1 months worth of dust equals 4% of lost production where I live.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Effect of dust on production
- Thread starter sorka
- Start date
nwdiver
Well-Known Member
.... same as 2, 4, 8 and 16 months
You're removing the dust that's keeping more dust off!This would make for a fun experiment. Clean one of those panels monthly and the other every 2 months. See what the difference is after a couple years.
Nice work!
I wonder what the detail data looks like. For example, what was the gap on the day after they were cleaned? Basically how fast does the dust/soiling start affecting the production and when does it start leveling off.
I've also wondered if the dust affects production in a small measurable positive way when the sun is perfectly perpendicular to the panel. Does the dust act as reflector for light coming in at off angles and reflect it back on to the panel thus reducing the penalty for soiling.
If the measurement were taken right after they were cleaned, it would be nice to see a comparison a day or two later. Especially if a moderate amount of water was used, the panels that were cleaned would be somewhat cooler than the uncleaned panels, as cooler panels can cause a few percent higher production for a short period as well.
The panels were cleaned the night before this measurement. I'll be looking at them every day after sundown to compare. We'll see how much dust prevents dust occurs vs being cumulative. I'm sure it's somewhere in between. I'll bet if I went two months that they would eventually converge and there'd be no difference between the panels that had been cleaned and ones that hadn't. But we'll see. I'm willing to run this for a while, but if the difference gets too large, I'll have to clean them all. In a few weeks, I'll clean a different two panels.
They were cleaned the night before. I'd never do this during the day when the sun is up. I also rinsed with deionized water and did so the night before as condensation was starting to form. This prevents even the soft water I rinsed with from forming any mineral deposits on the glass as all that water will be displaced through the night be condensation which is mineral free.
Also, I realize I misstated the loss. The dust free panels yesterday made 2.35 / 2.2592 = 4.02% more than the dust covered panels but that's not the same as the % lost from dust. The dust covered panels are making 2.2592 / 2.35 = 96.146% of the cleaned panels so the loss is 3.86% or 3.9% if you want to round up.
I love data! I unfortunately don't have module level instrumentation to perform this experiment. Thank you for doing this!
I would have preferred micro inverts to eliminate single points of complete failure but at least I'll know if a single panel or optimizer fails.
aesculus
Still Trying to Figure This All Out
It's tree pollen season where I live and I have noticed about a 10% reduction in production over the last week. So in a few weeks, once the trees are done doing their 'thing', I am going to wash all my solar panels off with soft water and the special 'water broom on a pole' device I bought last year for this reason.
I will then have a pretty good idea how much of a difference it makes.
I will then have a pretty good idea how much of a difference it makes.
As one who seems to have almost always lived in hard water areas, I am sure you are aware, but soft water can still have minerals in it. Rinse agents help. You might want to wait for a foggy night when you have condensing water to do a "final rinse", like @sorka.
All the best,
BG
aesculus
Still Trying to Figure This All Out
Silicon Desert
Well-Known Member
have found that after occasionally cleaning panels on my homes in different areas over the last 15 years, it really made absolutely no difference what water I used. And 4%, plus or minus a couple percent is about what I experienced with various levels of dust and dirt.
Ok, that's a long wait.In my area, Sierra foothills, I'll be waiting until November.
All the best,
BG
charlesj
Active Member
Well, the experiment is not so simple. The panel 2 and 5 have a difference of 1.3% in that picture. How do those two compare over historical past, always a 1.3% difference? Or there is a variance?
Then panel 3&4, the clean one have 2 decimal equivalence. Have they always be equal in production in history or there are time when they vary?
Number 5 is also on the outside, more ventilation, will that mess up the experiment?
Perhaps if that 4% difference can track over time, it may have some relevance?
Then panel 3&4, the clean one have 2 decimal equivalence. Have they always be equal in production in history or there are time when they vary?
Number 5 is also on the outside, more ventilation, will that mess up the experiment?
Perhaps if that 4% difference can track over time, it may have some relevance?
Nice, I'll be looking forward to more data over multiple days.
I have Enphase micro's (non-Tesla array), so have per-panel data including instantaneous temps, and do clean my panels about 3-4X a year. But I have such a variance in shading and temps on each panel over the course of a day, that I've never been able to consistently quantify soiling losses, even trying an A/B test like this. Since I can't use the full-day energy due to the shading differences, I have to compare the instantaneous readings only. But my panels instantaneous output varies by up to like 7% commonly.
I do perceive you might be getting a tad more cooling and slightly higher energy from the panels closer to the right side hip and from the top (esp if there is no skirt on the top edges). My right side panels are so close to the hip their corners even cross the hip line a bit, with 4" high standoffs - so I regularly see instantaneous outputs several % higher on panels on the hip/gable edge, since I think they can get the cross-breeze for better cooling.
I was just considering that. For the year, these two panels are already on the low side of production compared to the group as a whole. For the month, they're about even but that's because the moth screen shot includes 3 days of clean production. I have no shading even on January 1st so it appears the loss from dust in this case might actually be understated.
Similar threads
