shading penalty, +/- about 6-8 %/yr. or so, depending on weather. .

At current T.O.U. -SES tariff rates, the average $/kWh rate of your system's production before any shading allowance is about $0.274/kWh, again, before any shading penalty. Afternoon shading will tend to decrease that average.

That total annual offset, of course, will also change as tariff rates change. That, and the rest of the variables mentioned above and others, make all this a SWAG, but, it's maybe an educated number.

Your numbers line up with the spreadsheet I put together -- I took the PVwatts daily / hourly CSV file, and added columns to calculate revenue $ using the current summer / winter monts, and the peak, semi-peak, off-peak rates, i.e. I did the full calculations you suggested a long time ago in the TOU thread I started a couple of months ago.

I didn't chose DR-SES initially because I wanted to see how the actual numbers looked and in what time of day -- the fear was that if I ended up with high peak usage, then TOU would not be good. But looking at my consumption, and when A/C is likely to be run (when it is run), I'm now confident getting "peak" exported credit for the majority of the May to Oct months will be a plus for my usage pattern.

I'm surprised you, or anyone else, hasn't weighed in / commented on the installation thread I put up yesterday. Sorry for all the pictures, but I thought they might be of use to someone. In any event, to me, it was a nice, clean, professional looking installation. Some affirmation of that is always welcome Or even, some critique -- feedback is always welcome.

[J.P.M.]

It is a shame the 29th was so hot. Last year was cloudier, with a peak temp on my rooftop just uphill form the array of 73.2 deg F. This year it hit 87.8 deg F, on the same sensor. Total insolation measured (uncorrected) was 7515 W/m^2 for the day, adjusting for the palm tree shade that swipes across the sensor in the afternoon would give 7711 W/m^2 shade free.

Max array output on the 29th was 2577 W (0.826% of DC rating). It was higher on the 20th, at 2630 W. That was another fairly clear day but cooler, peak temp of 81 deg F. 2% more power from 3% less irradiance suggests the penalty for the high temp was about 4.8%. Published temperature coefficient of power for these panels is -0.43%/deg C, which would translate to a 5.3% power hit under these conditions. I'm happy to see that kind of agreement for this coarse level of analysis.

April 20th to 21st was actually a major shift in conditions, with the peak temp jumping from 81 deg to 93.9 deg F. Array power took a beating, as expected. There was some relief when the breeze picked up and shifted to be from the south (which is probably the most effective direction for cooling my array); the peak temp was at 1:25 pm, but rooftop temp dropped 10 deg F in 30 min after that, with array power boosting from 2396 W to 2501 W even as irradiance was falling.

(All power reported for 2017 is from the inverter, I don't have a revenue grade meter on this setup anymore. My estimate is up to 1% error for clear conditions)

Thank you. From what you report and similar to prior output comparisons between our two arrays (and Davis instruments), the agreement seems quite good. I appreciate the confirmation of my numbers. The agreement seems better than I might expect, but, actually and truth to tell, I'm not surprised.

As another, and what I believe important, point, I believe it's worthy of note that you report output for an array using Canadian Solar panels, and I report output for an array using Sumpower panels. The array orientations are similar, and the locations are about 20 or so miles apart. However, the output for most days, particularly "clear" days, in terms of kWh/day per STC Watt is quite similar.

All that lends further credence to the idea, and my sometimes adamant statements that most any decent panel, installed in a professional manner, be it Sunpower, Canadian Solar, LG, etc.etc., in the same location and orientation, will produce about equal output., making any initial premium for Sunpower equipment beyond maybe about 5 % or so to account for a perhaps lower annual degradation, a waste of money, except and until someone hangs a value on bragging rights.

My max. instantaneous array output, as reported by the inverter display on the 29th was 4673 W. At 1306 P.D.T. P.O.A. irradiance at that time is est. as 1,059 W/m^2 based HDKR model on a recorded and sensor temp. corrected GHI of 1,005 W/m^2.

Part of the reason my array's instantaneous output /STC Watt (4,672/5,232 = 0.893) is higher than yours at this time is probably due to the testing I'm doing just now. Between 04/06/17 and 06/12/17, I'm cleaning the array each day in the early A.M., and measuring a bunch of stuff at the time of min. incidence angle each "clear" day, that time being ~ 1306 P.D.T. on the 29th and shifting around a couple of minutes or so over those dates. The hope is to get at least 30+ days of instantaneous data at close to zero incidence angle (w/cos of the angle of incidence >099 ), get better or at least a better guess at some est. of the distribution of clean, instantaneous array performance as f(ave. instantaneous array temp.) while holding as many of the variables as steady as possible, and also use the data as part of the continuing quest for some better/updated SWAG as to the rate at which the array on my roof fouls, and also a bunch of other stuff too boring to describe.

[CharlieEscCA]

... I'm cleaning the array each day in the early A.M. ...

Would you be so kind as to share your cleaning methods? Just spray off? Any soap or cleaning solution? Do you squeegee or just let air dry? Water straight from the hose?

I just ordered a water fed pole and brush for my ground mount from Amazon -- trying to figure out if I need to squeegee or use a filter cartridge or just let the water dry after rinsing.

I'm going to do it once a month and call it good enough

Thanks.

[J.P.M.]

Would you be so kind as to share your cleaning methods? Just spray off? Any soap or cleaning solution? Do you squeegee or just let air dry? Water straight from the hose?

I just ordered a water fed pole and brush for my ground mount from Amazon -- trying to figure out if I need to squeegee or use a filter cartridge or just let the water dry after rinsing.

I'm going to do it once a month and call it good enough

Thanks.

I've been told I attempt to instruct folks on how to build a clock when they ask me what time it is. I'll attempt brevity here and probably fail at the brevity part..

Before I start - NEVER spray a hot array with water. I do mine in the early A.M., before 0700 hrs. or so, provided the sun is up.

1.) For a very soiled (fouled) array that's not been cleaned for several months or longer, I use dish soap - not dishwasher detergent - just mild soap.

2.) I have a soft cloth on a 12 ft. pole I got at Big Box.

3.) I rinse the array, hopefully removing the major dirt with simple tap water from a hose. Then, use the soap/H2O from a bucket and brush the array with the soft cloth.

4.) I rinse the array with tap water from the hose.

5.) I check observe for any missed/persistent dirt and repeat steps 3 & 4 as appropriate.

6.) I let the array air dry.

After measuring array performance for ~ 3 yrs. or a bit less, I've found my array fouls at a rate that, very roughly, decreases performance approx. 1% per week if it doesn't rain. Santa Anna winds or wind and high dew points with clear nite skies, or a lot of other galloping variables or weather conditions make that 1 %/week no more than a very rough appox. of average, not rainy, long term average performance with a lot of temporary variations/up & downs.

Rain will clean an array. How much is hard to quantify. A "good" rain, say 1/2 " or so or more, seems to restore ~ 2/3 to 3/4 of my array's clean performance. Sprinkles to ~ 0.05" - 0.10 " or so of precip. after no rain for,say, a month seem to turn accumulated dust that might otherwise be partially removed by wind turn into mud that gets new dust on top of it and makes fouling more persistent and increasing in rate.

Best I can figure, and the jury's still out on this one for me, if it doesn't rain, there seems to be some indication in my mind and observation, that the 1 %/week rate of fouling may tend to become asymptotic and begin to level off some, maybe after something like, very roughly and subject to a lot of very local conditions and terrain, about 7 - 10 weeks or so. I suppose that's also tilt dependent with horizontal panels becoming no more than mud bins and vertical arrays getting no dirtier than vertical windows.

As for cleaning on a regular basis, and for no more than an opinion and FWIW, for the hassle involved (or lack of it), for my money and time, if I ever stop farting around and stop trying to satisfy my curiosity, like you, I'll hose the array ~ 1X/month or so if it doesn't rain using water from a hose, not brush the array, and call it good enough. Some fouling is a fact of life.

Last point: Water spots. I've got data on daily instantaneous performance after a specific regimen of daily cleaning for ~ 40 clear days. The first day of the 40 clear days, I washed the array w/ soap and rinsed. no more, but did it 2X and rinsed each time, similar to the way as described in steps 1 - 5 above. The next 13 clear days, I simply hosed the array without brushing and let the array drip dry. The next 13 clear days, I washed/rinsed with the above described brush method - hose/brush/soap/hose, followed by rubbing to dry with no observed hard water spots. The final 13 clear days, I washed/rinsed with the brush and then cleaned the array with Windex and rubbed dry. I measured instantaneous performance each day, threw out the high and low calculated fouling estimate for each 13 day period and averaged each resulting 11 day's fouling estimate. Forcing the mean fouling of all 33 test days to == zero to compare the measured fouling difference among the 3 methods: The results were that between all the methods, the average of the measured fouling was 0.0035 for the first batch, + 0.0024 for the second batch and -.0053 for the 3d batch. The std. dev. of the measured/calculated fouling for all days was 0.0055.

Bottom line on hard water spots: Since I can't claim any accuracy at predicting fouling rates better than +/- ~ 0.008 to 0.01 or so (that is +/- about 1% or maybe a bit less), I'd say that, while the wash/Windex method seems to produce marginally better cleaning, given that the results from my efforts, I can't say that water spots cause a performance change that I can measure with much confidence. Furthermore, to the degree I can claim any precision of measurement, the impact of not removing water spots seems to be small.

Take what you want of the above. scrap the rest.