Whats with the insane heating spike at Mach 2?

[Crzyrndm]

The red line is convective flux (heat energy transfer rate) on a craft flying at 500m altitude. The sudden spike at Mach 2 is why planes all explode when you go much beyond 700m/s at sea level (the instant speed is > mach 2, flux/second increases by a factor of 70...)

Just wondering if there is any reason for there to be a sudden sharp change in heating characteristics passing through Mach 2

Edited April 28, 2015 by Crzyrndm

[Blu_C]

My guess would be it is simulating shock heating effects. As you go faster the air in front of you compresses more, and using the friendly formula of PV=nRT we can see that as Pressure goes up so too must temperature in order to keep the equation balanced (both n and R are constants). At some point the heat gained from friction is significantly less than the heat gained from shock heating.

How accurate that spike is I cannot comment on, however it is true that aircraft which are designed to travel at especially high speeds do have problems with heating. In fact, as I recall, the SR-71 was designed with many gaps in it's structure since, once in flight, the plates would expand as a result of heating and create a seal (as such, the thing was leaking fuel on the ground and had to be refueled shortly after take-off).

[NathanKell]

Good find.

[antbin]

What tools do you use to capture / plot these values?

[Yasmy]

It would be interesting to see these heat profiles on a per-part basis as a function of distance from the front. For example, have the front of a rocket be a stack of several empty liquid fuel tanks with nothing attached to them, and plot the heat fluxes vs. Mach number for each tank. (Might want to pick something with higher max temperature than a fuel tank, but something light and of constant cross section.)

Edited April 28, 2015 by Yasmy

[Crzyrndm]

The raw data of the kink:

[TABLE=class: grid, width: 150]

[TR]

[TD]Mach[/TD]

[TD]Flux[/TD]

[TD]Flux'[/TD]

[/TR]

[TR]

[TD]1.995[/TD]

[TD]2037.6[/TD]

[TD]2.17[/TD]

[/TR]

[TR]

[TD]1.996[/TD]

[TD]2039.8[/TD]

[TD]2.19[/TD]

[/TR]

[TR]

[TD]1.997[/TD]

[TD]2042.0[/TD]

[TD]2.18[/TD]

[/TR]

[TR]

[TD]1.998[/TD]

[TD]2046.3[/TD]

[TD]2.21[/TD]

[/TR]

[TR]

[TD]1.999[/TD]

[TD]2046.3[/TD]

[TD]2.14[/TD]

[/TR]

[TR]

[TD]2.000[/TD]

[TD]2048.6[/TD]

[TD]2.22[/TD]

[/TR]

[TR]

[TD]2.001[/TD]

[TD]2073.8[/TD]

[TD]25.3[/TD]

[/TR]

[TR]

[TD]2.002[/TD]

[TD]2215.8[/TD]

[TD]142.0[/TD]

[/TR]

[TR]

[TD]2.003[/TD]

[TD]2367.9[/TD]

[TD]152.0[/TD]

[/TR]

[TR]

[TD]2.004[/TD]

[TD]2506.9[/TD]

[TD]139.0[/TD]

[/TR]

[/TABLE]

Flux' (convective) stabilises at 150 (+/- 5) per frame

My guess would be it is simulating shock heating effects. As you go faster the air in front of you compresses more, and using the friendly formula of PV=nRT we can see that as Pressure goes up so too must temperature in order to keep the equation balanced (both n and R are constants). At some point the heat gained from friction is significantly less than the heat gained from shock heating.

Oh, I have no issue with the increase in heating. It is specifically the jump as you pass Mach 2 that I was wondering:

A) Is there any basis in reality

Any way to even begin to deal with it

C) Any settings that could smooth that out a bit

Good find.

Implying it's a bug I assume (although I guess there's not much chance of it not being one given how unmanageable gaining 100s of thousands of flux in seconds is).

What tools do you use to capture / plot these values?

Custom plugin => excel

It would be interesting to see these heat profiles on a per-part basis as a function of distance from the front. For example, have the front of a rocket be a stack of several empty liquid fuel tanks with nothing attached to them, and plot the heat fluxes vs. Mach number for each tank. (Might want to pick something with higher max temperature than a fuel tank, but something light and of constant cross section.)

In a stack, parts towards the rear have a delayed response almost exactly as you would expect. The parts that heat the fastest are small and stick out into the airflow (canards are really bad for this)

[NathanKell]

Try messing with machConvectionEnd in Physics.cfg.

[Yasmy]

Maybe they are using Mach 2 as something like a transonic/supersonic boundary, rather than a more conventional 1.2ish.

I wonder if there is a similar change in the heating model at a subsonic/transonic boundary and a super/hypersonic boundary (perhaps at Mach 3.5?).

From Physics.cfg: (thanks NathanKell)

machConvectionStart = 2

machConvectionEnd = 3.5

Edited April 28, 2015 by Yasmy
speeling

[Crzyrndm]

Can't really change those too much otherwise you end up killing re-entry heating off. The issue here is that the interpolation method used is completely linear, causing an extremely sharp kink as it cuts in

[Cinocal]

Not exactly and elegant solution.

[arkie87]

http://i.imgur.com/sYh563i.png

The red line is convective flux (heat energy transfer rate) on a craft flying at 500m altitude. The sudden spike at Mach 2 is why planes all explode when you go much beyond 700m/s at sea level (the instant speed is > mach 2, flux/second increases by a factor of 70...)

Just wondering if there is any reason for there to be a sudden sharp change in heating characteristics passing through Mach 2

Good work. Any chance i could get your plugin to output to excel?

I've noticed similar discontinuities/extreme sensitivities in my thead here

- - - Updated - - -

Oh, I have no issue with the increase in heating. It is specifically the jump as you pass Mach 2 that I was wondering:

A) Is there any basis in reality

Any way to even begin to deal with it

C) Any settings that could smooth that out a bit

According to this formula, it should be continuous starting from Mach 0: http://en.wikipedia.org/wiki/Stagnation_temperature but this only includes stagnation temperature and not viscous heating and/or other effects (though i'm not sure which effects KSP even attempts to simulate).

[Van Disaster]

Is there any modelling of thermal conduction anywhere? but really I think it's just a case of attempting to handwave & getting it wrong, rather like the odd drag behaviour of the new aero.