19th Aug 2010, 00:22
permalink Post: 23
Biggles78
I must admit, it seems excessive carrying 10,000kg in a trim tank, but this fuel system really was a study in elegance. Every single drop of fuel carried was usable by the engines, and the Mach Trimming was so good that you could fine-tune the process so as to achieve the minimum drag configuration for the aircraft of 1/2 degree down elevon in supersonic cruise. One rather amusing point about the fuel Mach trimming; the airworthiness authorities insisted that the aircraft also had a conventional Mach trimmer built into the electric pitch trim system. As the aircraft was mostly flown on autopliot, assuming the fuel trimming was being done correctly (it always was), the auto-trim would wind off this Mach trimming as it was applied, the net result of course being no change to the pitch demand. This really was a totally superfluous addition to the electric trim system. (If for any reason the aircraft HAD been hand flown during acceleration, the pilot would have to nudge the trim button nose down all the time as the A/C accelerated, in order to to oppose the nose up electric trim input).
The fuel, apart from 'lighting the fires' and trimming the aircraft was also used as a cooling medium for engine and IDG oil, as well as for the hydraulic system also. Where it was used to massive effect, was as a cooling medium for the air conditioning system. Here, at Mach 2 conditions, we needed air to exit the 'packs' (on Concorde these were called 'groups') at around -25 deg's C. By the time this air had travelled through the wing ducting it had risen to a sweltering 0 deg's C, at which temperature it entered the cabin. The astonishing thing is, that the air used for this, HP compressor delivery air, P3, was at around 550 deg's C as it left the engine. The ram air itself, used to cool the Primary and Secondary heat exchangers, had a total temperature anything up to 127 deg's C, and to complete this story, the fuel itself had an average temperature of around 60 deg's C. And surprisingly enough, it was a more or less conventional air conditioning system, using air/air intercoolers, an air cycle machine, with just the addition of the fuel exchanger (between the outlet of the secondary heat exchanger and the ACM turbine) to make it any different in concept to most other air cond' systems.
I must admit, it seems excessive carrying 10,000kg in a trim tank, but this fuel system really was a study in elegance. Every single drop of fuel carried was usable by the engines, and the Mach Trimming was so good that you could fine-tune the process so as to achieve the minimum drag configuration for the aircraft of 1/2 degree down elevon in supersonic cruise. One rather amusing point about the fuel Mach trimming; the airworthiness authorities insisted that the aircraft also had a conventional Mach trimmer built into the electric pitch trim system. As the aircraft was mostly flown on autopliot, assuming the fuel trimming was being done correctly (it always was), the auto-trim would wind off this Mach trimming as it was applied, the net result of course being no change to the pitch demand. This really was a totally superfluous addition to the electric trim system. (If for any reason the aircraft HAD been hand flown during acceleration, the pilot would have to nudge the trim button nose down all the time as the A/C accelerated, in order to to oppose the nose up electric trim input).
The fuel, apart from 'lighting the fires' and trimming the aircraft was also used as a cooling medium for engine and IDG oil, as well as for the hydraulic system also. Where it was used to massive effect, was as a cooling medium for the air conditioning system. Here, at Mach 2 conditions, we needed air to exit the 'packs' (on Concorde these were called 'groups') at around -25 deg's C. By the time this air had travelled through the wing ducting it had risen to a sweltering 0 deg's C, at which temperature it entered the cabin. The astonishing thing is, that the air used for this, HP compressor delivery air, P3, was at around 550 deg's C as it left the engine. The ram air itself, used to cool the Primary and Secondary heat exchangers, had a total temperature anything up to 127 deg's C, and to complete this story, the fuel itself had an average temperature of around 60 deg's C. And surprisingly enough, it was a more or less conventional air conditioning system, using air/air intercoolers, an air cycle machine, with just the addition of the fuel exchanger (between the outlet of the secondary heat exchanger and the ACM turbine) to make it any different in concept to most other air cond' systems.
19th Dec 2010, 18:24
permalink Post: 887
Autotrim
It wasn't quite as simple as that. The fuel transfer system really fixed long term problems like getting the elevator trim broadly at optimum throughout (and really at optimum in cruise of course). The Mach trim/autotrim really worked on a shorter timescale to maintain stability at constant CG.
Sure the autopilot made it superfluous to some extent, but to certificate the aircraft it had to be conventionally stable when flow manually, and applying a nose down command to get a speed increase is a basic airworthiness requirement for all aircraft.
Sure the autopilot made it superfluous to some extent, but to certificate the aircraft it had to be conventionally stable when flow manually, and applying a nose down command to get a speed increase is a basic airworthiness requirement for all aircraft.
19th Dec 2010, 18:50
permalink Post: 891
Clive,
Re the autotrim, tell us some more?
I wasn't directly involved with the control laws themselves, more with trying to assure those control laws were respected to well below 1%.
I thought it was the first
civil
aircraft, and that the Vulcan had already been there and done that...
I know what you're saying.....
Still, I think you'll agree that lessons were learnt, rather than totally ignored.
I would say Airbus can trace its history back to the lessons learned from Concorde.
CJ
Re the autotrim, tell us some more?
I wasn't directly involved with the control laws themselves, more with trying to assure those control laws were respected to well below 1%.
Quote:
|
Originally Posted by
CliveL
Sure, Concorde was the first aircraft to fly with FBW flight controls...
|
Quote:
| There are some (I am not among them) who would say that the Concorde project was a good way to learn how NOT to run a major international collaboration |
Still, I think you'll agree that lessons were learnt, rather than totally ignored.
I would say Airbus can trace its history back to the lessons learned from Concorde.
CJ
21st Dec 2010, 11:19
permalink Post: 912
quote:Re the autotrim, tell us some more?unquote
It is a little complicated, but let me go back half a step.
Concorde was not certificated to FARs or BCAR (the French code was essentially a straight translation of FAR) but to completely new set of requirements known as TSS (Transport Supersonique Standards). The old UK ARB had initiated discussions about these even before cooperation negotiations had started. The result was that young, junior engineers got to debate the basics of airworthiness rules with older, experienced airworthiness specialists. In hindsight it was wonderful training!
But to get to the point, it was this thinking that allowed us to ignore some of the older rules which, although great for the aircraft flying at the time they were written, had little or no relevance to SSTs. We could interpret that as trying to find out what the pilots really wanted the aircraft to do and then to try and provide it.
In the particular case of trim/speed stability it was quite clear that what they wanted was an aircraft that could be flown with minimal trim changes and which once trimmed would not go wandering off all over the place. We also knew that in some cases the 'elevator angle per 'g' ' could get as low as one degree/g in some cases and that the pilot could not tell exactly where his hands were positioned to that precision, although he would always know if he was pushing or pulling. So we could abandon the old rules for stick movement and instead supply classic stick force stability for deviations from the trimmed state.
All this had to be matched to the varying aerodynamics through the transonic region (where everything varies rapidly) and the fuel transfer system characteristics. The resulting Mach trim laws were quite complex and were not, in fact just Mach Number sensitive. We also had two airspeed (Vcas) terms, one of which had a variable gain which was itself Mach dependent and kicked in above Vmo = 5kts and the other was a straight nose up elevator command as a function of Vcas. The Mach trim itself was highly nonlinear. The best way to illustrate this is probably a diagram but now I've run into another gap in my knowledge of this thread - how do I do that?
Anyway, the result was that the fuel transfer held the trim setting variation down to between 2 deg down to 1.5 deg up through the acceleration from 0.95M up to 0.5 deg down at Mach 2.0. Without fuel transfer the trim at Mach 2 would have been closer to 10 deg. The trim between say 0.95 and 1.2 varies in a nonlinear fashion and the Mach trim law shows roughly similar variations.
But the best measure of our success is the comments we are getting here from the guys who actually had to fly it.
Clive
[IMG]file:///C:/Users/Clive/AppData/Local/Temp/moz-screenshot.png[/IMG][IMG]file:///C:/Users/Clive/AppData/Local/Temp/moz-screenshot-1.png[/IMG]
It is a little complicated, but let me go back half a step.
Concorde was not certificated to FARs or BCAR (the French code was essentially a straight translation of FAR) but to completely new set of requirements known as TSS (Transport Supersonique Standards). The old UK ARB had initiated discussions about these even before cooperation negotiations had started. The result was that young, junior engineers got to debate the basics of airworthiness rules with older, experienced airworthiness specialists. In hindsight it was wonderful training!
But to get to the point, it was this thinking that allowed us to ignore some of the older rules which, although great for the aircraft flying at the time they were written, had little or no relevance to SSTs. We could interpret that as trying to find out what the pilots really wanted the aircraft to do and then to try and provide it.
In the particular case of trim/speed stability it was quite clear that what they wanted was an aircraft that could be flown with minimal trim changes and which once trimmed would not go wandering off all over the place. We also knew that in some cases the 'elevator angle per 'g' ' could get as low as one degree/g in some cases and that the pilot could not tell exactly where his hands were positioned to that precision, although he would always know if he was pushing or pulling. So we could abandon the old rules for stick movement and instead supply classic stick force stability for deviations from the trimmed state.
All this had to be matched to the varying aerodynamics through the transonic region (where everything varies rapidly) and the fuel transfer system characteristics. The resulting Mach trim laws were quite complex and were not, in fact just Mach Number sensitive. We also had two airspeed (Vcas) terms, one of which had a variable gain which was itself Mach dependent and kicked in above Vmo = 5kts and the other was a straight nose up elevator command as a function of Vcas. The Mach trim itself was highly nonlinear. The best way to illustrate this is probably a diagram but now I've run into another gap in my knowledge of this thread - how do I do that?
Anyway, the result was that the fuel transfer held the trim setting variation down to between 2 deg down to 1.5 deg up through the acceleration from 0.95M up to 0.5 deg down at Mach 2.0. Without fuel transfer the trim at Mach 2 would have been closer to 10 deg. The trim between say 0.95 and 1.2 varies in a nonlinear fashion and the Mach trim law shows roughly similar variations.
But the best measure of our success is the comments we are getting here from the guys who actually had to fly it.
Clive
[IMG]file:///C:/Users/Clive/AppData/Local/Temp/moz-screenshot.png[/IMG][IMG]file:///C:/Users/Clive/AppData/Local/Temp/moz-screenshot-1.png[/IMG]
22nd Dec 2010, 20:13
permalink Post: 950
Quote:
|
Originally Posted by
exwok
Hazy recollection - effectively an additional autostabilisation input in the nosedown sense active at high alpha/low EAS.
Ultimately applied a further nose down elevon input (4 degrees????) if EAS was less than (140kts???? That's a VERY low speed). (Colloquially known as 'super-duper stab' on my course) |
To cover this case the 'superautostabiliser'was developed. It effectively restricts the rate of variation of incidence so that, if the pilot entered into an avoidance manoeuvre of sufficient magnitude to trigger the stick wobbler, i.e. about 1.5g, he would be able to recover easily without exceeding the maximum incidence demonstrated in flight (which was in fact slightly greater than the maximum steady incidence limit). This superautostab had gain scheduled against AoA and also included phase advanced pitch rate and speed terms. Finally, there was a 'yaw superautostabiliser which applied rudder as a function of lateral acceleration to restrict sideslip which (see below) could affect the maximum lift attainable. [Note that because of the dynamics of slender aircraft operating at high AoA it was readily possible to develop sideslip in a turn]
Hope that is clear.
Whilst talking about maximum lift etc. can I confirm the numbers quoted in an earlier posting for the start of vortex lift - about 6 or 7 deg AoA at low speed, and for the AoA at maximum lift - about 23 deg. This is where the pitching momemt curve vs AoA 'breaks'. It is not a stall in the conventional sense because of course the flow over the leading edge has been separated long ago. Instead it is the AoA at which the LE vortices become 'too big for their boots' and go unstable and 'burst'. This AoA is sensitive to sideslip and the leading wing half will go first.
CliveL
22nd Dec 2010, 20:40
permalink Post: 951
Before adding my own little bit to Clive's earlier reply about the autotrim, I will try to explain, for those not fully familiar with the subtleties of automatic flight control, the difference between "closed loop" and "open loop".
Closed loop
As an example, let's look (very simplified) at how the autopilot maintains a selected altitude.
On the one hand we have the desired altitude as selected on the autopilot controller (here 40,000 ft).
On the other hand we have the true altitude , as measured by the altimeter (let's say 39,000 ft).
We subtract the two to obtain the altitude error (in this case 39,000-40,000=-1,000 ft).
We 'multiply' the altitude error by a factor, the gain (for the discussion, let's assume this gain is 1 degree elevon per 1000 ft altitude error), and send the resulting elevon position command to the elevon.
So, the elevon moves 1\xb0 nose-up, the aircraft starts to climb, the altitude increases and the altitude error decreases until it becomes zero, by which time the elevon position has also returned to zero.
What we have now is a "closed loop" : any deviation from the selected altitude results in an elevon command in the opposite direction, until the deviation is again reduced to zero.
Another commonly used term is "feedback" : any error is fed back in the opposite sense until it's reduced to zero.
The significant figure here is the 'gain'.
If the gain is too small, the autopilot response to a disturbance (say turbulence) will be sluggish ; the aircraft takes too long to return to the desired altitude.
If the gain is too high, a small disturbance will cause the aircraft to start climbing too rapidly, and to overshoot the desired altitude, then descend to correct the new error, etc.
In other terms, the control loop is no longer stable, but starts to oscillate.
Both theory and practice show that the exact value of the gain is not all that critical, a few percent more or less do not markedly change the response of the loop.
Note: a "closed control loop" as described above can be implemented in just about any way you like.
It can be done purely mechanically, with a few clever clockwork mechanisms 'computing' the altitude error and controlling the elevator pneumatically or hydraulically. It's how the earliest autopilots worked.
After that came electromechanical systems, analogue computers and then digital computers... but the principle has remained unchanged.
Open loop
As already described in earlier posts, the situation with the automatic trim is the opposite.
We now need to compute a neutral elevon position from several data, such as Mach number or airspeed, but without any feedback as to whether our computations are correct.
We're now working in "open loop".
To complicate matters... that neutral elevon position is not a simple linear function of Mach and airspeed, but far more complex (see the earlier posted graphs).
And because of the large response of the aircraft to small changes in trim, in particular in the transonic regions, those computations have to be far more accurate : a one degree error is simply not acceptable.
In the end.....
The AICS (air intake control system) also uses several "open loop" functions.
The early development aircraft still had an analogue system, which proved all too soon to be inadequate, so, at a very late stage, it was replaced by a digital system (one of the rare digital systems on Concorde).
The "open loop" functions of the autotrim system initially had the typical "a few" percent" accuracy of the other flight control systems, which, for the autotrim, also proved inadequate.
We managed to "save the furniture" (as they say in French) by using 0.1% components in all the critical computing paths, so the autotrim computers remained analogue until the end.
But, a slide rule is not accurate to 0.1%... So that's when I had to buy my very first pocket calculator.
\xa342 at 1972 prices... just as well the firm paid.
CJ
Closed loop
As an example, let's look (very simplified) at how the autopilot maintains a selected altitude.
On the one hand we have the desired altitude as selected on the autopilot controller (here 40,000 ft).
On the other hand we have the true altitude , as measured by the altimeter (let's say 39,000 ft).
We subtract the two to obtain the altitude error (in this case 39,000-40,000=-1,000 ft).
We 'multiply' the altitude error by a factor, the gain (for the discussion, let's assume this gain is 1 degree elevon per 1000 ft altitude error), and send the resulting elevon position command to the elevon.
So, the elevon moves 1\xb0 nose-up, the aircraft starts to climb, the altitude increases and the altitude error decreases until it becomes zero, by which time the elevon position has also returned to zero.
What we have now is a "closed loop" : any deviation from the selected altitude results in an elevon command in the opposite direction, until the deviation is again reduced to zero.
Another commonly used term is "feedback" : any error is fed back in the opposite sense until it's reduced to zero.
The significant figure here is the 'gain'.
If the gain is too small, the autopilot response to a disturbance (say turbulence) will be sluggish ; the aircraft takes too long to return to the desired altitude.
If the gain is too high, a small disturbance will cause the aircraft to start climbing too rapidly, and to overshoot the desired altitude, then descend to correct the new error, etc.
In other terms, the control loop is no longer stable, but starts to oscillate.
Both theory and practice show that the exact value of the gain is not all that critical, a few percent more or less do not markedly change the response of the loop.
Note: a "closed control loop" as described above can be implemented in just about any way you like.
It can be done purely mechanically, with a few clever clockwork mechanisms 'computing' the altitude error and controlling the elevator pneumatically or hydraulically. It's how the earliest autopilots worked.
After that came electromechanical systems, analogue computers and then digital computers... but the principle has remained unchanged.
Open loop
As already described in earlier posts, the situation with the automatic trim is the opposite.
We now need to compute a neutral elevon position from several data, such as Mach number or airspeed, but without any feedback as to whether our computations are correct.
We're now working in "open loop".
To complicate matters... that neutral elevon position is not a simple linear function of Mach and airspeed, but far more complex (see the earlier posted graphs).
And because of the large response of the aircraft to small changes in trim, in particular in the transonic regions, those computations have to be far more accurate : a one degree error is simply not acceptable.
In the end.....
The AICS (air intake control system) also uses several "open loop" functions.
The early development aircraft still had an analogue system, which proved all too soon to be inadequate, so, at a very late stage, it was replaced by a digital system (one of the rare digital systems on Concorde).
The "open loop" functions of the autotrim system initially had the typical "a few" percent" accuracy of the other flight control systems, which, for the autotrim, also proved inadequate.
We managed to "save the furniture" (as they say in French) by using 0.1% components in all the critical computing paths, so the autotrim computers remained analogue until the end.
But, a slide rule is not accurate to 0.1%... So that's when I had to buy my very first pocket calculator.
\xa342 at 1972 prices... just as well the firm paid.
CJ
23rd Dec 2010, 00:39
permalink Post: 957
Superstab
Hmm. There was, I think, a raft of high-incidence (alpha) protection fitted.
Digging out the old BAe conversion course notes:
The "Anti-Stall" (SFC) 1&2 sytems offered:
Super Stab: Increased authority of pitch autostab as incidence increased above 13.5 degrees - proportional to pitch rate and incidence angle - and a nose down pitch trim with a Vc (CAS) deceleration with incidence > 13.5
Stick "Wobbler": the "unmistakable warning" - when incidence > 19 and Vc<270kts the control columns took a life of their own and tried to fling you into the forward galley. Served you right.
Some other high incidence stuff was fed from the ADC rather than the SFC, like:
The ">13.5d incidence" feed to the SFC
CAS (Vc) feed to the SFC
Incidence from 16 to 19 degrees (rate dependant) to get the SFC to feed in up to 4 degree nose down pitch command and the sticj wobbler trigger.
Increase of authority of yaw autostab as incidence > 13.5d
Autotrim inhibit > 14.5d incidence
Stick shaker >16.5d incidence
AP/FD disconnect > 17.5d incidence
There was loads of other technical stuff which engineers understood, but we had to learn by writing diagrams which made sense to us enough to pass the written exam. The bottom line was an aeroplane which flew beautifully, but which you had to understand well, and which you could not tease beyond its limits. If you ignored a limit or an SOP then you reached an unpleasant place far quicker than with the blunties - it was a challenge which rewarded as quickly and as deeply as it punished.
Digging out the old BAe conversion course notes:
The "Anti-Stall" (SFC) 1&2 sytems offered:
Super Stab: Increased authority of pitch autostab as incidence increased above 13.5 degrees - proportional to pitch rate and incidence angle - and a nose down pitch trim with a Vc (CAS) deceleration with incidence > 13.5
Stick "Wobbler": the "unmistakable warning" - when incidence > 19 and Vc<270kts the control columns took a life of their own and tried to fling you into the forward galley. Served you right.
Some other high incidence stuff was fed from the ADC rather than the SFC, like:
The ">13.5d incidence" feed to the SFC
CAS (Vc) feed to the SFC
Incidence from 16 to 19 degrees (rate dependant) to get the SFC to feed in up to 4 degree nose down pitch command and the sticj wobbler trigger.
Increase of authority of yaw autostab as incidence > 13.5d
Autotrim inhibit > 14.5d incidence
Stick shaker >16.5d incidence
AP/FD disconnect > 17.5d incidence
There was loads of other technical stuff which engineers understood, but we had to learn by writing diagrams which made sense to us enough to pass the written exam. The bottom line was an aeroplane which flew beautifully, but which you had to understand well, and which you could not tease beyond its limits. If you ignored a limit or an SOP then you reached an unpleasant place far quicker than with the blunties - it was a challenge which rewarded as quickly and as deeply as it punished.