Winch Launch Safety Study

In message , Derek Copeland
writes
At 23:53 18 March 2009, bildan wrote:
On Mar 18, 12:30=A0pm, Derek Copeland wrote:

2) Obviously you have never watched a dangerously steep, overpowered,
over-rotation on a winch launch. I have!

No, I haven't seen any of the weird things you describe. They all
seem to happen in your immediate vicinity. I wonder why.

Probably because I'm a gliding instructor at a large and very busy club
that does a lot of winch launching. I also visit other clubs that winch
launch.

I should add that probably 99% of winch launches go to plan (including
deliberate practice launch failures by instructors) without any
difficulties. About 1% suffer cable or weak link breaks, or other
technical failures of one sort or another. Given proper training these
should be just an inconvenience, rather than an emergency situation.

I have over 300 winch launches - not many by some people's standards but
still a fair number. I have had only 1 genuine launch failure (caused
by the club member driving the winch, not by me or by equipment
failure), so either I'm due a run of cable breaks and the like, or the
percentage of failures is better for some pilots/ glider types than
others... (or a bit of both of course)

snip

--
Surfer!
Email to: ramwater at uk2 dot net

On Mar 18, 11:39*pm, wrote:
Karl,
I agree that vehicles are not prone to make gliders do space shuttle
like launches, due to all the inertia of that set up. Though thanks
for sharing that even there it is possible, not just with winches.

search RAS on groups.google.com for phoebus c accident for just such
an accident

it's not entirely clear whether it was an all-flying tail stall or
over rotation into a stall

Frank Whiteley

bildan wrote:

On Mar 18, 12:30 pm, Derek Copeland wrote:

2) Obviously you have never watched a dangerously steep, overpowered,
over-rotation on a winch launch. I have!

No, I haven't seen any of the weird things you describe. They all
seem to happen in your immediate vicinity. I wonder why.
Hi Bill

Derek is not alone here.

As additional example - Std Cirrus can also be coaxed into too steep a
climb with too much power too fast.

In this case it is often a combination of
- poor seatbelt location, which allows the pilot to slide back on launch.
- full flying elevator - which is possible to stall if one holds the
stick full forward. (Low speed + big angular deflection)
- aft CG with a light pilot

Maybe if you only ever see launches of modern glass, or heavy two
seaters, with experienced winch drivers who know the different types,
you will not see this. I have yet to meet the perfect winch driver. At a
busy club, or where the winch driver is inexperienced with the glider
type or just inexperienced it is easy on a powerful winch to apply too
much initial power. And a little too fast on the throttle at the start
is very dangerous for a range of gliders.

Basically excessive acceleration can use most or all of the available
control authority. One more thing, like a gust / thermal / cable break
and you exceed the authority and have an accident. Best to avoid. Best
way to avoid is to train winch driving technique correctly. Denial gets
people hurt.

Power limited winches are no fun either...

Bruce

On Mar 18, 11:39*pm, wrote:
Karl,
I agree that vehicles are not prone to make gliders do space shuttle
like launches, due to all the inertia of that set up. Though thanks
for sharing that even there it is possible, not just with winches.

Bill,
The articles has an emphasis on heavy analysis that I am not going to
translate now, nor do I even have permission from Aerokurier to do
that. It's not my article. But trust me, it is very thorough and they
also calculated different scenarios. When there is and accident with
heavy damage or a fatality, the cause is investigated thoroughly and
that cause is listed in the tables in the article. There is no
statistical hoky poky. What for?

I would not presume to start with the assumption that the statistics
are somehow skewed, wrong, massaged or manipulated. If you want to
blast off on winch launch full power from zero on a 300HP (or more)
winch, vastly exceeding 1g and doing a space shuttle launch
simulation, that's your business. The analysis clearly shows that
doing that is 2.5-7 times more dangerous than accelerating under 1g
for the first few seconds. After that, the article says that tension
and airspeed can be higher than is the norm today with no problem,
provided a weak link is used that can handle it.

When I have some time, I'll post at least the accident tables and the
winch accident rate comparisons by winch type. It's easier to see what
I'm saying with the tables in front of you, rather than just my choppy
interpretation and translation of th text. It's very technical.

Speaking of experiences, flying all manner of gliders in Germany and
having participated and watched thousands of winch launches myself -
I've seen and participated in my fair share of space shuttle launches
"Kavalierstart", even with the stick full forward. Unlike Karl, I was
young and not so smart and I wasn't scared at all the few times it
happened to me and we criticized pilots when we saw it happen to them
(like they did it on purpose...free beer). Though I knew it wasn't
right and battled the nose down as soon as I could, there is nothing I
could do about it. That it was dangerous I knew, but I was thrilled to
be blasting off like that since it happened so fast and there was
nothing I could do about it anyway but push the stick full forward.

Simply put, if the rope tension in the first few seconds exceeds 1g
rope tension (not aircraft mass acceleration), the pilot is like a
passenger and can do nothing to correct the situation while the AOA is
pushed to the limit or much closer to the limit. Hopefully, the AOA is
not exceeded for any reason, that is the wing does not stall, because
if it does, the outcome is certain death within the first 7 seconds or
so.

Please understand that the 0.5g-1g tension is measured as rope tension
and not as felt inside the glider as mass acceleration. If you have a
strong head wind, it doesn't take much initial speed or mass
acceleration to create a space shuttle simulation (perhaps that's what
happened to Karl) as the maximum rope tension will occur at a much
slower drum (car) speed and set in much quicker than with no wind.
That is why tension controlled winches are safer as they easily
compensate for this - they deliver a certain rope tension and that's
it, regardless of wind and the drum speed will vary on it's own. It's
like the car rolling up hill - the tension will be the same, only
it'll roll slower up the hill and faster down or flat. RPM winches do
not compensate for any of this. On climb out it's fine, but not in the
first few seconds.

I didn't accuse anyone of "hoky-poky" and certainly not the publishers
of Aerokurier. I just said that statistics don't work very well for
accident analysis since the data set is too small to draw meaningful
conclusions. To the extent that you analyze anything with accident
statistics, it's pilot skill. That dominates the results not
equipment or procedures. Inconsistent and variable pilot skills
introduces way too much "noise" in the data. Statistics are simply
the wrong tool.

The right tool for this problem is engineering. Every single
parameter can be analyzed with measurement data. If excessive
acceleration does, in fact, reduce AOA margin, that would be very
straight forward to measure. Just install an AOA indicator,
commercial units are available, and simultaneously measure
acceleration with a tensiometer or with a simple video camera. Start
slow and work up. If the stall margin is getting smaller, you'll know
when to stop. Rock solid engineering data will also validate
mathematical models - or not.

I happen to think that the "Kavalierstart" you speak of is pure pilot
error. The pilot causes, or allows, the glider to rotate into a steep
climb before he has safe airspeed. Once over-rotated and stalled,
there is insufficient elevator authority to reverse the situation -
he's riding a kite not flying a glider. A good pilot can prevent this
at accelerations right up to the breaking strength of the weak link.
I know because I've done it.

Requite qualification: There MAY be a very small number of gliders
whose inertially induced pitch-up under hard acceleration exceeds the
elevator authority to prevent it. .

Bill,

Over accelerated "Kavalierstart" winch launches often have nothing to do
with 'pilot error'. If you are already holding the stick hard against the
forward stop to try and stop the glider over-rotating, what else can you
do, apart from pulling off? Even that would leave you in a very tricky
situation, with the glider very nose high and very close to the ground!

It is quite difficult to overpower a K21, a Grob G103 or a DG1000, mainly
because they are quite heavy and need a lot of horsepower to accelerate
them. Anything lighter than this, and fitted with a belly hook, is
potentially at risk with modern powerful winches, and I include K13
two-seaters in this. I would not recommend launching on a nose hook by the
way, as you will get a very poor launch and they normally don't have a
back release mechanism.

The problem can be avoided anyway by limiting the winch power or tension
setting to what the glider type can safely handle. The Skylaunch winch
does this very well, as long as the driver sets the throttle tabs
correctly.

Derek Copeland

At 15:18 19 March 2009, bildan wrote:

I didn't accuse anyone of "hoky-poky" and certainly not the
publishers
of Aerokurier. I just said that statistics don't work very well for
accident analysis since the data set is too small to draw meaningful
conclusions. To the extent that you analyze anything with accident
statistics, it's pilot skill. That dominates the results not
equipment or procedures. Inconsistent and variable pilot skills
introduces way too much "noise" in the data. Statistics are simply
the wrong tool.

The right tool for this problem is engineering. Every single
parameter can be analyzed with measurement data. If excessive
acceleration does, in fact, reduce AOA margin, that would be very
straight forward to measure. Just install an AOA indicator,
commercial units are available, and simultaneously measure
acceleration with a tensiometer or with a simple video camera. Start
slow and work up. If the stall margin is getting smaller, you'll know
when to stop. Rock solid engineering data will also validate
mathematical models - or not.

I happen to think that the "Kavalierstart" you speak of is pure pilot
error. The pilot causes, or allows, the glider to rotate into a steep
climb before he has safe airspeed. Once over-rotated and stalled,
there is insufficient elevator authority to reverse the situation -
he's riding a kite not flying a glider. A good pilot can prevent this
at accelerations right up to the breaking strength of the weak link.
I know because I've done it.

Requite qualification: There MAY be a very small number of gliders
whose inertially induced pitch-up under hard acceleration exceeds the
elevator authority to prevent it. .

Bill,
the statistics covers accidents and launches in the period from
2002-2008. Brought to an annualized figure that means there were an
average of 1,449,500 glider launches every year for the period. That
is, the analysis covers about 10,146,500 launches of all types in
Germany and the accidents that resulted from these. That's a pretty
big sample. The sample also covers 60 winches in France, listed
separately. (the only accidents in France involved electric winces)

In Germany, for the period (2002-2008), there was a yearly average of:
712,500 - Winch Launches
237,000 - Aero tows
500,000 - Motorglider take offs (fixed position motors)

The number of motorized glider take offs is not indicated, but the
number of their accidents is, which comes out to 5.3% of all glider
launch accidents involves this type of launch.

I don't think the sample is too small.

Your statement that one could simply install an AOA indicator to allow
the pilot to avoid exceeding the AOA misses the point entirely. The
point is that exceeding the AOA occurs because of what the winch does,
not the pilot - namely, excessive initial launch speed/tension/power
resulting in an uncontrollable excessive initial pitch up movement
that ends with a stall and flip into the ground at high speed within
seconds.

The only aircraft type pattern mentioned is that involving GROBs.
These types were never involved in flip ins during inititial launch.
Another anomaly is that in 45% of all accidents involving broken weak
links Grobs were involved - indicating the certified strength of the
weak links for Grobs are not strong enough.

I think the analysis is very thorough and makes strong
recommendations. What you are doing, strong initial acceleration
launches (which I presume exceed 1g rope tension), is what they are
saying should be avoided. Later in the launch phase, it's OK to
increase tension beyond 1g, but not in the initial danger zone.

On Mar 19, 12:26*pm, wrote:
Bill,
the statistics covers accidents and launches in the period from
2002-2008. Brought to an annualized figure that means there were an
average of 1,449,500 glider launches every year for the period. That
is, the analysis covers about 10,146,500 launches of all types in
Germany and the accidents that resulted from these. That's a pretty
big sample. The sample also covers 60 winches in France, listed
separately. (the only accidents in France involved electric winces)

In Germany, for the period (2002-2008), there was a yearly average of:
712,500 - Winch Launches
237,000 - Aero tows
500,000 - Motorglider take offs (fixed position motors)

The number of motorized glider take offs is not indicated, but the
number of their accidents is, which comes out to 5.3% of all glider
launch accidents involves this type of launch.

I don't think the sample is too small.

Your statement that one could simply install an AOA indicator to allow
the pilot to avoid exceeding the AOA misses the point entirely. The
point is that exceeding the AOA occurs because of what the winch does,
not the pilot - namely, excessive initial launch speed/tension/power
resulting in an uncontrollable excessive initial pitch up movement
that ends with a stall and flip into the ground at high speed within
seconds.

The only aircraft type pattern mentioned is that involving GROBs.
These types were never involved in flip ins during inititial launch.
Another anomaly is that in 45% of all accidents involving broken weak
links Grobs were involved - indicating the certified strength of the
weak links for Grobs are not strong enough.

I think the analysis is very thorough and makes strong
recommendations. What you are doing, strong initial acceleration
launches (which I presume exceed 1g rope tension), is what they are
saying should be avoided. Later in the launch phase, it's OK to
increase tension beyond 1g, but not in the initial danger zone.

I think understand, Tommyto is Derek Copeland using one of his
hundreds of aliases.

The statistics are interesting but prove little without actual
engineering measurements. What the statistics do suggest is that
pilot incompetence is sadly not rare and that someone should do some
actual certified engineering measurements.

To increase the weak link strength without engineering data to insure
that is in fact safe is grossly irresponsible - and illegal in every
country in the world except the UK. In any event, the final and only
authority on that subject is holder of the type certificate - Grob
itself. (US pilots note that weak link strengths are set under JAR-22
to a prescribed value with a +or- 10% tolerance as part of a types
airworthiness certificate - read your POH.)

To suggest that pilots are crashing due to the weak links being too
weak would be hilarious if not so tragic and I'm quite sure the LBA
and/or DAeC made no such suggestion.

Weak link failure accidents are 100% pilot error and 0% hardware
deficiency. Pilots must EXPECT wire or weak link failures and be
prepared to deal with them safely. Pilots who can't handle a launch
failure with big safety margins should be grounded for extensive
retraining.

I did NOT suggest using an AOA indicator to help the pilot avoid
"exceeding the AOA" [OF STALL?] although that is a great idea. What I
did suggest is using it as an engineering measurement tool to
determine if increased acceleration was causing an increase in AOA.

My measurements suggest the exact opposite - that increasing the
acceleration REDUCES the maximum AOA. Other measurements suggest that
even gliders with a strong inertial pitch up tendency will break their
weak link under strong acceleration long before running out of
elevator.

If you want statistics, they seem to show that over rotation leading
to stalls on the wire occur mainly on weak winches and auto tow -
always as the result of premature pilot induced pitch up - as in
counting 6 seconds and pulling up without consulting the ASI.

"I think understand, Tommyto is Derek Copeland using one of his
hundreds of aliases. "

Let us get at least this part straight - my name is Thomas Vallarino,
not Derek Copeland and I do not even know the gentleman.

"My measurements suggest the exact opposite - that increasing the
acceleration REDUCES the maximum AOA."

You can believe whatever you want to. You are free to disregard the
calculations in the analysis, the accident data and the experiences of
those here in this thread who describe specific cases of
uncontrollable pitch up movements, despite full forward stick - from
Karl's ASW-17 incident to the same experience of others including
myself with a variety of other models. It's not always the pilot's
fault when this happens as the pilot runs out of elevator authority.
The article calculated the pitch arm moment of many gliders and what
tension would be necessary to exceed the elevator authority at the
beginning of the launch. That's why they recommend a range of initial
rope tension of no more than 0.5-1.0g for singles and 0.7-1.0g for
doubles at the beginning of the launch.

Just think about it. Why does the tail go down into the ground at the
beginning of the launch, despite full forward stick? Because there is
little to no air moving over the elevator to overcome the pitch up
moment produced by the rope tension when the glider fist begins to
move forward at 0.1 seconds into the launch. This effect can continue
after the glider leaves the ground if the rope tension is too high.

If there is no stall and accident, then yes the airspeed will continue
to build up very quickly and so will elevator authority and the AOA
can be moved back to larger margins quickly by the pilot, so long as
tension doesn't continue to increase. However, there is a specific
rope tension for any airspeed within the first phase of the launch
(initial angle vector of rope VS horizontal), after which the maximum
AOA will be exceeded due to the pitch up moment and limited elevator
authority to prevent a stall.

In all my years, I have never even seen an accident on the winch,
where anything was damaged. Plenty of incidents, but everything
always turned out well. So what we're talking about here is to make it
even safer than it already is.
Thomas Vallarino
Manhattan Beach, California

On Mar 19, 3:14*pm, wrote:
"I think understand, Tommyto is Derek Copeland using one of his
hundreds of aliases. "

Let us get at least this part straight - my name is Thomas Vallarino,
not Derek Copeland and I do not even know the gentleman.

Well, that's incredibly good news for many reasons.

"My measurements suggest the exact opposite - that increasing the
acceleration REDUCES the maximum AOA."

You can believe whatever you want to.

Hard engineering data ALWAYS trumps calculations and predictions no
matter how logical they seem. I'll believe measured data instead of
calculated results any day.

"Just think about it. Why does the tail go down into the ground at the
beginning of the launch, despite full forward stick?"

And why does it stop there? It hit the ground, right? Now you have
hard acceleration and NO rotation. As the speed increases, the pilot
continues to hold full down elevator which increases in effectiveness
with the square of the airspeed.

As the glider leaves the ground, the inertial rotation will begin but
if the acceleration continues, the elevator effectiveness will also
continue to increase with the square of the airspeed. Rotation can't
happen instantly because the glider has mass and rotational inertia.
In fact, the pilot has to start backing off the down elevator to allow
the glider to rotate into the climb.

Now, contrast this with a slow acceleration. The glider staggers into
the air and the nose-up inertial couple starts the rotation but the
low speed and acceleration doesn't provide adequate control. The nose
rises as the pilot struggles to control it with inadequate
airspeed.....

I've collected stories on this type of accident for decades and they
ALWAYS happen with slow acceleration.