What's up with ATR 42/72 prop governor failures?

[Fordiesel69]

Been hitting the air crash investigations channels quite heavy lately and am quite confused by the number of accidents involving ATR 42/72 aircraft, and many other types of turboprop aircraft. All the crashes were related to the inability to feather the prop during an engine failure and another crash that there was a PEC (propeller electronic control) malfunction that did not give the crew good indication that the prop was fully feathered allowing the engine to undertorque but still be at full N1 indications on the cockpit instruments. And on another crash

Why is it that a design is allowed that will not allow the flight crew to have complete control over the prop pitch, and my second question is if you are going to make everything electronic, why is there not a failure mode to automatically feather the prop when the engine fails.

Here is another crash where a turbine blade failed (yes anything that spins that fast can fail) but the prop pin breaking on the GOV at the same time? I just cannot grasp this. Its like crappy automotive engineering has now found its way into avaiation.

Crashed PIA ATR 42 stalled and inverted after complex engine failure

Investigators have determined that the complex failure of a Pakistan International Airlines ATR 42-500's left-hand engine preceded a loss of control which developed into a stall, loss of altitude, and eventual fatal collision with terrain. At one point in the accident sequence the aircraft...

www.flightglobal.com

 

[Fordiesel69]

Stevo also had one that would not reach 2000rpm. Some kind of issues with prop gov....

 

[E365]

I flew the Saab 340 for many years. We were told that if you had an engine failure and the prop didn’t feather (or move closer to feather and “coarsen”) the plane will be impossible to fly.

I had a propeller speed sensor failure once. If I remember my systems correctly, during cruise the system maintained prop RPM with little fuel adjustments. So when the prop speed sensor failed intermittently (dip to zero for a second) the engine would add fuel to try to keep prop speed up. But then a second later the sensor would come back, and the system saw the prop was now too fast (because it added fuel unnecessarily), it reduced fuel. You end up with an out-of-control surging engine.

Thankfully this happened to other crews first and they published a memo on it. All that was needed was to “lockout” something called the Torque Motor which controlled secondary fuel functions like this fine-tuning of fuel flow at cruise speed. Problem solved - you just land with no reverse prop or “beta” mode while taxiing - Other things controlled by the torque motor. Previous crews understandably shut down the engine when it started surging before the issue was well known.

 

[Kuato]

Flew both the 340 and the ATRs. E365's description is a good one - in simple terms, shut down the automatic system that's causing the surge, the surge goes away, and you've got a good engine (but lose some capabilities on the ground).

Been awhile for the ATR, but IIRC there are 4 or 5 automatic safety systems that stop either an underspeed or overspeed condition. Without multiple failures, I don't recall a way for something to happen to make the plane unflyable; reduction of performance, yes. Maybe in the mountains with an unfeathered prop you wouldn't have enough climb performance, but I wasn't going to join that pay site to read the entire article so don't know where they were when it happened.

One thing I DO know for certain: pilots in that part of the world are put through mills (1) and don't have the stick and rudder skills that pilots in the US develop (2). So although they may be great with the automated systems, when things go bad they don't have the skills to actually fly the aircraft properly on one engine. FYI airline pilots in the US are required to have at least 1500 hours of flight time. Overseas, some regions allow you to be the First Officer with as little as 250(!) hours. So that A-350 from those regions could have somebody in the right seat with only enough hours to just have gotten their Commercial License in the US.



1) My instructor for the ATR flew it in Asia for 5 years. He related the story of something that he had witnessed, where there was a checkride for two pilots, with 12 observers in the back of the simulator. All 12 observers were also signed off as qualified - without actually flying the simulator. He also related that the instructional program for the pilots flying the ATR-72/600 for TransAsia ALWAYS had the instructor shutting down the SAME engine. So muscle memory took over in that accident about 6 years back.

2) Another aviation friend went to fly turboprops in India. He related the story that his First Officers were great at flying the airplane with automation, and were fantastic at programming the FMS(navigation) and autopilot systems. But he also said that in the first month there, he lost track of the number of times he had to take the controls from the First Officer during a visual and/or hand flown approach. When you are cleared for a visual approach in good weather, you generally click off the autopilot and hand fly the aircraft to landing. Often, the autopilot might not be able to be used safely (or quickly enough) to maneuver as required, so hand flying is necessary and the safest option. He said his First Officers' eyes would get big whenever they had to turn off the autopilot anywhere but on short final, and he had to give them a WHOLE lot of coaching for them to have a chance to make it safely.

 

[Kuato]

Sorry, typo, A340...

 

[mk378]

That is a pay site. The official report of the PIA crash https://www.caapakistan.com.pk/Upload/SIBReports/AAIB-386.pdf concludes these points:

The overspeed governor had been disassembled then improperly re-assembled with an internal pin out of position, which broke when the case was improperly forced together. The airline's records indicated that it had only ever been serviced at the manufacturer's authorized facility, not in the airline's shop, but the report decided that was likely not true. The internal condition was not immediately apparent because the OSG only acts during abnormal conditions, and the problem with the pin did not immediately make the unit non-functional. It did create stress on other parts which resulted in delayed failure during the crash flight.

The engine's power turbine blades were of the original design which in experience were found to have an unexpectedly short life. The manufacturer re-designed the blades and issued a sort of soft recall where engines were to be upgraded to new blades but it was based on a number of hours, not an immediate requirement. The engine that failed had been in the airline's shop with over the allowed number of hours on its original blades, thus a blade upgrade was required-- but it was not done. Sometime after that, on a flight before the crash flight, one of the blades broke but the situation was not detected. Running with part of a blade missing, the turbine was out of balance causing a turbine shaft bearing to fail and ultimately the engine stopped during the crash flight. The same oil system which lubricates the engine bearings also serves as hydraulic fluid for the propeller pitch control system. Debris from the failing bearing may have resulted in a pitch control malfunction. Additionally there was much other debris in the pitch control hydraulics which was assumed to have gotten there by having the parts improperly stored while off of the engine.

It was not a failure of electronics or automation, but a problem with the basic hydraulics that move the propeller blades. Regardless of the control scheme offered to the pilots, it was not going to move to the feather position.

As any official report will, it soft-pedals the idea that the pilots were not adequately trained for emergency operations, but that was clearly an issue here. Many pilot errors were committed including a major one in one engine out flight-- not bringing the good engine up to 100% power and leaving it there. The report concluded that with the un-feathered propeller in its worst-case position, the aircraft would still be flyable but it would not be able to maintain altitude, sinking at about 600 feet per minute.

 

[Kuato]

Great explanation mk378. I now recall the 600fpm descent with unfeathered prop number.

So as in any accident it was a series, or chain of events that led to the crash. First, manufacturer issuing soft rather than required fix. Then airline multiple substandard mx practices and finally the weak flying skills of the pilots in the crash. Remove one of these and the plane would likely still be in one piece.

 

[wpod]

Piloting a plane successfully is quite a skill.

 

[ms21043]

.

 

[ms21043]

Nothing wrong with propeller systems if properly maintained. In the report above I wouldn't fault the pilots at all. Control of the propeller was lost and as soon as blade angle went into ground range there isn't anything that could be done.

 

[Cujet]

Nothing wrong with propeller systems if properly maintained. In the report above I wouldn't fault the pilots at all. Control of the propeller was lost and as soon as blade angle went into ground range there isn't anything that could be done.

I think I understand your point, that a well designed, constructed and properly maintained propeller system is very reliable and safe. Especially when driven by a turbine engine.

However, it's good to know the move to jets was a huge improvement in overall operational capability. The reason was at least in part due to the elimination of the propeller itself. We could talk about mechanical complexity and the incredible reliance on oil pressure to adjust blade angle, the yaw created by a windmilling prop and so on, but those are addressable, both mechanically and through pilot training. One major issue is really that prop planes struggle in the poor weather conditions/altitudes they are often forced to fly in. Not only do the props ice up (despite the heated boots and various fluids used) but ice related loss of thrust can be made worse by wing and tail icing leading to loss of lift and thrust (compounding problems). Jets can often quickly climb above all that, with heated cowls, leading edges and so on.

We operate jets and turboprops. It's no surprise that in our Pilatus PC12 we sometimes get stuck at 20,000 feet and are unable to climb higher due to loss of thrust and lift. Leaving the plane in the thick of the weather and unable to climb above it. The fact that it can operate there does not mean it should be done, and does not mean it's safe. Descending against one's wishes down through the clouds as a vibrating block of ice is never a good operational method.

Here is a good example of prop icing:

 

[ms21043]

Yes, I was alluding to the subject of this thread, ATR 42/72 prop controllers. The prop systems on these airplanes are just about bullet proof. When there are problems, it's not surprising to find factors such as some unknown party monkeying around inside an OS gov as in that PIA accident report.
It is true that turboprops get to bump around in the weather below 250, but sending the right airplane for the job always comes down to money, just the nature of the beast I suppose.