|
|
 |
Essay on Plane Crash
| Date: |
02-09-99 5:46am |
| Subject: |
Technology |
| Word Count: |
4215 |
| Page Count: |
16.86 |
Plane Crash
Instructor: Greg Alston Abstract This paper examines the in-flight
separation of the number two pylon and engine
from a Boeing 747-121 shortly after takeoff from the Anchorage
International Airport on March 31, 1993. The
safety issues discussed focus on the inspection of Boeing 747 engine
pylons, meteorological hazards to aircraft, the
lateral load-carrying capability of engine pylon structures, and aircraft
departure routes at Anchorage International
Airport during turbulent weather conditions. Shortly after noon on March
31, 1993 the number two engine and
pylon separated from Japan Airlines Inc. flight 46E shortly after departure
from the Anchorage International Airport.
The aircraft, a Boeing 747-121, had been leased from Evergreen
International Airlines Inc. The flight was a
scheduled cargo flight from Anchorage to Chicago-O'Hare International
Airport. On board the airplane was the
flight crew and two nonrevenue company employees. The airplane was
substantialy damaged during the separation
of the engine but no one on board the airplane or on the ground was
injured. Flight 46E departed Anchorage about
1224 local time. The flight release and weather package provided to the
pilots by Evergreen operations contained a
forecast for severe turbulence. As fight 46E taxied onto the runway to
await its takeoff clearance, the local controller
informed the flight crew that the pilot of another Evergreen aircraft
reported severe turbulence at 2,500 feet while
climbing out from runway 6R. After takeoff, at an altitude of about 2,000
feet, the airplane experienced an
uncommanded left bank of approximately fifty degrees. Although the
desired air speed was 183 knots, the air speed
fluctuated from a high of 245 knots to a low of 170 knots. Shortly
thereafter the flight crew reported the number two
throttle slammed to its aft stop, the number two thrust reverse indication
showed thrust reverser deployment, and the
number two engine electrical bus failed. Several witnesses on the ground
reported that the airplane experienced
several severe pitch and roll oscillations before the engine separated.
Shortly after the engine separated from the
airplane, the flight crew declared an emergency, and the captain initiated a
large radius turn to the left to return and
land on runway 6R. The number one engine was maintained at maximum
power. While on the downwind portion of
the landing pattern bank angles momentarily exceeded forty degrees
alternating with wings level. About twenty
minutes after takeoff flight 46E advised the tower they were on the
runway. The aircraft was substantially damaged
as a result of the separation of the number two engine. Estimated repair
costs exceeded twelve million dollars. In
addition, several private dwellings, automobiles, and landscaping were
damaged by the impact of the number two
engine and various parts of the engine pylon and the wing leading edge
devices. The National Transportation Safety
Board (NTSB) determined the probable cause of this accident was the
lateral separation of the number two engine
pylon due to an encounter with severe or possibly extreme turbulence.
This resulted in dynamic lateral loadings
coming from many directions that exceeded the lateral load-carrying
capability of the pylon. It was later discovered
that the load-carrying capability of the pylon was already reduced by the
presence of the fatigue crack near the
forward end of the pylon's forward firewall web. As a result of this
investigation the NTSB made seven
recommendations to the Federal Aviation Administration (FAA),
including the inspection of Boeing 747 engine
pylons, the potential meteorological hazards to aircraft, an increase in the
lateral load capability of engine pylon
structures, and the modification of the aircraft departure routes at
Anchorage International Airport during periods of
moderate or severe turbulence. The NTSB also recommended that the
National Weather Service (NWS) use the
WSR-88D Doppler weather radar system to document
mountain-generated wind fields in the Anchorage area and
to develop detailed low altitude turbulence forecasts. In the course of the
investigation the NTSB explored virtually
every contributing factor contributing to the aircraft accident. These
included weather, mechanical failure, design
deficiencies, and human factors. The flight crew was properly trained and
qualified for this fight. None of the crew
members' Federal Aviation Administration (FAA) records contained any
history of accidents, incidents, or
violations. The flight crew and the mechanics who had worked on the
airplane before the flight volunteered to be
tested for the presence of alcohol and both lawful and illegal drugs. All of
the test results were negative. The
investigation revealed that the flight crew was in good health. The
airplane, registration N473EV, was a Boeing
model 747-121, serial number 19657. The airplane was manufactured in
June 1970, and was originally configured
to carry passengers. The airplane was acquired by Evergreen
International Airlines in December 1988, and was
subsequently reconfigured to carry cargo. The airplane had seating for the
three flight crew members and two
observers or passengers. The airplane was equipped with four Pratt &
Whitney JT9D-7 engines and appropriate
equipment for Instrument Flight Rules (IFR) operations. At the time of the
accident, the airplane had accumulated
83,906 flight hours and 18,387 cycles. The estimated economic design
life for the Boeing 747 is 20,000 flights,
60,000 hours, and 20 years. The number two engine, serial number
662812, had accumulated a total of 56,709.8
hours and 10,923 cycles since new and had accumulated 5,752.5 hours
and 1,200 cycles since overhaul two years
prior. The maintenance logs had no reports of severe e! ngine vibration on
the number two engine. The maintenance
records contained no deferred repair items regarding the number two
engine pylon structure. The airplane was
equipped with a Sundstrand Data Control Mark VI-J4 ground proximity
warning system (GPWS). In addition to
providing GPWS alerts, this system provides windshear caution,
windshear warning, and bank angle warning. The
system provides windshear warning and cautions between five feet and
1,500 feet during the initial takeoff and
between 1,500 feet and thirty feet during the final approach phases of
flight. The bank angle advisory indicates a roll
attitude that is excessive for the flight condition. Generally, above 1,500
feet, the callout occurs at forty degrees of
bank. The callout occurs again if roll attitude increases by twenty percent.
When roll attitude increases to forty
percent above the initial callout angle, the callout repeats continuously.
Below 1,500 feet, the callout angle is reduced
progressively. The windshear caution or windshear warning did not
activate because the turbulence was encountered
above 1,500 feet, well outside the warning envelope of the system.
However, the system did provide bank angle
warnings during the turbulence encounter. A significant meteorological
advisory (SIGMET) was issued at 1145 and
was valid until 1545. This SIGMET advised that moderate and frequent
severe turbulence could be encountered
from the surface to 12,000 feet. In addition, moderate and frequent
severe mountain wave turbulence could be
encountered from 12,000 feet to 39,000 feet within an area bounded by
Bethel, Johnstone Point, Sitkinak Island,
and Dillingham. The northern extent of the SIGMET area was about
thirty-six nautical miles south of Anchorage. A
correction to the SIGMET was made at 1342 adding the Anchorage area
to the list of locations within the advisory
area. According to an individual of the NWS forecast office at
Anchorage, the delay in issuing the correction (about
2 hours) was due to the workload. The delay caused the omission of
Anchorage from the SIGMET location points
to go unnoticed. The aviation weather forecaster also stated that
turbulence east of the airport was not an infrequent
event in the presence of a strong easterly flow near mountain top level. He
believed that in addition to the strong
easterly flow the turbulence was increased by an upper level trough
moving through the area, which, coupled with
heating, made the atmosphere unstable. He also stated that in the eighteen
years as a forecaster at Anchorage he did
not remember previously seeing as many severe turbulence pilot reports
as he saw that afternoon. Several other
pilots reported severe turbulence encounters about the time of the
accident. At 1210, a pilot of another Boeing 747
reported severe turbulence at 2,500 feet and moderate turbulence
between 3,000 feet and 10,000 feet during the
climbout to the north. The pilot of a U.S. Marshall Service Cessna 310
reported that he took off from runway 15 at
Merrill Field to perform a maintenance fight about noon. At 300 feet
above the ground, the airplane encountered a
downdraft and the airplane's air speed went from 120 knots to 90 knots
and lost about 200 feet of altitude. After he
emerged from the downdraft, the pilot turned the airplane to a heading of
120° and climbed to 900 feet. Shortly
thereafter, the airplane encountered an updraft. The vertical velocity
indicator pegged the needle at 4,000 feet per
minute upward and that despite reducing the throttles to idle the airspeed
would not fall below 160 knots. The pilot
stated that as he maneuvered the airplane back to the airport for landing,
the airplane encountered severe turbulence
with fifty-knot variations in air speed. The pilot concluded in his written
report that, in twenty years of flying, this was
the worst turbulence he has encountered. The NTSB also inspected the
navigation aids and communications within
in Anchorage area. No difficulties or problems were found. Damage to
the airplane consisted of the loss of the
number two engine and its pylon and the loss of most of the left wing
leading edge devices between engines number
one and two. During the investigation, the fuse pins holding the engine
pylons to the wings were removed from the
airplane. The two midspar fuse pins for the number two engine were
found to be deformed. The aft diagonal brace
fuse pin was fractured. The inboard midspar fuse pin for the number one
engine was found to be substantially
deformed. None of the other fuse pins on the airplane had any indications
of damage or deformation. Relatively
small areas of impact damage were also noted on the wings and trailing
edge flaps. The number two engine, all
portions of the number two engine pylon, and most of the leading edge
structure between the number one and
number two engines were recovered. There was no evidence of an
in-flight fire prior to the separation of the number
two engine. Several witnesses on the ground reported seeing a flash or
ball of fire as the engine separated from the
airplane. There were no reported fires on the ground as a result of falling
debris. Persons who first saw the engine
after it struck the ground reported steam rising from the engine.
Firefighters from the Anchorage Fire Department
sprayed water on the engine to prevent a possible fire. The pylon is
designed to carry the thrust and torque loads of
the engine as well as lateral, longitudinal, and vertical loads from
maneuvers and gusts. Lateral loads are ultimately
absorbed by the midspar fuse pins and side brace. According to Boeing,
the fuse pins can withstand an ultimate
lateral load of more than 2.8 G on the engine. Additionally, Boeing
reported that the portion of the structure of the
pylon that is critical under lateral loads is the firewall just aft of the
forward engine mount. The Boeing calculations
indicated that this firewall will fracture at a lateral load of between 2.35 G
and 2.88 G when it contains a fatigue
crack of the size found in this structure. The Boeing 747 airplane and its
pylon structure were designed in the
mid-1960's using the computer capabilities and analytical skills of the
time. Boeing's current computer modeling of
the pylon structure and the loads applied to it are considerably more
complicated and provide greater resolution of
the data than would have been possible with the techniques employed
when the airplane was designed. The use of
modern computer structural design programs allowed considerable
modeling of the pylon's response to various load
inputs with various structural failures. The number two engine pylon was
separated into four pieces as a result of
three principal fracture areas. These fractures were located just aft of the
forward engine mount bulkhead, among a
jagged vertical plane aft of the rear engine mount bulkhead, and around
the inboard midspar fuse pin fitting. The two
forward pieces of the pylon remained attached to the engine through the
forward and rear engine mounts.
Examination of the fractures around the perimeter of the break aft of the
forward engine mount bulkhead revealed
features typical of overstress separations, except for a small flat fracture
region in the firewall web. The flat fracture
area was approximately in the middle of the web on the outboard side of
the web centerline. The fracture was a
lateral fracture about two inches long through the thickness of the web
and was aft of the third transverse stiffener
behind the forward engine mount bulkhead. Investigators cut the flat
fracture area from the remainder o! f the firewall
and examined it in detail with a bench binocular microscope and a
scanning electron microscope. The mating
fracture faces had been heavily rubbed. Despite the rubbing, isolated
areas of contrasting color, indicative of
through-the-thickness propagation, was noted. Compression buckling of
the firewall web extended from the fatigue
crack area forward to the outboard side of the pylon at the second
transverse stiffener. Inspection of the other three
pylons on the airplane found no similar cracks. The fuse pin from the
underwing fitting for the diagonal brace was the
only one that was found broken. The outboard portion of the pin was
cocked within the underwing fitting. The
inboard piece of this fuse pin was recovered on the ground near the aft
portion of the pylon. The fractures on the
fuse pin and retainer bolt appeared typical of overstress separations. The
investigation found that all of the remaining
fractures and buckling of the structure were consistent with deformation
of the pylon structure in an outboard and
upward direction. Examination of the other fracture surfaces disclosed no
evidence of pre-accident damage or
cracking. All separations appeared typical of overstress separations.
Selected sections from the primary structures
of the pylon were returned to the safety board's materials laboratory for
examination. The material from the sections
was found to be within applicable manufacturer's specification
requirements for composition, conductivity, and
hardiness. The two fatigue cracks that were found in the number two
engine pylon structure were subjected to
metallurgical examinations. One of the fatigue cracks was a lateral fracture
about two inches long and was in the web
of the pylon forward firewall, just aft of the third transverse stiffener
behind the forward engine mount bulkhead. This
fatigue crack was lateral to the web. Although most of the features of this
crack had been obliterated by rubbing, a
few isolated areas of fatigue striations were found. The orientation of the
grain indicated that the cracking
propagated through the thickness of the web. The web material, a nickel
alloy, appeared to comply with
specification requirements. There was no evidence of damage or defects
that may have contributed to initiation of
the fatigue cracking. The pieces of the midspar web near the aft end of
the web had been deformed into a wave
shape, consistent with compression buckling. A fatigue crack was found
in this portion of the web, on the only piece
of the pylon structure that remained attached to the wing. Almost the
entire length of this crack was sandwiched
between portions of the inboard midspar fitting and other pieces of
structure at the aft end of the midspar. The plane
of cracking was oriented forty-five degrees to the fore-and-aft direction.
This is consistent with propagation under
tensile stresses from shear loading of the web. The cracking initiated from
both sides of a fastener hole. Additional
disassembly of the inboard midspar fitting and complete removal of the
web piece showed extensions of the fatigue
cracking. The overall length of the fatigue cracking area including the
extensions, was about three inches. There was
no evidence of any damage or defects that may have contributed to
initiation of! the fatigue cracking. Metallurgical
examination of the fracture in the fuse pin from the aft end of the diagonal
brace revealed evidence of a direct shear
overstress separation. The retention bolt for this pin was fractured as a
result of excessive bending and shear loads.
The maintenance records were examined at Evergreen's corporate
headquarters in McMinnville, Oregon. This
examination included a review of flight log entries, nonroutine work order
cards, work order cards generated by all
levels of routine checks and inspections, engineering orders, engineering
changes and repair authorizations,
mechanical reliability report files, airworthiness directive (AD) tracking
sheets, major alteration record lists, engine
logs, engine status reports, and engine trend monitoring sheets. The
records did not reveal any previous encounters
with severe turbulence. The three major alterations and repairs involving
the wing were either far outboard of the
number one pylon or were performed on the right wing. Two overweight
landings had been recorded since the
aircraft was put into service with Evergreen. In both cases, an inspection
of the airplane was accomplished in
accordance with the Boeing Maintenance Manual. A D maintenance
check was started in April 1992 and
completed in September 1992. During the check, a structural inspection
was performed on the number two engine
pylon. The inspection procedures called for the notation of any structural
irregularities, corrosion, loose or missing
fasteners, cracks, bulges, deformities, and delaminations. This check
specifically called for inspection of the torque
bulkhead, particularly in the area of the midspar fittings and diagonal
brace fittings. During the D maintenance
check, two cracks were found in the skin on the bottom of number two
pylon, just aft of the aft engine mount thrust
link. The cracks were stop drilled, and two doublers were fabricated and
installed. A third crack was found on the
diagonal brace upper end outboard clevis lug bushing. The diagonal brace
and lug were subsequently replaced. A
fourth crack was found six inches from the aft end of the outboard bottom
edge of the number two pylon internal
lower angle. A new internal lower angle was fabricated and installed.
During a B maintenance check performed in
November 1990, the entire number two engine pylon was removed from
the wing. During the time in which the
pylon was removed, extensive inspection and repair work was
accomplished on the pylon and its fittings. These
maintenance actions included the inspection and rework of the upper link
forward lug, the diagonal brace lug, and
the midspar attach fitting horizontal clevis, replacement of the upper link
fuse pins, inspection of the forward engine
mount bulkhead structure, replacement of tile forward support fitting
bolts, rework of the rear engine mount
bulkhead fitting, and rework of the midspar outboard attach fitting and the
inboard pylon attach fitting. The forward
engine mount bulkhead had been modified in order to prevent cracking in
the firewall web near the bulkhead. At the
time of the accident, the Boeing 747 Maintenance Manual did not
address inspection of the pylon forward firewall
web where the fatigue crack was found on the accident airplane. Boeing
had previously issued a service bulletin on
February 14, 1986, for operators to inspect for fatigue cracking of an
adjacent lower spar web. The service bulletin
reported an operator experiencing two cracks approximately six inches
long in the aft lower spar web of the number
one pylon after 8,500 flight-hours. Following the accident Boeing issued a
service bulletin that called for a detailed
visual inspection of the horizontal firewall of the inboard engine pylons on
Boeing 747 airplanes powered by Pratt
and Whitney JT9D-3A or -7 series engines. The service bulletin states
that airplanes with over 15,001 flight cycles
should be inspected within six months of the release of the service bulletin.
Airplanes with between 6,001 and
15,000 flight cycles should be inspected within twelve months, and
airplanes with less than 6,000 flight cycles should
be inspected at 6,000 flight cycles or within twelve months, whichever is
later. There have been no operator reports
of finding cracks in the forward web as a result of the inspections from
this service bulletin. Additionally, following
the accident Boeing requested selected operators of high time Boeing
747s to inspect their airplanes for cracks in
the forward web. Boeing reports that the operators found no evidencence
of cracking. The investigation found that
there were multiple separations in the number two engine pylon that
allowed the engine to separate from the wing.
There was evidence that the direction of separation was outboard and up.
This evidence included the lack of
damage on the inboard side of the pylon, the fractures and deformation in
the major structural members of the pylon,
and a piece of the wing leading edge structure that was embedded in the
rear of the engine. The examination of the
pylon structure also yielded sufficient clues to determine the sequence of
pylon fractures that resulted in the loss of
the engine. The rear engine mount fitting in the pylon was intact and, when
recovered, a major piece of the pylon
was still attached to the engine. However, the fitting was cracked and
heavily distorted in relation to the pylon
structure around it. This cracking and distortion indicated motion of the
forward end of the engine in the outboard
and up directions. This damage indicates that the pylon srtucture was
intact when the damage occurred. If the pylon
had been separated at any location aft of the rear mount fitting, the fitting
would not have been distorted as it was
because the pylon structure would have moved with the fitting as engine
motion attempted to generate the cracking
and distortion. The condition of the rear engine mount suggests that the
forward end of the engine separated from the
main portion of the pylon and moved in the outboard direction while the
remainder of the pylon was intact and
attached to the wing. The examination of the front of the pylon revealed
that the pylon structure was fractured just aft
of the forward engine mount bulkhead, and that a small piece of the
forward part of the pylon was attached to the
engine at the forward engine mount. The fracture on this part of the pylon
contained indications of overstress
separations except for the two inch fatigue crack in the forward firewall.
The firewall contained compression
buckling that extended to the area of the fracture. Overstress separations
from shear loading were found on both
sides of the fatigue area. These overstress separation areas probably
occurred immediately after the compression
buckling and was the start of the complete fracture of the pylon aft of the
forward engine m! ount bulkhead. The
front end of the engine was now free to swing to the left under the same
lateral loads that produced the initial
separation of the pylon. The movement of the front of the engine to the
left created the heavy distortion and cracking
in the rear mount fitting. As the front end of the engine swung to the left,
the pylon structure would have bent in the
outboard direction. At the same time, the engine would have been
producing thrust at an unusual angle The
combination of the bending of the pylon and the unusual thrust angle
would account for the damage found on the
midspar fuse pins, the large vertical fracture in the middle of the pylon, the
shear buckling of the mid spar web, and
the direction of fracture of the major structural members of the pylon.
Boeing performed a finite element analysis of
the forward portion of the pylon structure. This analysis showed that the
fatigue crack in the firewall would reduce
the stress capacity of the pylon by about ten percent. The computer
generated model predicted that in the presence
of the cracked web, the number two engine pylon would fail with a lateral
load of between 2.35 G and 2.88 G. The
separation of the number two engine pylon was due to an encounter with
severe or possibly extreme turbulence that
resulted in dynamic multi-axis lateral loadings that exceeded the ultimate
lateral-load carrying capability of the pylon.
The load carrying ability of the pylon was already reduced by the
presence of the fatigue crack near the forward end
of the pylon. The computer analysis found that encounters with severe
turbulence can produce enough lateral loads
to separate the pylon from the wing even without the presence of any
cracks in the pylon web. Encounters with
moderate and seve! re turbulence are considered relatively normal events
by pilots and controllers, and operations
are not curtailed by the forecast or pilot reports of severe turbulence.
Therefore, there is a safety-of-flight concern
regarding the lateral design loads for engine pylons during severe turbulent
conditions. However, diminishing this
concern is the fact that Boeing 747 airplanes, as well as many other
makes and models of airplanes, have been
operating successfully for many years without engines or pylons
separating from the wings solely because of
turbulence. ln general, it would appear that airline operating procedures
and pilots' actions have been effective in
avoiding operations into extreme or very severe turbulence that could
damage their airplanes. This is why no
structural modifications were required as a result of this accident.
However, the NTSB recommended that the FAA
should modify the design load requirements of 14 CFR Part 25 to
consider multiple axis loading and ! to consider
the magnitude of the loads that can be experienced in turbulence
conditions. The fatigue cracking found on the
midspar web probably resulted from sheet bending due to flexing or
vibration of the web material. The crack
probably would have been detected if there had been a requirement to
inspect this area. Therefore the FAA should
require all operators to inspect the entire pylon forward firewall web at
specific flight hour intervals. It is not
reasonable to suspend operations during turbulence because aircraft have
been able to operate safely during such
conditions. The most intense turbulence occurs near the mountains at low
altitude. Therefore, by staying away from
the mountains on departure, aircraft may lessen the chance of
encountering severe turbulence. The FAA should
consider modifying the departure routes of aircraft at Anchorage during
periods of moderate or severe turbulence in
order to minimize an aircraft's encounter with mountain-induced low level
turbulence. The NTSB conducted a very
through investigation of this accident They included areas that an average
person with any knowledge of aviation
would never have thought. Their final recommendations seem to be logical
and have merit. Common sense prevailed
and led to sound recommendations. Referrences Vogt, C. W., Coughlin,
S., Lauber, J. K., Hart, C. A., &
Hammerschmidt, J. (1993). Aircraft accident report, in-flight engine
separation, Japan Airlines Inc., flight 46E
(National Transportation Safety Board Rep. No. AAR-93/06). Oster, C.
V. (1992). Aviation safety in a changing
world New York: Oxford University Press.
Complete List:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
|