Panel cleaning

[J.P.M.]

fwiw, I got a long window washing tool from the cleaning supply store up the street for $80. Squeegee on one side and soft fuzzy cloth on the other. It did a pretty good job cleaning one panel with just the dew, but I haven't done any measurements.

You might consider a longer term experiment where you never clean one row of panels.
JPM's results make me think that no-clean might be an acceptable option; if each rain resets to 98% output, it's hard to justify the effort.

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.

[DanKegel]

I wanted to figure how often I should wash my panels since "it never rains in So Cal". I have the luxury of Enphase micro inverters so I washed half of my panels after 2 weeks of prior cleaning. What I discovered was the dirty panels had lost ~ 1.5% power over those 2 weeks. To verify, I then did a total cleaning and all panels returned to equal output. So I now feel I'm justified in washing them very two weeks. It's easy because I'm single story and can wash them from ground level.

I'm researching soft bristle brushes and long handles so I can do a more thorough cleaning perhaps annually.

fwiw, I got a long window washing tool from the cleaning supply store up the street for $80. Squeegee on one side and soft fuzzy cloth on the other. It did a pretty good job cleaning one panel with just the dew, but I haven't done any measurements.

You might consider a longer term experiment where you never clean one row of panels.
JPM's results make me think that no-clean might be an acceptable option; if each rain resets to 98% output, it's hard to justify the effort.

[J.P.M.]

I've had my panels since Feb this year and have cleaned them with soft water since the beginning. I bought a small portable softener made for using on RV's and my results have been positive. I wanted to figure how often I should wash my panels since "it never rains in So Cal". I have the luxury of Enphase micro inverters so I washed half of my panels after 2 weeks of prior cleaning. What I discovered was the dirty panels had lost ~ 1.5% power over those 2 weeks. To verify, I then did a total cleaning and all panels returned to equal output. So I now feel I'm justified in washing them very two weeks. It's easy because I'm single story and can wash them from ground level.

I'm researching soft bristle brushes and long handles so I can do a more thorough cleaning perhaps annually. I see a lot of dirty panels in my neighborhood and they have to be down 10-20 % production.

Thank you. That 1.5 % over ~ 2 weeks you report seems about in line w/ my 0.75 % - 1.0 %/week. Seeing as how most of this business on panel fouling is probably quite variable and not very predictable as to location, season, rainfall, and type of stuff floating around in the air, my opinion, FWIW, is settling in on an ~ 1X/month hose down of an array as the best tradeoff between a reasonably clean array that doesn't have a large performance penalty due to fouling, and getting on a roof with a bucket, etc. I'm still of the opinion that commercial cleaning of an array is a rip[ off and a waste of money.

On the idea of appearance being a valid indicator of fouling penalty, last year, my array, after about 7 weeks with no rain looked very much like most neighbors' arrays - dusty with the bathtub ring on the bottom 12" or so of all the panels. My performance at that point had dropped about 4-5 % from the clean condition.Then it rained. Performance came back to ~ 97-98 % of clean. Arrays in my neighborhood mostly never get anything but rainwater. If array fouling rate is not asymptotic at some point, at 1%/week, an array would cease to function at some point, and altogether at 100 weeks if the fouling rate was 1%/week. Skylights don't seem to block a lot of light. Of course, it does rain. And to perhaps belabor a point, this is not a precise endeavor.

Also to repeat, the eye is not necessarily the best tool to use to gauge how much an array's performance has deteriorated because of dirt/dust/water spots, etc. Again, qualitative info is scarce and people sometimes do no more than repeat what they hear or what seems common sense to them, but is not necessarily good/correct info. Looks can be deceiving.

Be careful of bristles/chemicals. They may interact with a panel's ARC coating and hurt performance by increasing panel reflectance. Such actions may also cause warranty claim denials or hassles. I use a soft cloth or microfiber and a light touch.

Thanx again for the info.

[SanteeFJ]

I've had my panels since Feb this year and have cleaned them with soft water since the beginning. I bought a small portable softener made for using on RV's and my results have been positive. I wanted to figure how often I should wash my panels since "it never rains in So Cal". I have the luxury of Enphase micro inverters so I washed half of my panels after 2 weeks of prior cleaning. What I discovered was the dirty panels had lost ~ 1.5% power over those 2 weeks. To verify, I then did a total cleaning and all panels returned to equal output. So I now feel I'm justified in washing them very two weeks. It's easy because I'm single story and can wash them from ground level.

I'm researching soft bristle brushes and long handles so I can do a more thorough cleaning perhaps annually. I see a lot of dirty panels in my neighborhood and they have to be down 10-20 % production.

[J.P.M.]

There is a lot of variation of people's experience to make me suspect that the kind of panels also plays a part.

Some, but not much. Sunpower panels get just as dirty as anyone elses. More of the variation is probably due to local or micro climate effects as well as panel tilt. Overall, trying to get some quantitative grip on what's usually a small number (several % or so) is difficult even under controlled conditions with decent equipment. This is a task that's about halfway between qualitative and quantitative and my opinion is that's probably good enough. As in : Do the panels look dirty ? If so, hose'em off about 1X/month or so if it doesn't rain. Commercial cleaning is a rip off.

Bottom line for measurements I've done: My array fouls at a rate of about 0.75% - 1.0%/week without rain. That's somewhat variable and non linear. That rate may decrease as the dry spell continues but I'm not sure about that. And, a decent rain removes ~ 1/2 to 3/4 of the effects of dirt.

That's probably about as precise as I can get and my guess is that rate is somewhat typical in So. CA, but by no means universal. It's probably an adequate description and about as precise or more precise than needed for the task.

Thanx everyone for the information/opinions. I'll report my results from ongoing measurements from time/time.

[huge]

There is a lot of variation of people's experience to make me suspect that the kind of panels also plays a part.

[nomadh]

After my panels were up for a year I washed two of the 40 panels. The two panels were on the ground at the time and I made sure I did a very good job cleaning them. I have micro-inverters and after the cleaning of the two panels, I could not detect a difference in the power output between the clean panels and the other panels. I live in a dusty area, but my panels are on the top of a 2nd story house. In my situation, I don't think cleaning my panels very often is worth it. When I look at the average peak power decline and the total power decline over the last year, I get a -1.3% change in both. It seems reasonable that each installation has different payout for cleaning panels. I can't see, but I expect there is some heavy dew on my panels occasionally that might help clean them.

The other thing to watch out for is making sure you don't let hard water dry on the panels and leave calcium deposits to block the energy.

I thought I saw my performance dropping and I have more of a farm atmosphere around my house than most anyone here. We dont have the horse any more but the emus and tortise keep much of the backyard 1/2 acre a dust bowl. Also the driveway is dirt and ours is a very fine powdery dust. Between cars and animals the dust is pretty thick in a month. I'd guess I have more dirt in a month than most standard suburban homes have in a year.
So I went up recently and hosed the panels down at dusk. I figured dusk as the best so we dont crack anything while hot and allowing slower drying. Hopefully having the hard water be partially wiped down overnight with a dew layer before the hard water was baked on during daylight. So I cleaned up 2 1/2 months of seriously dirty looking dust but watching output levels over the next week I didn't notice any noticeable output change. I suspect a 1-2 % change but without any kind of accurate automated constant monitoring I can't tell. And it seems for me the panels didn't look any more dirty after 1 month than 2 1/2 months.
I suspect I'd notice 5% and certainly 10% but it just isn't there. I may still clean them a few times a year just as a reason to check on them and keep my caustic dirt from pitting the metal and or maybe even the glass.

[philips]

Just to add another data point, I hosed off my panels the evening of July 8th and noticed a roughly 8% improvement based on looking at pvoutput data => http://pvoutput.org/list.jsp?userid=44718 The last time it rained was May 6th and the last time I cleaned them was April 2nd. When I cleaned them in April, it looked like recent rains had kept them fairly clean, they mostly had the yellow bee dropping spots on them that I scrubbed off.

My panels most likely soil faster than most in my area since they are horizontal and some neighbors have taken to drastic water saving measures that still get mowed each week by the gardeners creating a lot of dust.

[J.P.M.]

After my panels were up for a year I washed two of the 40 panels. The two panels were on the ground at the time and I made sure I did a very good job cleaning them. I have micro-inverters and after the cleaning of the two panels, I could not detect a difference in the power output between the clean panels and the other panels. I live in a dusty area, but my panels are on the top of a 2nd story house. In my situation, I don't think cleaning my panels very often is worth it. When I look at the average peak power decline and the total power decline over the last year, I get a -1.3% change in both. It seems reasonable that each installation has different payout for cleaning panels. I can't see, but I expect there is some heavy dew on my panels occasionally that might help clean them.

The other thing to watch out for is making sure you don't let hard water dry on the panels and leave calcium deposits to block the energy.

As for hard water stains blocking irradiance, as a subset of my array fouling estimates, I cleaned my panels on 2 consecutive "clear" days on a couple of separate occasions. The first day with tap water which does leave a residue, and the next day with distilled water, using a cloth to remove water drops for the distilled water cleaning which resulted in no visible residue. I was unable to detect a difference in the array's instantaneous performance from measurements taken at the 12 min. around the time of min. array incidence angle.That was done by measuring individual panel temps. as described in previous posts, measured GHI at the array, wind velocity at the array and air temp. at the array, converting the GHI to plane of array irradiance and calculating measured efficiency against theoretically clean efficiency. The difference in measured efficiency one cleaning method to the other - one with hard water spots and one without hard water spots was about .003, or 0.3%. Well below my instruments and methods precision of ~~ +/- 0.5% - 0.75% or so if I'm careful and lucky. If water spots impair performance, I can't measure it.

Not claiming perfection, but what I measured may me more than anecdotal.

Also, as it turns out, and depending on the nature of the light, and its frequency, the human eye is probably not a very good instrument to judge the transmittance or reflectance characteristics of various films with respect to solar radiation frequencies. In other words, hard water spots may very well look worse than they are with respect to how much they impair solar PV production. Sometimes things are not what they seem.

[Ward L]

After my panels were up for a year I washed two of the 40 panels. The two panels were on the ground at the time and I made sure I did a very good job cleaning them. I have micro-inverters and after the cleaning of the two panels, I could not detect a difference in the power output between the clean panels and the other panels. I live in a dusty area, but my panels are on the top of a 2nd story house. In my situation, I don't think cleaning my panels very often is worth it. When I look at the average peak power decline and the total power decline over the last year, I get a -1.3% change in both. It seems reasonable that each installation has different payout for cleaning panels. I can't see, but I expect there is some heavy dew on my panels occasionally that might help clean them.

The other thing to watch out for is making sure you don't let hard water dry on the panels and leave calcium deposits to block the energy.

[J.P.M.]

Oh but it can. Medical bills from falling off your roof or ladder can erase all solar savings for your entire lifetime. I'll wait on the rain.

But the benefit to society will be an improvement to the gene pool. Or, live in an area that gets 40-50 cc annual precip. Or, as I and others suggest, oversize the array by a few % to make up for fouling and consider the degree of added cost ineffectiveness as the insurance premium against gravity induced accidents.

[azdave]

Hosing things off before the sun rises too far can't hurt.

Oh but it can. Medical bills from falling off your roof or ladder can erase all solar savings for your entire lifetime. I'll wait on the rain.

[J.P.M.]

Update, 8:00 p.m. Today's total is 27.86 kWh. Much better than yesterday's 24.369 kWh and better than the previous highest August day so far this year of 26.397 kWh. Methinks washing the panels did some real good. As for the weather, it has been dry and hot this month so far with quite a few days over 100 degrees. Today peaked at 106 here in Bakersfield according the NWS and yesterday was "only" 101, so all things being equal, today's power generation should have been a bit less than yesterday.
Kirk

Hosing things off before the sun rises too far can't hurt.

[BakerPV]

Update, 8:00 p.m. Today's total is 27.86 kWh. Much better than yesterday's 24.369 kWh and better than the previous highest August day so far this year of 26.397 kWh. Methinks washing the panels did some real good. As for the weather, it has been dry and hot this month so far with quite a few days over 100 degrees. Today peaked at 106 here in Bakersfield according the NWS and yesterday was "only" 101, so all things being equal, today's power generation should have been a bit less than yesterday.
Kirk

[J.P.M.]

I washed my panels last evening before it got totally dark using Windex Outdoor Window Cleaner (the kind that screws onto the end of a hose) and my output at peak today was about 11% higher than yesterday (same weather conditions). It'll be interesting to see the total kWh I get at the end of today compared to yesterday. My system is fourteen 327 watt SunPower SPR-327NE panels and a SolarEdge SE5000, facing 110 degrees (East Southeast). They were installed at the end of April and have gotten some rain maybe twice. We have new homes going up near us and there is quite a bit of blowing dust from the empty lots. The panels had a very noticeable layer of dirt on them, and some spots of bird poop, and now they are actually blue/black again. I was thinking the lower output I was noticing over "just-installed" was due to the high temperatures we've been having lately (100+ degrees), but now that the panels are clean, I'm thinking dirt was the main culprit. Since I only have to stand on a ladder to get above a patio awning, and don't have to get on the roof, I believe I'll be hosing them off every other month or so now. I checked with SunPower before using the Windex and they said it was fine.
Kirk

Thank you for the info.

I've no problem believing the # you quote, but without an on site weather station (or even with one for that matter), its hard to separate weather effects from dirt. I'm SWAGing ~~ 0.75% - 1.0% performance degradation/week for my location with no rain, and I suspect that rate may become noticeably asymptotic at some point beyond 6-8 moisture free weeks, but I've got no data to support that suspicion. Your #'s seem to sort of agree on the rainless rate. This fouling business is far from an exact and predictable sort of thing anyway.

FWIW, I'm in the 10th day of a 13 day series of very thorough, daily (early A.M.) array cleanings, with instantaneous measurements/recording of irradiance and weather data, and important array parameters taken at solar noon on each day at the inverter. I do this around April 25 or so and also around August 11 or so when the incidence angle is closest to zero on the array and check performance degradation when glazing reflectance is probably at a minimum and approximating something perhaps close to STC incidence angle. I adjust for (average) array temp. as calculated using comparison of array voltage with measured array temps. done in the past, calc the POA irradiance using a consistent HDKR model, throw out the high and low resultant calculated % fouling number, average the rest, and call that the clean array performance. That ave. calculated fouling (assuming my 13 consecutive days of thorough cleaning gets things about as clean as they're likely to get) then becoming the cumulative performance degradation of the system. This year, I intend no cleaning until the winter solstice, but will continue to monitor instantaneous output in the usual fashion on very clear days as I've been doing for over two years.

One advantage of a micro inverter system for me might have been that it would be a whole lot easier to get a number for array fouling by daily or somewhat regular cleaning of just one panel. Too soon old. Too late smart. Just sayin'.