Re: Habsheim accident (was: Re: Airbus Safer?)

Date:         02 Sep 98 01:07:56 
Organization: InLink
References:   1 2 3 4 5 6 7 8
Next article
View raw article
  or MIME structure

Two points:
1) aerodynamics in high alpha flight (high angle of attack (high AOA))
2) engines in high alpha flight

In article <airliners.1998.1326@ohare.Chicago.COM>, kls@ohare.Chicago.COM
(Karl Swartz) wrote:
> >...I can certainly imagine
> >that when you've suddenly realized that trees are in your path and pushed the
> >throttles to max power that it'll seem like forever as the engines spool up
> >and you begin climbing. That subjective perception, however, need not imply
> >any actual problem. It's certainly not a secret that jets take time to spool
> >up.

High alpha flight causes several shifts in aerodynamic behaviour which can
become "contrary" to normal stick-and-throttle feel and required action,
just as inverted flight also requires "different" action. In hi-alpha one
must be aware of these and compensate, unfortunately an airline computer
probably doesn't have the same exception handling routines to switch over
gains and control laws to handle this.

Additionally, to recover from hi alpha positions there are several
options, some of which do not work on all aircraft. The AFTI F-16, F/A-18
hi aplha demo and F-15 STOL Demonstrator could hold their nose and power
out of this AND ADD CONTROLS. Several unique features (ex.
thrust-to-weight, active engine control, exhaust nozzle paddles,
additional control surfaces, etc.)  enabled this which are NOT in
airliners. Underpowered aircraft would require to drop the nose slightly
and add power but require altitude trade-off for more acceleration.
Turbofan engines also lengthen this accel period over turbojet or
afterburner/reheat engines.

IMHO the computers in the Airbus (or any other airliner) do not have the
specific aerodynamic look-up tables or equations to solve for this region
of flight, although one can fly into them at will (as Airbus and others
have demonstrated). In various NASA and defence program hi aplha flight
tests there was significant remodelling of the control laws and rewriting
of the code to allow acceptable, recoverable perfomance in these regions,
and some of the borders of these regions are not obvious!

Even with a recorder, the parameters captured probably do not correlate to
the actual vehicle attitude/performance without speculating on the
"unobservable" parameters which are missing. Missing not due to oversight
but because they are normally optimized for the EXPECTED flight region(s).

> True enough, but Asseline also claims that the elevators did exactly
> the opposite of what he requested...
> With regard to the engines, there's the pair of "booms" heard by many
> shortly before the crash.  I believe the DGAC report attributed these
> to impact with the trees, but both the timing and the sounds were wrong.
> There was speculation that they were compressor stalls -- the versions
> of the CFM56 used on the A320 were having some problems at that time,
> corresponding to conditions not unlike those just prior to the crash.

Hi aplha flight "can" run havoc with the inlet flow of an engine.
Particularly if the pilot demands maximum power while in an extreme angle
of attack. The normally well behaved air flow will separate and stall over
the lip on the inlet and turbulent or redirected air then impacts the
inlet blades at an excessively higher AOA. This has sometimes caused
decreased performance or compressor stalls or other "things". Many engines
designed to "go there" do have additional features, design and testing to
prove themselves worthy of that risk. It is possible that this disturbed
air caused the engine computers to delay the spool up even more to avoid
the increased stall probability.

A further decreased engine acceleration, the sluggish hi alpha aircraft
performance and the proximity to the ground and the oncoming obstructions
could cause quite a bit on concern in the cockpit. Imagine what you would

BTW: Engine manufacturers do their own investigations to both help the
agency and also to get the appropriate data to defend themselves and for
product improvement. This data is still available to law, should that be
deemed required, and they are required to share their findings. Many
engines are software controlled to the point that losing a computer
decreases thrust by a significant margin. The complexity of this higher
effiiciency engine can only be unravelled by the software engineer of that
company as software cannot be "observed" as can cams, pulleys and wires
that most investigators grew up on. (There are new or well retrained
investigators who can handle this too, but the limited numbers still make
the manufacturer the fastest and most sure respondent).