How do you deal with freezing rain and solar panels?

[coppercone2]

keep in mind the moisture seal. if you pump lots of amp into a frozen panel it will expand fast and possibly break a moisture seal. ramp it..

[Berni]

To lower the peak power consumption you could only turn on the heaters on one panel at a time. You only need to briefly get the panel above freezing to get the ice directly touching the panel to melt. So focusing all the available power to heat up one panel quickly is likely more energy efficient than using that same amount of power spread out to heat all of the panels slowly. The panels will be loosing heat to radiation and convection all trough the heating process, so the faster you heat it the less heat escapes into the environment rather than being used to change the ice back to water.

I would imagine you would need at least 1kW of heating power per panel to defrost it in a matter of minutes. Taking the worst case of 10 minutes to defrost you are looking at 0.17 kW/h of energy to defrost one panel. So its not that terrible.

Tho im not sure if you could heat the panel up with 1kW by reverse feeding power into it so you likely need to cover the back of the panel in self adhesive heating pads. But from the energy standpoint it probably makes sense to use energy to defrost them, its more of a question if the cost of the heating elements and extra wiring is worth the extra energy production (Especially since solar panels don't produce a whole lot at winter in these areas)

EDIT: By the way the defrosting here is only considering melting the ice directly touching the panel so that the rest can get unstuck and slide off by gravity. If the panels are not at enugh of an angle or the edges stick down too much you might still need to go up and nudge it to get it off. Melting all the ice and snow into water would require a lot more energy.

[apis]

If the panel is about 0.5 m2, a 1 mm ice sheet would be about 0.0005 m3 which is ~0.46 kg of ice.
Latent heat of fusion for water is 334 kJ/kg, so it would require about 153 kJ to melt it.
The heat capacity for ice is about 2.05 kJ/(kg*K) so you need about 940 J per degree C.
So if it's -10 degree when you start you need about 163 kJ to melt it. At 1 kW that takes 163 seconds or a little less than 3 minutes assuming perfect heat transfer. In reality you'll loose a lot of heat to the rest of the system, so you can probably expect 10 min or more at 1 kW per panel.

[Berni]

I think solar panel modules for large roof installations are usually about 1 to 2 m2 in area.

And yes it heavily depends on how much heat is lost to the environment. If you are loosing so much heat that you don't even reach 0°C then even leaving it running for a whole day wouldn't defrost it, hence by having a lot of power is important.

For example one way heat loss could occur is black body radiation. Glass and ice have a emisivity constant reasonably close to a black body(>0.9) and in such a case rising it 10°C above ambient on a 1 m2 panel would radiate off about 350W of heat. However the solar panel also has 2 sides do it so the bottom will also radiate just as much adding another 350W. But it gets even worse. Solar panels generally are pointed towards the sky and the atmosphere lets a lot of the infrared radiation pass out into space where it never returns. This makes the sky appear colder than the air temperature because its taking in more infrared radiation than it is radiating back. For this reason your car windshield can sometimes freeze even if the air temperature never goes below 3°C. The temperature varies a lot with atmospheric conditions but its usually about 10 to 30°C lower than ambient air. This could increase the heat power radiated off the front to rise above 500W bringing the total to 850W just due to radiation. However the smart thing is to do this on a sunny day where the sun provides you with an extra 200 to 400W of heating power for free. Covering the back of the panel with a low emisivity material such as polished aluminum could drop radiated loss on the back to as low as 11W. So depending on your setup and conditions the radiated heat loss alone is between about <50W (insulated back on a sunny day with favorable atmospheric conditions) to 850W (no insulation, at night)

The other major form of heat loss is trough convection. This could again vary a lot, i would imagine on a windy day you need >1kW just to overcome this heat loss.

So yeah 1kW per panel seams like it is the absolute minimum needed to defrost a panel in favorable conditions.

EDIT: Actually i got my numbers wrong for the solar irradiation. A sunny day can bring you about 600 to 1300 W/m2 of power.

[apis]

Yes. If you are lucky and there's a bit of "snow insulation" on top and around the panels that can probably help a lot. Worst case would be a thick layer of ice during a clear, cold and windy night.

I wonder if you couldn't put a sheet of transparent plastic on top of the panels. Then you remove the plastic and shake off the ice. Might require a bit of work to create a setup that is convenient. You would loose a little bit of efficiency of the panels but probably not too much and you only need it while there is risk of ice (not sure how much light a plastic sheet would block). Building plastic is pretty cheap.

[Berni]

Actually i did a bit of a miscalculation. Turns out you can get over 1kW just from a sunny day alone.

So if you are going to get a clear sunny day with no wind and -10°C ambient perhaps the only thing you have to do is get all the snow off it so that the panel can absorb most of the energy rather that just reflecting the light back into space.

We are going into the 2nd winter with our 38 panel array and so far we never had problems with ice and show. Yes the array produces very little power when snow covers it but we tend to soon enugh get a nice sunny day that warms things up enough for the snow so start sliding. Once one part starts to slide it tends to trigger all the rest on the same array with it making for a loud booming avalanche of snow suddenly coming down from the roof and revealing perfectly clean panels that once again start making proper power. The area in front of that array is not a place where people hang out, or anything of value is stored so we don't care about the sudden snow mountain appearing there, we have not installed any snow braces on purpose so that the snow would slide.

But if we get freezing temperatures flowed by rain im sure we could get into a similar problematic situation.

[apis]

On a clear day you get about 1 kW/m² perpendicular to the sun, so a panel would always get less than that from the sun, but a few hours of that might be enough if it isn't too cold and windy. The problem is that a layer of snow on top of ice would reflect away most of that solar energy (otherwise the panels would continue producing electricity as well).

[Red Squirrel]

Yeah I always figured the rule of thumb was around 1 kw of energy per square meter of sun light.  Of course that changes based on various conditions etc.  But if THAT does not melt snow (which it does not, hence why snow stays on the ground) then it means I need more than that to melt snow on the panels.  Though think I will indeed experiment with heating 1 panel at a time using a couple kw of ramped up heat.  Since I only need to melt under the ice and not the entire ice maybe it is indeed doable.  Something I can experiment with next year.      If this works and my only limit is battery capacity then suppose it's not a big deal to get like 4 or 8 golf cart batteries in there.   Can never oversize a battery bank, in fact for solar it's probably better, since if I'm always riding on 90%+ capacity it means they will last longer.

We have actual sun today - like actual direct sun light, which is very rare.   I went out a while ago to clear off the snow from overnight (still crusty snow left over, just skimmed the top soft snow) and I'm actually getting around 5 watts which is not bad considering they are pretty much all covered. It's at least enough to put a bit of charge in the battery.  With this sun light, maybe it will indeed cause the panels to heat up and then start melting the ice.   

But yeah think for a small array of this size where I don't have much property to do anything special like a ground mount setup that is tilted better, my best bet is to probably just experiment with ways to heat them.  Whether I do backfeeding, or put some nichrome wire on the back.   What is the best way to insulate nichrome wire, is there any kind of sleeve that is meant for that?  I'm thinking next summer I will look into putting nichrome wire on the back of the panels and then I can switch them individually.  I still need to design and build the control system anyway, like I want a voltage monitoring system and energy tracking system and what not, just have not gotten around to doing it. I can do the heating with that system too.  Basically just have a bunch of relays controlled by arduino.  Will do outside lights and stuff too.

[Berni]

Well more of the problem is that snow is really good at reflecting light. It can reflect 80 to 90% of light so that brings your 1000W/m2 down to only 100W/m2, the rest is just sent back into space. Perhaps once the snow is cleared away the crusty ice under it would not be as reflective so more of the light would end up absorbed to help heat the panel.

Snow is also very insulating so when it does get hot enough to melt only the very surface of the snow melts, the inside is protected by the insulating properties and the heavily heat consuming melting process keeps the outside edge from ever rising above 0°C. But if you want to melt snow on a cloudy day by heating the panel this could be a great advantage because the snow is insulating the panel from the cold sky and protecting it from wind. Helping you get that very bottom layer of ice to melt so the rest slides off the roof. But still i think you will need 1kW to do the job efectively and efficiently.

I would not recommend back feeding this much power into a panel as its way above the rated current. Nichrome wire is an option, you don't really need to insulate it, just make a pattern that never crosses the wires and use some sort of selfadhesive film material to stick it down to the back side of the glass solar panel. Ideally you want to snake the wire back and forth many times as one length of wire to spread the heat and minimize the number of wire joins required. It might actually become so long that it makes more sense to use copper or aluminum wire, since nichrome might need to be unpracticaly thick to have low enugh resistance to draw 1kW. You will also want to use a fairly high voltage to run it since 1kW at 12V needs 83A and that needs rather beefy cables.

One alternative are also silicone heating pads:
https://www.alibaba.com/product-detail/electric-water-boiler-heating-element-silicone_60471551007.html?spm=a2700.galleryofferlist.normalList.210.46e46a55Ofpbic
They aperantly make even massive ones that are 1x3m in size and rated at 13kW
So perhaps you can find ones that are about the right dimensions for your panels and stick them to the back. Or perhaps use one of them them as a temporary heating blanket that you manually go place on a panel and turn it on for a while to melt it before taking it off again or moving it to the next panel.

[apis]

The problem with the sunlight and snow is that snow can reflect as much as 90% of the energy. And the 1kW is for a surface perpendicular to the sun. Where I live the maximum sun-ground angle was 12 degree today. I think that means the ground gets sinus(12)*1kW/m² =  200 W/m², and if 90% of that is reflected it leaves 20 W/m2. Not nearly enough to melt the snow.

If you have a heater in direct contact with the panels you transfer the heat through conduction, not radiation, so most of the heat will go into the solar panel and the ice and snow on top of it (if the conditions are right as Berni pointed out). It still takes a lot of energy to do that though.

If you want to try heating them maybe something like floor heating wire would work:
https://www.ebay.com/sch/i.html?_from=R40&_trksid=m570.l1313&_nkw=Underfloor+Heater+Cable&_sacat=0

[Red Squirrel]

Yeah true I guess with the reflective properties the snow is not actually absorbing 1kw so that's why it does not melt.  Actually that would make for a neat experiment, if you simply spray paint snow black it would probably melt. Not that I want to go spray painting the snow on my panels but it would maybe work! lol

But yeah maybe playing with infloor heating wire, or nichrome wire is best bet can then tape it behind.  Just need a tape that won't unstick with the heat/cold cycles. 

Something to try next summer. 

[bicycleguy]

Although I now live in sunny southern California I did my time growing up in Wisconsin.  We used to make huge snow and ice forts.  Surprisingly the easiest way to burrow holes for vents or spying was with a garden hose and whatever the tap water temp was at the time.  Solid ice won't cut, but for regular snow plain water can do wonders.  It takes a lot of heat loss to freeze water.  The solar panel glass can't absorb much and the snow has no mass.  For a one time job, I think it's doable.

Of course you can screw-up, but the way your pictures looked, and if say >20F I think you'll be surprised.  The closer you can get the better.  Keep an eye on where that cold water is going.  Any puddling on the roof is a show stopper.  Start at the gutter and keep a flow and work your way up a panel.

Try it on your car first to get a feel for how it works and whether it's too cold out.  Make sure you have somewhere for the water to go or you'll regret it.  We used to pull a hose out of the basement laundry room.  Of course don't use an outside faucet without a way to drain it!

[Red Squirrel]

Would go well if I want to have a hockey game on my solar array.  The water comes out the faucet at around 0-1C so I would need to heat it up first.  Though I guess that would work.  Think I'm going to lean towards the heat tape though, might just throw it on top and plug it in to the house for a day, at least just to get it going.  For future I'll want to come up with something more self sustainable like the nichrome wire under the panels. 

[bicycleguy]

Obviously you didn't try it.  Take a pail of that cold water and poor a steady stream onto the snow in the yard and see how far it goes down.  I'll be surprised if you can't see grass.  If it was that quick to freeze you wouldn't need a Zamboni.

[edpalmer42]

Three more random ideas....  Implementation details are to be determined. 

Ice melting pellets that you normally put on the sidewalk?  I don't know how expensive or toxic they are, but they seem to be used a lot.

A black 'window shade' that you pull down over the panel.  When not in use it would be rolled up like a normal window shade or run around to the back of the panel to keep it from getting covered with snow.

If you decided to put hoses on the back of the panel for heating and cooling, you might be able to do the cooling for free.  Put a check valve on both ends of the panel so that fluid can only go one way.  When the fluid heats up, it pushes out in the only direction it can go.  But since it's in a sealed system, that pushes cool fluid in from the other end.  You've got pumping action without a pump.  Automotive block heaters that are installed in the heater hose use this idea.  Now add something like a car radiator to cool the fluid just with passive air flow and you're done.  I don't know how much heat you could remove with a system like this, but it wouldn't be too hard to set up if you've already got the hose in place.

Ed

[coppercone2]

careful with solutions so you don't get debris or animals making flammable structures in heating systems.

[Red Squirrel]

Oh yeah would not really have any temps that are hot enough to ignite stuff.  Maybe like 40C at most at the actual heat wire.

We just keep getting dumped with snow now so pretty much given up for the year.  I still go out in snow shoes to take it off as the weight might crack the panels but no way I'm getting down to the glass this winter.

[floobydust]

Locally I see many household solar arrays covered in snow. For some it rolls off and others people brush it off. It depends on the angle. I actually use 90 degrees for small (a couple hundred watts) and it works fine, if the tree line is not too high. Applying car wax on the panels helps the snow/ice slide off, like ski wax.
There are blocking diodes in all larger panels so injecting current would not heat up the cells I reckon.

[coppercone2]

are they like integrated into the panel or maybe you can put a relay across it?

[Red Squirrel]

Pretty sure they are integrated, probably in that black box where the wires go in, though it might be possible to open it and just bypass the diode?  But don't think I'll go that route though, I'll have better heat control if I just use a separate source of heat, and bonus is the panels will begin to produce right away as I won't be messing with their connectivity to the charge controller.  I can set the logic of the melt system to move to the next panel as soon as it sees a spike in power production as it means everything slide off.   Though I might not bother going that fancy, just stick a web cam so I can watch it and switch manually.

[Berni]

I don't think you need to disconnect the bypass diode.

The diode basically prevents the voltage across the terminals from going negative (Or at least under -0.7V) when the panel is in the dark but others are not. To push power into the panel you want to provide a voltage into it that is higher than the voltage it would produce from light. This turns the solar cell elements into a forward biased diode. This region of operation is probably a lot like a LED where the current has to be regulated. Except i don't think a solar panel can produce any light (Maybe it produces IR light?), just heat like a normal diode.

But in any case i don't think its a good idea to back feed the required 1kW into a solar panel that is rated for 200W, the interconnects might not be up to the job or the parallel cells perhaps don't current share very well. Perhaps hotspots can occur inside one of the cells? In any case i would first test this on a smaller panel and a thermal camera to see how well the heat spreads on a back driven panel.

Sticking heating elements to the back is likely easier anyway.

[duak]

Photovoltaic cells are forward biased when operating.  That's why each cell only contributes 0.6 V or so.  If the cells had a greater barrier potential, say like GaAsP at about 1.5 V, you'd need fewer cells.  The bypass diode would have to be bypassed itself to get anything other than photocurrent to flow.  Shortling out the panel shouldn't cause any damage and will cause any photocurrent to be converted to heat right at the panel.

Last year I had some ice frozen on the back steps.  They are oiled cedar and I didn't want to scrape them too hard so I used an old electric iron set to low heat to melt thru the ice.  Worked great - Not energy wise, but this was a safety issue.

Auto windshield de-icers contain alcohols of various types but they can and do cause aluminum to corrode.  Be careful if using any salts, especially alkaline, as it will attack anodized aluminum.

Cheers,

[Red Squirrel]

We have some sun in the forecast in the next few days so I'm going to try to get up earlier in the morning to go skim the overnight snow off.  With full direct sun all day, and -30's they might actually produce enough power to heat up on their own and melt the ice.  I'll see what happens.  The battery is reading -15 just now and it's at 12.7v so it's actually still holding a charge.

[Berni]

Its -30°C there? Well the sun is going to have its work cut out for it then.

On and at -15°C careful that you don't drain your battery too much since lead acid cells freeze a lot easier when discharged and must not be charged back up until they are thawed out again.

[Red Squirrel]

Yeah been keeping an eye on battery and it seems to keep itself charged even through the snow/ice, surprisingly.  I get about 5-10  watts on a good day. 
 
Eventually I want to setup a microcontroller and raspberry pi to monitor everything and run fibre to it so I can monitor remotely.