21

I was wondering about what happens if a professional pilot who's training to fly a major commercial airframe unintentionally crashes during flight time in a simulator? I'm not asking about some physical damage to the simulator itself, just within the simulation.

I would imagine there's a SOP for a post mortem analysis that includes logging/journaling, reviewing simulator data, and taking steps to study why the crash occurred in order to reduce the likelihood of another. Is that right? Can anyone run down the debrief after a sim crash?

Can professional pilots be reprimanded/disciplined or face any punitive consequences due to an unintended simulator crash? If so, under what conditions and what kinds of consequences?

cfx
  • 1,207
  • 1
  • 12
  • 14

3 Answers3

32

Crashing a Level D Full Flight Simulator during a training session is not good. They exist for:

  • Initial training, of licensed pilots who have not flown the type before and need to become familiar with the controls, systems, handling characteristics and procedures of the particular aeroplane. Usually a series of about thirteen 4-hour sessions.
  • Recurring training, for pilots licensed on the type, to demonstrate to the authorities that they are competent in handling emergency procedures.

During initial training, the trainees are pilots who are learning new procedures in unfamiliar circumstances, and a crash during training for an engine-out landing does not necessarily have serious consequences - if rapid progress in handling is shown in subsequent sessions. However, if the trainee is the only pilot out of 20 to experience a crash, they may not be hired by the airline.

For recurring training, a crash or failure to carry out an emergency procedure is not good. During the de-brief, the instructor will identify the actions that went wrong, and a re-assessment will be arranged. Failure of this may result in discontinued employment.

So the main consequence of professional unintentional simulator crashes may be professional discharge.

Koyovis
  • 61,680
  • 11
  • 169
  • 289
  • 1
    So do you guys train somewhere off the books before you go do it for 'real'? – Mazura Nov 18 '21 at 00:15
  • 6
    @Mazura yes especially for the pilots going to take part in the initial training sequence, they would book private sessions in the fixed base trainer. An investment into coming out at the top of the class. – Koyovis Nov 18 '21 at 01:59
22

Crashes, or at least wing tip strikes, are not uncommon on an initial jet type course during engine failure training on take off, on what are called "V1 cuts" (engine failure just before or after rotation speed) especially with pilots who are new to swept wing airplanes (if you are slow with the compensating rudder, the plane rolls very hard once the yaw angle gets large).

So a crash on take off during that type of maneuver would be no big deal on an initial, especially for a pilot new to jets. If a pilot is not able to get the hang of it and, you know, stop crashing, that would be a problem (you'll be expected to get the hang of it on the second or third try). If you can't be brought up to a safe standard fairly quickly, you'll just get washed out.

A pilot who crashes during V1 cuts on recurrent is a bigger problem. Everybody has "brain farts" and a single event that is followed by a normal skill demonstration may just be written off as a one-off (although it all goes on the pilot's training record). Skill deficiencies that are quickly rectified, and are not accompanied by other warning signs, won't hurt you too much. A flub followed by a competent demonstration will just mean a debriefing item after the session.

Skill deficiencies that are kind of borderline, but not bad enough to take someone off flying status, may just mean delays in being upgraded to captain. An experienced captain who has these sorts of problems on recurrent could face restrictions on operating as PIC or other restrictions as the causes are explored. It usually means something has changed in the person's life.

Pure skill deficiencies, whether they result in crashes or not, won't be punished professionally unless there is willful misconduct or severe personality issues. The training organization will work on correcting the deficiency up to a point. A much bigger problem is mental attitude deficiencies and personality problems - those will hurt you more than losing control during an engine failure in the sim.

One of the things the airline will be looking for when a new pilot is evaluated for selection is signs that the pilot learns and adapts quickly. During the evaluation, you are just told to perform some maneuvers and "do what comes naturally" without any help from the evaluator. They are looking to see how fast you catch on to things on your own.

One important facet of training is to "make things memorable", and the instructor may allow a crash to occur, say in something like a terrain avoidance maneuver, to make the lesson sink in (depends on the instructor and the training unit's policies - the instructor may just suspend the sim just prior to the final "crash" to avoid having to reboot the machine).

Also, you have to account for the basic realities of sim training. Sometimes you get pushed to the edge because if you are performing well, the instructor will keep adding pressure until you start making mistakes - it's a basic method of skill evaluation. At the end of a four hour session, you are pretty worn out mentally and physically, and if a new pilot you may be totally frazzled (I used to joke that it was like 4 hours spent riding in a cement mixer filled with bowling balls, if you had an instructor you enjoyed pushing you hard). A crash near the end of a session, like running into a hillside during a terrain encounter, may just be put down to late-session mental fatigue as you were (quite deliberately) loaded up to burnout.

John K
  • 130,987
  • 11
  • 286
  • 467
  • 4
    Isn't one of the reasons for simulator training so that you CAN push things to the limit - either yours or the aircraft's - without breaking expensive airplanes? – jamesqf Nov 17 '21 at 16:47
  • Yes but you generally avoid unnecessary crashes during a session mainly because it robs precious minutes waiting for the sim to reboot. Most the time an instructor will just be ready to freeze it the instant before the actual crash in scenarios where a crash is a potential result. – John K Nov 17 '21 at 17:23
  • 1
    Which feels worse to the simulator pilot, "crashing" into a mountain or pausing the simulation when the aircraft is 100 m away from "crashing"? – rclocher3 Nov 17 '21 at 22:40
  • 9
    Why does it take longer to reboot after a simulated crash than just before the crash? I would think either way it is just re-loading the initial conditions. – cxrodgers Nov 17 '21 at 23:52
  • 7
    @cxrodgers - it depends on how robust the flight model is. Some flight models barf and wind up generating a NaN that quickly propagates throughout. Others are designed to “handle it” and you can crash all day long without fear. – Jim Nov 18 '21 at 00:36
  • 2
    In the Level D sims I used to train in, a crash caused it to kind of shut down and the computers had to be rebooted. Which might take 5-10 min. If the instructor freezes the action, it remains active and just needs to be resumed with a different position loaded. – John K Nov 18 '21 at 00:39
  • 1
    @JohnK Yeah the older type ones, where any discontinuity in variables causes a jump in the flight model state, where it was hard to recover from (no code for re-alignment) and the whole model needed to be re-initialised. Not much of an issue anymore with modern day sims. – Koyovis Nov 18 '21 at 02:03
  • 2
    @JohnK That's..... indicative of some pretty crappy programming right there. It definitely indicates an inability to handle in-domain datasets. It surely does not boost my confidence in the reliability, predictability and accuracy of the simulator computers. (speaking as one that has been in computers, including design and implementation of realtime industrial process controllers, for 35 years) – PcMan Nov 19 '21 at 06:35
  • 6
    @PcMan, in a sense you are right, but it's not entirely fair. First, flight model is more complicated than any industrial controller. At the same time, it is certified to work exactly as prescribed, and it can be very difficult (i.e. costly) to guarantee it after an out-of-scope situation such as a crash. It may be "OK" 99% of the time, but that's not enough. Second, simulators often use real avionics (and other "controllers") which were never designed to handle crash (or sometimes even pause) in the first place. Finally, industrial software generally sucks in usability, as a rule. – Zeus Nov 19 '21 at 07:32
  • { it's quite bizarre that apparently older sims would "crash" when they ... "crash". that's just whacky. fascinating info! } – Fattie Nov 19 '21 at 13:23
  • @PcMan: Perhaps the issue is "overly defensive" programming. If a simulator is designed to prevent operation if there have been any numerical anomalies since the last boot, that will reliably ensure that no simulations produce results that are seemingly valid but wrong. Allowing smoother recovery from numerical anomalies may be more convenient, but ensuring that doing so would never cause the system could produce results that are seemingly valid but wrong would be harder than requiring a full restart after any kind of anomalies occur. – supercat Nov 19 '21 at 20:15
  • @supercat That is very likely, but sad. It's the same philosophy that microsoft windows uses. i.e. "have you tried rebooting it"? That does work, but is a very inefficient, lazy, and hazardous way to handle a complex system. As soon as parameters get wonky, GIVE UP is not a sound philosophy. And god forbid if one of those same programmers later develop a system for an actual airplane, or for ATC, where taking a few minutes off to reboot is not an option – PcMan Nov 19 '21 at 20:28
  • @supercat (and yes, I've seen exactly that in cockpit troubleshooting guidelines for airliners. "pull fuze XXX and reboot the autopilot" is the scariest line i've ever read in a 747 troubleshooting checklist!!) – PcMan Nov 19 '21 at 20:30
  • 2
    @PcMan resetting systems to clear faults is very common, where the faults are random and clear without coming back, and the overall system has sufficient redundancy. Some airplanes will have a reset guide that lists components that can be reset with CBs at the gate to clear a fault for dispatch. This has a huge impact on dispatch reliability with negligible impact on safety. – John K Nov 19 '21 at 20:37
  • @JohnK: Making systems fail-safe, and making systems operate reliably, are contradictory goals. If a fault has a 0.1% chance of occurring in ways that would be detectable but harmless, a 0.001% chance of occurring in ways that would be detectable, but intolerably bad if ignored, and a 0.000001% chance of occurring in ways that would not be detected until bad things happened, forcing a shutdown when a fault is detected would increase the probability of tolerable failure by 0.1%, but reduce the probability of intolerable failure from 0.001% to 0.0000001%. – supercat Nov 19 '21 at 20:45
3

If a pilot is doing a License Proficiency Check (LPC), i.e. an annual check of the skills required of a Type Rating (e.g. B757, A330, etc) as endorsed within that pilot's license, which is being conducted by a Type Rated Examiner (TRE) on behalf of the National Aviation Authority (NAA) and that pilot fails the LPC, the TRE is legally required to inform the NAA of the fail and that fail will then be logged against that pilots details as held by the NAA.

Conversely, if the pilot is doing an Operator Proficiency Check (OPC), i.e. a six-monthly check of the skills required of a Type Rating (e.g. B757, A330, etc) as endorsed within the pilot's license, which is being conducted by a Type Rated Examiner (TRE) on behalf of the airline that employs the pilot, and the pilot fails the OPC, the TRE is (typically) only required to inform the airline's Head of Training of the fail and it's then up to the airline's Training Department to decide how to proceed with the failed pilot.

That said, if the fail of an OPC is so egregious as to cause significant doubt that the failed pilot is not fit to operate the aircraft type on which they were being checked, the TRE is still duty bound to inform the NAA, for the latter to then follow it up with the airline that employs the pilot.

Fundamentally, failing a check (be it a LPC or OPC, or ones medical) can well indeed turn into a career limiting event.

Tony Layfield
  • 31
  • 1