Jet2 B738 at Kuusamo on Dec 1st 2021, takeoff with insufficient takeoff power

Last Update: October 6, 2022 / 19:52:05 GMT/Zulu time

Bookmark this article

Photo of Jet2.com G-JZHL, Boeing 737-800 — Jet2.com G-JZHL, Boeing 737-800 (Photo credit: cvtperson / Flickr / License: CC by-sa)

Incident Facts

Date of incident
Dec 1, 2021

Classification
Report

Airline
Jet2.com

Flight number
LS-2152

Departure
Kuusamo, Finland

Destination
London Stansted, United Kingdom

Aircraft Registration
G-JZHL

Aircraft Type
Boeing 737-800

ICAO Type Designator
B738

A Jet2.com Boeing 737-800, registration G-JZHL performing flight LS-2152 from Kuusamo (Finland) to London Stansted,EN (UK) with no passengers and 6 crew, was departing Kuusamo's runway 12 and became airborne just 400 meters/1400 feet short of the runway end and climbed out slowly. Subsequently the crew discovered they had left the thrust levers at engine run up power (70% N1) rather than takeoff power and corrected the power setting. The aircraft continued to Stansted without further incident.

The English AAIB was delegated the investigation by Finlands Onnettomuustutkintakeskus and released their final report concluding the probable causes of the serious incident were:

The aircraft took off with insufficient thrust set because the TOGA button was not pressed. It was not pressed because the co-pilot was startled by the aircraft moving as he commenced the run-up against the brakes. The aircraft started to move because insufficient brake pressure was applied. Human checks designed to detect the insufficient thrust were ineffective because both pilots were attending to other tasks. The commander was responding to a radio call from the FISO during the start of the takeoff roll. Neither pilot detected the low thrust until after the aircraft was airborne.

The AAIB analysed:

During a pre-takeoff engine run-up the thrust was increased to 70% N1
as required by the supplementary procedure. However, the thrust was not subsequently increased to the required 89% N1 and remained at 70% throughout the takeoff roll. This occurred because the co-pilot did not press the TOGA button and neither pilot checked the thrust was correctly set during the takeoff roll. Several factors contributed to these omissions, which are discussed below. The low thrust was not detected until after the aircraft was airborne.

The aircraft became airborne with 400 m of runway remaining. If the aircraft had suffered an engine failure after V1 it would not have been able to safely get airborne with the thrust
of 70% N1 on the operative engine.

TOGA button The co-pilot reported that he omitted to press the TOGA button because he was startled by the aircraft starting to slide and drift toward the snowbanks.

It is likely that the aircraft was rolling rather than sliding as the co-pilot was applying insufficient brake pressure to hold the aircraft stationary against the 70% N1. Co-pilots at this operator do not taxi the aircraft so are rarely required to use significant brake pressure.

This meant the co-pilot did not have any experience of applying significant brake pressure.

During simulator training after the incident, the co-pilot discovered the brake pedals have significantly more travel than he had been using.

After the incident the operator reviewed their FDM data and discovered several previous events where insufficient brake pressure had been applied during engine run-ups. The operator took safety action to alert pilots to the issue and is using FDM data to monitor further trends.

Whilst the subsequent analysis showed the aircraft was rolling, in the moment the co-pilot’s perception was that the aircraft was sliding. The pilots had briefed that the aircraft might slide and that, if it did, the co-pilot would release the brakes and continue the takeoff. So, when they perceived that the aircraft was sliding, they did as they planned and the co-pilot released the brakes. It is possible that because they were primed that the aircraft may slide, when it started to move, they were more likely to think it was sliding rather than considering insufficient braking. Briefing what may happen is generally very helpful but, in this case, it may have primed the crew to expect a particular outcome.

The co-pilot reported that he was startled by how readily the aircraft started to slide and by the aircraft starting to drift towards the snowbanks. A startle response can be defined as ‘a complex, involuntary reaction to a sudden unanticipated stimulus’. It is a ‘brief, fast and highly physiological reaction to a sudden, intense or threatening stimulus’. He perceived that the threat was the aircraft sliding towards the side of the runway and his attention was drawn to controlling the aircraft. NASA’s technical memorandum titled ‘Effects of acute stress of aircrew performance’ describes how a threatening stimulus can cause a pilot to focus their attention on addressing that threat and can lead to errors or omissions in other concurrent tasks. It is likely that, in the moment of the startle, suddenly faced with a threatening situation, the co-pilot’s attention was solely drawn to controlling the aircraft and this caused him to omit to press the TOGA button.

The operator’s SOPs require the co-pilot to remove their hand from the thrust levers ‘immediately’ after TOGA is pressed (the operator had added an additional note to the manufacturer’s standard takeoff procedure to state exactly when the co-pilot’s hand should be removed). This normally occurs at the start of the takeoff roll as the aircraft starts to accelerate, so the co-pilot would be used to having both hands on the control column as the aircraft travels down the runway. During this incident, as the aircraft started to move, with his attention focused on controlling the aircraft, it would have felt natural to move his hand from the throttles to the control column.

‘Thrust set’ check

The takeoff procedure requires both pilots to check that the correct takeoff thrust is set.

Once the pilot monitoring has checked the thrust, they are required to call ‘thrust set’.

During a normal takeoff, with the co-pilot handling, the commander will place their hand on the thrust lever as soon as the co-pilot presses the TOGA button. They can then watch the N1 gauges increasing until the actual N1 matches the target and then call ‘thrust set’.

However, during the incident takeoff, this normal sequence was interrupted.

As the co-pilot started to advance the thrust levers to 70% N1, the commander was expecting a 30 second stationary run-up. He had the secondary engine instruments displayed and was ready to start his timer. As the aircraft unexpectedly started to move, he had to clear the secondary engine instruments, start his timer and place his hand on the thrust levers. As this was happening, the commander made a radio transmission (as the aircraft accelerated though 6 kt) which was probably him confirming they were taking off. The FISO then made a further transmission to the aircraft asking them to confirm if they would be turning right after departure. By the time the commander had replied to the FISO the aircraft was passing 53 kt. The commander reported that it was this distraction that caused him to omit the thrust set check and call.

When an action is normally cued by a sequence of preceding events, if those events are changed or if a distraction occurs during those events, people are vulnerable to omissions.

Having missed the action, with the normal cues now passed, it is unusual for a person to remember to return to the omitted item, particularly in a time limited situation. A report published in Aero Safety World in December 200816 discussed how common this is and how it has caused previous accidents.

Having omitted the thrust set check after answering the radio call, the commander had seven seconds until the aircraft passed 80 kt, a further five seconds until V1 then a further 25 seconds before VR. During the takeoff roll the pilot monitoring would normally be monitoring the aircraft for any abnormal indications. There were several indications on the flight deck which might have alerted him that the thrust was not set correctly and that TOGA had not been pressed. The needles on the N1 gauges would not have been aligned with the target bugs and the digits would have been different from the target digits (Figure 8). With hindsight these indications may seem obvious, but it is common for humans to see what they expect to see or ‘look without seeing’ (Footnote 8). This is more common with an indication which is normally correct. Experienced pilots who have seen hundreds of takeoffs will nearly always have seen the thrust correctly set. On the one occasion when it is not set, it is possible that they will not see it, they will just see what they expect to see.

The PFD also had indications to tell the pilots that something was abnormal. The FMAs at the top of the PFD were subtly different to a normal takeoff and the flight directors were not displayed. There is no requirement for the pilots to check the FMAs during the takeoff roll so they may not have looked at them. Even if they did notice the FMAs were abnormal or saw the lack of flight directors, it may not have been immediately obvious why. The pilots had not seen a takeoff without TOGA being pressed before so it may have taken some time to understand why the indications were abnormal. The length of the trend arrow on the speed tape would also have been smaller than on a normal takeoff, but there is no requirement to check the trend arrow on the takeoff roll and this difference from normal may not have been sufficient to be detected. The trend arrow is also dynamically calculated and therefore can fluctuate significantly with wind gusts on the takeoff roll making it hard to interpret.

The co-pilot is also required to check the thrust is correctly set during the takeoff roll.

However, having been startled by the unexpected aircraft movement, his focus was on controlling the aircraft. His visual attention would have been outside the cockpit ensuring the aircraft was tracking down the runway centreline. It is likely that the effect of the startle and with his attention captured by controlling the aircraft, the co-pilot did not have the capacity to check the thrust.

Radio transmission during the takeoff roll.

During the takeoff roll the FISO made a transmission to the aircraft which distracted the commander. The FISO was expecting the flight crew to confirm their routing when they reported they were taking off. When the flight crew did not do this, the FISO felt she was required to ask for confirmation before the aircraft took off.

From the flight crew’s perspective, they had been cleared to route via a waypoint to the south of the airport and their charts included an instruction to turn right. Therefore, they intended to turn right but were not aware of any requirement for them to report this.

From the FISO’s perspective, there is no clearance between the airport and the first waypoint so an aircraft’s commander could take any routing at their discretion. The rules of the air state that aircraft should normally turn left after takeoff, but if they intended to turn right, they must report this intention. The requirement to do this is also stated in the Finnish AIP.

Therefore, the FISO expected the commander to report his intention to turn right.

The FISO felt she must obtain this confirmation before the aircraft was airborne because:
1. Kuusamo does not have radar, so once the aircraft was airborne, with limited visibility at night and in snow, the FISO would not have been able to determine the aircraft’s routing.
2. Once the aircraft is airborne the FISO is not able to issue any further
instructions to aircraft.
3. The FISO needed to confirm the restricted area to the east of the airport
along the FIR boundary would not be infringed.

The FISO was aware of the requirement not to call an aircraft during the takeoff, but because of the proximity of the restricted area she believed she was required to obtain confirmation of the intended turn direction.

The operator subsequently included in their OM C the requirement to report direction of turn to the FISO for all Finnish airports.

When a radio transmission is made to an aircraft on a takeoff roll it is difficult for flight crew to ignore the message. The message could contain vital information (for example, informing the commander of smoke coming from the aircraft, or a blocked runway), so the crew must listen to the message and understand what they are being told. It was therefore difficult for the commander, on this takeoff, to avoid being distracted by the radio transmission.

Lack of acceleration

Both pilots reported that something was not right during the takeoff but, at the time, neither could resolve what was wrong. They felt the acceleration and the cadence of the takeoff was slightly slow but thought this might be due to the runway contamination. The feeling was not compelling enough for them to abort the takeoff.

This is in common with many of the previous takeoff performance incidents that the AAIB and other SIAs have investigated. Human perception of acceleration in combination with the nature of the takeoff task means that pilots are generally not able to recognise when the acceleration is slower than required, even when the difference in acceleration is significant.

Takeoff Acceleration Monitoring Systems

During this incident, the takeoff performance was correctly calculated and correctly loaded into the FMC. The incident occurred because the planned takeoff thrust was not set. There is a barrier in place to detect this error, in the form of a human check, but this incident shows this check is vulnerable to distraction.

The AAIB and other SIAs have investigated many takeoff performance incidents which have resulted in aircraft taking off with insufficient thrust. The circumstances of each incident differ but the outcome is the same. The human checks currently in place do not always stop these incidents occurring. Learning from past events and research shows that, whilst they are effective in many cases, such checks are occasionally omitted or fail to detect errors.

Operational interventions maximise crew performance as far as possible but there is a limit to the reliability that can be achieved with any human task. Higher levels of reliability are likely to require a technological intervention. TAMS could detect these events and alert the flight crew at a low speed and enable them to safely reject the takeoff.