Has the pilot shortage arrived?

If the main control systems are not operating, then the computer can't control them, and it's down to whatever the crew can come up with.

A long time ago now, as part of some complex research into handling and flight envelope excursions, I did a lot of research testing in a number of Level D simulators, and the issues around control failure was one scenario that we explored in a lot of detail, in a sim, it's relatively easy to freeze the controls by pulling a number of circuit breakers, the challenge then is to work out how to keep the thing flying with what's left. We did a number of sessions in MD80, 737, 747(classic and 400), 757 and Airbus, and with varying degrees of difficulty it was possible to get them back on to the ground in a survivable manner. The MD was challenging, because the engines being close to the centre line and at the tail meant that getting it to turn was not easy, and the Airbus FBW in manual reversion was a pig, simply because the pitch trim took for ever to get to where it was needed if the power setting was significantly changed, so much so that one operator that we were working with stated that a manual reversion go around wasn't possible, because of the problems of keeping the pitch under control during the early stages of the climb out.

The computers in that situation are just weight, they can do nothing to help at all, and even if everything is working, in a lot of cases, the age of the computers is such that they can't have the size and complexity of software that is needed to fully analyse the issue and provide meaningful guidance, which is why some of the emergency check lists and the like are so long and complex.

It's scary how many systems are still using 386 processors, because the later ones were not certified for aviation use, there was a problem a long time ago with a floating point math calculation issue on 486 processors that stopped them from being used in critical aviation applications, and I'm not sure if it was resolved with the later 5 series processors, so much so that a few years ago, Intel had to do a special run to make 386 chips for Boeing, as they'd run out of the stock they had, and couldn't upgrade to a later device without a nightmare with certification, it was cheaper to pay for a special run to produce more 386 chips.

The other problem is that it used to be relatively easy for a flight crew member to spend some time in a sim to "experiment" with the extremes of the aircraft handling envelope, but "standard operating procedures", paranoia among the training community and bean counter restrictions on the use of simulator time now means that it's no longer possible for people to explore the handling of the aircraft that they fly every day, so when something extreme happens, there is a very good chance that none of the flight crew have ever had any chance to explore what the limits are, so that they know how to use them.

A classic example of this was a thing we used to do on the 747-4, you could use any runway, the scenario was from a "clear take off" position, start the stop watch, take off, climb to 10,000 Ft, and put it back on the runway again, shortest time wins. No other rules, other than don't break the sim! CAVOK, zero wind, any runway for the landing, so relatively easy in some respects. We tried it on a number of people, and the shortest time was about 6 mins 25 seconds, which needed some non standard handling to achieve it, but it was safe.

A few months later, one of my colleagues, who was a full time ground instructor on the aircraft, gave the same scenario to 3 type rated First Officers. 2 of the 3 broke the aircraft in the air, as they were lost without the magenta line to guide them, and didn't know the extremes of the envelope, so they over stressed the airframe. Scary stuff, and we've seen other examples with Air France and the like, the formalisation of training and procedures has had the very negative effect of making it very hard for people to know exactly where the limits are, and how to use the airframe up to those limits without danger. So many of the modern crew are put in the position where they handle the aircraft for maybe 10 minutes, and the automation then takes over, so their experience of some aspects of the handling is close to zero, and if the automation fails in a phase of flight that's unexpected, the crew are presented with a pig to fly in a situation that they've no experience or training for. (AF447 being the classic most recent example), and the result may be very much less than optimal!

I accept that the number of occasions where the requirement to use the extremes is very limited, but how are we to put a value on the difference between a bad outcome and a good outcome when something dramatic goes wrong in the air? Where commercial flights are concerned, while money is a factor, there is so much more that has to be factored in to that equation.

Then of course, there's the very real issue of the transition between a small single engine "spam can", and a modern full EFIS FBW commercial airliner that's landing at twice the speed that the spam can cruises at, it can be a very hard transition to make, if you've spent a couple of hundred hours flying analog instruments, and now have to adapt to the modern EFIS presentation, with all the extra bells and whistles that come with it, and are doing it at 4 or 5 times the speed that you're used to, that can be very challenging to adapt to, some of the airlines are not using small piston singles for training any more, as they don't see it as appropriate, the one down side of that being that if you don't fly a small single, your "solo" time may be very limited indeed, and it's only when there's no one else there to make a "committee decision" that you learn how to make critical decisions that are possibly life significant. It's been recognised by the regulators, some commercial licenses now are only valid for flight in multi crew environments, the holder can't hire and fly a small light aircraft, even though they may have flown a 777 or 380 into the exact same airport a few hours earlier.

A box the size of a large matchbox has the ability to navigate an aircraft with an accuracy of 20 metres to any point on the globe, but the technical challenges of making that fail safe, or even better fail soft, so that it can recover in the event of any form of failure is still beyond the capability of the computers, the most significant event I can think of was the uncontained engine failure on the A380, where it took an augmented crew a number of HOURS to work through all the warnings and errors that the systems threw up before they could make a decision on how to safely land it. The thought of trying to do all of that in the middle of a tropical storm doesn't bear thinking about, given the significant structural damage that the failure caused, trying to safely fly a damaged aircraft in extreme weather for that length of time while you work out if it's actually safe, or possible to land it is not a nice thought.

There was a report recently on AVHerald of a 380 that refused to land at Manchester, so they ended up diverting to Heathrow, as the runway was longer, but we don't have the details of why that happened yet, all we know is that something with the automation threw it's toys out of the pram, and gave the crew the issue to solve, as it couldn't do so. While that's still possible, unmanned flight isn't going to happen.

So, there will still be jobs at the sharp end of the aircraft for some time to come. As to the quality and longevity of that job, I can't predict it, I do know that it's changed massively over the last 20 years, for all the reasons outlined above, and it for sure will continue to change over the next 20 years.

smurfjed wrote: »

You really needed to shut down all 4 engines to make this more fun.....

Effectively, we did The fastest way was to close the throttles at about 8500 Ft, and zoom climb through 10,000 Ft, so it was on the verge of a stall, then throw everything out, flaps, spoilers, gear, and if you had to use power again before touchdown, you'd blown it.

I still have memories of doing a straight in over Stonecutters at about 6000 Fpm descent rate on the run that got me the best time of the day . It was all about looking out of the window, and working out how to get it back on effectively dead stick. Didn't bother to stow the spoilers, there wasn't time.

(And yes, on a different occasions, my "friend" did give me all four off on a 747-4 once, at about 5000 Ft downwind, which got my attention. Managed to (just) put it on the crosswind runway, albeit VERY close to the threshold.)

They were fun sessions, in the right sense of the word, but with a serious underlying intent, we learnt a hell of a lot about the extremes of the envelope of whichever type we were in on the day, which was what we were trying to achieve, so that the software we were producing was capable of accurately modelling the extremes as well as the "normal" performance.

Another trick my friend pulled on us was in a large twin, (like the 777), doing a descending turn at flight idle, and fail the inboard engine. It was surprising how many people ended up with some "interesting" bank angles not long after, the combination of the drag of the dead engine with the power from the good one during the transition from the descending turn caught quite a few out, and getting it back the right way up was a challenge, as it often needed the good engine throttled back to get enough control authority to be able to roll back, which didn't come naturally, especially at lower levels and speeds.