ATR 72 Icing & Loss of Control - Recovered
 

I cannot find any thread started at the time (almost exactly a year ago) of a serious incident involving an ATR 72, for which an AAIB report has just been released. .

the Conclusion was;

I think it's worth a read, not so much for the technical issues but for the Human Factors elements of why the crew did what they did.

If it's been done already in PPRuNe, apologies for not finding it, please delete...

PS I found a mention in PPRuNe, but buried in Airlines, Airports & Routes

FYI:

https://aviation-safety.net/database...?id=19941031-1

The ATR does not seem to like ice very much:

Incident: Jettime AT72 near Bergen on Nov 14th 2016, drops left then right wing in icing conditions

NNUTS. In anything other than a High Speed Jet Transport (or pehaps VERY high performance Turboprop), out climbing ice is a fool's decision. Always was and always will be.

With the potential of degraded performance, am I the only one that thinks they might have been better off actually flying the aircraft, or is this something peculiar to an ATR?

Bank angle limitations in severe icing conditions just released.

No, the QRH drill for severe icing - reproduced in the AAIB report - tells the crew to disconnect the AP (amongst other actions).

If I recall correctly the American Eagle ATR was in a holding pattern.

Flight Directors are an aid. To follow them blindly during inappropriate conditions is only asking for trouble. A universal problem it seems

Flying slowly at a high angle of attack in a holding pattern with a cold soaked aircraft in icing conditions can cause grief.

Although most of my turborop experience was not in the area where I live, the area where I live is a well-known location for severe icing conditions in freezing rain.

A retired test pilot that I know did Twin Otter in flight icing tests. His advice was that the worst of the icing was actually in quite narrow bands several thousand feet thick.

Of course, each situation is different but in a situation where one cannot descend to air warm enough to melt ice or clear of both cloud and icing(ie no freezing rain), I would consider climbing first. In the case of a warm front, you may get to warmer air.

With regard to climb/descend options......When in icing conditions, you will always have the capability to descend but the climb option may no longer exist if you chose the descend option first.

At night, you will never see blue patches and if there is a higher cloud layer, you won't see blue in the day either.

As I said, this is not based on significant personal experience in such conditions and I'm sure others could give theirs based on more personal experience.

I experienced severe icing once, in a Shorts 360 back in '86. The conditions had brought another one down some 48 hours previously, luckily with no fatalities. Airborne some 30 minutes, diverted, never reduced below climb power, and had the throttles firewalled coming over the hedge. Never try to climb above it, unless you are CERTAIN that you can clear the conditions.

Don't get slow in icing, if you do get slow don't turn, if you elect to turn use half bank until your not slow anymore.

The icing layer is rarely more than 3000 feet thick.

new OEB says HiBank only if IAS > icing bug + 30kts (30 !) in severe icing condition, otherwise LoBank - AND consider to descend to maintain the required speed.

Makes it even more likely for something to happen, Captain more or less single crew and watching what the other guy is doing, agree he should be used to being exposed those kind of situations but no one os infalliable.

I used to fly a type similar to the ATR, underpowered and a slow climber, even moderate collections of ice would cause large speed reductions and even require level offs to maintain min ice speed.

Its not an inherant problem with the ATR, if you fly a low powered aircraft with poor climb performance you are going to spend lots of time climbing through icing levels and if you get slow or don't follow procedures any aircraft in that catagory will bite you.

Not only against you but this whole is hard to believe.

Actually they did a test flight. They managed it to figure out the maximum level in ice =FL130. They stalled the aircraft at KIAS 165!

Manufacturer say >red bug speed< +10- what means 175 KIAS is recommended. That is 10 knots margin above actual stall speed.
The report cite a server ice CL- what demand minimum ice speed +10.. Is that to understand as (red bug speed/165 +10) +10 = 185 KIAS ?!
And dear Domi, never ever accept any >high bank< (what a term). Many flight directors have a small bug, omitting half bank in case pilot selecting >direct to<. You could choose pilot off FD or use a bank mode.

Generally is a lack in understanding the stall subject.
Lift oppose weight. But that is valid only under the premise, continuous unaccelerated flight.
The weight is mass multiplied by G-load. A = Mass x G-load
If you fly, like in this case, exactly at stall speed, any increase of G-load will immediate stall the wing. For that in ice, always fly above stall speed for ice, avoid any turns, avoid accelerations.Those is part of aviate, what comes before CL.
If a stall occur, reduce the actual stall speed. Push gently the yoke forward. That way you fly a ballistic curve, that reducing the G-load, that reduce the stall speed!
If a spontaneous roll occur, avoid big aileron inputs. That will increase the average angle of attack, of the aileron of outer/upper wing. That part will likely stall, aircraft bank to opposite side. (In our case reached 73°).
As you loose altitude/dive; airspeed will increase, so normal airflow should reestablish.
During recovering; likely to divert from given clearance and stall is to report as well as severe ice, if you anticipate over-stress aircraft need to report . But relevant call is PAN. Mayday is movie style as in Sullivan Movie, irl commander did no std distress call.

In conclusion, I would say, this report is to revoke.

It's definitely to be understood as 185.
Your minimum icing speed, half bank protected, is increased by ten knots, if you want to fly full bank or NAV you need 10 knots on top of the new red bug.

The ATR is a piece of crap, happy not flying it anymore...

If I remember right on the EMB-120 we used 210Kts min and autopilot off strongly recommended in icing. The autopilot off was to prevent control forces masking. We had similar loss of controls leading up to the min speed increase.

You forgot about the fuel problem. If you stuck in ice, your cruising level is decreed, like in the here mentioned case, from FL190 down to FL130, the fuelplanning is not correct anymore. If not prepared for this you could easily land below minimum diversion fuel. 2 other reasons why fuel planning in ice is delicate. Performance in ice is less than in clean weather, less IAS with same powersetting, more FF for same speed. Early descend is needed and higher speeds/pwr-setting in descend needed. Maybe idle descent not possible because bleed air is not enough for cabin and deicing.

But all airplanes have limitations. The problem starts if you fly your current type like another type. You have to consider actual stall speed in ice could be high as 165 KIAS. (BTW, you can stall any AC at any speed, it depends on G-load). I've seen blackbox printouts from actual stalls of EMB-120 in ice that happens at KIAS 155. In C421 I experienced few at KIAS130. Laminar airfoil have tendency to stall at much higher airspeeeds in ice than clean.
Secondly manufactor recommend Flaps 15 as seen in CL, what indicated Fuselage create to much drag when maintaining higher AoA. (I remember the same for SD3-60, in case of engine failure.) Despite of CL airframe (fuselage) of ATR is not antiiced, or?

I wouldn't say the ATR's crap.....But no doubt it's got more than its fair share of gotchas.

I had about 1,400 hours on both 42 and 72 spread over about 15 years, on and off, and it was never amongst my favourites.

I never experienced the iceing incident we're discussing here, but I do remember during the TR being warned about it. However, if you've never experienced it and, like me, about the best you can remember how to deal with it is the page number in the manual, then I can have some sympathies for the crew here when you're faced with a situation you've never experienced before and you do what seems ' appropriate ' while your colleague is desperately trying to find the page in the manual.

And just a thought....If airframers can create and install the automatics for detecting problems such as this one and warn you on the various displays, why can't they extend that functionality to either display the instructions written in the Ops Manual or to automatically initiate the required actions and then inform the crew what and why the machine's doing what its doing through the displays with a simple ' accept ' or ' reject ' for the crew.

Within the grand scheme of things, that wouldn't be an impossibly expensive fit into new machines and , depending on the actual age / spec of the machine involved, perhaps also not an impossibly expensive upgrade to older machines.