Underestimated Results of Cleaning Panels

[ButchDeal]

I bet the optimizers are changing the dirt deposition patterns, and, agreeing with you, having observed the same pattern on my panels at the wiring junction, I'd speculate that's happening due to the optimizers increasing the local panel top surface temp. around the optimizer, either from the optimizers possible (likely ?? - Butch ??) internal heat generation and/or increases in local temps when air circulation around the optimizer is reduced or cut off by the optimizer.

Optimizers do generate some heat though the PV surface is generally much hotter.
I suspect the majority if the difference in this location is due to the optimizer blocking some of the air flow (you you surmise).

They generally are mounted flat side up, leaving more room between the optimizer and the back surface of the modules though I have seen them mounted with the electronics up. This is not always possible due to the thickness of the module frame but is generally a bad idea. It is also one of the bad things about smart PV modules, the optimizers or micro inverters are directly attached to the back of the pv module, more directly increasing temps.

[J.P.M.]

Optimizers do generate some heat though the PV surface is generally much hotter.
I suspect the majority if the difference in this location is due to the optimizer blocking some of the air flow (you you surmise).

They generally are mounted flat side up, leaving more room between the optimizer and the back surface of the modules though I have seen them mounted with the electronics up. This is not always possible due to the thickness of the module frame but is generally a bad idea. It is also one of the bad things about smart PV modules, the optimizers or micro inverters are directly attached to the back of the pv module, more directly increasing temps.

I think I understand. Thank you for the response and information.

Perhaps to be clear, I'm not alleging anything bolted to the back of a panel is necessarily bad, just that any increased temp. as a result of something bolted to the back of a panel will TEND to make the driving temp. difference between the local panel temp. around the bolted object and the dew point greater, thus increasing the local dew evap. rate, and that greater local evap. rate being the mechanism that's causing the greater "dirt shadow" I see and Sensij seems to be seeing.

I'd suggest that while it's probably (but maybe not always) a better idea to leave a space between the device and the backside of a panel, doing so will still leave a temperature differential between the local plate temp. and the rest of the panel. It'll be a smaller temp. differential but it'll still be there, with the possible result, if my hypothesis has any relation to reality, that the "dirt shadow" will still form, but maybe take longer to be as noticeable. As for the space btweeen the object and the back of the plate always being a good thing, if the type of dust/debris blowing under an array is of a nature and a size that is amenable to getting into the open space between an object and the backside of a plate, it'll tend to find a home there more so than a flush mount with no gap. That accumulation site can lead to other problems over time, and will also have a thermal resistance which is responsible for ~ 1/2 to 2/3 or so of the local temp. differential.

I'm not saying I have a solution, or that one is even necessary. All this may be a bunch of solutions looking for a problem. But, while surmising that a gap will solve a heat created problem may be valid for most applications, I'd think it won't eliminate the situation of a dirt shadow as much as slow it down, and because, like it or not, stuff will get into an open space intentionally created with good intentions, it may perhaps, in the process of attempting to solve one problem, sometimes crate another.

Thanx again for the info.

[sensij]

Way to take one for the team.
Can you recharge your public solaredge site for everyone?

we should watch it till the clean ones rematch the dirty ones to see how long the cleaning lasts.

on the optimizers, were they mounted flat side up or the electronics up? About what is the gap from optimizer to the back of the modules?

The optimizers are mounted to the rail flat side up, and the frame thickness is 40 mm. I should probably have generated a map showing where the optimizers were physically located, or at least go back an identify how many optimizers are under each panel. I used the extra length of the optimizer leads to skip wire the section from 1.1.6 through 1.1.12, and also out at 1.2.13 and 1.2.14, so those are areas I'm sure have some panels with an extra optimizer and some without any at all. It might be easier to figure out with an IR camera at this point. (for the hypothesis building team... what I thought I saw is that the optimizer areas are *cleaner* than the surrounding areas. I'll try to take a picture tonight).

Back to the cleaning test, here are the weekly production numbers for the two panels being compared, over the life of the system. I can look a daily production numbers too, but going forward, anything difference more than 0.5% or so starts to resolve signal from noise.

Time	P1.1.7 E (Wh)	P1.1.10 E (Wh)
6/12/2017	6.25	6	Ratio
6/19/2017	11024.5	11000.25	100.2%
6/26/2017	13771.75	13756.25	100.1%
7/3/2017	12259.5	12222	100.3%
7/10/2017	12409	12412.5	100.0%
7/17/2017	12193.75	12185.75	100.1%
7/24/2017	11104.75	11109.75	100.0%
7/31/2017	10945.25	10972.5	99.8%
8/7/2017	12003.5	12022.25	99.8%
8/14/2017	10876.5	10915.5	99.6%
8/21/2017	8614.75	8641.75	99.7%
8/28/2017	10690.25	10684	100.1%
9/4/2017	10304.5	10287.25	100.2%
9/11/2017	8719	8708	100.1%
9/18/2017	9568.75	9595.25	99.7%
9/25/2017	3536.75	3527.75	100.3%

The public site is here.

The individual readings throughout the day are hard to compare because the optimizers report at different times, and with as much as 30 min between optimizer reports, even around solar noon when the irradiance change should be slowest the difference in timing can still complicate direct comparisons. In the public layout view, if you click the panel and then click the "i" that shows up in the menu bar, you can see how fresh or stale the data are for that optimizer.

I'm hoping that reports come in close enough together in time to allow both an "instantaneous" look at fouling at solar noon, which is at least comparable to what J.P.M. has shared in result (although clearly not as rigorous in method), and also daily production. If those two fouling numbers a re different, it could be as a result of method, but it could also be a sign that the effect of fouling is a function of angle of incidence (or something else that varies throughout the day).

[J.P.M.]

The optimizers are mounted to the rail flat side up, and the frame thickness is 40 mm. I should probably have generated a map showing where the optimizers were physically located, or at least go back an identify how many optimizers are under each panel. I used the extra length of the optimizer leads to skip wire the section from 1.1.6 through 1.1.12, and also out at 1.2.13 and 1.2.14, so those are areas I'm sure have some panels with an extra optimizer and some without any at all. It might be easier to figure out with an IR camera at this point. (for the hypothesis building team... what I thought I saw is that the optimizer areas are *cleaner* than the surrounding areas. I'll try to take a picture tonight).

Back to the cleaning test, here are the weekly production numbers for the two panels being compared, over the life of the system. I can look a daily production numbers too, but going forward, anything difference more than 0.5% or so starts to resolve signal from noise.

Time	P1.1.7 E (Wh)	P1.1.10 E (Wh)
6/12/2017	6.25	6	Ratio
6/19/2017	11024.5	11000.25	100.2%
6/26/2017	13771.75	13756.25	100.1%
7/3/2017	12259.5	12222	100.3%
7/10/2017	12409	12412.5	100.0%
7/17/2017	12193.75	12185.75	100.1%
7/24/2017	11104.75	11109.75	100.0%
7/31/2017	10945.25	10972.5	99.8%
8/7/2017	12003.5	12022.25	99.8%
8/14/2017	10876.5	10915.5	99.6%
8/21/2017	8614.75	8641.75	99.7%
8/28/2017	10690.25	10684	100.1%
9/4/2017	10304.5	10287.25	100.2%
9/11/2017	8719	8708	100.1%
9/18/2017	9568.75	9595.25	99.7%
9/25/2017	3536.75	3527.75	100.3%

The public site is here.

The individual readings throughout the day are hard to compare because the optimizers report at different times, and with as much as 30 min between optimizer reports, even around solar noon when the irradiance change should be slowest the difference in timing can still complicate direct comparisons. In the public layout view, if you click the panel and then click the "i" that shows up in the menu bar, you can see how fresh or stale the data are for that optimizer.

I'm hoping that reports come in close enough together in time to allow both an "instantaneous" look at fouling at solar noon, which is at least comparable to what J.P.M. has shared in result (although clearly not as rigorous in method), and also daily production. If those two fouling numbers a re different, it could be as a result of method, but it could also be a sign that the effect of fouling is a function of angle of incidence (or something else that varies throughout the day).

Thank you. More later. I'm getting ready for a reading at min. incidence angle (for me) @ 12:59 P.D.T. Solar noon for me today is at 12:39:13, P.D.T. FWIW, You're probably a few seconds. later.

[DrLumen]

Duh, I didn't think of just cleaning one panel to see the difference. When I get motivated, I will find a pair of panels that have the same relative output and just clean one. It may be a while though as we had a fairly good rain last night.

Now, if there was a way to reliably quantify how dirty a panel is... I guess I could wet vac one, let the water evaporate and weigh the dirt.

Would something like RainX applied to the panels void the warranty or hinder performance?

[sensij]

Thank you. More later. I'm getting ready for a reading at min. incidence angle (for me) @ 12:59 P.D.T. Solar noon for me today is at 12:39:13, P.D.T. FWIW, You're probably a few seconds. later.

Well, my results are nothing nearly as impressive as the OP's.

Peak power looks to be 246.3 W vs 237.5 W, which is 3.7% improvement.

It shows up about the same way in the raw electrical data. The data are a little bit noisier, but average current from 12 pm - 1 pm was 8.95 A vs 8.64 A, about 3.5% increase.

Voltage is not affected enough to be detectable here, about 27.3 V average for both over that hour but the individual measurements have only .125 V resolution and are noisy.

We'll see what the end of the day brings. If I'm looking at the current data correctly, the daily energy performance will end up being something less than the performance difference at peak output, maybe 2.5% - 3% ish. Cleaning.JPG

[max2k]

Well, my results are nothing nearly as impressive as the OP's.

Peak power looks to be 246.3 W vs 237.5 W, which is 3.7% improvement.

It shows up about the same way in the raw electrical data. The data are a little bit noisier, but average current from 12 pm - 1 pm was 8.95 A vs 8.64 A, about 3.5% increase.

Voltage is not affected enough to be detectable here, about 27.3 V average for both over that hour but the individual measurements have only .125 V resolution and are noisy.

We'll see what the end of the day brings. If I'm looking at the current data correctly, the daily energy performance will end up being something less than the performance difference at peak output, maybe 2.5% - 3% ish.

probably because your dirt is 'younger' or this is from your other system?

[ButchDeal]

Well, my results are nothing nearly as impressive as the OP's.

Peak power looks to be 246.3 W vs 237.5 W, which is 3.7% improvement.

It shows up about the same way in the raw electrical data. The data are a little bit noisier, but average current from 12 pm - 1 pm was 8.95 A vs 8.64 A, about 3.5% increase.

Voltage is not affected enough to be detectable here, about 27.3 V average for both over that hour but the individual measurements have only .125 V resolution and are noisy.

We'll see what the end of the day brings. If I'm looking at the current data correctly, the daily energy performance will end up being something less than the performance difference at peak output, maybe 2.5% - 3% ish. Cleaning.JPG

Ok so now how long before the match again?

[sensij]

probably because your dirt is 'younger' or this is from your other system?

Yeah, this is dirt from the new system, panels have been on the roof since mid-June.

Final numbers were 1.77 / 1.71 = 3.5% improvement.

Looks like mid-day instantaneous power fouling numbers are a decent indication of overall energy impact.

Maybe I just suck at cleaning.

[max2k]

Yeah, this is dirt from the new system, panels have been on the roof since mid-June.

Final numbers were 1.77 / 1.71 = 3.5% improvement.

Looks like mid-day instantaneous power fouling numbers are a decent indication of overall energy impact.

Maybe I just suck at cleaning.

you had less fouling before cleaning- OP accumulated his over 1 year. On a completely new system no cleaning would be able to improve its output.

[ButchDeal]

you had less fouling before cleaning- OP accumulated his over 1 year. On a completely new system no cleaning would be able to improve its output.

I seem to remeber others performing a test like this and the cleaning improvement was less than a week . Though that would Garry with dirt and weather, certainly a few months is enough time to bring them to their steady state.

we will know in a few days when sensij

[J.P.M.]

Duh, I didn't think of just cleaning one panel to see the difference. When I get motivated, I will find a pair of panels that have the same relative output and just clean one. It may be a while though as we had a fairly good rain last night.

Now, if there was a way to reliably quantify how dirty a panel is... I guess I could wet vac one, let the water evaporate and weigh the dirt.

Would something like RainX applied to the panels void the warranty or hinder performance?

That sounds like what Sensij is doing. More results from different locations using the same method can never hurt as long as the method is simple enough so that the output is a measure of the same thing.

My guess is that even with a 4 decimal place beam balance to weigh the sucked up dirt, panels would need to be pretty dirty to measure a difference.

While there is no easy way to ensure a panel or array has 100 % of it's performance due to lack of fouling, it is possible to define a panel as clean, such as immediately after a very thorough cleaning, measure performance immediately after and then use that performance as a benchmark. I'm quite sure comparing the output of 2 panels whose individual output is pretty well matched, as Sensij has done, and then monitored, either for instantaneous or period totals can be a very valid, accurate and reasonably precise method to evaluate array performance.

Rainex has been discussed here a few times. One of the mods (Russ - since retired) called Rainex and they said don't do it. The conclusion around here at the time was that Rainex may have unknown effects on any ARC coating and also void any warranty.

[azdave]

I still say you are far better off to let the rain do your cleaning and stay off the roof unless you have an issue that rainfall isn't going to remedy. Bones break when people my age fall off ladders so why risk it for a small reward? Would that 3-5% short-term gain cover my medical bills if I end up in the ER one day? Not even close. My ROI would be lengthened by years in pursuit of a minor gain in performance.

[inspron]

I still say you are far better off to let the rain do your cleaning and stay off the roof unless you have an issue that rainfall isn't going to remedy. Bones break when people my age fall off ladders so why risk it for a small reward? Would that 3-5% short-term gain cover my medical bills if I end up in the ER one day? Not even close. My ROI would be lengthened by years in pursuit of a minor gain in performance.

The key information for me to decide clean or not is knowing the residual improvements after ~2-3 months, if any. From hours of Googling, it appears there are no noticeable improvements after 1-2 months? So the effectiveness is short lived for most situations.

[organic farmer]

I was concerned about snow build-up. As I looked close, I saw that a layer of ice forms on the panels, then snow sticks to the ice. If it were only ice, the morning sun would burn through it quickly, but with an inch of snow on top of the ice, the sun can not get through.

I experimented with a thin layer of veggie oil and with PAM spray. Both seem to repel the ice from forming, but in a few months [half-way through winter] these oils oxidize into a yellow goo.

Rain-X seems to work best.