TUI B738 at Leeds on Oct 20th 2023, runway excursion on landing

Last Update: October 24, 2024 / 12:03:50 GMT/Zulu time

Bookmark this article

Photo of TUI Airways G-TAWD, Boeing 737-800 — TUI Airways G-TAWD, Boeing 737-800 (Photo credit: MercerMJ / Flickr / License: CC by-sa)

Incident Facts

Date of incident
Oct 20, 2023

Classification
Incident

Cause
Runway excursion on landing

Airline
TUI Airways

Flight number
BY3551

Departure
Corfu, Greece

Destination
Leeds, United Kingdom

Aircraft Registration
G-TAWD

Aircraft Type
Boeing 737-800

ICAO Type Designator
B738

Airport
Leeds Bradford Airport, Leeds

Airport ICAO Code
EGNM

A TUI Airways Boeing 737-800, registration G-TAWD performing flight BY3551 from Corfu (Greece) to Leeds,EN (UK), landed on Leeds' runway 14 at about 13:51L (12:51Z) but veered left off the runway in the final stages of the roll out and came to a stop on soft ground about 18 meters/60 feet left of the runway edge and about 410 meters/1340 feet short of the runway end. There were no injuries, the aircraft sustained minor if any damage.

The airline said: "We are aware of an incident at Leeds Bradford Airport this afternoon, in which upon landing flight TOM3551 slightly veered off while turning into the taxiway. There are no reported injuries, and our ground team are on hand to support passengers as they disembark."

On Oct 24th 2024 the AAIB released their final report concluding the probable causes of the serious incident were:

The initial phase of the landing roll was normal, with the aircraft touching down in the touchdown zone and meeting the target deceleration rate. The PF disconnected the autobrake and then stowed the reverse thrust early in the landing rollout and a constant deceleration was not maintained.

The deviation from the centreline, resulting from the strong crosswind from the left, required more right rudder input than was applied, in order to correct it. Additional use of differential braking to assist was also available. There was an unusual juddering from the nosewheel reported by the crew likely resulting from the failure of a nosewheel bearing. There was no mechanical defect identified by the investigation which would have prevented the crew from applying the additional right rudder that was available to keep the aircraft on the runway.

However, the crew’s actions may have been influenced by the nosewheel juddering.

The AAIB analysed:

Approach preparation

The crew carried out a thorough briefing of the expected approach at Leeds, the weather conditions and were well prepared for a diversion to Manchester Airport.

The landing performance was calculated and within limits to land at Leeds with a margin of 91 m when using auto brake max. The crew referred to limiting landing performance which prevented them from commencing an approach when the wind was from 060° at 19 kt, which was the basis for the pilot’s request to join the hold whilst on initial approach to Leeds. This performance landing limit could not be recreated by the investigation using the manufacturer’s performance tool and the reported environmental conditions.

Landing roll

The aircraft touched down in the touchdown zone and the initial phase of the rollout was normal, with the aircraft reaching the expected deceleration rate. After the PF cancelled the autobrake, there was a period of eight seconds with no wheel braking applied, during which the reverse thrust was also cancelled. The guidance on braking in the operator OMB and FCTM stated that pilots should ‘maintain deceleration rate with constant or increasing brake pressure as required until stopped or desired taxi speed is reached’ and ‘do not release the brake pedal pressure until the airplane speed has been reduced to a safe taxi speed’. The reduction in deceleration rate caused by disconnecting the autobrake and cancelling reverse thrust on a wet runway, with a significant crosswind, invalidated the landing performance calculated by the crew. However, there is no evidence that, as a result, the aircraft did not have sufficient stopping distance available or that the crew assessed this to be the case.

Runway surface

The ‘cleaned’ runway tyre markings are typically only associated with reverted rubber hydroplaning. However, no evidence of reverted rubber was found on the mainwheel or nosewheel tyres which requires locked wheels to occur, nor were there corresponding ‘melted’ rubber deposits on the runway. The runway marks suggest that the surface temperature underneath all the tyres was high enough to turn some of the water on the runway’s surface into steam but not high enough, nor combined with locked wheels, to cause rubber reversion. The aircraft deceleration rate and braking forces applied confirms that the aircraft was not friction-limited at any point and hydroplaning was not a contributory factor in this event.

The nosewheel tyre damage is consistent with abrasion scrubbing from being turned to an angle nearly perpendicular to the direction of travel whilst the tyres were rotating and in contact with the runway; also demonstrated by the runway marks left by the nosewheel tyres. The angle of the nosewheels could only be achieved by use of the tiller rather than rudder pedal steering alone, and this corresponds with the commander’s account of events.

Nosewheel bearing

It could not be confirmed when the nosewheel bearing began to fail, but it is likely that it suffered catastrophic failure during the rollout at LBA and was contributory to the vibration through the rudder pedals as felt by the commander. The failure of the bearing did not prevent the nosewheel from rotating nor did it affect the ability to steer the nosewheels.

At speeds above 60 kt the rudder is the primary method of yaw control. As the aircraft decelerates towards taxi speed the nosewheel steering becomes the primary control. There was no evidence that the bearing’s failure physically reduced or restricted the pilot’s ability to control the aircraft in yaw.

The PF recalled the nosewheel judder increasing as he increased the right rudder pedal, which caused him to reduce the input. Both crew members recalled the judder as being significant and unusual and the CVR recording contained a ‘juddering/rattling’ sound during this period. There was no evidence of a physical restriction of the rudder, rudder pedal movement or that the rudder position required to prevent the runway excursion was unobtainable.

Aircraft handling

The guidance for pilots on handling deviations from the runway centreline in high crosswinds and on slippery runways is to release the wheel brakes and cancel reverse thrust, but using this technique will increase the LDR. The runway was wet but there is no evidence that the aircraft skidded at any point, nor did the aircraft drift down wind of the runway centreline during the landing roll. It may have been that the pilot cancelled the reverse thrust earlier than normal SOP in an effort to apply a known procedure (Figure 11) in response to the nosewheel judder.

In this event, there was an adequate, but not significant, margin between the landing distance required and that which was available. The recorded data shows there was a significant amount of right rudder input available to the pilot which was not used. The simulation calculated by the manufacturer suggests that the available rudder would have been sufficient to correct the deviation and prevent the runway excursion without cancelling the autobrake and cancelling reverse thrust.

Based on analysis of the data from the event, the manufacturer stated: ‘Throughout the ground roll, there was additional directional control authority at the flight crew’s disposal, via rudder and asymmetric wheel braking, that was not utilised.’

Although a nosewheel judder would not necessarily be considered an ‘intense stimulus’, if it was perceived by the PF as such, it is possible he momentarily experienced a startle reflex. This would explain an interruption of the task of increasing the rudder input to prevent the deviation from the centreline. However, given that maintaining the aircraft’s nose on the centreline is a routine, simple and instinctive task, the interruption was likely to have ranged between 100 ms to three seconds. The rollout from touchdown until the aircraft stopped lasted just under 30 seconds, whilst the deviation from the centreline occurred approximately 18 seconds before the aircraft came to a stop. Therefore, it is more likely that, if he experienced a physiological response, that it was as a result of surprise, the effects of which typically last longer than startle.

Pilots are trained to deal with non-normal and emergency events and the aircraft deviation from the runway centreline may have been considered a challenging situation by the crew.

There was limited time for the crew to assess the cause of the judder and the practical impact it had on the directional control of the aircraft and surprise or startle may have been a factor. However, as the aircraft approached the side of the runway, it is not clear why the PF did not attempt to use all right rudder available, in spite of the judder, to prevent the runway excursion.

Related NOTAMs:
C5930/23 NOTAMN
Q) EGTT/QICAS/I /NBO/A /000/999/5352N00140W005
A) EGNM B) 2310201425 C) 2310202359
E) ILS RWY 32 U/S

C5926/23 NOTAMN
Q) EGTT/QFALC/IV/NBO/A /000/999/5352N00140W005
A) EGNM B) 2310201415 C) 2310202359
E) AD CLOSED

Metars:
EGNM 201350Z 07017G30KT 030V090 4000 -RA SCT003 BKN006 09/09 Q0981=
EGNM 201320Z 07018G31KT 4000 -RA BR SCT004 BKN005 09/09 Q0981=
EGNM 201250Z 07017G32KT 4000 -RA BR SCT004 BKN006 10/09 Q0981=
EGNM 201220Z 07019G33KT 4000 -RA BR SCT004 BKN007 10/09 Q0981=
EGNM 201150Z 07020G33KT 6000 -RA SCT005 BKN008 10/09 Q0981=
EGNM 201120Z 07021G35KT 6000 -RA SCT006 BKN007 10/09 Q0981=
EGNM 201050Z 07020G33KT 6000 -RA SCT005 BKN006 10/09 Q0981=
EGNM 201020Z 07020G33KT 6000 -RA SCT004 BKN005 10/10 Q0981=
EGNM 200950Z 07020G32KT 030V090 4000 -RA BR SCT004 BKN005 10/10 Q0981=
EGNM 200920Z 07017KT 4000 -RA BR SCT004 BKN005 10/10 Q0981=
EGNM 200850Z 07019G33KT 4000 -RA BR SCT003 BKN005 11/10 Q0981=