PSA: Solar panels generate less power in atmospheres

[Volix]

I've seen some threads talking about how solar panels now obey the inverse power law. I haven't seen anything about how they now produce less power in an atmosphere.

Here's a gigantor XL in Kerbin orbit:

It's aligned as best as possible for the 1.00 sun exposure, and produces 25.94 (whatever units). Here is a gigantor on the runway (forgot to switch off NavHud, ignore the green line down the middle):

It should be perfectly aligned since I placed the panels East-West. Sun exposure is still 1.00, but produces 21.62 whatevers. But wait, there's more! Here's the same craft at sunset:

Still a 1.00 sun exposure, but now only produces 6.14 whatevers. In my 0.90 save the panel generates 18.0 whatevers in all three cases.

This is how it works in real life too, the air absorbs some of the light energy leaving less energy flux at the surface. At sunrise/sunset the light must travel at a shallower angle through the atmosphere, and therefore takes a longer path. Passing through more air means more is absorbed before it reaches the surface, giving even less flux. Now I know why my Duna lander couldn't support the Remote Tech uplink on the surface when it was fine in space .

These screenshots were taken in my modded install, but I've confirmed the effect on a pure stock install as well.

[Zyffyr]

Nice info/testing.

[DarkGravity]

Good to know!

[Streetwind]

Probably temperature-related. Solar panel power output scales with the panel's temperature, and the atmosphere at Kerbin sea level at the equator is much warmer than space.

Hit Alt+F12 and activate the display of thermal properties in action menus, and then compare?

[SpaceChipmunk]

Could be temperature. Maybe atmospheric scattering??? Bit of a stab in the dark as to whether that is modelled or not?

[DarkGravity]

If it is temperature, see how they perform at night.

It's cooler at night.

[SpaceChipmunk]

If it is temperature, see how they perform at night.

It's cooler at night.

It is also a lot darker from a solar panel point of view.

[Volix]

Probably temperature-related. Solar panel power output scales with the panel's temperature, and the atmosphere at Kerbin sea level at the equator is much warmer than space.

Could be temperature. Maybe atmospheric scattering??? Bit of a stab in the dark as to whether that is modelled or not?

Interesting, hadn't considered temperature. Just did a quick test with the temperature showing as well. Fortunately the numbers update in time warp. I used the same craft on the runway.

First of all, the temperature variation throughout the day is around 10 degrees, not all that much. In the morning the panels are cool and warming up. As the panels warm the energy flow goes up. About an hour before the sun is straight up the panels start to cool but the energy flow continues to rise. After the sun is directly overhead the energy flow and temperature are both falling until night.

I noticed the temperature jumped a bit when beginning time warp, around 2-3 degrees, so I let it settle to equilibrium at 1x a few times during the day. The same general trend was observed, the energy flow would increase from morning to noon, but the temperature would first rise, then fall. I believe it's cooling near noon since the central tank is absorbing less heat as the sun hits it directly above, and it's also shaded on top by the probe core. This gives the heat from the panels somewhere to go. I also paid close attention to the minimum and maximum this time. Just before the sun sets/after it rises the energy flow dips as low as 1.41, while it's 21.66 at max (sun directly overhead).

So, it seems the energy flow has more going on than temperature and distance. Plus, I have a hard time believing a change of around 10-15 degrees would change the energy flow by so much. Atmospheric scattering is about all I can think of.

[Korizan]

Clouds - I know they are not shown but solar panels do have to deal with them and atmosphere scattering (dust / dirt on the panels and in the atmosphere and the sky is blue after all).

Temperature is should not be an issue on the planet.

Edited May 22, 2015 by Korizan

[KerikBalm]

If it were temperature, you'd expect them to produce less at mid day... but they produce less when the sun is low on the horizon (although I have no idea how well KSPs temperature system models this).

Do the same experiment, but with Eve!

Maybe my electric ducted fan plane won't fly so well there :/

More gravity means more power is needed... if the power/gravity ratio is less than for Duna (which the design can barely fly at)... the plane is doomed, and I'll have to buff my mod-part

[Volix]

Clouds - I know they are not shown but solar panels do have to deal with them and atmosphere scattering (dust / dirt on the panels and in the atmosphere and the sky is blue after all).

Temperature is should not be an issue on the planet.

Thanks to a mod, I do have clouds . This works in pure stock as well however.

I'm sure you were talking about real life though, in which case you're absolutely correct. Clouds will of course reduce power, and even a panel that tracks the sun will see a rise and fall of power through the day, depending on the amount of atmosphere the light passes through. Temperature is negligable next to both these effects.

Some numbers from real life, for those interested. In space there is about 1,500 W/m^2 (Watts per meter squared) of sunlight. Where I live, around 45 degrees north latitude, the peak power at the sun's zenith (in the summer) is just under 1,000 W/m^2. This number would be higher near the equator.

[Tex]

Ah, good to know! I never really gave it a thought until now... I love this system!

[Volix]

Do the same experiment, but with Eve!

Maybe my electric ducted fan plane won't fly so well there :/

I don't have time to try this until after work, but I'm curious as well. Eve is also much closer to the sun, so now that the inverse square law is properly modeled you might be okay, it's hard to say how much different the power output would be. I'd also like to see if this changes with altitude.

[MarkusF]

I have an unmanned lander on Eve, so I gave it a try.

When the sun comes in vertically (mid-day) the small 1x6 panels bring a power of 0.96 at 0.98 exposure. So it's safe to safe, they bring 1 when full exposed to the sun.

In the evening power output went down to 0.06 at 0.87 exposure before the sun had been blocked by Eve.

I don't have small panels in Eve's orbit right now, but the gigantor panels have an output of 43 when fully exposed to the sun in orbit.

[bitbucket]

For comparison, now we need to run a similar test on a body with no atmosphere (Mun or Minmus should do) and see if the same effect happens.

[Volix]

I have a lander on Ike already, so I checked it out. Unfortunately I botched the landing so it's on its side and I can't get good alignment. I found at local "noon" I got 0.74 energy flow with 1.00 exposure on a 1x6 panel (non-retractable). I next checked when the sun exposure was 0.50, which turned out to be near sunset, and got 0.37 energy flow. If we assume energy flow varies linearly with exposure (seems reasonable) this would indicate no change in power due to the sun being low, as I get half the energy flow at half the exposure.

[Starwhip]

If it is temperature, see how they perform at night.

It's cooler at night.

It is also a lot darker from a solar panel point of view.

:facepalm:

This is some interesting info, thanks for sharing!

[Spatzimaus]

I have an unmanned lander on Eve, so I gave it a try.

So do I. My little lander (nothing but a probe core, a small parachute, a 0.625m inline battery, four of the lightest legs, a thermometer, and an antenna) has four OX-STATs placed around the sides. Obviously these can't track the Sun, so they're not useful for this sort of test, but I was amazed to see how little energy they produced even at the optimum times of day. It took hours to refill the batteries from even a single thermometer reading.

-------------

As to the astrophysics of this, it's pretty realistic if you look at total solar energy. On Earth, the Sun produces 1367 W/m^2 of energy. When directly overhead, about 83% gets through the atmosphere. Three hours before or after (45 degrees off-angle) and you're down to about 80% of THAT amount, i.e. 65ish%. Five hours before or after the peak (75 degrees off, or 15 degrees above the horizon) and you're down to about 25% of the total amount (30% of the in-atmosphere peak). The Kerbin numbers in the OP match this almost exactly, and it does explain Eve's horrible surface numbers (very thick atmosphere), which makes yet another reason to ditch solar panels and use RTGs as soon as you can.

The catch is, solar panels don't absorb every wavelength of light equally well. Like a lot of other things (mirrors, digital cameras, the human eye, etc.), you'd want something that was optimized for the types of radiation that do get through our atmosphere with less absorption, i.e. visible light, radio waves, and to a lesser extent infrared. So, the amounts might actually go up quite a bit depending on the chemistry involved, especially as you approach sunset. Although, this'd be different for other planets with different atmospheres, so unless we want to assume the panels are "tuned" for each body's particular absorption spectrum we should probably just stick with the numbers we have now.

Still, it's yet another example of how the devs have done their physics homework.

[bitbucket]

Took some Gigantor XL panels to the Mun.

Early morning landing.

Midday high sun.

Late evening.

The last glimmers of twilight.

And it's gone.

Output stayed around 26 right until sunset. I think we can confirm atmospheric extinction is a thing that is.

[Volix]

That looks good bitbucket, and the variations are small enough that they could easily be attributed to temperature.

[vicwarrior]

Just to make sure, can someone test if solar panels produce more power if you heat them up, with a lets say, a few nukes?

It is probably just atmospheric extinction, which is anyways a very nice detail from Squad.

[Volix]

Still, it's yet another example of how the devs have done their physics homework.

This. So much this. When I couldn't transmit my pile of science from Duna, well, "ragequit" is the word that comes to mind. After I calmed down and really investigated I was pleased to find this was the reason rather than some bug. Just need to plan accordingly next time.

[Spatzimaus]

That looks good bitbucket, and the variations are small enough that they could easily be attributed to temperature.

More likely, it's that the aiming isn't perfect. To keep the internal simulation running at a reasonable speed, they probably do things like have the solar panels stick to integer numbers of degrees of rotation, and there's probably some rounding on the solar angles as well. So, even though it says 1.00 Exposure, it might be varying by a percent or so simply based on the fact that internally it's actually varying from 0.995 to 1.005 (yes, I know it's physically impossible to go above 1.00, but internal rounding CAN do that). That'd explain the 1% variation seen in those shots.

[Volix]

Just to make sure, can someone test if solar panels produce more power if you heat them up, with a lets say, a few nukes?

What kind of scientist would I be if I left a variable unchecked? So another quick test. Of course it takes some time for things to heat up, so the sun has moved from screenshot to screenshot, but not all that much.

The panels clearly lose efficiency as they heat, much more than I expected in fact. However, in my previous test held in thermal equilibrium the temperature never rose above ~320 and I saw the energy flow dip as low as 1.5 right at sunset. The panels would probably blow up before the efficiency dropped to those levels from heat alone.

EDIT: A little off topic, but as a new member to the forum I've greatly enjoyed this discussion. I haven't thought this much about KSP since my time played was in the double digits. Thanks everyone!

Edited May 22, 2015 by Volix

[sj176]

Interesting. I'd imagine that's how it should be.