Korean B773 at Tokyo on May 27th 2016, rejected takeoff due to engine fire

Last Update: August 20, 2018 / 10:44:50 GMT/Zulu time

Bookmark this article

Photo of Korean Airlines HL7534, Boeing 777-300 — Korean Airlines HL7534, Boeing 777-300 (Photo credit: Alec@B92 / Flickr / License: CC by-sa)

Incident Facts

Date of incident
May 27, 2016

Classification
Accident

Cause
Rejected takeoff

Airline
Korean Airlines

Flight number
KE-2708

Departure
Tokyo Haneda, Japan

Destination
Seoul Gimpo, South Korea

Aircraft Registration
HL7534

Aircraft Type
Boeing 777-300

ICAO Type Designator
B773

A Korean Air Boeing 777-300, registration HL7534 performing flight KE-2708 from Tokyo Haneda (Japan) to Seoul Gimpo (South Korea) with 302 passengers and 17 crew, was accelerating for takeoff from Haneda's runway 34R when unusual vibrations were felt throughout the aircraft and smoke was seen from the left hand engine (PW4098). The crew rejected takeoff at low speed and stopped the aircraft about 1300 meters/4200 feet down the runway. Large flames were seen from the left hand engine, the aircraft was evacuated. 9 occupants received minor injuries in the evacuation and were taken to hospitals.

The runway was closed for about 6 hours until the aircraft was moved off the runway and evidence was collected.

Japan's Ministry of Transport reported the tower controllers observed flames and smoke from the left hand engine and instructed the crew to stop. An investigation into the occurrence has been opened.

On May 28th 2016 the JTSB reported debris from the engines were found about 600 meters after the begin of takeoff run, tyre marks were seen in the following 700 meters.

On Jun 20th 2016 the JTSB reported that examination of the engine revealed that a turbine disk fractured and departed the engine through the engine casing. The aircraft rejected takeoff and was evacuated, 9 occupants received minor injuries. The occurrence has been rated a "critical incident".

On Nov 8th 2017 the JTSB released a press release reporting that the investigation into the "serious incident" is ongoing. So far it has been established that the disk of the first stage of the high pressure turbine partially fractured causing damage to the turbine case and engine cowl. The resulting vibrations also caused cracks to other engine components, in particular the engine fuel and oil heat exchanger, which resulted in a fuel leak with fuel pouring down the engine cover. The fuel ignited causing the fire outside of the engine.

On Aug 20th 2018 the JTSB released their final report concluding the probable cause of the accident was:

It is highly probable that the causes of this accident were the fracture of the high pressure turbine (HPT) disk of the No.1 (left-side) engine during the takeoff ground roll, the penetration of the fragment through the engine case and the occurrence of subsequent fires.

Regarding the cause for the 1st stage HPT disk to be fractured, it is probable that a step was machined exceeding the allowable limit when machining U-shaped groove on the aft side of the 1st stage HPT disk to manufacture the engine and from this step the low-cycle fatigue crack was initiated and propagated during running of engine.

Regarding why the step could not be found, it is somewhat likely that defects failed to be detected at the time of the inspection by the manufacturer during the production process. And as for the cracks that were not found, it is somewhat likely that those cracks failed to be detected at non-destructive inspection on the disk by the Company at the time of maintenance of the engine in use.

Regarding the fire breakout from the No.1 engine, it is probable that due to the impact forces generated by the release of the fragment from the ruptured rim part of the 1st stage HPT disk through the engine case and the engine rundown loads generated when the engine stopped suddenly, the cracks were developed in the outer case of the Fuel Oil Heat Exchanger and the fuel and engine oil leaking through these cracks contacted the hot area of engine cases of the No.1 engine to be ignited.

The JTSB analysed:

when this accident occurred (at around 12:38), wind blew at about 20 kt from the almost right abeam Runway 34R. Because the fire and smoke accompanying the fire breakout from the No.1 engine was blown by this beam wind to the lee side which was the left side of the Aircraft, it is probable that the fire did not cause that much effect on the Aircraft and the cabin.

Furthermore, using the right side slide of the Aircraft which was at the up wind side for the emergency evacuation, it is probable that there were not much effects by the fire and smoke.

However, regarding why the R5 slide at its deploying was slipped to under the Aircraft, it is somewhat likely that as described later in 3.7.5, the strong abeam wind from the right combined with the thrust still being produced by the #2 engine at the ground idle setting could have the effects.

With respect to causes of the fracture the JTSB analysed:

It is probable that the crack generated in U-shaped groove was propagating by the action of repetitive stress per a flight. The Company executed the FPI on the 1st stage HPT disk when disassembling the HPT module, but the crack was not found. As the result of analyzing the fractured surface, it is somewhat likely that the cracks in U-shaped groove of the 1st stage HPT disk may exist prior to the last inspection conducted at the engine repair shop on November 12, 2014. Based on these, it is probable that because the cracks were not be discovered at FPI conducted the previous inspection and due to the flight following the inspection, the cracks were propagating more to be fractured.

The JTSB analysed with respect to the engine damage, that first debris was found 680 meters from the threshold of the runway. The debris was found consistent with the ejection of the fractured 1st stage HPT disk rim. The damage caused fuel and oil leaks, the fuel and oil ignited when it contacted high temperature engine case parts. In additon, the fuel oil heat exchanger showed internal cracks although there was no damage from the outside. The JTSB wrote: "Therefore it is probable that those cracks were developed, when the 1st stage HPT disk rim had been fractured and released, and struck into the turbine case, a strong shock force was generated. In addition, it is also probable that engine run down loads, which were generated when the engine No.1 stopped suddenly following the disk rupture, was the contributing factor."

The JTSB analysed that the crew discharged the first fire bottle which resulted in the fire indication to extinguish for about 5-10 seconds, then the fire warning activated again, the second fire bottle was discharged, the fire warning again extinguished for 5-10 seconds then reactivated. The captain thus decided to evacuate the aircraft.

The JTSB analysed that following the decision to evacuate the checklist was not immediately handy, while the first officer looked for the checklist and because the captain felt the evacuation was needed in a hurry, the captain performed the checklist items to shut the engines down from memory. The FDR showed, that the right hand engine was shut down by moving the fuel cut off lever only 54 seconds after the evacuation order was given at 12:43:45. The JTSB wrote: "And the first door being opened was L1 door and the time was 12:43:17. Based on these, instruction of emergency evacuation prior to halting of the No.2 engine positioned in the evacuating direction was given, it is highly probable that the No.2 engine was stopped about 28 seconds later after the first door was opened. The wind velocity of the engine wake air flow is 55 km/h even at the time of idle thrust and when conducting the emergency evacuation prior to the engine stopped, there are potential threat that the passengers could be blown away by the engine wake air flow and others. It is necessary for the Company to revise the education and training in order to enforce the thorough compliance to emergency evacuation procedure." The JTSB further analysed: "when deploying R5 slide, the No.2 engine was at idle thrust. According to the materials of the manufacturer of the Aircraft, the predicted jet engine exhaust velocity contours at idle thrust is 7 m wide at the rear end of the Aircraft, extends to the approximately 40 m ahead while being tapered at the level of approximate 6 m high and the wind velocity is about 30 kt. Calculating the resultant wind from the wind of the jet engine exhaust blow at setting of the heading at 335º with the wind direction 060ºand velocity about 20 kt at the time of the accident, was resulting in the wind direction 007ºand the velocity about 37.5 kt. The velocity is 1.5 times of the allowable limit which is 25 kt for the velocity to make a slide have normal deployment."