Orbiter-Forum  

Go Back   Orbiter-Forum > Far Side of the Moon > Brighton Lounge
Register Blogs Orbinauts List Social Groups FAQ Projects Mark Forums Read

Brighton Lounge General off-topic discussions. Political or religious topics may only be posted in The Basement forum.

Reply
 
Thread Tools
Old 07-10-2016, 12:07 AM   #76
GLS
Addon Developer
 
GLS's Avatar
Default

Quote:
Originally Posted by Thorsten View Post
 Big moment... after some sweat and frustration, the AP just brought the first Shuttle home after an engine failure 30 seconds into the flight via an RTLS abort - from liftoff right to the turn into final under full automatic control.

The flight history shows the elongated flightpath reaching some 355 miles out before turning back, then an S-turn during TAEM to dissipate energy, followed by the HAC.

{image}

The most vicious phase is the drop from ET-sep into the lower atmosphere with Nz-holding active, making sure acceleration stays within structural limits while qbar ramps up to dangerous values (here it peaked just around at 650 psf, which is about equal to launch stresses).

It's quite a ride to sit through this...

The AP does a hell of a job during the powered RTLS phase because it has to deal with a lot of potential trajectories (dependent on just when the engine failed the underspeed is different, same for payload, there's the throttle mapping to 105 and 109% of nominal value on SPEC 51 to consider, the abort isn't always called at precisely the same time,...) and nudges and focuses them all over the next minutes into a reasonable MECO solution with the right amount of horizontal and vertical speed to survive the plunge, a manageable range to landing site, the correct qbar limits for ET-sep and a reasonably empty tank.

Mathematically it's a very interesting problem, far beyond the run of the mill PID controller.

And admittedly from a thrill factor, RTLS beats everything else the Shuttle does, the moments from powered pitchdown followed by ET sep and 'the plunge' are pure adrenaline.
If you keep doing stuff like this, SSU might just have to hire you.
GLS is online now   Reply With Quote
Old 07-10-2016, 06:32 PM   #77
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

Quote:
Hey Thorsten,are you able to see the reentry plasma from the cockpit windows?
Not yet... it's halfway done, there's a mesh to project the plasma onto, and my plan is to write a dedicated GLSL shader to visualize it, but I haven't gotten around to actually do it.

I think you might get to see raindrop splashes on the cockpit window though if the weather is poor... (not 100% sure whether we assigned the effect, but it exists).

Quote:
If you keep doing stuff like this, SSU might just have to hire you.
:-)

You'd have a hard time convincing me to leave Linux though... If Orbiter were OpenSource and multi-platform...

Btw. I don't want to give the impression I'm working on this all on my own - there is a whole team working on various aspects, and there's other people contributing to the infrastructure.

So let me take this opportunity to credit Wayne Bragg who is patiently animating switch by switch and wiring everything up with the systems model - here's a technical rendering of what switches and talkbacks are active as of today (in yellow):



Thanks to his work, we not only have a complete thermal and electrical control panel, but also nice touches like actually working brightness controls for all MDUs and the HUDs - as well as the panel backlighting at night.

There's also now (by our forum user chris_blues) hires versions of planet textures with the encoded city lightmaps which look really nice with the cockpit night visuals and a touch of aurora at the horizon:





So while I'm kind of a project coordinator, I'm by no means doing everything by myself (my 3d modeling skills suck...) :-)
Thorsten is offline   Reply With Quote
Thanked by:
Old 07-18-2016, 06:05 PM   #78
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

Teaching the Shuttle G&C to fly contingency aborts. From an aerodynamical perspective, they're very interesting and challenging stuff, and there's a whole new can of worms opened.

The avionics can now note the failure inertial speed of engines and determine the contingency procedure to be executed (here's a two engine failure in the BLUE region as an example).

The AP knows how to do the procedure for BLUE and GREEN in powered flight (the others are to come) and can compute dynamically optimum Nz after handing over to unpowered flight).



My current vision is to be able to do an East Coast / Bermuda abort eventually. And perhaps to support bail-out by using the FG parachutist model piggy-backed on the Shuttle sim.

Single engine flight dynamics is bitchy in more than one way. First, the fact that the trajectory predictors inevitably drop like a rock is slightly depressing. Second, it's really crucial to get the roll channel off the engines - I spent two hours figuring out why I had no pitch authority before I figured out that the DAP was still trying to utilize a thrust vectoring roll.

Even single engine roll control (SERC) isn't without pitfalls, because it needs to be active right when the OMS fuel dump happens - the solution is to do mirror-imaging, rather than, say, firing the left pod up thrusters, one inhibits left pod down thrusters for a moment (and vice versa).

It's very subtle (the upward flame is missing on the left) but this is rolling the Shuttle under SERC with a fuel dump going on:



I absolutely love implementing all these little details about the Shuttle. It's ingenious how it all works together.
Thorsten is offline   Reply With Quote
Old 07-22-2016, 06:11 AM   #79
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

The handover from powered contingency guidance to OPS 602 entry software has been a headache (unprotected integrator, thoroughly would up during SERC...) but now the flight software can fly the required Nz holding and alpha transition after a contingency abort.

Here's a test flight with a two engine failure early in the GREEN zone (that's a nasty region, a 3-engine failure is not survivable there).



About halfway into the trajectory it's seen to bend - that's yaw steering vectoring the Shuttle closer to the coast by doing a 45 degree yaw relative to the current course. Doesn't make much of a dent, but then according to the manual that's how it is.

Pullout for that one has been severe, a bit over 3.8 g (that's actually in the blackout simulation region, so I couldn't read the meter at all times) with qbar maxing out close to 680 psf (and here I was thinking the RTLS is rough on the Shuttle...)

Real maneuvering capability (the turn at the end of the trajectory) only comes at the end in the atmosphere. In the event, there was no cross-range to reach any landing site (wasn't expected for a really early GREEN, in addition I had a stuck speedbrake in open position...) so this is a bailout scenario where the Shuttle is flown a bit closer to the coast before we all jump...

Another billion of virtual dollars gone...
Thorsten is offline   Reply With Quote
Thanked by:
Old 07-23-2016, 06:18 PM   #80
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

Actually made it to Bermuda - the most expensive holiday trip ever I guess:

Regular low inclination launch, TAL abort after the first engine failure some 3 minutes into the flight, down-moded to contingency abort after the second engine failure.

Going through the contingency abort procedure (that's an original cue card available in-sim for reference):



Kicking off the ET - in the GREEN region, it's actually disconnected via high rate sep.



What follows is a 3.4 g pullout, some following phugoid oscillations as the Shuttle flies the alpha-transition pattern starting at Mach 6 through the upper atmosphere, taking over the roll axis to CSS to aim at Bermuda (to S-turn or to acquire site during alpha transition is a bit of a tricky decision and not implemented yet), then letting the AP fly again.

Bermuda, here we come (I've added plenty of random shipping traffic around the islands for enjoyment, so there's yachts watching me landing):



And, a short-field touchdown - yikes, not much space here:



Here's the trajectory:



First wiggle is the TAL trying to steer northward to Spain, second (very small one) is the contingency yaw steering trying to steer back south towards Bermuda, and the third major one is my CSS takeover during supersonic flight. The HAC is too small to be seen on the scale, but I basically made it right into the TAEM pattern without any S-turns and was delivered into a perfect final from 12.000 ft at the end.

Very satisfying to meet such a challenge.
Thorsten is offline   Reply With Quote
Old 07-26-2016, 12:20 PM   #81
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

Some nice weather during bailout-tests. The Shuttle has now the AP channel to hold her steady and slow during crew bailout if no runway can be reached - which is why we're just heading down into the sea here.





Thorsten is offline   Reply With Quote
Thanked by:
Old 07-31-2016, 06:18 PM   #82
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

Looking a bit into rendering night launches, I realized I had the solution to a cool effect right at my fingertips - here's the SRB flame lighting up the cloud layer in a radius of a few miles:

Thorsten is offline   Reply With Quote
Old 08-10-2016, 09:11 AM   #83
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

I suspect the following is really for fans of the Space Shuttle avionics and all its intricacies. I've now as good as completed the MEDS layer (as much as this is possible - you can't patch files or command memory dumps, because it doesn't run internally on the historical MEDS code, so these operations make no sense in the context of the sim, but you can do pretty much everything else).

The MEDS menu shows now primary, secondary and selected port for every MDU, as well as flight-critical bus and port reconfiguration mode. In addition, the self-diagnostics checks for MEDS failures and fills the MEDS fault line with them if it detects any.



That means there can now be two faults flashing on one screen, one to be acknowledged by keyboard, the other via edgekey. MEDS faults are always a property of one IDP, so there's four lists of them being maintained, and the message gets bumped to a DPS screen under the command of the current IDP.

The maintenance display shows the current configuration for the active IDP



all items under the command of the IDP shown in blue, all items the IDP can listen to (because it's connected via the non-active port) show data, all else is blanked. Failed self-tests or equipment that doesn't communicate is shown in red. Using the config option allows to change ports on the current MDU or to change reconfiguration mode.

Finally, it's possible to access the fault summary page where all the faults the current IDP has detected are displayed.



I haven't come across any place that has a comprehensive list of faults/warnings which are detected and announced, so right now it knows just two categories (I/O ERROR and PORT CHANGE) for a specific device. If anyone knows where to find that info, please let me know.

It's clearly not something the casual user will ever need... but it's nice to have in the simulation for those who are interested i dealing with more exotic failures (broken ADC, problem on a flight-critical data bus,...)

Last edited by Thorsten; 08-10-2016 at 09:13 AM.
Thorsten is offline   Reply With Quote
Thanked by:
Old 08-14-2016, 05:12 PM   #84
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

More avionics fine-print - the system now responds correctly to failures of data transmission.

If the IDP can't get in touch with a GPC running the required software, it crosses out the DPS part of the screen and throws a poll error. It's possible to intentionally isolate an IDP by assigning GPC 0 via the keypads.



If the commanding IDP of an MDU goes non-operational, the MDU proudly proclaims its independence:



If no automatic reconfiguration is set, you can then change ports manually and get back to a working IDP, if this is set to AUTO, it happens on its own but leaves a MEDS message about the port change in the system.

(If an MDU fails or the main power bus it connectes with goes below operating voltage, it simply blanks)
Thorsten is offline   Reply With Quote
Thanked by:
Old 08-16-2016, 10:05 AM   #85
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

Getting the interior lights installed...





Wayne has been busy putting the 3d model into a raytracer to render the direct and indirect illumination from every cockpit light into lightmaps which we're now bringing into the cockpit rendering scheme. Using a raytracer for the job just gives an accuracy that no real-time technique can hope to match.
Thorsten is offline   Reply With Quote
Old 08-18-2016, 05:50 PM   #86
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

Quote:
Hey Thorsten,are you able to see the reentry plasma from the cockpit windows?
Well, now you do... Some first tests rendering the plasma sheath from within - this is a night scene for better contrast.

Onset of plasma builtup:



Max heating phase:



There's movement and structure to this which looks quite compelling (it runs via a dedicated GLSL shader), the light has a pulsating quality to it (and the alert observer might have spotted that the cockpit ambient light is influenced by the plasma glow).
Thorsten is offline   Reply With Quote
Thanked by:
Old 08-18-2016, 07:51 PM   #87
Donamy
Beta Tester


Default

You should be able to see the interior with a orange glow. IMO
Donamy is offline   Reply With Quote
Old 08-18-2016, 11:38 PM   #88
Interceptor
Orbinaut
 
Interceptor's Avatar
Default

Here you go Thorsten,you have probably seen this video already,but if not check it out it should give you an idea about the colors of the reentry plasma from the cockpit windows of the shuttle.
Interceptor is offline   Reply With Quote
Old 08-19-2016, 05:38 AM   #89
Thorsten
Orbinaut
 
Thorsten's Avatar
Default

Quote:
You should be able to see the interior with a orange glow.
I suspect you may be overestimating the intensity of the light emission - the plasma isn't all that bright (the illumination is there when compared with a pure night shot, but not all that prominent).

The section 'sensory cues' of the crew manual has:

Quote:
During the night portion of an entry, the glowing plasma sheath around the vehicle can
be observed through the orbiter windows. The pulsing light has not interfered with tasks
inside the vehicle. The plasma glow is less visible during daylight portions of entry and
may not be noticeable.
(emphasis by myself).

You can go through the physical intensity numbers of daylight sky at high altitude and ambient lit surfaces in daylight (Wikipedia has a few numbers, I've done my own tables with a series of controlled photographs to get my rendering framework onto solid grounds), you can ask what emission intensity is barely/not visible under these conditions, you translate that from physical intensity into perception space for rendering, and you get brightness value approximations from that statement.

Also, I've worked some time in plasma physics labs (MPI Garching and JET Culham) when I was a bit younger, and I've seen plasma discharges live, so I think I have a reasonable idea of how they look like.

Quote:
Here you go Thorsten,you have probably seen this video already,but if not check it out it should give you an idea about the colors of the reentry plasma from the cockpit windows of the shuttle.
I'm afraid as far as color references are concerned, the video is unusable. You can easily see that it's overexposed to capture the rather dark cockpit visuals (you can already see the glow sticks de-saturate).

I frequently seem to get into these discussion, but in low light the perception of eye and camera is completely different and even high-quality still photographs often do not convey the right impression. One has to take care to select reference material where the exposure time is known/can be inferred, one has to translate the visuals into what the eye would see under the current lighting conditions,...

The current colors/intensities are based on a) my own experience seeing the glow of various plasmas b) a handfull of still photographs from the Shuttle cockpit where the exposure time can be inferred c) some images from the NASA mission simulator

I'm not claiming I'm doing it correctly 'as is', but there's a lot of experience in light and perception modeling behind these visuals as well as a decent amount of research, so if you believe they're off, I'd like to see a solid case.

Last edited by Thorsten; 08-19-2016 at 05:43 AM.
Thorsten is offline   Reply With Quote
Thanked by:
Old 08-19-2016, 06:18 AM   #90
Interceptor
Orbinaut
 
Interceptor's Avatar
Default

No,you seem to be misunderstanding me,I am not questioning your judgement,I was just trying to help.

Last edited by Interceptor; 08-19-2016 at 06:27 AM.
Interceptor is offline   Reply With Quote
Reply

  Orbiter-Forum > Far Side of the Moon > Brighton Lounge


Thread Tools

Posting Rules
BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
Forum Jump


All times are GMT. The time now is 06:15 PM.

Quick Links Need Help?


About Us | Rules & Guidelines | TOS Policy | Privacy Policy

Orbiter-Forum is hosted at Orbithangar.com
Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2019, vBulletin Solutions Inc.
Copyright 2007 - 2017, Orbiter-Forum.com. All rights reserved.