News The Space Shuttle for Flightgear 3.6

[Gingin]

Awsome :thumbup:
I almost managed to set up dev version with Windows Git.
It works like SVN

Let's Torque those IMU

Ah nicely looking the SPEC 21, some Delta degrees to Torque.
You know by how much can degrade the IMU in the sim so far if we don't periodically align them?

 

[Thorsten]

You know by how much can degrade the IMU in the sim so far if we don't periodically align them?

Well, the IMU workbook has finally something like data on the drift rates (have been looking at those for a while) - I've used the fault detection threshold time evolution to model them.

Then it's a question of how long you wait. The IMUs will also degrade and drift faster with too high temperature (not done yet). The system is easy to code if you assume it works nominally, but there's a lot of combination of failures and bad crew input that need to be caught. So I'm still not quite done with interfacing, although the worst is probably done now.

***

I've just received a hands-on lesson in redundancy management procedures. I've tried a TAL test for a high-inc launch from Vandenberg to Mataveri Airport (Easter Island).

First that uncovered a small bug in the TAL AP which prompted me to take CSS during late ascent.

No big problem, but I was then careless and engaged Aerojet DAP without nulling beta first (I had nonzero beta from vectoring towards Mataveri which I hadn't managed to get quite perfect). To make matters worse, I was at that time looking down and busy closing ET umbilical doors and finishing the MPS fuel dump.

By the time I noticed, the nonzero beta had led to some pretty bad yaw oscillations outside the atmosphere and the entry AP was on the verge of losing control, so I re-engaged CSS and stabilized attitude first. I barely managed to pitch up to a thermally safe attitude but entered the atmosphere with a severe bank, thus creating a pretty hefty heating peak which almost fried the TPS (I think I came within 20 F of the limit...) - by the time I had that under control I had dropped quite a bit below nominal trajectory.

Since I didn't dare re-engage AUTO, I flew the rest of the entry manual in both axes (haven't done that for a while...) and managed to recover the energy I had lost - before mis-judging the handling in the late entry and climbing quite a bit to a high-energy condition.

I arrived at TAEM interface with above Mach 4, which the TAEM AP couldn't handle gracefully, so I was busy flying the Shuttle while hacking a different TAEM guidance in (changing the runway direction as well as down-moding to straight-in and MEP did the trick). Somewhere inbetween I also managed to deploy air data probes, incorporate TACAN, set altimeter and make use of air data and set the ELS speedbrake option.

Unfortunately I had had too little time to look at the instruments, and by the time I came out of the HAC I had zero spatial orientation, no view on the runway due to cloud cover and (without MLS) a not quite clear navigation state.

Coming through the clouds I was then some 400 m laterally off the runway. I actually managed to correct that as well, but the approach was pretty messed up already (and Mataveri is not a large runway) and I then broke the landing gear upon touchdown and had quite a hard belly-landing.

Lessons learned:

* There's a lot of safety nets which open in case something goes wrong
* There's a good reason there's both a commander and a pilot on the flightdeck - under stress, the workload is high and for a single person trying to manage it all, this quickly leads to bad decisionmaking
* The ability to pilot manually is important to survive if things start to go wrong

 

[Gingin]

Very nice story !

I love the fact that you can have significant offset from the runway if you don't integrate very precise kalman filters like MLS or GPS.
That's making the simulation very immersive and challenging.

Yes, doing every thing in single pilot operation, especially abort or contigency is very tight and demanding time wise.

I can confirm you that to be 2 in a cockpit is essential for the workload and confirmation, decision taking, and monitoring what the other is doing.
It's like synergia decouples our strenght.

I am gonna test some similar scenario tomorrow, with not incorporating all filters to see the divergence from Real and Perfect State vector.

Well done for your manual entry and TEAM, looks like it was tight temperature wise

 

[Thorsten]

On a personal note...

I suspect everyone is far more fond of posting success stories (don't we all like to be dashing space pilots?) But I've learned far more from my simulated failures, and for that reason alone I enjoy less than perfect stories.

I remember one sim flight with a single prop aircraft in the Alps, with weather deteriorating. And I was 'Well I know the area, I'm going to make it.' Poor visibility, wrong turn into a valley, not enough maneuvering room to turn around, crash - you get the picture. Brought the virtue of playing it safe home better than a dozen lectures could have.

Or one of my first IFR approaches with the Concorde - I managed to lock onto a secondary lobe of the ILS signal, which went crazy as I approached the runway and I barely managed to save the plane. Told me that there's a reason there's approach charts and that one should stick to the correct location when intercepting ILS...

I continue to be amazed by reports of people who let the simulated autopilot kill them - they apparently observe that things go wrong but persist in a belief that the AP will know and distrust their own better judgement that things look bad - till it's too late. It goes a long way to explain some real life accidents (not so long ago, I had a discussion with someone who lost a simulated plane on a transatlantic trip when the AP disengaged as it couldn't handle a situation while he was sleeping - he was pretty annoyed when I argued that that's the reason there's got to be someone in the cockpit in a real plane at all times...)

Anyway - I'll freely admit I'm not the best pilot around and I have crashed a lot of virtual Shuttles and planes, often due to my own mistakes. But usually I've learned something interesting from it...

 

[Gingin]

I managed to lock onto a secondary lobe of the ILS signal,

Ahah you were on a 6° glide with the concorde?

that's the reason there's got to be someone in the cockpit in a real plane at all times...)

So true, and that's why I will never take a plane without pilots if its gonna exist
Human factor is too much important to let a piece of engeneering do everything alone.

they apparently observe that things go wrong but persist in a belief that the AP will know and distrust their own better judgement that things look bad

Yep, and there are still somme incidents/accidents every year about that stuff.
That's why its really important to still fly manually and keep our competences, be confident and thrust our "Human" judgement and experience.

And I crashed so many Shuttle also ahah, 90% of them during Manual RTLS or Contigency abort.

 

[Blake]

Interesting conversation. I would recommend the book Digital Apollo by David A. Mindell. The book is about the Apollo Guidance Computer, but the deeper theme explored throughout the book is the conflict between man and machine and what it means to be in control of an airplane or spacecraft.

BTW Thorston, I always love coming to this thread. I appreciate the passion and interest with which you approach this project.

:thumbup:

 

[Urwumpe]

Interesting conversation. I would recommend the book Digital Apollo by David A. Mindell. The book is about the Apollo Guidance Computer, but the deeper theme explored throughout the book is the conflict between man and machine and what it means to be in control of an airplane or spacecraft.
:thumbup:

And then you have the F-8C Crusader project of NASA, which used a AGC for testing fly-by-wire fighter aircraft with it.

The reports about NASAs experience with FBW since the X-15 go into the same field of questions: What a pilot expects and how the aircraft behaves.

 

[Gingin]

Wow crazy stuff to test the AGC with a Cruz :thumbup:
Thanks for the reading recommendation


Let's go for some Tal abort to ZZA

 

[Thorsten]

but the deeper theme explored throughout the book is the conflict between man and machine and what it means to be in control of an airplane or spacecraft.

That sure sounds interesting! Need to have a look.

Illustrating that theme of control is actually the reason I've left the 'raw' control modes in the Shuttle FDM (by 'raw' I mean stick directly commands gimbal angles during ascent or airfoil positions during entry - like for a simple GA plane - rather than going through the rate controllers of the DAP). It's very educational to see how the 'feel' changes (and how little 'feel' is related to the aerodynamics) and what work the controller loops actually do.

I can barely manage a survivable entry with raw airfoil control but I can't do ranging (it's insane - initially you need HUGE movements because even full deflection hardly generates any moments, in the end you hardly need to move the stick...) - and I cheated somewhat because there's still a controller in to tame the yaw instability - that I never managed to control by hand. Likewise, I can fly into orbit with direct engine gimbal control, but not achieve any sort of precision in the inclination. But it illustrates the role of the atmosphere very drastically, flight characteristics are completely different inside and outside, which is very well hidden by the rate controller CSS utilizes.

 

[Gingin]

A nice test of how State vector computed can differ from reality if no update is done from MCC or from Kalmann filters like GPS, MLS, ADTA

Roughly 2000 feet upset on -Y axis, 1000 f on -X and 500 on +Z axis :thumbup:

 

[Gingin]

I tried a nice scenario this afternoon, some failures with Fuel Cells and Main Bus, very interesting. I had a look to non normal checklist concerning those kinds of failures, quite time consuming, but interesting to have an idea of what need to be done and what system will be concerned by associated Buses.

Basically, a loss of a main Bus will lead to a shutdown of the associated Fuel Cell later on. Some Bus Tie are available to link a Good Bus to the bad one and provides power like if nothing happened. ( 1 Fuel cell powering 2 Main Buses)

Consequences of the loss of a main bus will have impact also on AC Associated bus, which can impact SSME an APU powering ( double redundance "only", not triple like systems linked to Essential bus)

So loss of main is almost negligeable, loss of two begins to be critical and decision making will be essential.



Let's start our journey, by a warm evening on the East Coast, almost no clouds in the sky, beautifull sky and some Altostratus waiting to be ripped off by the Shuttle.
Launch was delayed by some false warning indications now cleared.
Final countdown begins, final adjustement are made, APU starts and IMU pre launch alignement are completed. PASS software transfers his State Vector to the Backup System, let's hope it will not be activated.

A view from camera 1, everything looks good so far,

View from the pilot side is quite astonishing, let's have a good deep breath before the stress of the 3 G's

T-6 s, SSME ignitions, nominal thrust reached, SRB activation, We have a LiftOff.

Everything is looking good, bucket is approaching and maximal dynamical pressure is now Behind.
Master Caution comes alive, Main Bus Undervoltage, AC Voltage, and some Black MDU, Let's have a look to the SM BFS System page

All right, seems we have some issues with Main Bus A, FC 1 is still alive, and AC 1 is lost so far. MCC confirms our diagnostic, but no rush, everything else is looking good, no abort so far.
Most of the systems have double redundancy, some even triple redudancy.
Let's have a look and reconfigure the systems that are working on MnA or AC 1.

Some concerns about Avionics Bay one fan which are working on AC 1, temperature is rising.

Let's switch it on AC 2, Temperature is decreasing now, good.
Next move will be a BUS Tie on MCC call ( it's normally not recommanded before MECO, but you know curiosity...), we gonna connect Main B to Main A, switching off AC 1 and inverter switches before Cross Tie and putting them back then.

It seems to work, MDU are back, and MnA and AC 1 are alive again, taking load from FC 2

Everything is back under control, everyone is sweating, and SRB separation happens, we are 2 mn10 sec into the flight.

And here we are again, the worst case scenario is happening, another Master Caution Alarm and blank MDU. What now? Cockpit is looking dark and stress is rising.

Let's have a quick glance on SM page: Main A and B, AC1 and AC 2 are lost, Essential Bus are still showing good voltage due to triple redudance.
FC 1 and FC 2 are shutdown as per Checklist, and no Bus Tie available.

Fuel cell 1 and 2 secured in Off position

And to top it all, as we lost two AC sources, two controllers of one engine are lost, so is the engine and one APU is also lost, Phew. Situation is rapidly degradating

Here are the SSME controllers, we lost AC 1 and AC2, so center engine is lost

Also, on SM page we can see that APU 1 is out of service, which means a thrust lockup on center engine, hopefully, no lock up on available engines.

So we are approaching 3mn40, MCC took a decision between RTLS and TAL , two Main Buses Down mean a RTLS, It's a highly time critical failure, no time to cross the Atlantic and to wait for a complete Electrical failure.

Commander took time to select the correct Abort Mode, here we are, abort initiated and OPS 6 is alive.

Ok, now it's time to see what is working or not, so no Main A and B, no AC 1 and 2, let's have a look to the Environmental and Thermal Panel on L2 side.

So we lost for sure some systems, but let's try to save what we can.
IMU Fan is switched on C ( AC 3), Avionics Fan from bay 2 and 3 are switched to AC 3, Avionics Bay Fan 1 is lost, temperature is rising a lot

Let's continue our Diagnostic, One Water Loop is lost, hopefully Loop 2 is switched on AC 3 and Freon Interchanger will work
We lost also one Freon Loop, the other is working well.( on the picture above, Freon Flow from loop 2 sensor is alive, good )
On the next one, we see no flow on loop 1

Primary A FES is powered by Mn C, so it works. For later in the entry, Ammonia Boiler will be available on Primary B controller, good.
Air System 2 is also activated in case system 1 fails .

So, let's sum up.
We still have 2 engines and 2 APU, no lock up.
Thermal systems works, one H2O loop, one Freon loop and at least one FES and Ammonia boiler work.
IMU Bay air circulation is ok.
Air system and pressurisation ok.

So vital systems are working so far.

It's gonna be a normal RTLS, with some limited access to CRT and MDU, that's gonna be the complicated part, switching between SM and OPS 6/Spec 50/Spec51

We are heading back to KSC, MCC is on fire:

It's becoming very dark inside, Guidance convergence is optimal and PPD approaches

Alpha recovery phase

It's time for the Glide Return To Launch Site, tremendous view

Time for the probes to be out, we can see on Spec 51 what the probes are measuring under ADTA ( Altitude, angle of attack and Mach) and compared it to what IMU is displaying on ADI, then we can intregate via Spec 50 the ADTA value to the guidance loop and Altitude coming from ADTA to navigation filter to affinate the State vector.

 

[Gingin]

You can notice some difference in Alpha angle and Altitude, Probes gonna affinate our Guidance value

I tried a Nav Delta update on Spec 50 ( Right Under NAV DELTA, we can enter an offset to the current state vector on X, Y ,Z axis and on speed, then load it with item 16)
Those offset will be given by MCC

We are approaching from the ground, looking for the runway

Monitoring is not easy task with just 2 MDU and on CRT


I was short for Runway 15, I changed to 33 with Straight in and Minimal Entry Point to Shorten a maximum, and when turning on final runway axis, following perfectly the guidance, I had the good surprise to be way closer to the Runway than expected
I forgot Kalmann Filters ( so much things to do), and State Vector was very degraded ( He thought runway was 3 km further on X axis ahah)

I managed an acrobatic non conventionnal landing, not very proud of it, but it worked out.


I was very impressed by the fidelity included into Electrical Systems. Very nice to do some scenarios like that and to have some very realistic effects on systems, especially Engines and APU's.
I have to say that the addition of the BFS systems and especially the SM function is really great and usefull for those kinds of scenarios, thanks again and great work on it in such a small time period !!!!


I hope you enjoyed my sum up, it was a lot of fun to prepare this unfortunate scenario

 

[Thorsten]

First test of automated droop guidance kicking in for relatively late 2nd stage dual engine failure.

After the failure, the guidance automatically pitches the vehicle up (likewise, SERC comes on automatically). The prognosis is okayish...

... but activating 109% of thrust improves it somewhat.

(currently the prediction tends to be a tad pessimistic...)

Guidance stopped the descent soon after (at 342 kft), but the remaining propellant was of course insufficient to reach orbit, so this could perhaps be down-moded to a single-engine TAL (I didn't try). This class of scenarios always ends with a regular OPS 3 entry and not an OPS 6 contingency N_z holding entry, propellant is sufficient to stabilize the trajectory and build speed.

 

[Gingin]

Looking good :thumbup:

So after, plan is to be able to do some sort of low energy TAL or Bailout after a " normal entry" ?

 

[Gingin]

Everything started, as usual, with a nice morning in Cap Canaveral, astonishing view from the pad and Sunrise far on the Horizon.

Some nice lighting on the Pilot Panel, quite eyes candy :

A bit of Rain is coming, coherent with Metar and tempo of showers of rain, nothing to worry MCC, still Good to go

Ignition time, we have a liftoff, tower is cleared and Shuttle is now breaking the clouds at Mach 1

First stage looks good, SRB separation at nominal time

What a nice view of Florida waking up

A nice view from the rear cabin, Switches shining in the sunlight

We are now almost 5 mn into the flight and master caution with Red failure message appeared on CRT, loss on center engine

We are too far into the flight for a RTLS, Negative return happened around 4 mn / VI of 8000 ft/s.
It's gonna be a TAL abort, to Banjul in Gambie, almost located on equator and easily reachable with our low inclinaison launch.
TAL Traj 2 appears on CRT, guidance is coming alive.

A nice view of a feature I love, The Flash Evaporator System, which evacuates the heat from Freon Loop after Water vaporisation cooling, to maintain temperature of Freon under 60°F

Right, we have initiated TAL abort, some work to do then.
First is to get rid of all Fuel and Oxydizer from OMS tank to have a correct Center of gravity for re entry.

To do that, we go to Spec 51 Override ( Item 6 7 and 8), and initiate an OMS dump via the RCS thruster, the famous RCS to OMS interconnect
Hopefully, we don't have to do manually the cross feeding between Fwd RCS and OMS, it's done automatically in those time critical abort.

And some visuals on the dump, every RCS thrusters firing simultanously to avoid any moment on the Shuttle

Dump completed, propellant remaining less than 1 %

MECO time approaching around Mach 25, ET separation, Umbilical doors closure, LH2 and LO2 dump initiated.
Look at those almost 3000 pounds of LO2 going directly outside through LO2 Main Valves and SSME nozzle !!

Everything happens fast in a TAL abort, we are almost at entry interface, 350kfeet less than 2000 Nm from Banjul
First stage of entry gonna be steeper than usual one, angle of entry is higher due to engine failure, Thermal Constraint gonna be critical in the next minute, but nothing to worry about, normally...

Nice angle of attack

Here we are, temperature rising a lot, Orange Plasma glowing around the window is frightening, hopefully, Cabin is well insulated against heat, otherwise, we would be transformed into some fried Shrimps

Phew, thermal phase done, back to a normal entry, and roll reversal begin to control Drag required via roll and rate of descent.
Great work from Entry DAP and Pitch, Roll, Yaw law, like in the book

Closing in, almost in upper atmosphere and less than 500 Nm from Africa

Approaching Mach 5 , time for Probe Heat and some work in Spec 51 and 50 to affinate State Vector.
First on Spec 50, we can update Altimeter QNH on item 9, and via item 19 42 22 allow automatic selection of GPS, TACAN , Drag model which will correct the state vector with some ground based Beacons, GPS and Drag model vs altitude.
Also in Spec 51, below ADTA we can see value from the Pitot and Static Probes and include them into Guidance and Control software via item 28 on Spec 50
One of the part I prefer, so realistic.

Coast in sight, we are overflying Cap Verde, last step before Gambia

Approaching TAEM interface, below 100kft , almost straight in for Runway 14, Half of a HAC is forecasted, a bit low on the path

I decided to go in straight in option via item 6, and suddenly , we are too high, time for some S turn and energy bleeding, if time allows

Really high on speed to catch the path, 380 kt, far above 330kt VNE, but NASA said Loss of Control shouldn't be expected below 470 kt, let's hope ahah

Touch high in speed and too far from aiming point...

And of course, as I could have guess during final approach, runway excursion at low speed , bu still

One last nice view from Tropical trees and Savana, we will have to explain ourselves with Nasa Chief Pilot John Young

I think I will have to practice more TAEM and final approach, I am still not really accurate with Flare and estimated distance.

 

[Thorsten]

I think I will have to practice more TAEM and final approach, I am still not really accurate with Flare and estimated distance.

The new approach trainer function ought to help... You fly your landing (here to Bermuda)...

... and after wheels stop, you get to see an evaluation (how your touchdown and rollout parameters were, how the gear looks like)...

... then you can change the weather or set different auto-SB or aim point options for the Shuttle, press 'Retry' - and all damage gets repaired, guidance re-set and you're back in the air and can re-do it all again.

Right till you get sick of it

 

[Thorsten]

I managed to get a lot of nice screenies testing the approach trainer function - it's really neat how one can set up different weather and environment conditions between approaches and train all sorts of scenarios.

Approaching Bermuda International Airport in somewhat poor visibility (I got a view on the runway only after clearing the cloud base):

Evening touchdown on Bermuda:

A fair evening, still on Bermuda:

LeTube Airbase, France (TAL site):

... and drag chute out (note that the next FG release will have these nice volumetric grass blades rendered on the airport keep - very neat effect if you have a new GPU, based on a geometry shader)

Same approach in sketchy weather - slight drizzle:

I didn't actually know we had spray implemented, Wayne must have done this at some point...

And a night landing at LeTube, touchdown zone illuminated by floodlights:

 

[Wolf]

Man, on more of these posts and I'll install FG on my PC

 

[Gingin]

Wow very nice, thanks for the trainer addition.

Wolf , you don't have Enough to be convince ahah ?

 

[Thorsten]

Setting up approaches is where a software that does flight simulation by nature can really shine - you can have it all, crosswinds, turbulence, windshear, poor ground visibility, rain, ice, various clouds, thunderstorms and lightning if you really dare - even taxiing traffic (I suppose I need to do a landing on Boston Logan - that's really on the level of payware addons for FSX these days).

Though it's sort of fun watching the yachts cruising around Bermuda while landing.

---------- Post added 08-24-17 at 12:23 PM ---------- Previous post was 08-23-17 at 05:04 PM ----------

I couldn't resist at least trying this and running some nastier weather simulation...

There's a good reason NASA requires the landing site clear of T-storms - it's basically not doable. Not even a bit (which would establish that I'm definitely not that guy from 'Space Cowboys'...).

You keep thinking that it doesn't matter that you don't see the runway because you have the virtual overlay, but turbulence is so heavy it can throw you so much off-course the virtual runway is no longer in the HUD. By the time you reach the boundary layer and turbulence subsides somewhat, you still have not much in terms of visibility and it's too late to correct the approach, usually you are a hundred meters laterally off already, so you might barely make a touchdown, but the first gust that catches you under the wing will throw you off the runway - and the drag chute becomes a nasty piece of work in crosswinds.