Efficiency drop of panels per year?

[RonRN18]

I have a SolarEdge system that was designed for 7.32 kWp, which was installed at the end of 2015. I never go up on my roof, as I have a very tall house, but I can see my panels from a distant and they appear a little dusty. It got me thinking on whether I needed to clean them or not. I decided to look at numbers before doing anything. I was trying to figure out the best way to compare apples-to-apples over the past several years, so I decided to download daily power generation for every day in July of 2016, 2017, 2018 and 2019. I then found a site that calculated the number of minutes of daylight there was for every day involved, per my latitude/longitude. I then divided the power generated per day and divided it by the minutes of daylight, giving me the average Watt-hours per minute for the day. I then took the best five days of July from each of the years to compare. I figure, July is usually clear and hot, but there is very little variation between the top five days. Here is my results:

• 2016: Average of 48.51 Wh/minute of daylight
• 2017: Average of 47.37 Wh/minute of daylight
• 2018: Average of 45.52 Wh/minute of daylight
• 2019: Average of 44.25 Wh/minute of daylight

Another way to look at this is the average Wh/minute of daylight dropped 2.34% between 2017 and 2016, 3.91% between 2018 and 2017, and 2.79% between 2019 and 2018. I find it odd that there was a larger drop between 2017 and 2016 than in other years. The overall drop from 2019 and 2016 is 8.78%

My question is, is this within the normal drop in efficiency per year or could it be due to the panels being cleaned? My wife doesn't want me to climb up there to clean the panels, so I'd have to hire someone to do so. I've looked around to see what it would cost to have the 24 panels cleaned and have been quoted $150. Based on a cost of $0.1435/KWh, I'd need to generate an additional megawatt to pay for the cleaning. Below is my MegaWatt production per year for the first three full years:

• 2016: 12.132 MWh
• 2017: 11.910 MWh
• 2018: 11.710 MWh

The drop is far less than a megawatt/hour of difference, therefore, I'm thinking that paying someone to clean my panels would not be a valuable expenditure. I know I could clean it for much less, but my wife doesn't want me to go up there (it's a bit over 20 feet from the eaves to the ground).

Any advice?

[inetdog]

It is also possible that some of the loss in output is simply the result of aging of the cells, but not close to 3% per year.
Without actually climbing onto the roof, you may be able to use a low pressure water stream with detergent added, followed by clear water or maybe a mild rinse aid. Just be sure to do this when panels are not sun heated to avoid thermal stress.

[JSchnee21]

Welcome RonRN18 !

It's certainly possible that your array's production has decrease some due to panel degradation and soling. 8% seems a touch high. But it is not impossible. But, even though it looks like you might have a trend, the reality is that weather patterns could be varying enough to confound your observations.

You didn't mention what brand your panels are nor where you live. Most (but not all) panels show a higher year 1 initial LID (light induced degradation rate). This can be as high as a 3% drop in the first year. Best case it's more like a 1-2% drop. But during subsequent years the LID loss is generally only 0.5% or less.

Additionally, soiling can contribute up to a 5-10% total loss under very heavy dusty, salty, bird or guano conditions. Finally the weather will vary from year to year (clouds, haze, smoke, dust, pollution). One last factor to consider -- trees grow. If you have any shading at all due to tree, etc. this will get worse each year.

Looking at the data for my system (Solar Edge / Panasonic / significant tree shading), my average production actually increased slightly from 2017 to 2018. And so far from 2019 this trend seems to be continuing.

https://pvoutput.org/aggregate.jsp?i...=52103&v=0&t=y

But I live in Central NJ which actually has very clean air. And we have frequent heavy thunderstorms and snow so my array stays pretty clean. If you live in the dry, dusty midwest, somewhere with lots of new home construction, California with smoke and haze, or at the beach, you may have more significant soiling.

Certainly check you SolarEdge data, make sure all panels are producing, and assuming you don't have any shading or weird orientations, each panel should be producing roughly the same kWh per month/year. That is to say, verify that you don't have a panel or two that has gone bad (unlikely).

[NEOH]

ng. Below is my MegaWatt production per year for the first three full years:

• 2016: 12.132 MWh
• 2017: 11.910 MWh
• 2018: 11.710 MWh

The drop is far less than a megawatt/hour of difference, therefore, I'm thinking that paying someone to clean my panels would not be a valuable expenditure. I know I could clean it for much less, but my wife doesn't want me to go up there (it's a bit over 20 feet from the eaves to the ground).

Any advice?

I do not agree with your "daily" analysis.

2016 was a leap year with 366 days, normalized to 365 days ..

• 2016: 12.099 MWh
• 2017: 11.910 MWh = 98.4% or 1.6% drop
• 2018: 11.710 MWh = 98.3% or 1.7% drop

This is reasonable.

Rain water is more like soft water than hard water, so when it is dry and dusty ...
I use "softened" water to wash my ground-mounted solar panels.
I do not use hard water, because calcium spots are difficult to remove.

[ButchDeal]

I have a SolarEdge system that was designed for 7.32 kWp, which was installed at the end of 2015. I never go up on my roof, as I have a very tall house, but I can see my panels from a distant and they appear a little dusty. It got me thinking on whether I needed to clean them or not. I decided to look at numbers before doing anything. I was trying to figure out the best way to compare apples-to-apples over the past several years, so I decided to download daily power generation for every day in July of 2016, 2017, 2018 and 2019. I then found a site that calculated the number of minutes of daylight there was for every day involved, per my latitude/longitude. I then divided the power generated per day and divided it by the minutes of daylight, giving me the average Watt-hours per minute for the day. I then took the best five days of July from each of the years to compare. I figure, July is usually clear and hot, but there is very little variation between the top five days. Here is my results:

• 2016: Average of 48.51 Wh/minute of daylight
• 2017: Average of 47.37 Wh/minute of daylight
• 2018: Average of 45.52 Wh/minute of daylight
• 2019: Average of 44.25 Wh/minute of daylight

Another way to look at this is the average Wh/minute of daylight dropped 2.34% between 2017 and 2016, 3.91% between 2018 and 2017, and 2.79% between 2019 and 2018. I find it odd that there was a larger drop between 2017 and 2016 than in other years. The overall drop from 2019 and 2016 is 8.78%

My question is, is this within the normal drop in efficiency per year or could it be due to the panels being cleaned? My wife doesn't want me to climb up there to clean the panels, so I'd have to hire someone to do so. I've looked around to see what it would cost to have the 24 panels cleaned and have been quoted $150. Based on a cost of $0.1435/KWh, I'd need to generate an additional megawatt to pay for the cleaning. Below is my MegaWatt production per year for the first three full years:

• 2016: 12.132 MWh
• 2017: 11.910 MWh
• 2018: 11.710 MWh

The drop is far less than a megawatt/hour of difference, therefore, I'm thinking that paying someone to clean my panels would not be a valuable expenditure. I know I could clean it for much less, but my wife doesn't want me to go up there (it's a bit over 20 feet from the eaves to the ground).

Any advice?

I love how people come up with their own measue like Wh/minute of daylight...
daylight is kind of meaningless they way you have it figured. daylight is not the same as solar producing and would have to do with azimuth and tilt as well as major shadows of your particular system.

Common would be kWh/kW/day..

in any case Have you checked that all of your pv modules are performing? Are they all producing every day? are they all producing roughly about the same as the ones near to them (non are 1/3 or 2/3 less than neighboring modules).

Are there any significant dirt or bird droppings on any modules that you can see from the ground or with binoculars?

Have any trees grown to increase shadows over the years?

First year has the most degradation and drop due to initial soiling.

[RonRN18]

I love how people come up with their own measue like Wh/minute of daylight...
daylight is kind of meaningless they way you have it figured. daylight is not the same as solar producing and would have to do with azimuth and tilt as well as major shadows of your particular system.

Common would be kWh/kW/day..

in any case Have you checked that all of your pv modules are performing? Are they all producing every day? are they all producing roughly about the same as the ones near to them (non are 1/3 or 2/3 less than neighboring modules).

Are there any significant dirt or bird droppings on any modules that you can see from the ground or with binoculars?

Have any trees grown to increase shadows over the years?

First year has the most degradation and drop due to initial soiling.

The reason I came up with my own unit of measurement was to take into consideration that some days are longer than others and I was trying to compare the power output of the five highest power generation days. Nearly every day in July is hot and cloud-free, but there may be the occasional cloudy day, so I didn't want to put that into comparison. What I had to make measurements was Wh/day and then I found an NOAA tool that gave exact minutes of daylight based on my latitude and longitude for every given day. I figured then, that dividing the Wh by the length of the day. Given what I had to measure, I figured it was more accurate than just comparing months, as I was focusing on only the highest producing days and comparing them based on length of daylight. Granted, the difference of a couple of minutes is likely not going to be statistically all that significant, I was trying to even the playing field.

In my setup, all 24 of the panels are LG LG305N1C-B3 panels, going into my SolarEdge SE6000 inverter. There are no taller trees or structures than my house within several hundred yards of my house, so the panels get unimpeded sun all day. Of the 24 panels, 17 are at a tilt of 24 degrees and the azimuth of 185 degrees while the other 7 panels are also a 24 degrees tilt but the azimuth is 275 degrees.

[J.P.M.]

I'd agree with most of what JSchnee and Butch write. There are lots of variables you are unaware of that cause annual PV output to vary, some of them by an order of magnitude more than you're trying to measure when you styart considering how array fouling affects output.

Array fouling, its nature, origins, rates, both increrasing and decreasing are quite site and array specific and dependant.

As an example only, if it doesn't rain, my array fouls such that efficiency rolls off at a rate of approx. 0.75 % to 1.0 %/week. I believe that rate tends to become asymtotic after about 8-10 weeks or so of no rain, but my data is not able to confim my suspicions in that regard. Your site and situation will most likely behave differently, but perhaps with some similarities.

As for your methods of estimating production as f(various parameters), I'd seriously encourage your curiosity but suggest you consult some academic sources that handle parametric analysis of the solar resource and its variability in more meaningful ways. Right now, your efforts remind me of a powerful locomotive with no flanges on the wheels. One source of flanges: Duffie & Beckman's "Solar Engineering of Thermal Processes" has a thorough but still concise treatment of the solar resource and good explanations of applications to both solar thermal and PV applications.

One thing I'd most certainly agree with is your surmise and estimate that paying for panel cleaning is a ripoff.

Welcome to the neighborhood.

[azdave]

I've long been a fan of not cleaning them either professionally or on your own if they are in a place with a fall hazard. One trip to the ER could set your ROI back dozens of years. Paying for cleaning is rarely going to come close to being a good investment return-wise.

[ButchDeal]

Nearly every day in July is hot and cloud-free, but there may be the occasional cloudy day, so I didn't want to put that into comparison.

In my setup, all 24 of the panels are LG LG305N1C-B3 panels, going into my SolarEdge SE6000 inverter. There are no taller trees or structures than my house within several hundred yards of my house, so the panels get unimpeded sun all day. Of the 24 panels, 17 are at a tilt of 24 degrees and the azimuth of 185 degrees while the other 7 panels are also a 24 degrees tilt but the azimuth is 275 degrees.

part of the issue is the heat in the summer. Higher temperatures cuts the performance of the PV modules and is why the ones on the south face higher up (1.1.8 and 1.1.9) are showing slightly lower performance for that day as the heat rises under the rows and the ones at the top are not getting as much cool air to cool down like the bottom row.

From the image I would say that the south array is looking pretty reasonable with the only exception and it is slight would be 1.2.4. at the bottom and end it should have good cooling but might have some slight shade, the difference could be be just variance in modules as they have some variance of a small amount (varies by manufacturer).

You should look at the west array production latter in the day or use the playback feature to see them all in high production and compare them for outliers.

[solarix]

I have several dozen systems that I get to see the monitoring on, and am amazed how predictable they run. One system in particular I have watched for over 8 years and its efficiency loss over that time is within the statistical margin of error. Weather variation is certainly a bigger factor than panel degradation. I think I saw a study published in SolarPro a while back that over a large number of solar systems concluded that actual panel degradation is proving to be substantially less than the 0.8% per year that has been the industry expectation.
Don't wash your panels with hard water. PV panels use a special glass with an anti-reflective texture on the front. If you get calcium deposits down in that texture, its difficult to remove...