Space Shuttle

[Thorsten]

If anyone is interested, the next version of LEO targeting is out. This adds the ability to simulate 3rd body gravitational perturbations to an orbit.

Since that's a tiny effect for Shuttle orbits and I'm frankly not sure whether it's in FG at all, it doesn't really add much for anyone who computes for the Shuttle - but it can do the Apollo trajectory...

[abassign]

Thank you, it really did a nice job In Linux I compiled the program directly in the src directory with the command: "make leo_targeting" and tested the program with the configuration file: shuttle_insertion.cfg but extending the time to 50000.

The diagram obtained is truly spectacular and clearly highlights the variations in the altitude for the single orbits.



Obviously I would like to see this function integrated into FGFS, as it is present in many other space simulation programs. It would be extremely useful to see the future effects of a maneuver or to perform a proper orbiting of a spacecraft. I think I use it to define the correct parameters for the programming of the on board sequencer in the Vostok 1 rocket.

[Thorsten]

Obviously I would like to see this function integrated into FGFS, as it is present in many other space simulation programs.

I actually don't think that's true - at least the Orbiter community is fairly interested in the software, because while many in-sim tools for Orbiter can do predictions and targeting based on analytical 2-body orbital mechanics, there's really not much tools to do numerical targeting including the various perturbations.

The only other numerical tool I've seen so far is NASA's mission design software (which however is quite a bit heavier and of course more capable still...).

Anyway - as far as I'm concerned, that's a mission control toolkit, not something flown on-board, and it's not written to be integrated in a per-frame environment, it assumes it can block the CPU till the task is done.

[Thorsten]

I wonder - has anyone tested the de-orbiting via a PEG-4 target yet?

I'm currently in the process of updating the targeting routines to support the insertion burn in PEG-4 as well (getting the launch site as ThetaT reference is a bitch...) and I have some concerns about breakage - so if anyone has a tested de-orbit scenario and would verify that it all still works, that would be nice. So far I've just benchmarked the routines against the output of LEO targeting.

In other matters, looking over my website statistics it seems the Shuttle Manual gets 40+ downloads every month - so there seem to be quite a few people interested. For a vehicle that complex, that's not a bad number of users (probably not all of them delve into it in depth, but still...)

[Thorsten]

Theoretically PEG-4 targeting should now work for OMS-1, OMS-2 and de-orbit burn (assuming the parameters aren't too far off a reasonable solution, the fit likely won't converge for things outside the Shuttle's capability).

I haven't tested all the details yet, and the apsis preview (which shows instantaneous analytical orbital elements) won't be more accurate than the 10 miles or so this is different to J3 gravity - but at least the basic tests went through.

[GinGin]

Wow, a lot of changes.
I am doing a clean install.

If I understand well, I need now the new version of FG ( 2018.1.1) and I will pull the development branch for the Shuttle again, then the stand alone LEO targeting to put some value into the game ?

Exciting time

[Thorsten]

You don't need to update FG - the changes are all Shuttle-side.

[GinGin]

Ok thanks.
Wow I installed new FG version and moved everything on a fast M2 SSD and a very large pagefile

Don't know from where it is coming , but I went from 30 fps to 70 average fps.

[Catalanoic]

"from 30 fps to 70 average fps.": Can anyone check this out?

[wlbragg]

@GinGin, when I moved to an SSD I was shocked at the performance increase in FlightGear. There must be a lot of data i/o going on in the background.

[Thorsten]

I suspect some other factor... I have FG installed on an SSD under Windows, and I don't see the Shuttle improve by such a margin. Neither do I see anything else scale in a way that isn't consistent with the improved GPU performance from my older to my current machine. Neither do I even see a dramatic effect on 3d model loading times (the actual loading time seems to be small compared to the optimization phase later).

If disk access has been the choke-point before, then an SSD would help - but FG does not usually have a dramatic disc I/O in any installation under any OS I have had.

then the stand alone LEO targeting to put some value into the game ?

Just for clarification - you can enter PEG-4 targets into MNVR right now, and when you load the plan, the system will crunch in the background and eventually give you the corresponding PEG-7 targets. You'll just have to guesstimate/know a reasonable one - if you enter an unreasonable target and are lucky, Delta vz will come out large, if you are unlucky the fit won't converge within the set iterations.

Or you can run LEO targeting outside FG to get your burn solution (if you do, you might as well use the PEG-7 targets this spits out, as internally PEG-4 is only used to create PEG-7, different from how the real thing works). For this, the state vector can be exported with a timestamp in the dialog, and you can feed the numbers right into LEO targeting - and you can punch the solution LEO targeting produces into MNVR. LEO targeting is somewhat more robust against bad targets as it allows for more iterations by default (but won't do grossly unreasonable fits either...)

Or you can simply ask MCC for a PEG-7 solution to de-orbit, it won't be as accurate as what LEO targeting produces, but it will work.

[GinGin]

I was on a "normal SSD " before also.

Only thing that I changed was to put the game on a Samsung 960 M2 SSD ( 5 times faster than a normal one, with speed around 2500 mo/s for written/reading data) and to increase the page file ( virtual memory) to 32 Go for an other game that is in Alpha with massive memory leaks.



Other than that same config than before, but I have now twice the FPS with high graphical settings .

Thanks for explanation Thorsten, I will try PEG 4 targets with official Nasa one, and also read again some stuff about that guidance ( I have to find again some good readings about that )


I tried a RTLS, nice new keyboard




I found the guidance much more accurate and into the profile







Then a manual Nz hold phase, I lost my skills and finish into the ocean

[Thorsten]

Thanks for explanation Thorsten, I will try PEG 4 targets with official Nasa one

I'm not 100% sure how close to real life MECO targets we are - obviously the OMS-1 and OMS-2 burn targets are not independent of that... Basically we cut thrust once the apoapsis is at a given value, but thetaT depends on the vertical speed you permit at that point. So you might have to fiddle with the MECO target as well to get to a real mission (?)

(Bad luck with the Nz holding... It is one of the trickiest bits of an abort...)

Oh - you might enjoy playing around with leakage scenarios... I suppose it might be fun to figure out the valve config that keeps you alive (see the fault dialog).

[GinGin]

Cool for the leakage , love those stuff

Forgot how cool was that weather tool

[GinGin]

I noticed lot of very small ajustements that are very nice


Like the right decount for Body axis velocity






Also MCC deorbit solution is very accurate, I found almost the same than what was advise by MCC.
Very easy now to deorbit where you want to go