Jump to content

[old thread] Trajectories : atmospheric predictions


Youen
 Share

Recommended Posts

This mod is well worth the wait.

Sit back, drink coffee (maybe decaf if you are tooooo hyped) and have some bacon.

and, finish that station around Jool in 0.90

Heh, I'll probably land everything on Duna in my 0.90 save, and only then continue to 1.0.

Just a heads up, Mechjebs Landing Autopilot has a similar function to it now

Can it predict aerocapturing trajectories?

Link to comment
Share on other sites

Maybe? Seriously though, I'm sure it's a matter of days for most mods. I'm sure most authors are *well* aware that 1.0 is now out and they should check on their creations. ;)

D'oh! I should check the forum more often. I wasn't aware version 1.0 was out :blush:, didn't play for some time... I'll take a look.

- - - Updated - - -

I just had a really impressive but strange occurrence with Trajectories. I set up for an aerocapture around Duna, but it didn't work, even though trajectories said it would. So I reloaded and tried again, nope, still didn't capture. I couldn't figure out what it thought was happening but I definitely wasn't capturing before leaving the atmosphere. Finally it dawned on me...

So, what was happening is that I just happened to be lined up so perfectly that with the proper aerobrake, I would have left Duna's SOI, hit my apoapsis around the sun, and come back down into Duna's SOI a second time, right into the planets atmosphere and then captured. It was calculating a second aerobrake only after leaving and re-entering the planets SOI. I don't know if this can be fixed, but I can't imagine anyone planning their trajectory beyond the planets current SOI. Maybe you could have the mod stop calculating there.

I think what actually happened is that in some cases the mod doesn't see the trajectory will leave the SOI, and so even if it's highly elliptical, it would come back (if the ship wasn't entering the sun SOI) and that second passage would be an aerocapture. At least I'm pretty sure I saw that once. Didn't have time to fix it yet.

- - - Updated - - -

Well, if this worked with NEAR/FAR this mod should still work similarly in stock after some tweaks to it.

The problem I've already got with stock is that it's not open source, so I don't know what they do and can't predict it. This is why stock wings are not predicted for example.

- - - Updated - - -

Oh oh, this might be a problem...

From KSP 1.0 release notes :

The flight model has had a complete overhaul, meaning the lift is now calculated correctly to all lift-generating parts, which includes lifting bodies. The drag simulation has also been completely revised, and uses automatically pre-calculated data based on the each part’s geometry, to be finally applied based on not just the orientation of parts in flight, but also taking other parts into consideration. Stack mounted parts are now occluded from drag by neighboring parts, and lift induced drag is also properly simulated. Both the lift and drag are dependent on air density and the speed of sound, which are calculated from temperature and pressure. Be careful when flying aircraft in this new update: stalls are now properly simulated as well, and spontaneous craft disassembly during high-G maneuvers is now a very real thing

Looks like they do exactly the kind of things FAR does... Excepted if I don't have access to the source code I won't be able to predict any of it. So if it's as complicated as it sounds, Trajectories won't be compatible with stock aerodynamics anymore.

Link to comment
Share on other sites

Looks like they do exactly the kind of things FAR does... Excepted if I don't have access to the source code I won't be able to predict any of it. So if it's as complicated as it sounds, Trajectories won't be compatible with stock aerodynamics anymore.

Well, Squad did make a point of making the new aerodynamic model easy to override so that people like Ferram can continue to provide FAR. So they appear to be keeping modability in mind. Hopefully Squad has some new functions one can use to make predicting aerodynamics possible? I seem to recall it being possible to pull up aerodynamic and thermal info via the debug menu.

You might even want to ask Squad directly if stock Trajectories would be possible with their recent changes, and if so, how.

Link to comment
Share on other sites

Looks like they do exactly the kind of things FAR does... Excepted if I don't have access to the source code I won't be able to predict any of it. So if it's as complicated as it sounds, Trajectories won't be compatible with stock aerodynamics anymore.

What about getting in touch with one of the devs to see if they can reveal any of how they calculate aero? Given that some of them are modders themselves, they might be sympathetic. If not, they'll face our collective outcry because this is one of my must-have mods.

Edited by Booots
Grammar
Link to comment
Share on other sites

Frankly I'd be happy if you could publish a vacuum-only 1.0 version for now. Just something that can compensate for rotation. Hitting a survey landing spot on a rotating Mun is tricky - I landed 4km away after several adjustments.

Link to comment
Share on other sites

As @djnattyd said, it looks like the MechJeb team figured out at least some of the new aero stuff. Looking at the mechjeb commit here https://github.com/MuMech/MechJeb2/commit/5926cba83850e13b4b150822104f0b7fc524e074 in SimulatedPart.cs and SimulatedVessel.cs there are examples to simulate lift/drag. I've been lurking the boards since 1.0 seeing who is figuring out the aero - MechJeb and the AeroGUI are the best examples I've seen.

Link to comment
Share on other sites

Not sure if it is helpfull, or if you found it allready, but in the debug window theres a tab for physics, and theres also a reference to physics.cfg for the database. No idea if there is any usefull info for you there, but wanted to point it out

Link to comment
Share on other sites

Well I have seen other use the new data in their mod Sarbian e.g. . Maybe you should contact him about the numbers (and maybe he knows a way that KSP will tell your mod the numbers needed instead of hard coding them)

Link to comment
Share on other sites

Well I have seen other use the new data in their mod Sarbian e.g. . Maybe you should contact him about the numbers (and maybe he knows a way that KSP will tell your mod the numbers needed instead of hard coding them)

The trajectories could become more accurate if you have atmospheric sensors on the ship!

Link to comment
Share on other sites

I hope the mod authors are working hard on this. I can't play KSP without my atmospheric trajectories, especially now that you have to consider the heating of re-entry and using a low profile re-entry trajectory.

Since 1.0.1 you don't really have to consider reentry trajectories as much anymore. But you're right – this plugin is essential.

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
 Share

×
×
  • Create New...