Panel cleaning

[J.P.M.]

I recently purchased a pole mounted brush for my panels and have used it twice. The first time was just with water and I didn't see a huge improvement. However, that was after about a year of not cleaning with any mechanical means (I had done a hose spray a couple of times). I think I incurred a pretty significant penalty due to ash from local brush fires, one of which caused me to evacuate (side note, monitoring my solar production remotely let me know that my house had not burned down). When looking at the panels after they dried I could definitely see that the panels were still dirty.

My latest clean was a couple of weeks ago and I dragged a bucket with some Dawn up on the roof with me and used that to scrub the panels down and did see a decent increase in production as a result. It looks like my peak production went up about 250W with very similar solar radiation readings. Total production went up a little over 2kWh on the day of cleaning (on my 6.27kW system). The panels looked much cleaner and definitely had a shinier appearance as compared to using just water.

After no cleaning for a long time, like months, or no significant rain, I've found my array benefits from mild detergent and some elbow grease, but unless the gunk is caked/cooked on by a lot of repeated cooking/dew cycles My experience is that a simple hosing restores about 2/3 or so of the lost performance due to dirt if done, say 1X/month or so. 6 months of no rain/no cleaning takes some soap/elbow grease and a soft cloth. That observation applies to my array only, and probably similar in my neighborhood and I'd think in similar environments.

I sure don't doubt your reported improvement but I'm confused about your reported numbers. Your peak production went up ~ 250W. If so, what were that an instantaneous readings ? That is, assuming they were instantaneous readings, what was the POA irradiance and what was the array output before cleaning ? What were those same numbers after cleaning ? How did you determine the irradiance levels ? How far apart were the before/after readings in days ? Were the readings done at the same time of day ?

Thank you.

[jasonvr]

I recently purchased a pole mounted brush for my panels and have used it twice. The first time was just with water and I didn't see a huge improvement. However, that was after about a year of not cleaning with any mechanical means (I had done a hose spray a couple of times). I think I incurred a pretty significant penalty due to ash from local brush fires, one of which caused me to evacuate (side note, monitoring my solar production remotely let me know that my house had not burned down). When looking at the panels after they dried I could definitely see that the panels were still dirty.

My latest clean was a couple of weeks ago and I dragged a bucket with some Dawn up on the roof with me and used that to scrub the panels down and did see a decent increase in production as a result. It looks like my peak production went up about 250W with very similar solar radiation readings. Total production went up a little over 2kWh on the day of cleaning (on my 6.27kW system). The panels looked much cleaner and definitely had a shinier appearance as compared to using just water.

[dannieboiz]

I agree with what everyon say here. My panels has not been cleaned since installation about 3 years now. I decided to climb the roof to scrub it with simple water and a car was brush then squeegee it off a couple days ago. I definitely see improvement. The weather in the Bay area has been pretty consistent the past couple days.

I washed the panels late morning on the 22nd and look how it improved the past two days.

solar.png

[J.P.M.]

I've observed that quite a bit of dew collects on my PV panels in the early morning, which then picks up dust and dirt. My area is very dusty, so quite a bit accumulates in just a few months. Spraying by itself is not enough to get the caked on dirt off, so I purchased a wide window washer/squeegee tool with an extendable pole long enough to reach the panels in the middle of the array.
Cleaning Tool.JPG

The production results were pretty significant before and after cleaning. The weather was pretty much identical with clear skies all day so the primary difference should be due to the cleaning.

Here's the side by side daily production from the day before and the day of cleaning:
Total production 38.52 kWh vs 43.37 kWh (12.5% increase). Peak 5.16 kW vs 5.71 kW
Daily Production Before and After.PNG


The weekly view shows the big jump.
Production Before and After Cleaning.PNG
And this is with 5 of the 21 panels not cleaned. The sun was coming up and I was getting hungry so I still need to get the last 5 cleaned. There would likely be a couple more percent increase if all of them were cleaned.
Panels Semi-cleaned.JPG

Thank you. for the input.

Given that you had several months worth of dirt, and that works out to about maybe 12 - 15 weeks or so, an 11% to, say, maybe 15 % or so, that fouling penalty sounds about in line with my number of ~ 0.75 - 1.0% performance loss/week due to fouling without rain or intentional cleaning. And still, keeping in mind that array fouling can be highly dependent on location and lots of other factors.

As for the crud caking on from dew/cooling and then cooking on the next day cyclic type activity, I too have seen a fair amount of that due to copious marine layer and low nitetime radiant sky temps. that will often produce the equivalent of about 0.03"-0.05" of precip. in the Davis rain gage that's located about 4 ft. north of the array. I agree that left for several months, some mechanical means is necessary to get rid of it, most of which, for my array anyway is what seems to be what's responsible for the bathtub ring at the bottom 6" - 12" or so of the panels. But, so far and, I stress, for my experience only, hosing about every 3-4 weeks seems to also restore most array performance due to that mechanism as well, with the eye again being not necessarily the best judge of performance.

I haven't been able to get accumulated fouling data on my array since 06/30/18 as I've hosed, mopped (with a brush identical to yours BTW), and hosed again every morning at 0700 hrs. I then measure instantaneous output and input, voltages and currents from the inverter, and a bunch of other parameters the Davis spits out 1X/minute, and get an estimate of instantaneous input and output at the minute of minimum incidence angle. That'll continue until 09/04/2018.

Question: How do you avoid tile breakage ? Looks like they're all intact.

[Kendalf]

I've observed that quite a bit of dew collects on my PV panels in the early morning, which then picks up dust and dirt. My area is very dusty, so quite a bit accumulates in just a few months. Spraying by itself is not enough to get the caked on dirt off, so I purchased a wide window washer/squeegee tool with an extendable pole long enough to reach the panels in the middle of the array.
Cleaning Tool.JPG

The production results were pretty significant before and after cleaning. The weather was pretty much identical with clear skies all day so the primary difference should be due to the cleaning.

Here's the side by side daily production from the day before and the day of cleaning:
Total production 38.52 kWh vs 43.37 kWh (12.5% increase). Peak 5.16 kW vs 5.71 kW
Daily Production Before and After.PNG


The weekly view shows the big jump.
Production Before and After Cleaning.PNG
And this is with 5 of the 21 panels not cleaned. The sun was coming up and I was getting hungry so I still need to get the last 5 cleaned. There would likely be a couple more percent increase if all of them were cleaned.
Panels Semi-cleaned.JPG

[J.P.M.]

You can refer to the article posted at http://www.reonenergy.com/how-often-...-solar-panels/ .. It shows clear video for cleaning and maintaining solar panels.

Probably not bad advice and I can't disagree with too much of it. However, my experience, measurements and data indicate that for my array (and I'd make an educated guess and say others' arrays, but not everyone's) hosing without wiping is sufficient except for really caked/cooked on stuff, provided the array gets hosed about 1X/month or after 1 month without a decent rain. That is, I've been unable to detect a difference in performance improvement over rinsing and wiping then rinsing again, over simple rinsing. Added to that, if my rate of fouling is such that my array's output drops ~~ 1% per week without rain/hosing, which is about what it is, the small difference in cleanliness between simple hosing and hosing/wiping/rinsing becomes moot after a few days or less.

And to repeat, spots from hard water content do not seem to reduce the restored performance of my array over using D.I. or distilled water in any way I can measure, and I've been looking for a difference off and on for going on 4 years. Maybe I/m wrong and/or my methods are incorrect, but That's what my experience and data are telling me.

I'd like to see other data that would improve my knowledge of the subject.

As for guano, I don't have much (on my array anyway) but for the most part it comes off with hosings/rain.

Aside from that, the folks at Furnace Creek Ranch in Death valley continue to tell me they wash their 1 MW array about every 2 years to remove droppings from the large flocks of desert starlings that live on/around the adjacent golf course and it's relatively abundant water supply. But aside from that, they (facility maint.) don't seem to be worried about array performance, although I don't believe they have all they need to correctly interpret their monitored data.

Bottom line for me: Hose the array every 4 weeks or so if it doesn't rain. Use ~ 3/4 gal./panel. Do it early before the array heats up.

IMO, for most folks in SO CA or deserts, more than that won't improve things much.

As before, just based on my findings. Your results may and probably will vary, but my guess is not by much.

Around my area - San Diego county, I look at PV arrays as I'm driving around the area and walking in my neighborhood. The last measureable rain at my house was 06/06/16, and that was only 0.03" - just enough to cake the dust on my panels. Many/most arrays I see are getting a nice and quite noticeable bathtub ring with the remainder of the panel getting pretty dusty looking.

From a bunch of measurements I've done and continue with, as best as I can give an educated guess, my array's performance degrades approx. somewhere between 0.75% and 1.0% per week as a result of dirt buildup on the panels if they are not washed. Without rain or washing, I believe that rate starts to get asymptotic after about 6-8 weeks or so in some way I can't yet estimate, quantify or model.

A decent rain or hosing the array down seems to restore about 1/2 to 2/3 or so of the clean array performance. So, if my array is fouled to the point of, say, a 6 % decline in performance due to dirt,
hosing it off will probably improve the array's performance by about 3 or 4 %.

A mostly south facing array around here will produce about 150 - 175 kWh/month per nameplate kW this time of year. 3% of that is ~ 5 kWh per nameplate (DC) kW. If in tier 4, or prime time T.O.U. that's ~ $0.35 or so per kWh --->>> approx. savings per installed kW ~ = $1.75/mo. So, for those with arrays that have seen no moisture in a few months, a few minutes with a hose and probably what amounts to less than 50 gal. of H2O, someone with a 5 kW array might well reduce their current electric bill by something like ($1.75) X (5) = ~ $8 or $9. per month or ~ 30 day billing period.

Caution: before you run out with a hose: ONLY HOSE AN ARRAY WHEN THE SUN IS NOT ON IT, PREFERABLY IN THE MORNING, AND NEVER AFTER ABOUT 8 A.M. AT THE LATEST.

REASON: THE PANEL GLAZING MAY CRACK FROM THERMAL SHOCK.

Just sayin'. No guarantees expressed or implied. Your mileage may vary.

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

You can refer to the article posted at http://www.reonenergy.com/how-often-...-solar-panels/ .. It shows clear video for cleaning and maintaining solar panels.

[J.P.M.]

Soap, water and some scrubbing certainly gets things cleaner than just dribbling water on them. A heavy rainstorm would do just that. I try to wash mine once a year. I have a pole mounted brush.

Have you measured the difference using some quantifiable method ? I have for about 2+ years now.

Long story short, my array fouls at a rate such that the performance deteriorates about 0.075 % to 1.0% per week if It doesn't see rain or a cleaning. A rain of some decent quantity, say, 1/4" to 1/2", will restore very roughly, about 2/3 - 3/4 of the clean performance, sort of depending on how much/how long it rains, pretty much regardless of how fouled the array is before the rain.

I generally agree that it seems washing with mechanical scrubbing would restore almost all of the clean performance, and it may well do that.

But, the surprising thing I believe I've found after doing mechanical scrubbing and also simple hosing about a dozen times each, and measuring before/after performance each time as well as a lot of measurements between cleanings, including hand measurements of individual panel temps from under the array on about 200 days, and a recent series of cleanings on sunny days somewhat around the dates of min. solar incidence angle as sunny days allow (31 total) to compare 3 different cleaning methods to see if one is superior to the others in terms of how clean they will get an array, is that:

Aside from some minor scrubbing to remove things like owl skrock, I've found hosing my array from the top down 3 consecutive times such that the rate of H2O use is about 10 l/m^2 of array, a process that takes about 10 min., restores about as much of the original performance as mechanical scrubbing.

Not only that, but at a fouling rate of ~~ 0.75 % /week, and for comparable results, at least as close I can SWAG it using a +/- tolerance of about 0.75 % on my ability to estimate the performance impairment due to fouling level, hitting the array with a hose 1X/month will probably result in an array that's over time only slightly less fouling impaired than one that's been scrubbed, and with a lot less hassle.

As for how the fouling rate proceeds as f(time) I'm still unsure the nature of the fouling rate I'm seeing so as to be able to do a better job of estimating when best to hose the array. That is, is it a linear rate, with the rate being unchanging at a rate of ~ 0.75% -1.0%/week, or, as I continue to suspect, does it tend toward some type of asymptotic behavior that becomes more apparent after 8 or so weeks, or is it some wildly galloping function of weather and local conditions ? Or ...????

BTW, for those wondering, The remaining water spots as a result of hosing and drip drying do not seem to impair performance compared to scrubbing, squeegeeing and hand drying to remove water spots in any way I've been able to measure.

[nate379]

Soap, water and some scrubbing certainly gets things cleaner than just dribbling water on them. A heavy rainstorm would do just that. I try to wash mine once a year. I have a pole mounted brush.

[J.P.M.]

The panels are MUCH tougher than that. A car wash brush won't harm anything. When the snow gets deep on my panels I use a plastic roof rake to clear them off. (Not to make power... almost no sun in the winter, but because 3-4ft of snow is heavy!)

Yea, they probably are, but I'm not sure mechanical cleaning actually gets things that much cleaner. That's part of the reason I've got data from 08/04 - 08/20 - a long story that involves comparison of consecutive daily mechanical cleaning vs. consecutive daily simple hosing. Besides, a water stream is usually easier to get up to a roof than a body.

[nate379]

The panels are MUCH tougher than that. A car wash brush won't harm anything. When the snow gets deep on my panels I use a plastic roof rake to clear them off. (Not to make power... almost no sun in the winter, but because 3-4ft of snow is heavy!)

[J.P.M.]

My production went from 43 to 35 kWh per day over the last 2 months. At first I thought it was high temperatures, less sun hours or most likely increased shading, as I thought. Recently I went to the neighbor's house and looking down to my house I was shocked at how dirty the panels were. Another unexpected thing I noticed was that there was actually less shading of the south panels than 2 months ago. After reading this thread, I was sure I found the culprit.

The next day, what I did was hose down half of the panels for about a minute. The panels were much cleaner looking. I monitored the panels for the next 2 days and here is what I found. The production for both those days was the same as the previous few days prior to hosing, at 35 kWh. The temperature was about 5-10 degrees F hotter on those 2 days, so even though the production did not change, we can assume hosing the panels did make a small difference, which I would estimate at 1-2%, otherwise there would have been a drop in production from the higher temperature.

Here is the real interesting point. I did not find a statistically significant difference between the clean and dirty panels' production. I proceeded to hose down the other half, and for the last week, production has been static around 35 kWh on a clear day. I'm not sure what to make of this. The panels look much cleaner but the production did not change. My hypothesis is that there are other environmental factors not accounted for. The only thing I can think of is humidity. It has been more humid lately. I only considered clear days into my calculations and ignored days with scattered clouds.

Edit for the scientists: the west panels were the control group. Half of the south and half of the East side by side panels were the experiment groups.

Thank you for the info.

1.) Without knowing your arrangement, I'd tend to agree with your ideas on temps., less daylight and more shading, as well as the additional factor of less POA irradiance as the sun moves south would account for most of the difference in output.

2.) I'd also guess that 2 months of no rain would reduce output on the whole array by very ~~ 6 - 8% or so over that period, such that hosing half an array would remove, again, very approx., about 1/2 of the dirt from less than 1/2 of the array, thereby (maybe) restoring a bit less ~ 1/2 X 1/2 of the fouling caused performance decrease, or something like the 1% - 2% you SWAG.

3.) On how the array looks: One more time - The human eye is not necessarily a reliable tool when used to judge how much a PV array's performance is impaired by what look like a bad situation with respect to dirt. My limited experience is that it's not usually as bad a situation as it may appear, in spite of what the panel cleaning companies might want to stampede owners into.

None of this fouling business is as exact, precise, or as amenable to analysis as we'd all like it to be.

Bottom line: Hose the array about 1X/month in the early A.M. if it doesn't rain.

Thanx again for the info.

J.P.M.

Edit: On the humidity: That would probably not have an impact on the electronics or the panel temps. in a direct way. However, increased water vapor in the atmosphere, as well as other atmospheric constituents like dust, ozone, particulates, mixed gasses, etc. can have an effect in atmospheric transmittance and therefore how much irradiance the array sees. The eye is again, somewhat the culprit. All "clear" days are not the same.

Example: Over a recent 17 day period (08/04 - 08/20), around the time of min. incidence angle on my array, which happened between 1315 and 1312 hrs., P.D.T. for those dates, and when the sky was cloudless at that minute (10 days of the 17),the instantaneous GHI at my house, corrected for pyranometer temp., varied from 925 to 970 W/m^2, with the calculated POA values varying from 970 to 1025 W/m^2 respectively. The sky at all those times, on those days appeared "clear" with no clouds. The measured irradiance varied by about 5-6% or so.

FWIW, the estimated GHI from the HDKR clear sky model I use for those dates and at that time is about 998 W/m^2 for my location. Most days are not as clear as the model suggests.

[huge]

My production went from 43 to 35 kWh per day over the last 2 months. At first I thought it was high temperatures, less sun hours or most likely increased shading, as I thought. Recently I went to the neighbor's house and looking down to my house I was shocked at how dirty the panels were. Another unexpected thing I noticed was that there was actually less shading of the south panels than 2 months ago. After reading this thread, I was sure I found the culprit.

The next day, what I did was hose down half of the panels for about a minute. The panels were much cleaner looking. I monitored the panels for the next 2 days and here is what I found. The production for both those days was the same as the previous few days prior to hosing, at 35 kWh. The temperature was about 5-10 degrees F hotter on those 2 days, so even though the production did not change, we can assume hosing the panels did make a small difference, which I would estimate at 1-2%, otherwise there would have been a drop in production from the higher temperature.

Here is the real interesting point. I did not find a statistically significant difference between the clean and dirty panels' production. I proceeded to hose down the other half, and for the last week, production has been static around 35 kWh on a clear day. I'm not sure what to make of this. The panels look much cleaner but the production did not change. My hypothesis is that there are other environmental factors not accounted for. The only thing I can think of is humidity. It has been more humid lately. I only considered clear days into my calculations and ignored days with scattered clouds.

Edit for the scientists: the west panels were the control group. Half of the south and half of the East side by side panels were the experiment groups.

[J.P.M.]

If you have individual panel monitoring and they are all on the same plane, you can measure performance improvement by cleaning only a portion of the panels and do a reasonable comparison of the clean and dirty panels.

Correct. I mentioned as much in a prior post - see the last paragraph of my 08/14/20106 missive posted at 1435 hrs.

Just get a good, clean baseline for comparisons to be able to correct/adjust for small differences between individual panels. After that, I'd respectfully suggest monitoring one cleaned and one noncleaned adjacent panel that have the closest performance and obviously the same shading characteristics, and avoid inintentional cleaning of the "dirty" panel w/H2O overspray or anything that the "clean" panel might get.

[SanteeFJ]

As for rain, my (maybe) slightly better than SWAG on cleaning is : Depending on how much and how long it rains, a "decent" rain will remove about 1/2 - 2/3 or so of the fouling. Thus, a, say, 8% fouled panel will be very approx. 3% to 4% fouled after a decent rain, +/- some depending on how hard and how long it rains. Additionally, how much stuff the rain removes is also f(what the crud actually is and its' composition/particle size distribution and a bunch of other stuff.)

As for trying to get some quantitative estimate of how fouled an array is (or clean performance, which will be needed before you'll know how dirty it is) partly because of the small %age that fouling represents relative to the rest of the variablers involved and the uncertainty/accuracy of measuring those variables, without a pyranometer that's pretty close by, and a way to convert its output to plane of array irradiance, as well as some reasonably consistent way to determine an "average" or representative array temp., your wasting your time thinking you will be able to get a reliable quantitative result for how dirt is affecting array performance.

Even with a pyranometer and a way to estimate array temps. and a lot of very careful and precise/consistent procedures, corralling all the variables is a lot like herding cats. I've done it a few hundred times at my array's min. daily incidence angle. According to SWMBO, the result of having too much time on my hands.

Don't shoot me, I'm only the piano player.

If you have individual panel monitoring and they are all on the same plane, you can measure performance improvement by cleaning only a portion of the panels and do a reasonable comparison of the clean and dirty panels.