SunPower Solar Cell Discoloration

[newnmex999]

I wouldn't put a whole lot of faith in such a method.

- It is very likely, more like almost a certainty - that the prior 4 or 5 months are more sunny or less sunny, and other weather inputs like temps., wind speeds, etc. are different than the TMY data for corresponding months, with the result that the actual system output for those months will most certainly be different than PV Watts will calculate. This is because all those inputs have a very pronounced influence on system performance and output.
- If the current month (or so) is an anomalous weather period and/or much sunnier or cloudier, colder, or windier, etc. than immediate prior months, output will be different than prior months relative to the PV Watts output. How will system problems, if they exist, be segregated ? Put another way, if it's 10% more (less) sunny in any month, you'll never know it by eyeball. Trust me on that. If it is 10% less sunny, and you don't know it because its not being measured, you will not know if a decrease in system performance is caused by panel problems or weather, something else, or a combination of things.
- According to the NREL Solar Radiation Manual, Albuquerque solar irradiance on a south facing, lat. -15 deg. surface has monthly variations of +/- 11% for Aug., that variation generally increasing to about 20% in Dec/Jan., w/ standard dev. probably something like about 5-10% respectively. That says nothing about other weather variables that also affect system output.
- What if you have sunnier than TMY weather for a month or two as can happen ? Will the increased output mask any problems ? What if prior months were sunnier than the TMY data and current month is close ? Without current input data , that scenario will (erroneously) make current the current month's output seem deficient and lead to suspicion of a problem where one might not exist..

- To measure a change of performance in a system, measured, ACTUAL and current input to the system is needed, that is, weather data including, and primarily, solar irradiance data and other weather parameters for the time period under consideration, not the weather for the TMY month used, which BTW is from a different year for each TMY month and therefore inappropriate for this application. It is measured but historical, not measured and current.

- Sizing tools for solar electric systems rely on weather data that is better described as climate data. Climate is what you expect. Weather is what you get. The two are not the same and will lead to different output, for different purposes.
- If there is a problem with the panels and the deterioration is of the same order of magnitude as the weather variation, such deterioration will be masked to a large degree by the noise of the weather data variation. If there is a panel problem and it is progressive, it may be still harder to find.
- According to the NREL website, PV Watts estimates of actual system production can be different by as much as 40% for individual results and up to 20% for any complete year. That alone, and as the NREL website suggests, makes prediction of performance over shorter periods much less reliable, if at all.
- To have faith in the accuracy of any estimate of performance and changes in that performance, reliable measurement of the pertinent and controlling input and output variables for the time period under consideration is a requirement. TMY information is not that type of data. It is historical.

FWIW, to help solve this problem will take some measurement of current weather data including solar irradiance. Also, converting the irradiance from a horizontal surface to the array position can be something of an art in itself and also open to some interpretation.

Don't shoot me, I'm only the piano player. Not everyone who tells you unpleasant stuff is your enemy. Not everyone who tells you what you want to hear is your friend.

J.P.M.-thanks much for your input. You seem very knowledgeable in this area and I am a newbie. While I do understand most of what you have explained, my conclusion based on it is that there are so many damned variables that it will be next to impossible for me to determine what, if anything, is going on with my system! Would you say that is a fair assessment?

Given my situation, how would you approach this issue?

Thanks

[Ian S]

Given my situation, how would you approach this issue?

Thanks

If you have access to the Sunpower monitoring website for your system, you should download 5 minute interval production data all the way back to system startup and keep doing so from now on. Based on my understanding of my access to the Sunpower monitoring site, you can only download the 5 minute interval data for a time period of 30 days or less so you can't do it all at once but rather 30 days at a time. If you have access to the advanced monitoring (from the basic monitoring site), you can also get the 5 minute interval power data but only for the past 60 days. This data all comes in spreadsheet compatible format so it will allow you to keep detailed track of your production over time.

IMO, the discoloration you are seeing will have very subtle if any effect on your production so without a year or more of data, you'll be hard-pressed to identify an effect especially if this issue occurs very gradually. At least you are in New Mexico where I assume you have a large number of perfectly sunny days so comparison among the data is a bit easier.

From what you claimed your maximum hourly production is lately, it seems reasonable. Assuming a typical panel angle with horizontal and an ideal southern azimuth, on an ideal cool sunny day somewhere around noon with the sun's rays perpendicular to the plane of the panels, you should see nameplate production or higher - I know I have.

[J.P.M.]

What I proposed is a quick and dirty way to do a sanity check if you don't have any of the actual current month's data from the previous year as a reference like the poster said, but you do have historical production data from previous months before you notice the discoloration. Nothing exact or scientific. More like statistical comparison, that's all. Take it or leave it for what it's worth. No need to over analyze it. Obviously it's better than nothing.

The whole point, maybe you don't get it, is about relative comparison. If your "before discoloration" production is on the average about 10% higher than PVWatts's estimate, and you have a small enough standard deviation of this grouping, then statistically speaking, you should be around 10% compared to PVWatts result this month, too, if discoloration does not affect production. If you're well below the normal 10% deviation from PVWatts results as a reference, then maybe you should look into it more. But if you're on the mark at around the same deviation as other months, then maybe you're OK. It's not exact comparison. It's all about relative comparison against something else as a reference.

Of course nobody is dumb enough to not realize that previous months' weather is different than current month. No need to give a long winded lecture about it and still end up not suggesting anything better to the OP.

I believe I understand what you are saying and I respect your opinion, and you, but because I lack the eloquence to avoid long winded lectures I'll leave it that I respectfully disagree with your proposed method. I believe it's ill conceived, lacks a working knowledge of solar design tools, their strengths, weaknesses and limitations, TMY and weather data and statistics, and is may be dangerous in that it will perhaps lead other unsuspecting users to equally erroneous conclusions. It's just wrong. Nothing personal.

[J.P.M.]

J.P.M.-thanks much for your input. You seem very knowledgeable in this area and I am a newbie. While I do understand most of what you have explained, my conclusion based on it is that there are so many damned variables that it will be next to impossible for me to determine what, if anything, is going on with my system! Would you say that is a fair assessment?

Given my situation, how would you approach this issue?

Thanks

You're Welcome !
Close to Sandia Heights huh ? Nice neighborhood.

I never tell people what to do, but speaking for myself only:

1.) I'd Think about what I want to get out of this. As in, is this really a problem ? Seems funny colors are on my new panels ?
2.) If I said "maybe", the first call I made would be to the vendor. I rag on Sunpower stuff a lot, but not because it's poor quality, just overpriced for most cost effective applications, but it's a real Mercedes, to quote myself. I'd chance that a reputable vendor peddling Sunpower would at least be curious. One thing I would not do is not call the vendor, and I would call the vendor first before calling Sunpower as a respectful, and practical gesture. Dance with who brung ya'. A reputable vendor really can be an asset, and deserves the courtesy, respect and a chance to help. There really are vendors like that out there. You may have used one. That's why they sometimes extract a premium. I paid a lot of $$'s for my 327's. I'd be surprised if my vendor would leave me twisting in the wind - I picked him from the cream of the crop. Before I called the vendor, however, I'd do 3 and 4 below and get my ducks and questions lined up.
3.) I'd look around the neighborhood and see who else had Sunpower stuff, take a 6-pack , bang on their door(s), introduce myself, have a chat and exchange info.
4.) I'd sit down and look, and I mean really look, at the array. Then clean it on a cool day or early A.M. and see what happens. Colors change ? More/less discoloration ? Anything else going on with it ? Dimensions changing ? Migrating/stray sandhill cranes from the bosque passing through leaving steamers on the array ? Wiring screwed up/missing ? Small animals chewing on stuff ? Etc. I have an infrared thermometer that is real fun to see the difference in temp. over individual cells in a panel. About $40 at an electronics store. I'd measure the discolored spots vs. the other parts. Maybe good info. can't hurt.
5.) When I got my array, I knew I would want to monitor system performance and performance changes in the system as a function of time and changing weather variables. Now, let me tell you, I'm so cheap I take my glasses off when I'm not looking at anything. I'm so tight I squeak. Do you think it would make any sense for me, or my self image, knowing what I think I might know about the subject, to spend about $1,000 for a weather station and put it about 3ft. from my array to measure a bunch of weather variables if there was ANY way I could accomplish the same result for nothing out of pocket with a design tool/piece of software, more than a few of which I've written for my own use, besides all the NREL stuff that's available ? I wish there was. I'd be $1,000 richer, which is about 200 times a lot of money by my reckoning.
6.) So, my solution, good for me only, was to get equipment to measure the input to my solar array. If you want a definitive answer to the performance question, that's about the best and most accurate one I'm aware of and it's still only +/-5% at best. The precision may be a bit better however. I posted some results on this forum about a month ago, suggesting my array's performance improved something like 1% SWAG by washing off 3 months of dirt. If I'm wondering about discoloration of my array, and I am(none at this time BTW) and how it affects performance, I can get a SAWG of actual performance by measuring input and output just like I reported last month and like I do several times/week. If I'm VERY careful and a good data taker and real lucky, I might get within about +/- 5 % or so of reality, maybe a bit better day/day. That's probably about the best I can claim and that's probably optimistic. Knowing that, I bet if my performance deteriorated due to discoloration, or perhaps more accurately, the discoloration was a symptom or consequence of some equipment malfunction, because of instrument limitations, I wouldn't be able to measure (notice)it until it got at least greater than my instrument accuracy => 5%. If it was that large, I bet I'd notice it in my increased usage that I got billed for.
7.) To that end, one thing I've done since the mid '70's was to read my utility meters 1x/day, about the same time and record the readings. I learned a long time ago, if you want to use less of something, measure it and keep track. I swear it works. Applied to this situation, if my panels are losing performance, and I read and record my meter reading and thus use every day, I'm quite sure if my usage goes up due to performance problems, I'll spot it fairly quickly. Just another tool and info source.
8.) I'd study, learn and think, and use PV Watts and even better, a product called SAM from NREL. It's like PV Watts on Steroids, and a PITA to learn but really worth the effort. Before I did that I'd buy a "Dummies" (no offense intended) book about solar energy and read, study and learn. I've got a lot invested in my array and I want my money's worth.

None of the software design tools are useful for predicting or backing into performance for short periods. That is not what they were designed to do. Honest. I wish they were. I'd probably aggravate fewer people if that were the case.
Finally, I'd try to remember that sometimes a dirty or discolored array is just that : a dirty or discolored array. It may be more, so I'd enlist the vendor and use their expertise to look for a solution or use their resources to monitor the situation.

Regards,

J.P.M.

[silversaver]

999,

After all the reading, have you found the answer?

[J.P.M.]

J.P.M.-thanks much for your input. You seem very knowledgeable in this area and I am a newbie. While I do understand most of what you have explained, my conclusion based on it is that there are so many damned variables that it will be next to impossible for me to determine what, if anything, is going on with my system! Would you say that is a fair assessment?

Given my situation, how would you approach this issue?

Thanks

Thinking on this a bit more, I wonder if some of the observable brownish stuff/discoloration has anything to do with the antireflective coating that Sunpower uses and if/how/maybe it interacts with the rest of the system - hot cells/cold cells, etc. ??

[J.P.M.]

Wondering if someone can explain to me what kind of output I should see on a perfect solar day at high noon from an 18 panel 5.88 E20 327 Sunpower system? Currently my hourly output at this time of year is in the mid 4kWh. I am getting this info from SunPower monitoring website. As we get closer to the summer months what output should I expect to see from my system based on the 5.88kWh capacity. Should I expect to see hourly output in the mid 5's? Or am I looking at this entirely incorrect?

999: Check your P.M.'s

[J.P.M.]

If you have SunPower solar system, most likely you will also have the SunPower monitoring system. The Kiosk mode let you share your solar system performance. just basic info as outputs. Here is an example:

https://www.sunpowermonitor.com/residential/kiosk.aspx?id=2B0B7671-2507-4702-AB24-DA5805D4FF86&type=address

That is a 6.075kW Sunpower system facing about South 190 deg

Silver:

Do you know about how long this system has been in service and approx. zip ?

J.P.M.

[silversaver]

Silver:

Do you know about how long this system has been in service and approx. zip ?

J.P.M.

Sorry, the link is from one of the member here install in March 2011. It is a 30 X 225W SP SW220 with 20 deg pitch. I use it as refererce since the solar has late afternoon shade just like my system. If you need a good reference SP 327 panels system in SD area, I think I have a few links at other PC. I'll pm you

[J.P.M.]

Sorry, the link is from one of the member here install in March 2011. It is a 30 X 225W SP SW220 with 20 deg pitch. I use it as refererce since the solar has late afternoon shade just like my system. If you need a good reference SP 327 panels system in SD area, I think I have a few links at other PC. I'll pm you

Thank you. I'd appreciate that.

Regards,

J.P.M.