Cloudy day impact on solar production & sizing

[arghx7]

I live in Michigan and we only have consistent sun in the summer months. The rest of the year there are a lot of cloudy days. My question is, what kind of decrease in solar are you guys experiencing on overcast days? How much would you need to upsize your solar system to account for this? Assume you have two powerwalls and maybe a smaller to medium sized house (1500 to 2500 square foot, which I'd call medium by 21st century American standards). Let's say you want to survive off grid for 2 days

Most of my big appliances run off natural gas (oven/stove, drier, heat) and I only need significant A/C maybe 2 months out of the year. I also charge my Model 3 for free at work. I'm using between 300 and 500 kwh per month of actual electricity righ tnow, and I pay about 15 cents per kilowatt hour (fixed price I think).

Where I live solar + powerwall makes no financial sense, so this is just curiosity at this point.

 

[Ampster]

Snow removal is going to be a big issue in Michigan. One week of panels covered in snow could be like a month of overcast days. It would be hard to get more granular than that. I would definitely over panel. I am in California and have an DC to AC ratio of 1.50 to 1.

 

[Patrick66]

Best way to look at it is to go to pvwatts.nrel.gov. I did a quick check for Ann Arbor, and in December you would get an average of 2.5 hours of sun there if your panels are sloped 50 degrees pointed due south. For your 500kWh/month value, that would mean you need at least 7kW of panels for “average” conditions for the month, and likely 10-11kW if you wanted to be confident that it would be enough.

That would mate pretty well with two powerwalls.

The economics wouldn’t be too good unless maybe you installed it yourself— about a 10-year payback with current federal incentives.

 

[Dan123]

I live in Michigan and we only have consistent sun in the summer months. The rest of the year there are a lot of cloudy days. My question is, what kind of decrease in solar are you guys experiencing on overcast days? How much would you need to upsize your solar system to account for this? Assume you have two powerwalls and maybe a smaller to medium sized house (1500 to 2500 square foot, which I'd call medium by 21st century American standards). Let's say you want to survive off grid for 2 days

Most of my big appliances run off natural gas (oven/stove, drier, heat) and I only need significant A/C maybe 2 months out of the year. I also charge my Model 3 for free at work. I'm using between 300 and 500 kwh per month of actual electricity righ tnow, and I pay about 15 cents per kilowatt hour (fixed price I think).

Where I live solar + powerwall makes no financial sense, so this is just curiosity at this point.

You can put your address into PVWatts Calculator and it will give you a pretty good estimate (by month or by day/hour) of your solar production, and it takes into account clouds.

 

[arghx7]

Thanks. I checked that site and yeah they are estimating maybe half or less sunlight, and that doesn't take into account snow.

Does anyone here have any real world data to share on their own systems as to how much energy they lose on a cloudy day?

 

[SoundDaTrumpet]

Thanks. I checked that site and yeah they are estimating maybe half or less sunlight, and that doesn't take into account snow.

Does anyone here have any real world data to share on their own systems as to how much energy they lose on a cloudy day?

I dropped from 15kWh to 1kWh from Sunny to Rainy. I imagine snow to be 100% loss =). All kidding aside, just get friendly with a neighbor who has solar and have him text you a monthly screen shots from his/her solar phone app. Solar owners tend to be enthusiastic about talking about their systems. Good luck!

 

[arghx7]

Thanks for the input. That's a pretty drastic drop.

There is literally nobody in my neighborhood with solar. Not a single one out of about 100 houses, for the economic reasons described above (Natural gas for all appliances and heat, limited need for A/C).

 

[MorrisonHiker]

Thanks. I checked that site and yeah they are estimating maybe half or less sunlight, and that doesn't take into account snow.

Does anyone here have any real world data to share on their own systems as to how much energy they lose on a cloudy day?

Like @SoundDaTrumpet , we see about a 90% drop in production from about 15 kW to 1 kW if it is cloudy. We are able to even have some solar production in a light rain. Fortunately, that's usually enough to cover the base household load but it doesn't really allow us to charge up the Powerwalls or export to the grid.

In Colorado, we have over 300 sunny days a year so typically snow doesn't last long. Fortunately, a third of our panels are on a steep roof and the snow usually slides right off of them once the sun comes up. We also got a roof rake to help clear the snow off another 30% of our panels on two other arrays. We had about 10" of snow earlier this week but only about 40% of our panels are currently producing since I haven't gotten out to clear the "back 9" panels off yet. Normally we'd be producing about 14 kW at this time of the day but the highest I've seen so far today is 7 kW.

 

[Dan123]

Thanks. I checked that site and yeah they are estimating maybe half or less sunlight, and that doesn't take into account snow.

Does anyone here have any real world data to share on their own systems as to how much energy they lose on a cloudy day?

Thanks. I checked that site and yeah they are estimating maybe half or less sunlight, and that doesn't take into account snow.

Does anyone here have any real world data to share on their own systems as to how much energy they lose on a cloudy day?

You can see here:

Residential System

It is really a difficult question to answer, because almost every day is cloudy. It just depends on how cloudy, what type of clouds, and what time of the day.

 

[ajdelange]

Does anyone here have any real world data to share on their own systems as to how much energy they lose on a cloudy day?

Yes, but it won't be of much value to you. You will get what you will get and the best you can do is, as others have suggested, look at the NREL data for your area. That data is, presumably, averaged over several seasons but keep in mind that next season might not be like this season and can differ appreciably from the average.

To give you some idea, NREL says I should get 4.71 kWh/m^2/day in my locale (near D.C.) for the month of October. I got an average of 2.53. October was unusually cloudy here. But note that the reduction includes not only clouds but siting effects (azimuth, tilt). My panels are affixed to my roof planes which are oriented the way they are oriented and fixed. At 1 kW/m^2 my system should deliver 13.92 kW. The maximum I saw in October was 81.4% of that so I have orientation loss of AT LEAST 18.6% and I should expect, from that alone, production based on 0.814*4.71 = 3.83 kWh/m^2/day. Compared to the observed 2.53 that still says October was relatively cloudy,

On the sunniest October day here I got 3.91 kWh/m^2, on a cloudy day I got 1.04 and on an even cloudier one I got 0.28 so, as is plain, there are cloudy days and really cloudy days. But even on the cloudiest you will get some insolation even if it is, as here, only 7% of a sunny day's worth.

The following are the percentages of our electrical load covered by the solar system in the months we have had it. Note that we are not here during the summer but the house is air conditioned to tolerable level nonetheless. Heat in the cold season is electric (heat pumps).

Jun 101.19%
Jul 78.48
Aug 80.85
Sep 85.76
Oct 57.61

We get 95.18% of 1 kW/m^2 at the summer solstice which says that the panels are effectively aimed at the sun at that time. It's pretty clear that performance is going to be much worse near the winter solstice (cos(43° ) = 0.695 and many fewer hours of sunshine). I am estimating only 20% of our load will be covered and wouldn't be surprised if it is worse than that. The only solution is to add more panels but note that were I able to do that with orientations as favorable as I have now - the best spots are taken - doubling the number of panels would only get me, at best, 40%. Solar ain't so great in the winter.

 

[charlesj]

While my system is small 4kWh compared to others I am reading here, the best day produces about 26 kWh, not many of those, and the lowest I had on overcast day was about 2 kWh.

 

[NuShrike]

From what I've seen here in SoCal, you lose 80% of your rated production on a cloudy day, so size it that at least the 20% covers your most minimal baseline.

So summary:
-5% for dirt/dust
-10% for DC to AC
-33% for winter season (roughly)
-75% for clouds (roughly)

 

[doghousePVD]

I have a system in Maine. Last few cloudy days I got less than 10% of rated capacity. In heavy rain it’s always zero.

December will.be worse. YMMV. I am off grid, so I watch every photon in the wintertime.