Polar Cargo B748 at Tokyo on Jul 15th 2017, overran runway on takeoff

Last Update: June 11, 2019 / 14:56:38 GMT/Zulu time

Bookmark this article

Photo of Polar Air Cargo N852GT, Boeing 747-8 — Polar Air Cargo N852GT, Boeing 747-8 (Photo credit: masak2_ukon / Flickr / License: CC by-sa)

Incident Facts

Date of incident
Jul 15, 2017

Classification
Incident

Airline
Polar Air Cargo

Departure
Tokyo Narita, Japan

Destination
Shanghai Pudong, China

Aircraft Registration
N852GT

Aircraft Type
Boeing 747-8

ICAO Type Designator
B748

A Polar Air Cargo Boeing 747-800, registration N852GT performing freight flight PO-213 from Tokyo Narita (Japan) to Shanghai Pudong (China), departed Narita's runway 16L (length 2500 meters/8200 feet), however, the main gear did not become airborne until about 85 meters/280 feet past the runway end already on soft ground. The aircraft climbed out to safety and continued to Shanghai without further incident.

On Aug 10th 2017 Japan's JTSB reported the scenario was consistent with a runway overrun and opened an investigation into the occurrence.

The airline reported a wrong thrust setting was being used and was recognized and corrected close to the runway end.

In Jun 2019 the JTSB released their final report concluding the probable causes of the serious incident were:

It is probable that in this serious incident, the aircraft commenced a take-off roll by using the take-off thrust lower than the thrust required for the Aircraft to take off, causing it to take a longer take-off roll distance to lift off; and its lifting off in the vicinity of the end of departure runway resulted in a case equivalent to runway overrun.

It is probable that the aircraft commenced a take-off roll by using the take-off thrust lower than the thrust required for the Aircraft to take off, because the Captain did not correctly change the FMC settings for the take-off thrust at the time of take-off from the runway different from what the Captain and the FO had assumed, the Captain did not correctly change the FMC settings for the take-off thrust, in addition, the Captain and the FO did not ensure to verify the take-off thrust by the time when they commenced the take-off.

The JTSB reported the captain (60, ATPL, 14,896 hours total, 5,178 hours on type) was pilot flying, the first officer (42, ATPL, 13,586 hours total, 408 hours on type) was pilot monitoring.

The crew had anticipated and prepared to depart from Narita's runway 16R (length 4,000 meters/13,123 feet) due to past experience of taking off runway 16R all times as well as the shorter taxi route to that runway.

The JTSB summarized the captain's testimony:

Assuming take-off from runway 16R, by using FMS CDU (hereinafter referred to as gCDUh) the Captain programed the FMC because he had been often instructed to use runway 16R as a take-off runway at the Airport, and the taxiing time from its parking spot to runway 16R was shorter.

The Captain thought that all procedures including the change in the FMC take-off data entries following the runway change would be able to be completed during taxiing to runway 16L because the parking spot was away from runway 16L, even if he was instructed to use runway 16L as its take-off runway.

In addition, as being aware of runway operation procedure (see 2.13 ) during the hour from 21:00 to 23:00 at the Airport, the Captain thought it would be possible to take off from in terms of the aircraft performance when the Aircraft at parking spot obtained a clearance and was instructed to use runway 16L as its take-off runway; therefore, the Captain did not request the Narita Delivery to issue a clearance for take-off from runway 16R, and was thinking of the complicated taxi route to runway 16L.

At the parking spot, before pushback, the Captain changed the settings related to the take-off by himself in the changes of FMC setting associating with the change from runway 16R to runway 16L, and instructed the FO to change the settings related to the flight route including standard departure procedures.

The Captain changed only Assumed temperature (see 2.9) without changing to the Rated Takeoff Trust (see 2.9) from the De-rated Take-off Thrust (see 2.9) that was selected on the THRUST LIM page (see 2.12. 1 (1)). After changing the necessary FMC settings associating with the runway change, the Captain briefed the taxi route to runway 16L and others, but did not brief on the verification of the take-off data by using CDU; and the Captain instructed the FO to verify that there was no discrepancy of take-off data between CDU and FDP (see 2.10).

The taxi route to runway 16L from the parking spot of the Aircraft was so long and complicated that the Captain carefully taxied the route instructed by ATC while checking it on a chart.

After gaining clearance for take-off from the Narita Tower and commencing a take-off roll from runway 16L, the Captain verified the engine indications of EICAS* 6 for the N1 values* 7; and he did not see the values as abnormal because they were almost the same with those for take-off from runway 16R.

During the take-off roll, the Captain felt that the acceleration of the Aircraft was a little slow but not so abnormal, while recognizing that they were approaching the end of departure runway, and performed the normal take-off procedures. The Captain felt that the Aircraft lifted off in the vicinity of the end of departure runway and the take-off path became lower, therefore, he discussed with the FO what he felt that the take-off path was low during the flight.

The JTSB reported that the aircraft continued the flight to destination and flew for more than 25 hours following the occurrence until the occurrence was discovered. It was therefore clear both CVR and FDR had been overwritten, both recorders were thus not removed from the aircraft.

According to the QAR recordings the aircraft passed the runway end at a radio altitude of 16 feet. The aircraft disconnected sensors for the intrusion detection system about 150 meters and 450 meters south of the runway end during departure, the aircraft did not receive any damage though.

The JTSB analysed:

It is highly probable that the FMC of the Aircraft was set before departure by assuming a takeoff from runway 16R. (See Figure 6 (1).)
After that, the Aircraft was instructed to take off from runway 16L at the time of receiving the clearance, therefore, the Captain as the PF had entered assumed temperature 38Ž after selecting the TO on the THRUST LIM page, and had to compare the N1 values to the FDP data. (See Figure 6 (2).)

However, as described in 2.1.2 (1), the Captain changed only Assumed temperature without changing to the TO from the TO2. In addition, as described in 2.12.1, if the Captain had selected the TO on the THRUST LIM page, any previously entered Assumed temperature would have been removed; however, it is somewhat likely that the Captain changed only Assumed temperature from 40Ž that remained on the display to 38Ž that is described in 16L of Table 1. Based on the above, it is probable that the take-off thrust set in FMC was D-TO2 (38) that was lower than D-TO (38) required for take-off from runway 16L. It is also probable that the Captain, afterwards, entered takeoff speeds corresponding to take-off flaps and D-TO (38). (See Figure 6 (3): Boxes highlighted in white indicate the take-off data that were not changed correctly).

In addition, it is probable that because FMC settings had not been changed correctly and the take-off thrust had been D-TO2 (38), as described in 2.11, at the time of take-off when the Aircraft reached the airspeed of 80 kt, the N1 value came to 89.1 % which was lower than 97.2 % required for take-off from runway 16L as shown in Table 1.

...

Judging from the facts that it was confirmed that the Aircraft had no damage during the aircraft inspection after arriving at the destination, and that only some of the intrusion warning sensors were disconnected, it is somewhat likely that the disconnection was caused not by the contact with the Aircraft but by its engine blast.

Metars:
RJAA 151430Z 17003KT 9999 FEW020 BKN/// 23/22 Q1012 NOSIG=
RJAA 151400Z 16004KT 9999 FEW020 BKN/// 23/22 Q1012 NOSIG RMK 1CU020 A2990=
RJAA 151330Z 17004KT 9999 FEW020 BKN/// 24/22 Q1012 NOSIG=
RJAA 151300Z 18004KT 9999 FEW020 BKN/// 24/22 Q1012 NOSIG=
RJAA 151230Z 18003KT 9999 FEW020 BKN/// 24/22 Q1012 NOSIG=
RJAA 151200Z 16002KT 9999 FEW020 BKN/// 24/22 Q1012 NOSIG=
RJAA 151130Z 18001KT 9999 FEW030 BKN/// 25/22 Q1012 NOSIG=
RJAA 151100Z 13003KT 9999 FEW030 BKN/// 25/22 Q1012 NOSIG=
RJAA 151030Z 14003KT 9999 FEW030 BKN/// 26/22 Q1012 NOSIG=
RJAA 151000Z 14004KT 9999 FEW030 BKN/// 26/22 Q1011 NOSIG=
RJAA 150930Z 14003KT 9999 FEW030 BKN/// 27/22 Q1011 NOSIG=