Fedex MD11 - Subic Bay

[Shore Guy]

03 Jun 02
1600Z
DCA02WA039
At Subic Bay
Philippines
MD-11 of FEDEX
Reg: N588FE

Elevator damage during escape maneuvers after ground proximity warning (GPWS) alert. Approaching Subic Bay at 8000 feet in heavy rain at the time.

Outboard portion of the left elevator had almost entirely separated and a small area inboard of this location was
missing.

^^^^^^^^^^^^^^^^^^^^^^^^^

Anyone have more information on this incident?

[Huck]

DC-10's and MD-11's are designed to shed the outer panels of their elevators when exceeding load limits.

[Hand Solo]

A GPWS warning at 8000feet!:eek: Whats the MSA round there?

[dallas dude]

Hand Solo,

Most Enhanced GPWS (latest version) use a "3-D"database to compute terrain.

Original GPWS (Mk I) usually uses a feed from the Wx radar and the return is mathematically calculated to be cumulo nimbus or cumulo granite etc..

Certain size droplets, usually in very heavy rain, can have the effect of fooling the GPWS (Mk I) into the cumulo granite choice.

The problem is typically that the EGPWS (new version) does not wholly replace the original GPWS. The original version was usually just "overwritten", i.e. not deleted, and therefore occasionally the original system thinks it has something to say!

I do NOT know the details of the MD-11 event. I'm sure the situational awareness of the crew was such that they reacted as required to a GPWS alert in the conditions noted.

I've had the same problem in very heavy rain at 11,000 over Dallas (twice).

Considering you can stand on a six pack in East Texas an "see" El Paso, it's a non event.

Don't know the topography this crew was over though.

Cheers,

[NigelOnDraft]

I have never heard of any GPWS system trying to determine Cumulo Granite from the Wx radar...

Is this specific to the MD-11? Or was it an early EGPWS?

All the GPWS I have flown with (A340, B757/767) are RAlt / VSI based, and the EGPWS we now have are terrain database dependent...

NoD

[invalid entry]

Within 25NM of apt highest MSA is 6800' to NW

[Waldo]

I,m not sure about the feed from Wx radar into the GPWS I have however had a false GPWS "pull up" whilst descending into Corfu through heavy rain at around 10,000 ft. At 3AM it certainly concentrates the mind!

[bblank]

"DC-10s MD-11's are designed to shed the outer panels of their
elevators when exceeding load limits."

Bad things may happen when a load limit is exceeded but is there
any information that supports the notion that the separations that
occur are designed?

First, there may be differences between the two types mentioned
because of materials: composites are used for the MD-11 elevators
but I don't think that they were ever used in the DC-10 elevators.
Not sure about the latter, though.

I don't know anything about this particular incident but in the past
there have been several cases in which large portions of MD-11
elevators have been lost to fracture, delamination, and disbonding
of composite materials. There was an earlier Fedex MD-11 incident
(to N582FE) at Manila very much like this one. In that incident on
30 June 1999 a flight from Narita to Manila experienced a GPWS alert
while descending through 9500 ft on the approach in heavy rain. The
captain performed a GPWS escape maneuver. The event resulted in a
structural failure in the left elevator outboard counterbalance area.
The outboard end of the elevator and counterweight were found hanging
vertically on the outboard hinge bolt. The elevator lower surface
cracked and the upper surface delaminated.

Damage to elevators as the result of intentional stalls was known
from Douglas test flights. The earliest operational losses of portions
of the MD-11 elevators go back to incidents of Alitalia, KAL, and CI in
1992. The latter, mentioned in connection with PIO in the JAL 706 thread, (http://www.pprune.org/forums/showthread.php?threadid=55495)
resulted in this recommendation:

"Require Douglas Aircraft Company to advise MD-11 operators of
the potential for damage to the composite elevators if the airplane is
operated beyond the limits of the design buffet boundary."

I did not see anything in the reports to suggest that Douglas was able
to precisely state what damage might occur or under what circumstances
that damage might occur. For instance, the language used went like this:

"DAC engineers believe that an MD-11 needs to exceed its buffet
boundary by a margin of 20 percent or more in order to sustain
elevator damage. The extent of severe damage is believed by DAC to
be related to some combination of the degree of exceedence above the
20 percent margin and length of time exposed to buffet. DAC has been
unable to determine the exact relationship.
...
DAC engineers were unable to determine exactly where the elevator
damage occurred or whether the amount of time outside the buffet
boundary exacerbated the elevator damage."

With regard to elevator safety design, Douglas did not mention
anything about benign separation - only redundance.

It would be interesting to know what forces were involved in
the two Fedex GPWS escape maneuvers. (Well, I'm not a pilot -
that's what is interesting to me.)

[woderick]

Hate to disagree but ..............
GPWS Mk1 inputs Rad Alt, Vertical Speed, Baro Rate.
GPWS Mk2 inputs Rad Alt, Vertical Speed, Baro Rate. just refined software and a couple more sub modes.
EGPWS inputs Rad Alt, Vertical Speed, Baro Rate, and PPos from FMC, FMGEC or whatever is knocking about. this is compared to what I believe is a fairly rudimentary terrain Db sort of like FS2000! hey it does the job.
Wx Radar haas never been an input to GPWS. I've installed enough of the various stages to know.
What concerns me a little is where someone can get the idea that Radar is an input. Is someone teaching this ?

[411A]

"Is someone teaching this?"

Yes, probably a pilot, who is under the impression that...."pilots know everything". See the thread "I am an army of one" for proof of this phenomenon.

The terrain around Subic rises quite steeply to the east and southeast and pilots NEED to have situational awareness there...apparently some do not.

[SimJock]

It is the Rad Alt signal that gets confused during heavy rain not WX.

Airbus brought out a FCOM Bulletin for A319/320/321 about this,

quote "Under heavy rain condition At 2600ft at least one of the radio altimeter delivered a height of 480ft during 13 seconds, The warning L/G not down was displayed"

Clearly this same phenomenon causes incorrect GPWS warnings also.

[dallas dude]

Of course, most of the above posts are also correct as I didn't specify my example applied to EFIS/FMS (Incl ex AA MD11) and certain manufacturers' GPWS systems.

For example, the Honeywell system GPWS/EGPWS as fitted on AA F-100's uses the radar transmitter/antenna for terrain mapping, along with the ppos database. If the wx radar isn't switched on and terrain selected ONLY the standard samplings i.e. VSI, rad alt etc. will provide warnings.

Other flight envelope samplings also occur depending on the aircraft's configuration.

For EGPWS on AA F-100;
turn Wx radar ON and push terrain button ON which, hopefully, turns a pleasant shade of blue. This is where the "signal" attenuation is sometimes mis-read and misinterpreted in heavy rain.

Radar alt on F-100 isn't "supposed" to work above 2500' agl.

However, along with the spurious GPWS warnings at 11,000' the rad alt can also read 1000' (in a nice green hue) in heavy rain. Honeywell is working on a fix at this time.

Cheers,

[brokepilot]

Yes, Dallas Dude
Question? Does the GPWS on the F-100 tell the crew to pull out or pull up?

[ICT_SLB]

Beleive we're getting a little confused about basic GPWS, the "Enhanced" EGPWS and the display systems needed to show the EGPWS terrain maps and visual alerts to the crew.

First, basic GPWS is not much more than a smart Radio Altimeter. The various other inputs, Baro Alt, Vertical Speed (= closure rate), flaps & gear only modify or even inhibit the "Pull Up" or "Terrain" calls. Radio Altimeters only read to 2,500 feet agl by design (as opposed to the high altitude military Radar Altimeters which go up to 50,000 feet or more). At 2650ft, they actually set "cruise mode" to avoid false indications.

Second, Enhanced or EGPWS always includes the basic function that must work independently and as if it was a GPWS alone (it's a requirement). EGPWS provides additional alerts by comparing aircraft height against a terrain database and Obstacles alerts by position & heading (the Obstacle must be within 4 degrees of heading to be shown if memory serves).

If a display is available, EGPWS will show a colour coded picture of surrounding terrain (if that display mode is selected by the crew). When they do this, they remove the WXR bus & replace it by the EGPWS data over a similar bus. This is why you have to have the WXR powered on some aircraft to get the display. Some aircraft also switch in TCAS the selection mix all on the MFD.

[jetboy]

Well said ICT_SLB. With all the misinformation regarding avionics systems out there it's no wonder the most common phrase in the flight deck is "what's it doing now?" with the second being, "what's WHAT doing now?". Low Range RADAR altimeters-Geesh!

[L84Wrk]

I have heard this EGPWS/radar theory before. Have seen it in action and have been given ground instruction on its operation and inputs, but, could it be that simple people like myself, jump to this conclusion when the EGPWS terrain depiction looks likes wx radar ground returns and are on the same sreen as wx radar returns.
Didnt know about rad alt being effected by wx, thanks will watch out for that.

[mcdhu]

In simple terms, GPWS uses the (downward looking ) Radalt to see how fast you are closing with the ground below and, depending on what stage of flight/configuration you are in, gives you a clue as to how to extricate yourself from the incipient mire. The EGPWS, however, uses the Radalt to see how fast you are closing with the ground below and, using the navigational accuracy of GPS and a digital database which gives it the contours of the ground ahead, can predict whether or not you will hit the ground ahead and will therefore know whether or not to serve you up an appropriate warning. Hence the various names given to EGPWS eg TAWS(Terrain Awareness and Warning System), Predictive GPWS etc.
With apologies to those more technically minded than me!
Cheers,
mcdhu

[QNH1013]

Flew in to Subic Bay a few times. There is extensive high terrain around the airport and the main approach is over the sea. Non Precision, VOR/DME approach and it is an Offset final approach course due to terrain anyway. At missed approach point or prior you are to be visial with the runnway and make final heading corrections to line up with the centreline.

Can't remember off hand what the MSA is there but there are 4 different high MSA altitudes centred around the Subic VOR. Any GPWS warning in IMC would be definitely well respected regardless of location anyway. But Subic is certainly a place to take not of the terrain even for the Missed Apporach procedures or you'll fly in to the side of a volcano. I'll post more accurate airport details when I look at the chart next time I'm at work.

[Localiser Green]

6,800 is the worst MSA.

http://b757.users.btopenworld.com/SUBIC.jpg

[Clipper811]

Dallas Dude said;
If the wx radar isn't switched on and terrain selected ONLY the standard samplings i.e. VSI, rad alt etc. will provide warnings.

Most likely off on this one mate.
With the switch (terrain) off and radar off on
the 737, 757, 767, and 777's I'm flying the
database warning will
occur, aural that is. No time on the F100 but
that's the way the EGPWS is wired to the above
jets I've flown (AAL that is.)
Cheers, "California Dude"
:cool:

[mcdhu]

Further to Clipper811's post above ref Boeing EGPWS display, on the Airbus, even if you do not have 'Terrain' selected on the ND, an EGPWS warning will put the terrain on the ND whether you like it or not. The catch is that the EGPWS does not put the terrain warning on the ND at full brightness so you need to have the wx radar/terrain brightness knob - as opposed to the ND brightness knob - turned up to see the warning; you will, of course, hear it at full volume from 'Pierre' anyway!
Hope that is clear.
Cheers,
mcdhu

[downfourgreen]

A very important issue about this incident is:
MD11 has a fuel tank in the horizontal stabilizer. Usually it's empty during approach but ballast fuel may be carried. So, it's unacceptable a structural failure after an EGPWS or GPWS recovery maneuver.

[GlueBall]

In the mid 70s an AeroMexico DC-10-30 during climbout (over Spain) on A/P in VS mode encountered a high altitude stall, followed by an apparent violent recovery maneuver (according to FDR data) which also had resulted in substantial damage to elevators and stabilizer, including some outer surface panel separations. The damage was discovered on the ramp at MIA during a scheduled stop en route to MEX.

[kennedy]

I have experinced a few in the DHC-8 and one in 737, spurious, but in IMC, and rain, (certainly gets the heart pounding) and from the general discussions with engineers, and watching the RADALT display wind down very quickly, they have suggested that water gets into the radar head or receiver and spoofs the GPWS.

In fact, sitting in Fragile rock (GUR) loading pax, in very heavy rain I got both "TOO LOW TERRAIN" and "WHOOP WHOOP PULL UP" a number of times. Only time I found it funny especially seeing the pax faces in the cabin.

In the air, however it is the Terrain escape manoever for Warnings in IMC every time, and "what the f***!" for alerts:eek: :eek:

[Shore Guy]

Expanded narrative from the NTSB site…..interesting that a GPS input into the EGPWS allows for an altitude reasonableness test against the RA’s to prevent false warnings off a radar altimeter spike.


HISTORY OF FLIGHT

Expanded narrative recently released by the NTSB…..interesting that a GPS input to the EGPWS does an altitude reasonableness test off GPS derived altitude to prevent false warnings from RA spikes…..


On June 3, 2002, about 1600 hours Universal Coordinated Time (UTC), a McDonnell Douglas MD-11, registration N588FE, operating as Federal Express cargo flight 5181, experienced structural damage to the left elevator during an abrupt maneuver while descending to land at the Subic Bay International Airport, Subic Bay, Philippines. The abrupt maneuver was initiated by the flight crew in response to a ground proximity warning system (GPWS) alert that occurred as the airplane was encountering heavy precipitation at an altitude of about 9,800 feet above mean sea level. The airplane landed uneventfully at the Subic Bay International Airport and both flight crewmembers, the sole occupants, were not injured. Night and instrument meteorological conditions prevailed at the time of the accident. The flight had departed from Kuala Lumpur, Malaysia, and was destined for Subic Bay. Under the provisions of Annex 13 to the Convention on International Civil Aviation, the investigation was delegated to the U.S. National Transportation Safety Board (NTSB) by the Air Transportation Office (ATO), Department of Transportation and Communications, Republic of the Philippines.

The flight crew stated that they performed the GPWS escape maneuver as per the Federal Express procedure, and the warning subsided. They also stated that they had no problems with aircraft handling after the escape maneuver. This information is consistent with information from the digital flight data recorder (DFDR). The DFDR data indicate that the maximum recorded pitch of the airplane during the escape maneuver was 22.5 degrees nose up, and the maximum recorded g-force was 2.2 g's. In addition, the radio altitude signals for both altimeters indicated an abrupt drop to below 1,000 feet at the time of the GPWS "Pull-up" warning. The false radio altitude lasted for approximately two minutes. There were also several stall warnings during the maneuver and recovery.

Information recorded at the time of the escape maneuver from the 2-hour cockpit voice recorder (CVR) was overwritten; the CVR recording began during final approach to Subic Bay, after the GPWS warning.

PERSONNEL INFORMATION

The captain, age 56, was acting as the pilot-not-flying on the accident flight. The captain held type ratings in the MD-11, Boeing B727 and Douglas DC-10. The captain reported that he had 13,537 hours of total flight time, including 8,643 hours in type. The captain's first class medical certificate was issued on January 3, 2002 with no limitations or waivers.

The first officer, age 48, was acting as the pilot flying on the accident flight. The first officer held a type rating in the MD-11. The first officer reported that he had 7,096 hours of total flight time, including 4,188 hours in type. The first officer's first class medical certificate was issued on May 20, 2002 with no limitations or waivers.

METEOROLOGICAL INFORMATION

The weather at the Subic Bay International Airport at 1600 hours UTC was reported as wind variable at 3 knots, rain with broken skies at 1,800 feet AGL and overcast skies at 8,000 AGL, 4 miles visibility, temperature 75 degrees Fahrenheit, dew point 73 degrees Fahrenheit and altimeter 29.89 inches of Mercury.

WRECKAGE INFORMATION

Post-flight inspection of the airplane revealed that the outboard portion of the left elevator, including the balance weight, was partially separated from the rest of the elevator. A 12 to 18 inch section of the elevator was missing. The right elevator appeared undamaged and underwent a detailed examination to confirm that it was undamaged.

RESEARCH AND TESTING

According to Boeing, elevator damage to DC-10 and MD-11 airplanes can occur due to the dynamic response of the elevator when it is driven by the separated flow from the wing that results from a maneuver beyond the stall buffet boundary. During these events, the outboard elevators respond in their first torsion mode of about 10 cycles per second. The elevator horn balance weight contributes to this effect. The DC-10 and MD-11 elevators are equipped with an outboard damper to reduce the horn balance response, but this can be overcome by forces that occur when the stall buffet boundary is exceeded.

Analysis of the DFDR from this accident revealed that the airplane's design aerodynamic stall buffet boundary was exceeded during the abrupt GPWS pull-up maneuver.

The airplane was equipped with an upgraded Allied Signal enhanced GPWS and two upgraded Rockwell Collins LRA-700 radio altimeters. The enhanced GPWS receives inputs from two radio altimeters instead of just one, like the previous model, and is capable of sensing false radio altimeter tracking in airplanes like the MD-11, assuming that at least one of the radio altimeters is providing accurate data.

The radio altimeters, model LRA-700, P/N 622-4542-222, were tested by Rockwell Collins in Melbourne, Florida, on August 22, 2002. The no. 1 radio altimeter, S/N 4133, had no fault data logged. The no. 2 radio altimeter, S/N 4130, had faults logged from two previous legs. Both faults were internal faults listed as "lra-1 transceiver" and occurred prior to the incident flight. There were no faults listed for the incident flight. Both units passed the normal operation, return to service (with the exception of the 2500-foot sensitivity test), and temperature testing with no anomalies noted. The Melbourne facility lacked the support equipment to perform the 2500-foot altitude test so Rockwell Collins in Forth Worth, Texas, performed the remaining testing. Both units were tested and showed a sensitivity of 125 dB at the 2500-foot altitude. These measurements were within limits.

Both radio altimeters had been modified by Boeing service bulletin (SB), MD11-34-094. The SB modified the sensitivity of the LRA-700 altimeters in an attempt to reduce the frequency of false warnings during heavy precipitation. The modified altimeters were identified with a -222 part number, and their sensitivity was decreased by 6 dB per the SB.

The enhanced GPWS unit, P/N 965-0976-003-212-212, S/N 3930, was tested by AlliedSignal in Redmond, Washington. The GPWS unit passed the full Acceptance Test Plan with no anomalies noted. There was no examination of the antennae or the attaching hardware for the radio altimeter and the GPWS unit.

According to representatives of Federal Express, the airplane also had incorporated a Global Positioning System (GPS) multi-mode receiver, but the receiver was not interfaced with the GPWS. When interfaced, this unit provides an altitude that is derived from the global positioning system (GPS) and a terrain database. The altitude is used as a reasonableness test for the radio altimeters.

ADDITIONAL INFORMATION

Previous Federal Express Elevator Damage Occurrence

A similar accident involving a Federal Express MD-11 occurred on June 30, 1999, when, on an approach to Acquino International Airport in the Philippines, the airplane experienced a GPWS "terrain-terrain" alert while descending through 9,500 feet in heavy precipitation. The pilot executed a GPWS escape maneuver, which resulted in similar damage to the elevators. The left outboard elevator and balance weight was completely separated from the rest of the elevator. The right elevator sustained substantially less damage. The airplane landed safely.

The airplane had a previous model GPWS unit that was capable of only accepting inputs from one radio altimeter. The airplane also had two previous model radio altimeters prior to a Boeing service bulletin that desensitized them.

According to Rockwell Collins, radio altimeters are susceptible to false warning due to reflectivity from ice crystals, heavy precipitation, or aircraft that fly underneath the airplane. The altimeters are designed to be sensitive and accurate enough for the auto-land mode, while minimizing false warnings.

History of DC-10 and MD-11 In-flight Elevator Damage.

According to information presented by Boeing in the 1999 GPWS incident and a previous NTSB investigation of a China Airlines MD-11 in-flight turbulence encounter in 1991 (NTSB/AAR-94/02), the damage found on this Federal Express airplane was similar to damage found on other MD-11s that experienced aerodynamic stall buffet.

In similar instances involving high-altitude upsets, damage to MD-11 elevators occurred with no reported loss of control. In the China Airlines incident report, the Safety Board concluded that the elevator buffet damage in that incident and previous incidents was caused by overstress but did not create an unsafe condition. In the 1999 GPWS incident, the captain stated that she had no indication after the escape maneuver that the elevators were damaged, and that the airplane's flight characteristics were "not irregular."

Integration of Global Positioning System with GPWS.

Prior to this accident, Boeing developed and released Service Bulletin MD11-34-116, allowing operators to install two GPS multi-mode receivers in place of the Instrument Landing System (ILS) receivers. Honeywell then developed Supplemental Type Certificate (STC) ST00536SE to allow the GPS information to be inputted into the GPWS computers to provide a reasonableness test for the altitude reported by the radio altimeters. The accident airplane had the GPS multi-mode receivers installed on February 8, 2002 but these receivers were not linked to the GPWS computer. As a result of this accident and other operational considerations at Federal Express, all MD-11s in the FedEx fleet have had the necessary equipment installed and the connections made to allow the GPS information to be used by the GPWS computer as of December 6, 2003.