Tailplane icing in aircraft like the Bombardier Q400

The safety statistics point to the fact that for every crash there will be many more aviation incidents that thankfully go unreported by the mainstream media. The Aviation Herald web site at http://avherald.com/ is one source I go to to keep an eye on these (if there are better public sources that give global coverage I would be interested to know).

At the time of writing, of the 14 occurances for Thursday, 12 and Friday, 13 February 2009 the Herald is listing three involving DH8D aircraft (as well as the Colgan crash, a Malev engine failure and Austrian Arrows gear retraction problem) and one other for Colgan, a runway excursion at Augusta the previous day.

Sadly sometimes they make it all the way through James Reason's Swiss cheese. :-(

Wow.

If i'm right in thinking that this NASA research was from 1998, then it seems to describe almost exactly what must have happened to flight 3407.

Tailplane icing is a problem with turbo-powered aircraft. (NASA) has done extensive studies and testing on tailplain icing. It is a killer and unless the pilot is acutely aware that he has tailplain icing, he might react incorrectly. The use of an auto pilot is acceptable in light and moderate icing condition according to FAA procedures. However the (NTSB) strongly disagrees with FAA. The (NTSB) believes that aircraft,should and must not use the auto pilot in any icing conditions. The (NTSB) believes that the crew will not be able to distinguish between wing icing or tailplain icing with the aircraft on autopilot. The emergency correction for these types of icing are unfortunately the opposite control response to recover the aircraft and prevent a crash.

see (NASA) tape from 1998.

http://video.google.com/videoplay?docid=2238323060735779946&hl=en

That tailplane icing video was for a different style of airplane. It affects traditional tail airplanes not T-tail like the Dash 8. The reason Bombardier actually raised the horizontal stab out an away from the main wing and prop shadow. The dash 8 does not have this problem, especially not the Q400 with hydraulically actuated elevators.

Now for the ugly truth, it was pilot error, arrogance, inexperience in icing conditions.

First, the Buffalo accident bears none of the classical ice contaminated tailplane stall (ICTS) characteristics. The flaps were not approaching the "full" position, and the aircraft speed, so far as has been published, was well below the limit speed for the flap selection. The activation of the shaker and pusher indicates that a genuine main wing stall was encountered (or at least the air data and angle of atack probes thought one was). The main wing stall does not generally coexist with ICTS; the former requires a slow speed, the latter tends to favor higher speed. Further, the impact attitude does not resemble the ICTS signature, as it appears to have been flat rather than a straight nose down impact.

However, none of this evidence rules out the crew mistaking a shaker/pusher for an ICTS event, particularly if they had not been well educated in ICTS characteristics.

I would like to point out that the T-tail is no guarantee against ICTS. In fact, there are a number of ICTS events recorded involving the MD-80. Fortunately, landing flaps on a jet are usually selected farther from the runway than manu turboprops, and thus more room is available for recovery.

Within a 120 event dataset of air carrier icing events dating back to 1940, the average total flight time for the pilot in command involved was approximately 7300 hours. Within the set of all US multi engined turboprop icing accidents from 1978 forward (24 events), the median flight time for the PIC in an air carrier event is approximately 7000 hours. For the same dataset, the median time for the PIC of a non-air carrier multi-engined turboprop icing accident (14 events) is approximately 5700 hours. I have published both numbers in two AIAA papers, in 2002 and 2006, respectively.

As can be seen, the inexperience argument just won't wash. That has long been used to marginalize these cases, to the detriment of all.

Thanks Steve, very interesting. I don't suppose your papers are on the web at all. If by any chance they are perhaps you could post the links.

Kieran,

Both are located on the Skybrary bookshelf:

http://www.skybrary.aero/bookshelf/books/531.pdf

and

http://www.skybrary.aero/bookshelf/books/532.pdf

I retained the copyright for both at the time of publication, so you are not in violation of AIAA property rights when downloading them.

You can also view a powerpoint of an updated version of this analysis at

http://www.sae.org/events/icing/presentations/2007s30green.pdf

This was done at the 2007 SAE conference in Seville. The web version has some corruption of a couple of slides; still usable but mildly irritating. This covers the same material, with no changes to conclusions. However the data is a bit more detailed.

What impact would what appears to be a very short wing cord and tail plane cord have on tail plane stall in icing conditions? Does wing design for the absolute best aerodynamic effeciency and fuel economy increase the resk of tail plane stall in icing conditions? Food for thought.