Accidents resulting in aircraft overturning

This forum is for safety-related discussions. Be safe out there!

Moderator: drseti

Post Reply
User avatar
dstclair
Posts: 1092
Joined: Thu Mar 06, 2008 11:23 am
Location: Allen, TX

Accidents resulting in aircraft overturning

Post by dstclair »

I asserted previously that a high-wing was more likely to overturn than a low-wing in similar circumstances due to the higher center of gravity. I had a little time to test my theory so I pulled real data from the NTSB database. My query was restricted to Part 91 operations, substantial damage, reciprocating engine, not rotary or gyro, and factual report containing the words "flip" or "invert". This returned 84 records. In a spot check of 10 records, 8 of them were accidents that did indeed result in the aircraft turning over while 2 mentioned the aircraft nearly flipped over. The nearly flipped over were low-wings (Cirrus). Since this is not a scientific study, I opted to decide that the spot check was sufficient to validate the query and included all the data.

Results:

* 69 of the 84 accidents were high-wing aircraft: 83%
* 15 of the 84 accidents were low-wing aircraft: 17%
* 13 of 84 accidents included one or more fatalities: 15%

The unscientific data implies a high-wing aircraft is 5 times more likely to overturn than a low-wing aircraft.
dave
Merlinspop
Posts: 999
Joined: Mon Apr 08, 2013 2:48 pm
Location: WV Eastern Panhandle

Re: Accidents resulting in aircraft overturning

Post by Merlinspop »

How many of those high-wings were taildraggers that the pilot hit the brakes too hard and it stood on it's nose and then on over? Of those, how many were Legend Cubs? Their CoG is pretty far forward and they have effective brakes.
- Bruce
User avatar
drseti
Posts: 7227
Joined: Sat Nov 28, 2009 6:42 pm
Location: Lock Haven PA
Contact:

Re: Accidents resulting in aircraft overturning

Post by drseti »

dstclair wrote:The unscientific data implies a high-wing aircraft is 5 times more likely to overturn than a low-wing aircraft.
In the scientific community, Dave, we refer to what you just did as a "pilot study." (The name has nothing whatever to do with the guy who drives the plane.) I would say, from your results, that you have demonstrated the validity of your methodology sufficiently to justify advancing to the next stage, a full controlled study. Any graduate students here, who want to make this the basis for a thesis?
The opinions posted are those of one CFI, and do not necessarily represent the FAA or its lawyers.
Prof H Paul Shuch
PhD CFII DPE LSRM-A/GL/WS/PPC iRMT
AvSport LLC, KLHV
[email protected]
AvSport.org
facebook.com/SportFlying
SportPilotExaminer.US
MovingOn
Posts: 632
Joined: Fri Dec 20, 2013 5:34 pm

Re: Accidents resulting in aircraft overturning

Post by MovingOn »

.......
Last edited by MovingOn on Thu Aug 14, 2014 12:02 am, edited 1 time in total.
User avatar
dstclair
Posts: 1092
Joined: Thu Mar 06, 2008 11:23 am
Location: Allen, TX

Re: Accidents resulting in aircraft overturning

Post by dstclair »

All good questions and I encourage you to do similar queries.

Excluding taildraggers really doesn't really have statistic significance. Going with a null hypothesis that there is no relationship yields a standard deviation of 4. This means that we would expect between 38 and 46 overturning events for each class. Exceeding one standard deviation invalidates the null hypothesis and the high-wing number exceeded 6 standard deviations. The data then concludes that a high-wing does overturn more often in accidents involving substantial damage.

I eyeballed approximate 20 taildraggers with 2 being RVs. This changes the high-wing number to 51/64 and low-wing to 13/64. The standard deviation is 5 with an average of 32 and the accepted range 27-37. High-wing numbers are still well outside a standard deviation so the conclusion is the same although the number of standard deviations is reduced to 4.

I don't believe the total number of aircraft type is relevant although it would be interesting to know the total number of high-wing accidents and total number low-wing accidents. I leave this as an exercise for the reader :)

Also, I'm absolutely sure there are a lot more accidents involving overturning but it would take a manual review of the unparamaterized data. Would make for a good senior project in college.
dave
MovingOn
Posts: 632
Joined: Fri Dec 20, 2013 5:34 pm

Re: Accidents resulting in aircraft overturning

Post by MovingOn »

.......
Last edited by MovingOn on Thu Aug 14, 2014 12:01 am, edited 1 time in total.
User avatar
MrMorden
Posts: 2184
Joined: Fri Aug 17, 2012 7:28 am
Location: Athens, GA

Re: Accidents resulting in aircraft overturning

Post by MrMorden »

dstclair wrote:
The unscientific data implies a high-wing aircraft is 5 times more likely to overturn than a low-wing aircraft.
However, scientific data (physics) suggests in cases of inverted aircraft occupants of high wings are much more likely to be able to egress quickly... :mrgreen:
Andy Walker
Athens, GA
Sport Pilot ASEL, LSRI
2007 Flight Design CTSW E-LSA
User avatar
dstclair
Posts: 1092
Joined: Thu Mar 06, 2008 11:23 am
Location: Allen, TX

Re: Accidents resulting in aircraft overturning

Post by dstclair »

MrMorden wrote:
dstclair wrote:
The unscientific data implies a high-wing aircraft is 5 times more likely to overturn than a low-wing aircraft.
However, scientific data (physics) suggests in cases of inverted aircraft occupants of high wings are much more likely to be able to egress quickly... :mrgreen:
No argument here -- well maybe... Perhaps not :)

Interesting that fatalities are approximately the same: 20% for low-wing and 17% for high-wing. And, no, I didn't try to analyze what would have happened if egress was faster or what the cause of death was.
dave
User avatar
dstclair
Posts: 1092
Joined: Thu Mar 06, 2008 11:23 am
Location: Allen, TX

Re: Accidents resulting in aircraft overturning

Post by dstclair »

MovingOn wrote:It's an interesting subject, but not one I am going to research. My questions were simply suggesting that if 80 percent of the aircraft flying were high-wing (example), then maybe you would expect 80 percent of the accidents to be high-wing. Then, you might also expect 80 percent of the flip-over's to be high-wing. This might indicate, no significant difference in flip-overs high-wing vs low-wing. I'm not suggesting that's the case, I don't know. Flip-overs might be more dependent on where they landed.
It looks like the piston population is pretty close to 50/50. According to the highly reliable source of Wikipedia there are:

* Cessna - 200K
* Piper - 150K (22K are tripacers and cubs)
* Misc (Beechcraft, Grumman, Mooney, Cirrus, Diamond and RV) - 55K

Assuming all Cessna's are high-wings and the rest are low-wings (except those noted) yields 183K low-wings and 222K high-wings for 45/55 split.
dave
Post Reply