Board index FlightGear Development Spaceflight

Space Shuttle - Development

Discussion about development and usage of spacecraft

Re: Space Shuttle - Development

Postby Thorsten » Sun Jan 13, 2019 6:44 pm

Wow, this actually (kind of) works - in a single shot from 120 km distant, the Lambert solver got me as close as 2 km to the desired intercept point for a sane solution request.

Not sure what accuracy can be expected in reality, but with subsequent burns to refine the solution, this should be actually feasible...
Thorsten
 
Posts: 10979
Joined: Mon Nov 02, 2009 8:33 am

Re: Space Shuttle - Development

Postby Thorsten » Mon Jan 14, 2019 1:40 pm

Not all is working yet and there's plenty of quirks to be ironed out, but basically it's there.

The Lambert targeting code having computed a 45 min transfer solution:

Image


Burn data is automatically forwarded to MNVR:

Image

... and some 40 minutes later, we can see a fairly decent intercept coming up on the planning tool.

Image

If someone wants to try, after updating FGData (I've needed to modify the simulation of the orbital target to include the leading J3 terms) from --lat=0.0 --lon=5.07 --heading=44.66 one has a decent phasing orbit to do the intercept.

ISS as 3d model is automatically placed when approaching < 5 km, but not yet at the correct position and the correct rel. velocity. ISS as scenario by default close to the Shuttle is no longer available - now you have to work...)
Thorsten
 
Posts: 10979
Joined: Mon Nov 02, 2009 8:33 am

Re: Space Shuttle - Development

Postby legoboyvdlp » Mon Jan 14, 2019 3:01 pm

What's the technical barrier that doesn't allow us to launch to the ISS? Is simulating its orbit at the same time as FlightGear's the problem?
User avatar
legoboyvdlp
 
Posts: 7007
Joined: Sat Jul 26, 2014 1:28 am
Callsign: YV-LEGO
Version: 2018.3.1
OS: Windows 10 HP

Re: Space Shuttle - Development

Postby Thorsten » Mon Jan 14, 2019 3:09 pm

As of today you can also launch to ISS if you like (provided you know how to compute the launch window, the phasing orbits, the plane alignment burns etc. and you have the kind of time to wait it out) - that was kind of the point of the exercise, and I believe it would work okay (I've just tested the last hour of the approach of course...)

There's just no in-sim tools to plan such a mission, and you'll find that without such tools and a lot of knowledge, you'll not get anywhere.
Thorsten
 
Posts: 10979
Joined: Mon Nov 02, 2009 8:33 am

Re: Space Shuttle - Development

Postby GinGin » Mon Jan 14, 2019 3:16 pm

Spec 34 in game wow, brillant Thorsten.
I will have a thorough look to all that soon :)

Argh, indeed, time :)
Let's go for a quick phasing then for the first test.

Would it be possible to have distance in feet / Nm in the Nasal drawing of approach? to be coherent with Spec 34 and 33 Unit


Once again, Awsome work.


I was thinking to which tools would be really useful for a complete rendez vous

I think something like the Orbit Mfd and Align Plane Mfd like in Orbiter to have Orbit parameters from both Target and chaser and to know the node and relative inclination between the two orbit to null that with a Plane Change burn.
Combined with spec33 for rendez vous and spec 34 for final targeting burn, that would do it for a complete rendez vous from ground to ISS or whatever :)

Image


Image
GinGin
 
Posts: 741
Joined: Wed Jul 05, 2017 10:41 am
Location: Paris
Callsign: Gingin

Re: Space Shuttle - Development

Postby Thorsten » Mon Jan 14, 2019 3:36 pm

Would it be possible to have distance in feet / Nm in the Nasal drawing of approach?


There'll be a button to change the unit system eventually.

I think something like the Orbit Mfd and Align Plane Mfd like in Orbiter to have Orbit parameters from both Target and chaser and to know the node and relative inclination between the two orbit to null that with a Plane Change burn.


Actually, since these are MCC tasks, I guess you'll have to use LEO targeting for the time being to compute the required burns and just pretend someone uplinked them. There'll be an option to export the target state vector for that purpose soon.

I don't actually remember how the Orbiter MFDs handled the fact that inclination is not a constant in J3 or how the drift of the ascending node is taken into account....
Thorsten
 
Posts: 10979
Joined: Mon Nov 02, 2009 8:33 am

Re: Space Shuttle - Development

Postby GinGin » Mon Jan 14, 2019 4:22 pm

Actually, since these are MCC tasks, I guess you'll have to use LEO targeting for the time being to compute the required burns and just pretend someone uplinked them. There'll be an option to export the target state vector for that purpose soon.


Alright nice :)
So we will be able to know main parameters of target orbit ( apogee, perigee, LAN, Inclination )?


For Orbiter, it depends.
Some MFD don't handle very well nodal precession and non shperical gravity sources ( we can uncheck those parameters, the sim can run with either Spherical gravity only or more complex one,non spherical gravity sources)



But there are some MFD which take into account Harmonics and precession, like those one :

http://users.kymp.net/p501474a/Orbiter/Orbiter.html from Jarmo Nikkanen

He is on Orbiter forum and very active, you can maybe contact him if you want to have more info on how he implemented J3 and precession in his mfd :)
GinGin
 
Posts: 741
Joined: Wed Jul 05, 2017 10:41 am
Location: Paris
Callsign: Gingin

Re: Space Shuttle - Development

Postby Thorsten » Mon Jan 14, 2019 5:42 pm

So we will be able to know main parameters of target orbit ( apogee, perigee, LAN, Inclination )?


You set the target orbit the mission file (currently only perfectly circular orbits work...)

But the more interesting option is to export a state vector - then you can import that in LEO targeting along with the Shuttle state vector and run a numerical node finding and run plane-changing burns. Should be reasonably good in practice - the Lambert solver can easily do a (few) km offset in the final stage.
Thorsten
 
Posts: 10979
Joined: Mon Nov 02, 2009 8:33 am

Re: Space Shuttle - Development

Postby eatdirt » Mon Jan 14, 2019 8:35 pm

Woo!! That is Christmas AGAIN, unbelievable!!! Thorsten, you have no limit!!!

@Gingin, I am starting to read a bit the SCOM, and as I was already following your advanced tutorial, I was wondering if you would not fancy devising an "Advanced" check list for the Shuttle? I could try to help, although I have yet no clue how to code this and I am still a newbie as far as nominal procedure are involved :P
eatdirt
 
Posts: 360
Joined: Wed Aug 15, 2018 2:06 pm

Re: Space Shuttle - Development

Postby legoboyvdlp » Mon Jan 14, 2019 9:29 pm

GinGin, eatdirts question just reminded me, I hope you don't mind, a couple of days ago I made an edit to the advanced tutorial to correct one thing. Hope that is ok :)
User avatar
legoboyvdlp
 
Posts: 7007
Joined: Sat Jul 26, 2014 1:28 am
Callsign: YV-LEGO
Version: 2018.3.1
OS: Windows 10 HP

Re: Space Shuttle - Development

Postby eatdirt » Mon Jan 14, 2019 11:50 pm

@lego, cool, I should read them again then.

@Thorsten and others: guys, I was making that timelapse for atmospheric braking, and actually, being above the Shuttle, I am now understanding why I am getting many troubles doing the nasty Mac > 30 re-entries.

aerobraking_test.mkv

Look at the timelapse (the beginning is a bit long), but pay attention as soon as I am switching on the APUs:

1) The elevons of the Shuttle take a very weird initial position, maximal roll on one side. RCS thrusters are actually firing to compensate this at the beginning!?

2) Consequence -> as soon as air becomes thick, the Shuttle gets a roll kick, I am jumping on CSS roll manual, see my nice driving to compensate afterwards :oops:

3) I also noted that when the message "switching to aero" appears, the elevons go back to normal

PS: I suspect (to check), we have this as well in normal entry, but that does not matter because the entry is smooth and the roll kick is not enough to do anything bad. But at Mach 31, it does.

4) Finally, I am losing the ROLL control after the bounce, impossible to get it back in either CSS or AUTO when back to OPS 301. I suspect there is some no-reinit from OPS 304, but only on the ROLL control!?. Maybe that's because I have switched it to CSS to save my life? That would be cool Thorsten if you could have a look to reinit that one too, if possible (I have the feeling this issue only appears when there is a transition to AERO, because sometimes, I do not get any problem, especially when the atmospheric braking is so gentle that the AP never switches to that AERO mode).

Hopefully I am gonna make a nice timelapse at some point :)

Cheers,
Chris.
eatdirt
 
Posts: 360
Joined: Wed Aug 15, 2018 2:06 pm

Re: Space Shuttle - Development

Postby Thorsten » Tue Jan 15, 2019 7:16 am

From the description at least, this sounds like a thing that is quite real - the different systems (elevons, rudder, jets,...) all have their independent control loops, and since they're all engaged on the same tasks (managing roll and sideslip), there's situations where they might end up fighting each other, leading to potential loss of control (you can read about this in the workbook where the Aerojet DAP design is described).

Basically elevons are commanded not only to roll but also to fight beta excursions by inducing adverse yaw - if they do so, they act counter to the current roll command - that would prompt RCS jets to fire to support the current roll command. So a situation where jets fire against the elevons is in fact fairly common - usually then the elevons control beta via adverse yaw, the jets control roll. There's also the opposite situation where the elevons roll the Shuttle and the jets fire to manage beta and kill th adverse yaw induced by the roll.

The real Shuttle had the issue during aborts that the DAP was well used outside the normal envelope, leading to bad gains, this could potentially develop into a fight of rudder and elevons over beta. I can well imagine the same thing happening between jets and elevons - a Mach 31 entry is rather well outside the nominal profile and the gains would be similarly off.


4) Finally, I am losing the ROLL control after the bounce, impossible to get it back in either CSS or AUTO when back to OPS 301. I suspect there is some no-reinit from OPS 304, but only on the ROLL control!?


When roll is way off, you're likely suffering from integrator windup of some sort (?) What was the control mode (shown in the upper left of the FG-native HUD overlay that you get by pressing 'h') - did that properly switch to an orbital DAP?

Basically the whole control problem of the Shuttle during entry is a vastly complicated affair, and, sorry - I'm not delving into adjusting the gains for a Mach 31 entry, if you want to play with that, you'll have to dive in yourself.
Thorsten
 
Posts: 10979
Joined: Mon Nov 02, 2009 8:33 am

Re: Space Shuttle - Development

Postby eatdirt » Tue Jan 15, 2019 9:06 am

Thanks for the clarification, that's actually amazing!

I'm not delving into adjusting the gains for a Mach 31 entry, if you want to play with that, you'll have to dive in yourself.


Sure, I naively thought it could have been a bug. In any case, there is still the option to stay in CSS for roll all over the bounce, but I'll inquire what happens when transitioning back to orbit, it is not the first time I am loosing it. It goes back with a save an resume though.

Cheers.
eatdirt
 
Posts: 360
Joined: Wed Aug 15, 2018 2:06 pm

Re: Space Shuttle - Development

Postby Thorsten » Tue Jan 15, 2019 12:35 pm

As I said, it's likely an integrator windup, but I have no real idea how and when it arises.

Usually the PIDs should be protected, so for this to happen requires that for an extended time the DAP can almost control the vehicle, but not quite, so the integrator keeps adding error.

At the same time, we'd likely see a channel saturation.
Thorsten
 
Posts: 10979
Joined: Mon Nov 02, 2009 8:33 am

Re: Space Shuttle - Development

Postby GinGin » Tue Jan 15, 2019 7:23 pm

@Legoboyvdlp: No problem for the wiki, it is make to be corrected :) What did you change?

@Eatdirt: Concerning advanced checks in game, it was discussed couple of months ago. I think that in game Checklist are sufficient, showing the vital actions to be done to suceed a Shuttle mission without overwhelming the player with advanced stuff that are not vitals and that can be a game breaker.

You already have the advanced checks, based on real FDF, that I split to just have the nominal actions to do and not all the pages related to off nominal situations that may arise. ( Available in the wiki)


@Thorsten: Concerning the Spec 34, what is working and what is not? I see on your screen that item 17 through 20 where you specify where you want to be relative to the target wasn't fed. Did you make some entry and then after item 28 it disappears ?
I see that Face did the .exe for windows Leo, that is nice :)
GinGin
 
Posts: 741
Joined: Wed Jul 05, 2017 10:41 am
Location: Paris
Callsign: Gingin

PreviousNext

Return to Spaceflight

Who is online

Users browsing this forum: No registered users and 5 guests