UH-1H missing from Coffs Harbour 6th Sept 2019
Originally Posted by [email protected]
The age of the helicopter might be a factor in turning mast bumping into rotor separation.
It it would explain the sudden disappearance from radar.
Impossible I would have thought.
I would imagine that a tumbling fuselage will induce all manner of pressure errors and the VSI (and IAS etc) data will be invalid.
the vertical rates (those are barometric VSI derived, not GPS) are not indicative of a free-fall event
Minus the Rotor head and in a tumbling motion that is about the order of magnitude which I would expect to see.
As some already wrote before: teetering head in up to 70kts of gusts at night over water in bad weather with questionable avionics suit for that purpose and unclear NVFR/IFR experience of the pilot.
What could possibly have gone wrong here?
Very tragic. But for sure I don't expect any novel cause of an accident in this particular case.
As some already wrote before: teetering head in up to 70kts of gusts at night over water in bad weather with questionable avionics suit for that purpose and unclear NVFR/IFR experience of the pilot.
What could possibly have gone wrong here?
Very tragic. But for sure I don't expect any novel cause of an accident in this particular case.
Physicus - your ADSB data doesn't seem accurate
In several places the Alt reading remains the same for 3 or 4 readings in a row whilst the VS is off the clock - they can't both be right so I wouldn't read too much into them other than a 9000'/min RoD is not indicative of a controlled situation, as others have said
UTC Time +/- 10s: 2019-09-06 08:14:17: TRK: 48.0 GSP: 127.0 ALT: 1450.0 V/S: -9024.0
UTC Time +/- 10s: 2019-09-06 08:14:07: TRK: 48.0 GSP: 127.0 ALT: 1450.0 V/S: -9024.0
UTC Time +/- 10s: 2019-09-06 08:13:57: TRK: 48.0 GSP: 127.0 ALT: 1450.0 V/S: -9024.0
UTC Time +/- 10s: 2019-09-06 08:13:47: TRK: 48.0 GSP: 127.0 ALT: 1450.0 V/S: -9024.0
UTC Time +/- 10s: 2019-09-06 08:13:37: TRK: 121.0 GSP: 152.0 ALT: 2200.0 V/S: -8768.0
UTC Time +/- 10s: 2019-09-06 08:13:27: TRK: 121.0 GSP: 152.0 ALT: 2200.0 V/S: -8768.0
UTC Time +/- 10s: 2019-09-06 08:13:17: TRK: 121.0 GSP: 152.0 ALT: 2200.0 V/S: -8768.0
UTC Time +/- 10s: 2019-09-06 08:14:07: TRK: 48.0 GSP: 127.0 ALT: 1450.0 V/S: -9024.0
UTC Time +/- 10s: 2019-09-06 08:13:57: TRK: 48.0 GSP: 127.0 ALT: 1450.0 V/S: -9024.0
UTC Time +/- 10s: 2019-09-06 08:13:47: TRK: 48.0 GSP: 127.0 ALT: 1450.0 V/S: -9024.0
UTC Time +/- 10s: 2019-09-06 08:13:37: TRK: 121.0 GSP: 152.0 ALT: 2200.0 V/S: -8768.0
UTC Time +/- 10s: 2019-09-06 08:13:27: TRK: 121.0 GSP: 152.0 ALT: 2200.0 V/S: -8768.0
UTC Time +/- 10s: 2019-09-06 08:13:17: TRK: 121.0 GSP: 152.0 ALT: 2200.0 V/S: -8768.0
From Physicus' post with the data:
(updated ADS-B data with 10s interval data. A note on that: not every interval provides a full complement of the data. The only reliable indicator that data is fresh is if it has changed from the last timestamp. Keep that in mind when inferring conclusions from this.)
Robbo Jock - good point but that means very little of that information is reliable and he may as well have not posted it.
I flew in that area for many years and when it is dark it is very dark. There is also plenty of high ground along the coast which produces a lot of turbulence. I would not want to be flying there in that weather at night in a poorly equipped helicopter with no IFR experience and (if the rumours are correct) very little night experience either.
If you sieve out the data points where nothing changes you can certainly see it was a bumpy night and there was very little straight and level flight going on - he seems to have been trying to maintain about 3500' and a SSWesterly heading but there are some significant variations in height from a fraction under 4000' down to 2400' and changes in vertical speed of almost +/- 2000'/min at times.
The last few data points though are significantly different.
The last few data points though are significantly different.
There always are two camps regarding tracking data release and accidents. I'm in the "some data is better than no data" camp - it provides a reference frame going beyond the obvious "it dropped off the radar". As [email protected] points out, the data does allow for valid statements to be made about the flying conditions prior to the incident. But gulliBell and 212man also have valid points that much of the last seconds of the data must be viewed cautiously due to the nature of how the data is collected. Static and pitot systems may not have been aligned for the designed air flow for correct measurements. The ground speed however is inferred from the GPS ground positions (in the on board GPS, not from the broadcast ground positions in post processing, it is an ADS-B field). It is clearly evident from the plots below that the aircraft turned with the wind, resulting in a corresponding increase in ground speed, which aligns with the ground track data.
Most importantly, it is evident from the pre-incident data points that both track hold and altitude hold were significantly perturbed, either due to the conditions encountered, or a lack of experience/workload on the pilot in dark conditions, resulting in significant departures in altitude and a steadily decreasing ground speed. Whether that was due to the aircraft penetrating increasing headwind and turbulence, or because of handling problems, is not possible to infer from this data. In the vertical speed plot, I derived the vertical speed based on the updated altitudes and they are in very good agreement with the ADS-B reported V/S. I've updated the plot with the last datapoint (not included above), showing the last altitude inferred V/S was -14400 fpm. This amounts to about 80m/s, which is approximately the terminal velocity for a body with a drag coefficient of 0.25 (see https://dspace-erf.nlr.nl/xmlui/bits...pdf?sequence=1). The sharp left-hand turn at 8:13:07, along with this new V/S datapoint, would lend evidence to the possibility of a main rotor separation. The sudden loss of main rotor torque would lead to a left yaw moment caused by the tail rotor on a counter clockwise rotating main rotor.
Note that because these data are derived from a physical process, prior data points are causal to the next points, therefore connecting them with smoothed lines is not inherently wrong is it might be when connecting randomly moving parameters such as stock prices. But: do not use the smoothed lines inbetween data points for interpolation (as is evident between 7:55 and 7:58)!
Most importantly, it is evident from the pre-incident data points that both track hold and altitude hold were significantly perturbed, either due to the conditions encountered, or a lack of experience/workload on the pilot in dark conditions, resulting in significant departures in altitude and a steadily decreasing ground speed. Whether that was due to the aircraft penetrating increasing headwind and turbulence, or because of handling problems, is not possible to infer from this data. In the vertical speed plot, I derived the vertical speed based on the updated altitudes and they are in very good agreement with the ADS-B reported V/S. I've updated the plot with the last datapoint (not included above), showing the last altitude inferred V/S was -14400 fpm. This amounts to about 80m/s, which is approximately the terminal velocity for a body with a drag coefficient of 0.25 (see https://dspace-erf.nlr.nl/xmlui/bits...pdf?sequence=1). The sharp left-hand turn at 8:13:07, along with this new V/S datapoint, would lend evidence to the possibility of a main rotor separation. The sudden loss of main rotor torque would lead to a left yaw moment caused by the tail rotor on a counter clockwise rotating main rotor.
Note that because these data are derived from a physical process, prior data points are causal to the next points, therefore connecting them with smoothed lines is not inherently wrong is it might be when connecting randomly moving parameters such as stock prices. But: do not use the smoothed lines inbetween data points for interpolation (as is evident between 7:55 and 7:58)!
Last edited by physicus; 12th Sep 2019 at 21:11.
would lend evidence to the possibility of a main rotor separation. The sudden loss of main rotor torque would lead to a left yaw moment caused by the tail rotor on a counter clockwise rotating main rotor
I was addressing a main rotor separation mostly in the context of what others have suggested, I personally think that is an unlikely scenario. But other than being a professional physicist, a greenhorn helicopter pilot, but seasoned fixed wing pilot having flown many different types from small to heavy (and corresponding professional level human factors involvement), I have little experience to contribute other than the physics behind this, and the (likely) human factors involved. The V/S rate deduced from the data points would indicate that main rotor separation remains a possibility though, but there are many ways to make an intact helicopter descend at a corresponding V/S in a more or less controlled state of flight. I have no idea what the site evidence says other than a news article stating the main body of the helicopter was found floating with the tail sticking out of the water.
Are there any non-Robinson aircraft included in your list of main rotor separations that resulted in sheared off tails?
Are there any non-Robinson aircraft included in your list of main rotor separations that resulted in sheared off tails?
Last edited by physicus; 15th Sep 2019 at 12:22.
18 Aug 1981, a Huey had a separation, the blade came through the cockpit and bisected the left-seater, took off the left side fuel tanks (it was a Bravo) and then removed the tail boom. Freefall from 1500'. I had been the last person to fly this bird and survive. And no, I didn't break it - the T/R pitch change cable came off its pulley, wrapped around the driveshaft and put a ridiculous amount of pitch onto the T/R. It flapped and struck the vertical pylon, one blade came off, then the whole gearbox and remaining blade separated.
The cg change from losing 30kg from 30 feet back and the rapid yaw/roll caused the mast bump. Three fatalities.
The cg change from losing 30kg from 30 feet back and the rapid yaw/roll caused the mast bump. Three fatalities.
One reason I supported the MR separation idea was this paragraph from the linked newspaper report in post #19
it seems strange that they use the phrase main fuselage without commenting on the main rotor when they have specifically mentioned the tail rotor.
A TR assembly won't float by itself but could if still attached to the tail boom - if the tail boom has detached from the main fuselage and there is no mention of the MR then it supports the idea of MR separation.
I still think that a combination of the poor references and the hideous turbulence are a perfect recipe for overcontrolling and creating a negative G situation that is so bad for teetering head helicopters.
Just before the data drops off a cliff there is a rapid climb which could result in a 'pushover' at the top leading to the aforementioned mast bumping.
A major air and sea search operation spotted the tail rotor of the aircraft before 9am Saturday and the main airframe was sighted 30 minutes later, the Australian Maritime Safety Authority (AMSA) said.
A TR assembly won't float by itself but could if still attached to the tail boom - if the tail boom has detached from the main fuselage and there is no mention of the MR then it supports the idea of MR separation.
I still think that a combination of the poor references and the hideous turbulence are a perfect recipe for overcontrolling and creating a negative G situation that is so bad for teetering head helicopters.
Just before the data drops off a cliff there is a rapid climb which could result in a 'pushover' at the top leading to the aforementioned mast bumping.
Whatever theories might prevail until the wreckage is found, one thing is virtually certain. If the pilot had landed before last light instead of pressing on into an environment where he wasn't qualified to operate, in an aircraft that wasn't appropriately equipped to operate, we wouldn't be having this discussion.
Join Date: Jul 2008
Location: Australia
Posts: 372
Likes: 0
Received 0 Likes
on
0 Posts
MRB separation history
I have waited to respond to this thread until I saw some suggestions of MRB separation.
I am a specialist in adhesive bond failure forensics, having worked in the past as a scientist for the ADF. Prior to 2007 I was involved in a collaborative investigation with the NZ DTA into a severe cracking problem with one RNZAF UH1H MRB. I sincerely hope that this is not the case with this event.
The primary cause of the RNZAF occurrence was disbonding of the adhesive layer between the blade upper surface and the steel grip pad where the blade attachment fittings clamped onto the blade adjacent to the attachment pin. There were significant concerns in relation to how the adhesive bond design concept was developed. The grip pad is approximately 12mm thick and the end terminates at sharp angle of approximately 45 degrees. There is a common perception that adhesive bonds distribute load over the entire adhesive layer, with a uniform shear stress spread over the joint. Of the order of 76% of bonded structures in the US were found in 2004 to be designed on this basis. This has been known since 1938 to be incorrect. In an adhesive bond the shear stress peaks at the ends and all of the load is transferred just at the ends of the joint with no load whatsoever being transmitted in the centre of the joint. The larger the stiffness imbalance between the adherents, the higher the4 shear stress peak at the end of the joint. In the case of the grip pad bond, there is a very large and rapid increase in thickness and a large difference in elastic modulus (steel vs aluminium) meaning that the shear stresses at the end of the joint are very high.
In the case of the RNZAF blade, post event investigation showed that there was clear evidence of adhesive bond fatigue and environmental disbonding of the adhesive resulting in the entire bond failing as far back as the rotor attachment pin. The consequence of this disbonding was that all of the load that should have been transferred by the adhesive bond was now transferred by increased bearing loads at the attachment pin, and these increased bearing loads initiated fatigue cracking in the MRB. The cracking extended through about 90% of the D spar and extended well into the blade structure itself. It was only because of visual observations of the cracking by the flight crew in their walk around that disaster was avoided..
The RNZAF and ADF managed the problem by visual inspection of the paint layer at the outboard end of the bond. Any cracking in the paint was cause for further investigation. I am unaware of RNZAF findings from these observations but I am aware that a number of suspect blades were removed from service by the ADF and subjected to strip-down investigations that confirmed the presence of disbonds.
Like I have said, I just hope that this is not the cause of the Coffs Harbour incident.
Regards
Blakmax
I am a specialist in adhesive bond failure forensics, having worked in the past as a scientist for the ADF. Prior to 2007 I was involved in a collaborative investigation with the NZ DTA into a severe cracking problem with one RNZAF UH1H MRB. I sincerely hope that this is not the case with this event.
The primary cause of the RNZAF occurrence was disbonding of the adhesive layer between the blade upper surface and the steel grip pad where the blade attachment fittings clamped onto the blade adjacent to the attachment pin. There were significant concerns in relation to how the adhesive bond design concept was developed. The grip pad is approximately 12mm thick and the end terminates at sharp angle of approximately 45 degrees. There is a common perception that adhesive bonds distribute load over the entire adhesive layer, with a uniform shear stress spread over the joint. Of the order of 76% of bonded structures in the US were found in 2004 to be designed on this basis. This has been known since 1938 to be incorrect. In an adhesive bond the shear stress peaks at the ends and all of the load is transferred just at the ends of the joint with no load whatsoever being transmitted in the centre of the joint. The larger the stiffness imbalance between the adherents, the higher the4 shear stress peak at the end of the joint. In the case of the grip pad bond, there is a very large and rapid increase in thickness and a large difference in elastic modulus (steel vs aluminium) meaning that the shear stresses at the end of the joint are very high.
In the case of the RNZAF blade, post event investigation showed that there was clear evidence of adhesive bond fatigue and environmental disbonding of the adhesive resulting in the entire bond failing as far back as the rotor attachment pin. The consequence of this disbonding was that all of the load that should have been transferred by the adhesive bond was now transferred by increased bearing loads at the attachment pin, and these increased bearing loads initiated fatigue cracking in the MRB. The cracking extended through about 90% of the D spar and extended well into the blade structure itself. It was only because of visual observations of the cracking by the flight crew in their walk around that disaster was avoided..
The RNZAF and ADF managed the problem by visual inspection of the paint layer at the outboard end of the bond. Any cracking in the paint was cause for further investigation. I am unaware of RNZAF findings from these observations but I am aware that a number of suspect blades were removed from service by the ADF and subjected to strip-down investigations that confirmed the presence of disbonds.
Like I have said, I just hope that this is not the cause of the Coffs Harbour incident.
Regards
Blakmax
Are there any non-Robinson aircraft included in your list of main rotor separations that resulted in sheared off tails