.. 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.