On the evening of 9 May 2015, a Boeing B737-8FE aircraft, registered VH-YID and operated by Virgin Australia Airlines Pty. Ltd. (Virgin), was on a scheduled passenger service from Sydney, New South Wales to Adelaide, South Australia. During a high-speed descent, the crew responded to aircraft indications that they were approaching airspeeds greater than desired by extending the speed brakes. While the speed brakes were still extended, the airspeed continued to increase towards the aircraft’s maximum speed, the result of which would have been an overspeed. In an effort to prevent an overspeed, the first officer overrode the autopilot by pulling back on the control column until the autopilot entered a secondary mode known as control wheel steering-pitch mode. This was followed immediately by an abrupt release of the control column, after which one cabin crew member sustained a minor injury.
What the ATSB found
The ATSB found that the crew selected a descent speed of 320 kt, which they routinely used for air traffic control-initiated high-speed descents. However, the increased risk of an overspeed in changing wind conditions had not been adequately considered by the crew.
The ATSB also found that, whereas Virgin’s training included a focus on the management of overspeeds, the crew had not yet completed this training. This increased the risk that the guidance provided through other sources would not be followed correctly.
Additionally, the flight crew had initiated the cabin preparation for landing earlier than usual due to the expectation of turbulence later in the descent. This likely reduced the risk of more serious injury to the cabin crew as they were in the final stages of securing the cabin than had they commenced preparations for landing at the normal time.
What's been done as a result
Prior to this occurrence, Virgin had implemented improved crew training and guidance on managing overspeeds. This included the addition of a cyclic simulator training session that focused on overspeed management on descent.
This occurrence highlights the increased risk of overspeed when conducting high-speed descents in conditions of varying winds and any associated turbulence. Identifying and discussing the risks associated with high-speed descent increases the likelihood that crew will select a lower descent speed and/or consider the best way to deal with an impending aircraft overspeed before the descent is initiated.
On the evening of 9 May 2015, a Boeing B737-8FE aircraft, registered VH-YID and operated by Virgin Australia Airlines Pty. Ltd. (Virgin) as ‘Velocity 436’, was on a scheduled passenger service from Sydney, New South Wales to Adelaide, South Australia. The first officer (FO) was the pilot flying.1
At about 1915 Central Standard Time2, the flight crew conducted an approach briefing in preparation for the descent. The crew discussed the ‘fairly strong’ westerly wind forecast in the area that would affect their descent, and decided they would instruct the cabin crew to prepare the cabin for landing at about flight level (FL) 2703. This was reported slightly earlier than usual and was intended to reduce the risk of turbulence-related injuries. Soon after, air traffic control (ATC) requested the crew to conduct a high-speed descent into Adelaide, which the crew accepted. The FO changed the planned descent speed from 280 kt to 320 kt, which was reported by the crew to be routinely used during ATC-initiated high-speed descents.
The descent was commenced about 4 minutes later and, passing about FL 270, the captain made the public address ‘cabin crew prepare for landing’. This was the cue for the cabin crew to commence the cabin preparation procedure by securing all loose cabin equipment, ensuring passengers were in their seats with seatbelts fastened and then securing themselves in their jump seats. Seven minutes later the flight crew switched the seatbelt signs on, as a cue to the cabin crew to finalise this procedure and take their seats.
During the descent the FO controlled the aircraft’s vertical profile using various autopilot descent modes. Passing through FL 250, the FO selected the VNAV PATH mode (see the section titled VNAV PATH descent mode).
As the aircraft passed through 10,000 ft, the airspeed started to increase above 320 kt and at 8,400 ft the message ‘drag required’ displayed on the flight management computer (FMC) scratchpad as per system design. In response, the FO extended the speed brake, making an effort to do so slowly and smoothly. The captain observed this action, initially concerned that the FO may extend the speed brake too quickly. However, the captain was satisfied that the speed brake was appropriately extended and that the increasing airspeed trend was being managed. The captain’s attention was then turned to other tasks. Over the next 6 seconds the airspeed continued to increase, and the FO recalled seeing the speed trend vector on the primary flight display (PFD) extend beyond the aircraft’s maximum certified speed (VMO).4
In an effort to avoid an overspeed, the FO pulled back forcibly on the control column in order to raise the nose, overriding the autopilot and activating the control wheel steering – pitch (CWS-P). This technique was routinely used to manage overspeeds on descent. On this occasion the FO reported feeling greater than usual resistance when raising the nose due to the already low pitch angle. The FO then recalled feeling a ‘pinch’ as the autopilot reverted to CWS-P, along with a sudden pitch change and a high g loading5 on the aircraft. In response, the FO abruptly released the back pressure on the control column, rapidly unloading the g loading. At that moment, the cabin crew had almost completed securing the cabin and were about to take their seats. They reported experiencing what they thought to be turbulence, and two cabin crew in the rear galley lost their footing. This resulted in one of the cabin crew impacting the galley floor heavily, sustaining an injury to their knee. The FO stowed the speed brake, selected the LVL CHG (level change) descent mode and reduced the selected speed to 250 kt. The captain checked the status of the cabin crew and was informed of the injury to the crew member..
Sequence of events on descent, including the flight and cabin crew actions plotted against altitude and time (ATSB Report)
The forecast meteorological information that was available to the crew included:
• an Aerodrome Forecast for arrival into Adelaide that predicted a light south-westerly wind and showers of rain with intermittent periods of less than 30 minutes of south-westerly winds at 20 kt, gusting up to 30 kt
• no SIGMETs7 that were relevant to this flight
• winds from 264 °M at 54 kt at FL 350, from 250 °M at 38 kt at FL 250, from 258°M at 40 kt at FL 150 and from 270° at 41 kt at 5,000ft.
The flight data was examined to derive the aircraft’s computed airspeed (CAS) and the wind affecting the descent as they pertain to the crew’s actions. This examination showed the following:
• By 10,468 ft in the descent the aircraft reached the selected descent speed of 320 kt. Over the next 30 seconds, as the aircraft descended to 8,136 ft (when the speed brake was extended) the headwind increased from 30 to 39 kt, and the airspeed increased from 320 to 333 kt.
• Over the next 5 seconds, as the aircraft descended from 8,136 ft to 7,790 ft, the headwind increased from 39 to 48 kt, and the airspeed increased from 333 to a peak of 339 kt. At this point the back pressure on the FO’s control column was recorded as 41.5 lb, and CWS-P engaged with a peak g loading of 2.14 g.
• One second later the back pressure on the FO’s control column reduced to 2 lb, consistent with a release of the back pressure, with a minimum g loading of 0.066.
• The speed brake was stowed about 10 seconds later at 7,492 ft.
Overspeed occurrences on descent
Between 2012 and 2015, 51 occurrence reports were submitted to the ATSB that involved an overspeed in a B737 aircraft on descent. Of these, two occurred during high-speed descents. Of the other occurrences, three reported the use of CWS-P This included investigation AO-2012-138 below.
ATSB investigation AO-2012-138
The crew of a Boeing 737-800 aircraft was conducting a flight from Adelaide, South Australia to Canberra, Australian Capital Territory. Just prior to commencing descent, ATC cleared the aircraft to conduct a high-speed descent. The aircraft descended below the 7,000 ft altitude clearance limit and, after being alerted to this by ATC, the flight crew climbed the aircraft back to 7,000 ft and continued the approach to land. The ATSB found that the combination of auto-flight system mode changes (including inadvertent use of CWS-P) and the management of the airspeed during the descent resulted in a high workload environment. This occurrence highlighted the need to continually monitor descent profiles and airspace limitations in relation to the aircraft’s position.
High-speed descent and potential overspeed
As part of their consideration of the effect of the ‘fairly strong’ winds on the approach, the flight crew discussed the possible effect of turbulence on cabin safety. Similarly, the crew considered the aircraft’s turbulence penetration speed when accepting the high-speed descent. However, the captain recalled that whilst they were aware of changing wind conditions, there was no forecast of severe turbulence, and therefore a descent at 320 kt was considered by the crew to be appropriate. Although likely influenced by the routine use of 320 kt for high-speed descents, the descent at that speed, when the maximum certified limiting speed was 340 kt, increased the risk of an overspeed.
In considering the influences on crew decision making, Orasanu (2010) stated:
What constitutes an appropriate course of action depends on the affordances of the situation. Sometimes a single response is prescribed in company manuals or procedures. At other times, multiple options exist from which one must be selected.
Poor decisions may…arise when a flight crew is aware of conditions that require a decision, but underestimates the level of risk associated with the conditions…Another arises from pilots’ routine experience. If similar…situations have been encountered in the past and a particular course of action has succeeded, the crew will expect to succeed the next time with the same response.
Likewise, Sitkin (1992) as cited in Orasanu (2010) stated that uniformly positive experiences provide no baseline by which to determine when a situation is becoming more dangerous.
The crew reported that selecting 320 kt for a high-speed descent was routine, indicating that this course of action was expected to be successful. Therefore, the likelihood that the crew would consider the risks of an overspeed in this case were harder to identify.
The Flight Safety Foundation (2014) recommended that the pilot monitoring role should include monitoring the aircraft’s flight path and immediately bringing any concern to the pilot flying’s attention. In this case, the captain was monitoring the aircraft’s speed, before focussing on the first officer (FO) as he extended the speed brake and then other operational tasks associated with the descent. Whilst this precluded the captain’s ability to detect the FO’s reaction to an increasing speed trend vector, it was reasonable that the captain felt that the situation was under control.
Orasanu (2010) outlined that the development of expertise contributes to decision making in different ways. This included the development of ‘stored condition-action patterns’, where decision makers interpret a cue pattern as being of a particular type and match it with an action according to a routine (Klein, 1989 and 1993 cited in Orasanu 2010).
In this case, the FO identified an immediate need to prevent an overspeed, and did so by pulling back on the control column and activating control wheel steering-pitch (CWS-P). This had previously been successful for the FO in addressing an impending overspeed, but the difference in this case was that the force required was larger than experienced by the FO in past situations. The flight data recorded a 41 lb back pressure on the FO’s control column with a resulting 2.14 g loading on reversion to CWS-P. Given the altitude at the time, reversion to CWS-P was considered contrary to Virgin Australia Pty Ltd’s (Virgin) guidance and training in overspeed management. However, the use of the CWS-P mode was reported common among some pilots and, in this case, had possibly become a stored condition-action pattern.
Surprise is a cognitive-emotional response to something unexpected. It results from a mismatch between the individual’s mental expectations and what actually happens around them. Experiencing surprise is a combination of physiological, cognitive and behavioural responses (Rivera and others 2014). If a pilot is not expecting things to go wrong, then the level of surprise can result in taking no action, or the wrong action (Martin 2012).In this case, the FO reported feeling a ‘pinch’ when CWS-P activated. This, combined with feeling a high g loading, led the FO to abruptly release the amount of back pressure 1 second after its application.
A combination of a sudden increase, followed by a sudden decrease in g loading would have first pushed the occupants of the aircraft towards the floor, followed shortly after by a feeling of weightlessness. This would have increased the difficulty of moving around the cabin and the risk of injury.
Flight management computer data entry procedures
In this occurrence the QNH variation, temperature deviations and descent winds were insufficient to contribute to an inaccurate VNAV PATH construction. In addition, at the time of the occurrence there was no procedural requirement after pre-flight to enter QNH and temperature deviation data, nor to update descent wind data in the descent forecast page of the flight management computer. There was also no guidance on the benefits of entering that data into the computer to produce a more accurate calculated vertical flight path. In some circumstances, the use of pre-flight data would reduce the accuracy of the calculated vertical flight path and result in increased crew workload in managing the energy state of the aircraft.
Training in the management of overspeeds
At the time of the occurrence, Virgin’s Flight Crew Training Manual and other guidance material included information on how to effectively handle an impending overspeed. However, the simulator training session in which crews practiced overspeed management (including on descent) had not been completed by the crew. This simulator session was subsequently completed by all Virgin B737 flight crew, and additional guidance has also been provided.
Early preparation of the cabin for landing
The procedures for preparing the cabin for landing required their commencement no later than FL 200. In this case, the flight crew identified that turbulence on descent may increase the risk of injury to crew and passengers and therefore decided to initiate the cabin preparation at about FL 270. Flight data for the descent showed that this provided an additional 2 minutes before the occurrence for the crew to secure galley equipment and prepare the cabin, including seating all passengers. This reduced the risk of more serious and numerous injuries to the cabin crew and passengers.
From the evidence available, the following findings are made with respect to the flight path management occurrence involving a Virgin Australia Airlines Pty. Ltd. B737-8FE, registered VH-YID, on descent into Adelaide Airport, South Australia on 9 May 2015. These findings should not be read as apportioning blame or liability to any particular organisation or individual.
• Although routinely used for high-speed descents, the selected speed for the descent of 320 kt was close the aircraft’s maximum speed of 340 kt and, in combination with the increased headwind as the aircraft descended through about 8,000 ft, increased the risk of an overspeed.
• After identifying the unexpected rapid increase in airspeed, the first officer applied sufficient back pressure on the control column to override the autopilot with the intention of avoiding the impending overspeed. This increased the g loading on the aircraft to an extent where safely walking in the cabin would be difficult.
• In response to feeling the increased g loading, the first officer abruptly released the back pressure on the control column, resulting in a sudden pitch change and reduced g loading that led to the cabin crew member losing their footing and sustaining a knee injury.
Other factors that increased risk
• The Virgin Australia Airlines Pty. Ltd. procedures did not require flight crew to update the QNH, temperature deviation or the descent winds in the flight management computer after departure. In some circumstances, this would reduce the accuracy of the calculated flight path and increase crew workload in managing the aircraft’s energy state.
• At the time of the occurrence, the Virgin Australia Airlines Pty. Ltd. training included a focus on the management of overspeeds. However, the associated simulator training session had not been completed by the crew, increasing the risk that they would not react in accordance with the published operational guidance material.
• The earlier-than-usual preparation of the cabin for landing meant that, at the time of the occurrence, the cabin crew were in the final stages of that activity. This reduced the risk of more serious and numerous injuries to the cabin crew and passengers.
Data Source: Australian Transport Safety Bureau
Full report here