Old Facts, New Insights – Lessons from A-320 Part 2

The details of the automation surprises faced by the pilots participating in the questionnaire survey is presented hereafter [2].

  • Failure to activate the approach. By “activating the approach” (a single button push) the pilot arms the automated cockpit system to manoeuvre the aircraft to chosen approach configuration. In case the “pilot forgets to activate the approach before engaging managed speed”, instead of slowing the aircraft, automation instead increases thrust “to return to the last active target speed” which is most likely too high for the desired and chosen phase of approach.
  • Loss of constraints after entering change. These were the situations where pilots were assigned revised approach on a new runway by ATC after the data for the originally assigned approach and runway was entered by the crew into the multi-function control and display unit (MCDU). Once the change in runway is entered in MCDU, automation deletes all previously fed data, including speed and altitude restrictions which may still apply. For the pilot, only one parameter of approach required a change (preserving other settings and constraints) but automation interprets it as a requirement for defining an entirely new approach.
  • Indirect mode transitions. Here automation changes its behaviour without any command from the pilot, which at times may be contrary to the intent. Many times such occurrences may be to ensure that the commands of the crew do not conflict with the designed-defined safe flight envelope.
  • Exceeding an airspeed of 250 knots below 10000 feet. Federal Aviation Regulation Section 91.117 stipulates slower speeds to 250 knots during descent through 10000 feet MSL [6]. Automation does not caution pilots about such violations of regulations if the pilot selects air speeds more than 250 knots while descending through 10000 feet MSL.
  • Automation strategies in managed vertical navigation. Automation may allow little deviations from its target altitude to ensure target airspeed, by minimising thrust variations, as a design requirement for better fuel efficiency. This is unlike manual flying.
  • Failure to immediately detect a failure of the flight management and guidance computer. There were odd situations of only one or a subset of possible indications were available, with corresponding autopilot and flight director not engaged. This happens when failure of the single flight management and guidance computer (FMGC) leads to loss of redundancy, where an “input from either MCDU is sent to the one remaining operational computer, and any entry on one MCDU is transferred to both MCDU”. But information can no longer be exchanged and cross-checked between the two FMGCs. Instead of an aural alert, helping to focus the attention of crew, this degraded mode of operations may appear as different indications in different cockpit locations, which may even go unnoticed at high workload situations in the cockpit.
  • Unexpected airspeeds during a go-around. A go-around at 100 feet without flight director guidance has caught the crew by surprise when “advancing the thrust levers to the takeoff/go-around position does not automatically arm the autothrust system”, thus failing to maintain appropriate airspeeds.
  • Decrease in airspeed when leveling off in the “open descent” mode. In an open descent mode, when the aircraft was flown manually with at least one flight director on, if the pilot failed to follow the flight director bars, aircraft continues decelerating till V1 (Critical engine failure recognition speed). Only at this point does the autothrust system revert to the speed mode so also the vertical speed mode.

As far as flight controls are concerned, thrust levers and sidesticks in A-320 do not move under automation, unlike some other aircraft. The feedback from pilots revealed the following [2]:-

  • Nonmoving thrust levers eliminate one source of feedback on automation behaviour.  Most (n = 85) of the pilots adapted to the nonmoving thrust levers. But 77 pilots felt that such design increases the monitoring load, where to compensate for the absent throttle movement, they need to pay more attention to the N1 indications on the centralised aircraft monitoring system, engine noise or airspeed indications on the PFD. In fact a smaller group (n = 30) would prefer moving throttles specifically while flying at low altitude, especially during final approach, or instances when automation reduces power to idle or near idle. This is to compensate for visual attention which is outside the cockpit during final approach rather than on engine indications. Five pilots were concerned about disconnect procedure for the auto thrust system, it being cumbersome.
  • Uncoupled sidesticks eliminated one source of feedback on pilot input and performance.  Controlling the aircraft using sidesticks, instead of a yoke, was a novel concept introduced in A-320, which most of the pilots found easy to adapt. But 10 pilots voiced concerns about the lack of coupling of the two sidesticks in the cockpit. They felt that this deny them opportunity to monitor performance of the other pilot during manual flying, particularly during situations like strong crosswinds conditions or flying close to ground. Another concern raised was that simultaneous but uncoordinated inputs by both pilots is combined, which may at times aggravate (rather than resolve) the situation.

Crash of Indian Airlines flight 605 on 14 Feb 1990, though blamed on human error, did have elements of doubt about the automation related issues about then newly introduced Airbus A-320 and inadequacies of training for the same [7]. This is further proof of violations of pilots’ expectations about behaviour of the automated system, reflecting three kinds of breakdown in coordination, where the automation [2]:-

  • Takes an unexpected action
  • Fails to take an expected action, or
  • Carries out an action in an unexpected manner.

This leads to the question as to what is it about the A-320 automation that leads the pilots’ to be surprised at times?


Read More


Acknowledgement  Image courtesy Wikipedia

2 pings

  1. Old Facts, New Insights – Lessons from A-320 » Aviation Medicine :: Aerospace Medicine | Aviation Medicine :: Aerospace Medicine

    […] Old Facts, New Insights – Lessons from A-320 Part 2 » […]

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>