automation...civilian vs military attitudes
 

"What's it doing now?", and essentially "I wish I was mechanically connected to the flight control surfaces", is something I have heard often from some civil aviators. But never from military aviators. Do I know too few of them? (I count 3). Any ideas?

Cause fast jet pilots know if they were "mechanically connected to the flight surfaces" they would be dead! Without the automation they are unflyable. They are designed to be aerodynamically unstable.

Military pilots can bang out if it starts going awry!

The Tornado was a transition aircraft, fly-by-wire, stable, with mechanical back-up links.

The principle behind the Tornado FBW system was very simple - the stick acted as a manoeuvre demand device. Simply put, the more you moved it, the higher the rate of turn in that plane the system tried to give you. If the system detected motion the pilot hadn't asked for, e.g. a wind gust, it tried to remove that. It was smooth,but perhaps lacking in 'feel'. The mech. back-up was pretty rudimentary (e.g.no rudder) and intended only as a 'get you home'.

The auto pilot was very simple and only used occasionally whilst the pilot was busy doing something else.

The real problem the civilians have is with the automation e.g. autopilot, not the FBW, in my opinion.

Essentially civilians can still disconnect all the auto systems and fly it manually, just like mil. do, but they don't

Why not?
Company policy
Unwillingness to fly manually
Inability to fly manually

Your question should be directed towards the those points, which will vary considerably from company to company and pilot to pilot.

Thats pretty broad brush to state that civilians have a problem with automation, Fox. Perhaps you'd like to justify that?

In most airlines, the stress is on using the APPROPRIATE level of automation. Most of us are fully capable of hand flying a complex STAR amongst busy traffic at 4am after 13 hours on duty. But would it be the best professional decision to choose to do so?

No insult intended ShotOne. The OP refers to those who do have problems, and that is where my comments are directed. Large numbers don't have problems and use automation,as you say, correctly and to enhance Flight Safety, not diminish it.

A percentage of ex-mill pilots I work with struggle with mode confusion when disengaging or reinstating automatics and poor hand flying abilities with raw data and visual flying. It's not to the extent of the civilian world where experience levels and training seem to be less than what they were a decade ago and MPL like candidates have zero flight time experience to fall back on. So if some experienced ex-mil pilots struggling I'd suggest this points to airline training system modelling themselves on manufacturer recommended minimums. Airbus is probably a good example.

Having seen a number of airline training systems, by far the best for automated aircraft was starting an endorsement process and building up the automation as the course progresses. Perhaps issues aren't totally eliminated such as mode confusion, but the pilot should have the skills to confidently disengage or engage automation, put the aircraft in its desired state and proceed with the flight phase accordingly.

All the types I have flown have recommended maximum use where possible, but with an expectation that you maintain handling skills from time to time.


We have very sop driven use of automation though

Following the magenta line.......

https://www.youtube.com/watch?v=fbQB72-4WQ0

A very candid self appraisal by an airline training pilot.

Well those USAF pilots who managed to turn a perfectly serviceable C-12 in the cruise at 12000ft into a smoking hole in the ground in Afghanistan about 2009 weren't too hot on automation or hand-flying either. In the space of three minutes they managed to:
1. Select the wrong autopilot mode for climb (VS instead of FLC)
2. Failed to monitor decreasing airspeed and increasing AOA.
3. Failed to apply correct recovery when stall warning sounded (they kept autopilot connected and tried to recover using the autopilot)
4. Failed to apply correct recovery technique when aircraft flicked into a spin (they pulled back on the stick).

The USAF report is available online - you can google for it as well as I can if you are interested to read the full report.

Bit unfair on those comments...

and:

Familiarity - or lack of it - with the aircraft is one thing but most of those points are basic flying skills pertinent to any platform.

Both the mishap pilots were ex-multi-heavy, so I don't think there was confusion possible over stall/spin recovery actions (some fast jets do require stick fully back for spin recovery)

From the accident report:
Horsecrap. I taught at an airline training school in AZ, and we overcame this problem by deploying to the California coast to fly real IMC in the fog. Beale is only 100 miles from said coast, they could do it if they wanted.

"Do I know too few of them"...? Well, aside from the pilots of the MC12 above, there's also the crew of the RAF Voyager not so long ago. Setting aside the serious lapses of professionalism which caused the event, the actions of both pilots during the recovery displayed very poor knowledge of the flight control system -to the extent that many days after the event, the Captain apparently remained convinced he'd done nothing wrong!

It's a pity that the OP chose to frame this as a military versus civilian issue because I strongly agree with his underlying point. These issues are not always well trained. Simulator sessions which would be a perfect opportunity to do so are mostly used for regulatory box-ticking with the emphasis very much on testing rather than training.

As a former sim instructor, I agree completely. We worked very hard in our sims to bring in realistic training which provided feedback in a blame-free manner, but that was 20 years ago...

Trim Stab
 

Well Trim Stab, what a completely fatuous and pathetic argument.. do you think that by quoting one (military) case study then you have proved a point? You are very wrong.

How about AA 965 in Cali in 1995..?? The report is online, as you are so keen to point out. A sad case of automation confusing the civilian crew, causing the deaths of 189 people with a perfectly serviceable aircraft. I could quote more but won't sink to your pathetic depths. Believe it or not it has happened to both civilian and military pilots and will continue to do so.

Instead of trying to score points by quoting military incidents, why not agree that all pilots from all backgrounds have suffered from mode confusion, it happens but awareness and education can help to to reduce the number of occurrences??

Automation isn't the problem, it's understanding what it is doing and how to react to that, the problem with automation, and with the best will in the world is it will only do what man has programmed it to accomplish, it cannot differentiate between that which the pilot needs to know and that which they do not in abnormal circumstances.

Circumstances can be thrown up where normal fault reporting to the crew which in normal operations is acceptable suddenly could become an overload factor in critical situations.

A classic example of that, which took them 50 minutes to go through, and then they had to adjust the info the inputed to allow the aircraft to land.

http://www.flightglobal.com/blogs/wi...other-stories/

http://www.flightglobal.com/blogs/wi...th-qf32-pilot/

Lockstock, That crash was mentioned by Trimstab to refute the OP's contention that this is a military vs civil issue. You claim to agree with him on that point (as I do). In fact the Cali accident had nothing whatever to do with automation or flight mode confusion. There are lots of others, military and civil, which any of us could have cited. The fact you picked that one suggests you either don't know much about the issue, or were just grabbing at any incident involving fault by a civilian flight crew, in your words, "to score points".

Some have tended to merge the issues of "what's it doing now" mode confusion with that of hand flying skills. In my opinion these are two separate issues; I've flown with pilots with excellent hands who are not good at reading the FMA and (less often), vice versa.

The crew weren't confused by automation. They lost situational awareness during a short notice change to an approach. They accepted the new approach without any appreciable consideration, leading to rushed and faulty procedures. Contributing factors were the lack of approach radar, two NDBs in the same area with the same frequency and ident and there not being a requirement to retract the speed/air brakes during the reaction to a GPWS alert. There last would probably have allowed them to clear the hilltop.

Better to cite the Air France crash off Brazil - not only was there a lack of basic flying skills (select an attitude and power) but they didn't understand the logic in the sidestick controllers and the process for handing over control from co-pilot to pilot that prevented corrective action as one stick was being pulled back whilst the other vwas being pushed forwards.

The designers don't help - my current helo has a 3-axis autopilot where IAS, alt and vertical speed (all mutually exclusive) are managed through the cyclic channel - that means that with VS mode engaged (or with coupled glideslope on ILS) you are controlling speed with collective which is completely unnatural for a helicopter pilot.

Yes, you can train to understand it but if it wasn't designed in such a poor way in the first place you wouldn't have to - no wonder pilots get lost in automation.

San fran 777 crash a more likely candidate for equivalence.

Really..? The crash was investigated by the Aeronautica Civil of Colombia with assistance from the US NTSB who reported that the number one cause of the accident was most likely..

"The flight crew's failure to adequately plan and execute the approach to runway 19 at SKCL and their inadequate use of automation."

Yes, really. The Cali crash doesn't in any way support a case about over-reliance on automation. Quite the reverse in fact. Indeed back when aircraft were less automated, accidents like this, sadly, were far more common.

Not the view of the Colombian CAA and NTSB... but hey, what do they know?

Try actually reading the report rather than just plucking out the word "automation"

Re. "Better to cite The Air France crash off Brazil". The trouble with that is so much went wrong it's hard to come to any meaningful conclusion. The CM2(co pilot) action in pulling back the stick was clearly terribly wrong, but I don't see how you can possibly know he did so because "he didn't understand the sidestick logic". I suggest it's far more likely that in the pitch-black turbulence and hail of the tropical storm cell he was simply overwhelmed by a mass of contradictory information, crucial parts of which(airspeed) were wrong and a cacophony of alarms and warnings.

Lockstock, the automation didn't confuse them. What confused them was the fact that they didn't plan their initial approach, they then accepted the offer of a shorter approach without giving adequate time to consider whether it was a good option or not. The PNF(Capt) entered a new (incorrect)waypoint in the FMS without checking it's Lat/Long or asking the PF to confirm his selection. The new course line, which gave them a sharp left turn, wasn't checked and confirmed. They didn't pick up on this turn initially, but when they did and turn back towards the airport, they didn't realise that they'd crossed into a different valley. They didn't follow the approach plate. They asked to cut short the published approach. They didn't recognise the ident of a beacon, because they hadn't briefed the approach. They started their descent without passing over a specific beacon.

Yes, they were confused, but not by any automation.

And that was part of the problem, the beacon that they wanted the aircraft to steer towards had the same frequency and identifier as another beacon 132nm away and in another valley. If they had 'recognised the ident of a beacon' it would have been the same anyway. The logic of the FMS made the aircraft turn towards the nearer (wrong) beacon. As you say, the new course wasn't checked and confirmed as it should have been but subsequently the FO asked 'where are we?' and 'where we heading?' and the Captain's reply was that he didn't know.

The aircraft on autopilot, following the FMS had put them in a place where they had lost situational awareness. The report says that over-reliance on automation was a factor.

Shotone - no, the misunderstanding of the logic occurred when the other pilot finally tried forward stick on his side - this didn't work because the co was still pulling back on his and hadn't actually relinquished control. The control logic wouldn't allow one sidestick to override the other as it doesn't know which is the valid input.

?? That's not quite how it works, crab. Theres no such thing as a "valid input" as far as the aircraft is concerned. If the side sticks are moved in opposite directions they will be algebraically summed, I.e. Full fwd/full back inputs would cancel each other out. The only way one would be overridden is if the takeover push button was pressed (which it wasn't at any point as far as I know)

And the other side of computerising everything and the pitfalls.

App fail on iPad grounds 'a few dozen' American Airlines flights | Technology | The Guardian

Lockstock: As others have pointed out to you, you are missing the point of my post. There is no simplistic "military v civvy" argument - I was merely pointing out that there seems to be an increasing capability gap between automatism and hand-flying skills in all aspects of aviation.

Shotone - I get that, so full fwd and full aft would not effect a recovery from a deep stall. And they didn't operate the takeover button so they clearly didn't understand the system.

www.vanityfair.com/news/business/2014/10/air-france-flight-447-crash


Very good article.

It's one thing to make the obvious observation that there was confusion on the flight deck, but quite a stretch to single out one particular action and state it was carried out because they misunderstood the function of a specific system."They didn't operate the takeover button":who's "they"? Which one should have been pressed? Your statement pre-supposes one pilot had accurately diagnosed the problem. I don't believe they did -and for much of the fatal descent were trying to make sense of conflicting alarms and information(the stall warning only sounded for a relatively short period but, confusingly for them, recurred when the stick was pushed forward). While we can agree that many things went terribly wrong, where does that leave us in respect of the OP's question? One could pluck aspects of this tragedy to support or refute many opinions. Certainly I'm not aware of any operator, military or civil,which practices stall recovery for large transport aircraft.

I am surprised no one has mentioned the RAF Tristar Autoland incident where the crew sat and watched it autocrash.

There are folk, like shotone, who operate highly automated aircraft and understand them and their limitations. There are others who operate and don't understand them, they would be as inept on steam driven machines as on highly automated ones.

And then, of course, there are those who don't operate them and consequently don't understand them.

We tend to forget that an inept pilot (and they do exist) will be inept whatever type he/she flies. Which goes back to the idea of good training, good assessment and a willingness not to accept low standards of airmanship in ourselves and in our colleagues.

NutLoose wrote: SEngO 101, sitting in his office, felt the impact through his chair!

The first words on the CVR after the crash-and-go were "It shouldn't have done that....."

They then flew round the circuit losing about 8 tonnes of fuel - a welder in the industrial estate near Witney was rather concerned at the kerosene rain he was experiencing. After it landed, there was much consternation as to whether it could be towed in. Having seen the photos of daylight through the massive crack in the main wing spar, I think that they were very lucky to have survived. I'm not sure whether it was that DeathStar, or the one the ground crew tried to set on fire by holding in a reluctant CB, destroying most of the wiring, but one of them had evidence of a previous heavy landing which the vendor hadn't disclosed.

Nothing wrong with highly automated aircraft provided that the crews have been adequately trained and have remembered how to apply basic flying skills.....:hmm:

Assuming they have those basic flying skills...

Shotone - I agree there were, as ever in such accidents, a plethora of factors contributing to the crash; but, as that excellent article highlights, simply using the controls correctly (they both pressed their takeover buttons in rapid succession increasing the confusion about what control they had over the aircraft) would, at pretty much every stage, have recovered the aircraft to unstalled flight.

It would appear that the problems were correctly diagnosed, both the loss of airspeed info and the subsequent stall - the pilots just did not take the required corrective action.

Stall recovery is about the most basic FW technique taught (I still remember it from Chipmunk and JP 30 plus years ago) so the only explanation of why the pilots in this case didn't perform the recovery is that the level of automation, its complexity and its behaviour in degraded modes overwhelmed them.

Now is that due to poor training, poor quality pilots, CRM breakdown or poor design and display? I don't know.

All 4 I should say.
I feel just about every form of education has been nibbled to death by ducks. The causes are:
A 'bite-size' approach to instruction, with not enough holistic work.
Emphasis on passing the student, not on attaining a standard.
Training syllabi which are easy to deliver and examine, not easier for the student to learn from or recall.
A minimisation of ongoing training by companies on cost grounds.

Part of the problem is financing. Students all over the World in many industries are now obliged to self-finance. Inevitably, this means weaker students are attracted to educational establishments which will pass them, whatever. For-profit educational establishments make higher profits if they have more students (and if they cut instructor salaries, which is easier when they need less qualified/experienced instructors, which is easier with 'canned' courses), so a race to the bottom has ensued.

Furthermore, everybody has a vested (i.e. financial) interest in keeping quiet about dropping standards.

Now and again, a few with standards make a stand,e.g.
A&M-Galveston prof says flunking entire class was the right thing to do | Dallas Morning News

There remain a few beacons of excellence, but the standard of the majority has been declining gradually for 20 years.

I have instructed/taught at Uni, high school, flying, offshore sailing, scuba diving,and a bunch of others, on 3 continents, and I have seen this situation developing pretty much universally.