How do Heat Radiators work?

[spinomonkey]

I get that they are used to bleed off heat and cool the vessel. But it seems like they don't have a large effect? Or maybe I'm using them incorrectly? What conditions are they best used for? Is there an optimal setup? Do they work as well in atmosphere vs space?

Thanks in advance.

Edited July 8, 2015 by spinomonkey

[Geschosskopf]

They seem totally bugged and I strongly recommend never, ever using them. They will make your rocket explode for no damn reason.

KSP radiators violate several principles of thermodynamics. The most obvious and annoying is that they embody Maxwell's Demon, which is kinda impossible in a system such as a KSP rocket. Saying this less esoterically, they expel heat into space at a much lower rate than they suck heat from throughout the rocket into the part to which they're attached. The heat from the entire rocket thus being concentrated in 1 part causes this part to explode even though nothing else in the whole rocket was even showing a temperature gauge.

You can avoid this by spamming lots of radiators ATTACHED TO DIFFERENT PARTS, so the heat of the whole rocket has more places to go and hopefully won't blow up any single part. And if you do enough of this, then the radators can serve a useful (but exploity) purpose in reducing reentry heating of the rocket by eliminating some of the heat as it's coming in. However, this shouldn't be possible because reentry heating apparently works by the process of "forced convection", which functionally is very analogous to direct conduction. That means that the surrounding air is hotter than the rocket (so heat moves from the hotter air to the colder rocket). Heat only moves downhill, after all. So radiators also rely on being hotter than their surroundings, so heat moves from them into the surroundings. If the air is hotter than the rocket, then radiators should not be able to expel any heat and in fact should increase the severity of the rocket absorbing heat.

Plus, the radiators look ugly and take up deck space you need for other things.

So all in all, radiators provide no benefit unless used in significant numbers to exploit bugs in both the radiators and the heat model. If you prefer not to play that way, then there is no benefit to using them and in fact their bugs will do more harm than good. Best thing is to avoid severe aerocapture/braking passes so nothing gets too hot to begin with. And know that even if you use very gentle, low-temperature aerobraking, if you have just a couple of radiators on your ship then parts of your ship will explode by the time you reach Ap after the pass.

Note that radiators really shouldn't be in the game at all anyway. Back in 1.0 - 1.0.2, LV-Ns had a stupidly unrealistic tendency to overheat, which basically made them useless. Nuke engines don't get that hot in reality, so this was just wrong. However, lots of people don't know this so starting whining that we needed radiators so they could use LV-Ns again, instead of lobbying to fix the real problem of unrealistic overheating. So in 1.0.4, Squad did both. They removed most of the LV-N's overheating problems AND added useless, buggy, radiators as well.

Edited July 8, 2015 by Geschosskopf

[sal_vager]

Funny, they work fine for me, protip, don't run KSP.exe via a shortcut, click the file directly so KSP can find the Physics.cfg in the folder where it is started, rather than on your desktop.

The static radiators will absorb heat from the part they are attached to and radiate it away, which in turn draws heat from the parts the first part was attached to, and so on.

The folding radiators act as if they have coolant pipes allowing them to draw heat from your hottest parts anywhere on the vessel, this is the "Maxwells demon" mentioned above, so it's not really violating principles of thermodynamics, but then it is just a game.

As for the bug, there is one, but it's caused by cargobays, try to keep things away from the doors as much as possible

[Geschosskopf]

Funny, they work fine for me...

Strange. For me, they (fail to) function as I described.

This little ship was coming home from Mun early in a career game. It had 2 small, fixed radiators attached to its 1.25m SAS unit.

The image on the left shows it just after emerging from Kerbin's atmosphere after a braking pass at 40km. As you can see, the entire ship is the "cold" color and no part is showing a temperature gauge. All is fine.

So I warped around the orbit to do another gentle braking pass. As soon as I quit warping at about 100km so I could line up retrograde and furl the solar panels, the SAS unit exploded, tearing the ship in half. This is shown in the center and right images. The temperature of the ship is still all cold and no temperature gauges exist.

I removed the radiators and did this again. No explosions.

Conclusion: radiators suck. In fact, they unrealistically suck heat into the part to which they are attached. This is the embodiment of Maxwell's Demon. There wasn't enough heat, when dispersed throughout the ship, to cause any problems or even register on the displays we have. But there was still enough, when concentrated in 1 spot, to blow up the SAS unit.

[Laie]

Repost:

"Radiator panels" (the fixed type) assume the temperature of their parent part without noticeable delay. In effect, they increase the surface area of the parent. Compared to fuel tanks of 16t and above, even the large panel doesn't provide much extra area, though. And anyway, radiation depends very much on temperature: in order to radiate well, a part has to be very hot. Which in this case means that the panel's parent has to be very hot -- a situation you probably want to avoid from the outset. Still, a panel or two may make the difference between a part going poof, or merely running dangerously hot. But as I said two sentences ago: if a part does not become hot, the panel attached to it will also have little to do.

"Thermal control systems" (the extendable swivel type) gather heat from all over the vessel. They can easily be the hottest part on the ship and thus also radiate disproportional amounts of heat. For purposes of heat managment, it does not matter where you place them. However, they will break if they collide with other parts of your vessel, so you better make sure they have enough elbow room.

So far, it seems as if the fixed radiators were nearly useless. They may have extra benefits inside an atmosphere, I haven't tested this.

@Geschosskopf: I suspect that the part that's killing you is once more the service bay. The whole heat system is known to spontaneously create heat out of nowhere at times, and parts in closed cargo bays have no way to shed heat. Attaching heat sinks to a bay may increase the likelyhood of such... incidents, but I don't think it's the radiator's fault.

Edited July 8, 2015 by Laie

[spinomonkey]

I appreciate all the feedback and you guys have settled some suspicions I've had a little while. Namely, that that fixed panels only affect the part that it is mounted to. Haven't tried the extendable type, I didn't realize there's a dramatic difference. Either I didn't read the description or it failed to mention it. That's sorta the catch 22 with the silly descriptions- sometimes you miss vital info that's included.

I haven't had any issues with panels causing random explosions, however I have found using air breaking during decent solve most of my problems thus far- and honestly they're more convenient in my mind anyways. I can mount them in a way that half or so of the airbreak hangs off the tail of the vessel, freeing up valuable space that otherwise would be used for panels. Plus... they look cool haha.

[Skorpychan]

I've had randomsplosion issues as well. Does this issue happen with the radiators built into solar panels as well?

[Laie]

I have found using air breaking during decent solve most of my problems thus far

Be careful! There may be some corner cases where they provide a small benefit, but by and large, panels don't work if your vessel is engulfed by hot plasma during re-entry. Besides, the extendable type will simply break.

Right now, there is not much use for heatsinks because since the latest LV-N workover there's not much heat to shed. However, if you insist on placing many LV-Ns in a dense cluster and/or want to visit the surface of Moho, panels will make that much easier. I also have an inkling that in the future, drills may again produce heat.

[Anglave]

When I built my first craft using LV-Ns in 1.0.4, I didn't know their heating had been fixed. I also saw the radiators and thought "Neat, new parts to solve the heat dissipation issue!"

I put a ring of radiators (the extending kind) around the main tank (Mk3 liquid fuel tank) of my interplanetary stage. They seem to work (in this application) as intended. The radiator gets hot and radiates a lot of heat. The rest of the craft gets less hot than it would have without the radiators.

Given that it's a large area, lightweight, flexible structure, I had never considered deploying it during aerobraking, assuming it would simply shear away. If you did deploy such a thing during atmospheric heating, I'm not sure what I'd even expect the behavior to be. It's a huge surface designed for efficient heat transmission. Would it absorb a lot of heat from the surrounding atmosphere and concentrate it into the part it's attached to? Possibly.

IMO, the radiators are intended to radiate waste heat more efficiently into deep space, and they work well enough for that purpose, though they may not be strictly necessary now that the LV-N's heat generation has been corrected. Using them during atmospheric braking is off-label and may possibly have unanticipated results.

[sal_vager]

I'd be interested in seeing the following from anyone suffering heat issues, as there's a known bug that resets them to incorrect values.

If your heat values don't match up then yeah, you're going to have some problems.

[spinomonkey]

I've found I can rip into the Kerbin atmosphere at around 3,000mps at a gentle-ish angle and be alright. The airbreaks will get pretty hot, but I haven't had any explode or rip off and it kills your airspeed so quickly that typically the worst of the heat doesn't linger for very long anyways. Now, full disclosure I do have panels nearby- but I'm not sure if they have as much effect as just as killing my airspeed rapidly.

EDIT: I'll post a picture when I get home.

[Edax]

When I built my first craft using LV-Ns in 1.0.4, I didn't know their heating had been fixed. I also saw the radiators and thought "Neat, new parts to solve the heat dissipation issue!"

I put a ring of radiators (the extending kind) around the main tank (Mk3 liquid fuel tank) of my interplanetary stage. They seem to work (in this application) as intended. The radiator gets hot and radiates a lot of heat. The rest of the craft gets less hot than it would have without the radiators.

Given that it's a large area, lightweight, flexible structure, I had never considered deploying it during aerobraking, assuming it would simply shear away. If you did deploy such a thing during atmospheric heating, I'm not sure what I'd even expect the behavior to be. It's a huge surface designed for efficient heat transmission. Would it absorb a lot of heat from the surrounding atmosphere and concentrate it into the part it's attached to? Possibly.

IMO, the radiators are intended to radiate waste heat more efficiently into deep space, and they work well enough for that purpose, though they may not be strictly necessary now that the LV-N's heat generation has been corrected. Using them during atmospheric braking is off-label and may possibly have unanticipated results.

I use the deployable radiators to cool down my reuseable rocket tanker. They can be deployed safely at 30,000 m, as they are necessary given my tanker has the "boars" for engines, which produce a LOT of heat just getting into orbit. Without the radiators deployed, the thrusters at the bottom of the rocket on the "boars" will explode from the heat, which would make aligning docking ports extremely difficult.