1.1 Extreme Thermal Testing!

[Archgeek]

For the past month or so, I've been doing crazy thermal extremis testing in 1.1!  I built a little testbed of a z200, two little solar panels, an OKTO core, a nuclear engine, and an NCS adapter with only 8 units LF in it, then attached various cooling structures to it, saved those as new craft, and launched them to orbit on a modified aero-equis to test how well they can radiate the heat produced.

Methodology was to turn on infinite fuel, burn to an internal temp 1100 in the probe core, record burn time by comparing MET to MET at start of burn, then cut throttle to see if it can coast cool before the core hits 1200 and explodes.  Most burns were angled to escape Kerbin retrograde, so as to delay NaN-spam from hitting a solar escape.  A few actually reversed solar orbit and hit escape anyway, and were turned around to re-capture and restore performance.

Tested where a pair each of Structural Wing Type Ds, wing strakes, gigantor solar arrays, Large, Small, and Edge static radiator panels, a pair of Small thermal control systems, and single front-mounted Medium and Large thermal control systems.

On to the results!  Keep in mind these are tests of a tiny, mostly-engine ship with a tank perpetually almost out of fuel, and thus a terrible heat sink.  As such, the given times are to be regarded as impossibly extreme lower bounds.  Also, to explain the odd numbers to any flux-accountants, the active radiators would steal internal flux from the engine.  It was indeed producing 7471.27kW for all tests, though.

Has anyone done similar tests using drills/refineries as heat sources?

[Streetwind]

The thing about drills and refineries is that they don't produce heat like the LV-N does.

- Engines generate an internal heat flux, adding directly to to the part's internal temperature depending on the thermal mass of the internals.
- Drills and ISRUs use the "core heat" mechanic, which works differently.

From my experience, a part's core internally determines its temperature based on its own rules and maths. It interacts with active radiators based on its own rules. It does not interact with any other parts besides active radiators. The only part it can interact with is the part that carries the core itself. In other words, if the core is hotter than the part that carries it, it can create a heat flux that increases the part's internal temperature. How high this heat flux is is governed by the core's own rules, not the "normal" heat conduction mechanics between parts or between internals and skin. Generally, this "core leakage" is very small, and there exists no scenario where a core in a drill or ISRU will stabilize its temperature based on leakage alone.

If an active radiator is present (and within range, for the fixed panels), and if the core is also at or above its target temperature (if the core is not at its target yet, it will continue to heat up regardless of radiator presence until it hits the target), the core will lower its own heat output by the radiator's "core transfer" value. If enough radiators are present that their combined "core transfer" number exceeds the core's internal heat production, it is canceled out entirely, and the core is stabilized (stays at its target temperature indefinitely).

If you read the above carefully, you will notice that at no point, there is any statement about the ability of a radiator to provide cooling. That is because for core heat, there is practically no such thing. There is only the "core transfer" value, which is an extra stat on the radiator, independent of its cooling capabilities, which will interact with the core directly and bypasses all heat simulation. Now, if you test this ingame and look closely, you might notice that a radiator engaged in "cooling" a core actually gains a small internal flux value, as if it was indeed removing heat from something... but that flux is very small. Like 1/100th of the (already small) core transfer or something, I forget the precise value. The point is, it's so small as to be completely meaningless - it's probably there for debug purposes only, or perhaps to allow something like the radiator failing to provide core transfer in the presence of extremely high ambient heat (like being close to the Sun).

To make things more complicated, though: a radiator engaged in providing its full core transfer to one or more core(s) is regarded as being under full load, despite it handling no (appreciable) amounts of heat at all. But, should there be a part on the ship that requires actual cooling, the radiator will try to deal with that as well, splitting its capacity. This means that less than the full core transfer value is available to the drills and/or ISRU, meaning that they will start overheating.

[moogoob]

Thanks for the LV-N tests! I'll be sure to reference this next time I use them!

[msasterisk]

Use 2 engines facing each other to cancel the thrust.

[moogoob]

55 minutes ago, msasterisk said:

Use 2 engines facing each other to cancel the thrust.

That could work - you'd have to double the amount of radiating parts, though.

Though it might be a bit more useful to have them facing opposite each other, so their exhausts aren't heating the other engine.

[Archgeek]

58 minutes ago, msasterisk said:

Use 2 engines facing each other to cancel the thrust.

Hahah, where'd the fun in that be?  Far more amusing to have probes all over on insane eccentric orbits, and watch them cancel and reverse solar orbit as their tests progress.

1 minute ago, moogoob said:

That could work - you'd have to double the amount of radiating parts, though.

Though it might be a bit more useful to have them facing opposite each other, so their exhausts aren't heating the other engine.

HEH, I think that's what they meant, there.  Something like retro-engine <= mostly-empty NCS adapter <= cooling system under test <= OCTO or HEKS + Ox-STAT <= z200 <= mostly-empty NCS adapter <= more cooling system under test <= prograde engine.  I must admit such nuclear fire satelites do sound amusing in their right, upon imagining them.