Why is automation dependency encouraged in modern aviation ?

[Check Airman]

Originally Posted by vilas

I think all the discussion is only centred around short to medium haul flights. If someone needs to do a dozen approaches to practice raw data then spare some thought to long distance guys. They hardly get to do 8 to 10 landings total in a month some they got to share also.

That’s very true. That doesn’t mean they can afford to be any less vigilant. They’ll probably be more tired, and be going into less familiar airports, but they’re prone to the same mistakes narrowbody guys are. With fewer sectors per month, I’d argue they need to be more proactive in pushing the red buttons.

As an aside, if I fly a widebody and do 8-10 landings a month, something’s wrong. That’s more than I do on a narrowbody! Some of my more senior colleagues do 4.

[vilas]

As an aside, if I fly a widebody and do 8-10 landings a month, something’s wrong. That’s more than I do on a narrowbody! Some of my more senior colleagues do 4.

That is more correct. I was being over generous.

[hans brinker]

About a decade ago, when I flew corporate, sat at the bar with a retired legacy US legacy carrier pilot. He had always been able to do long haul, but most not senior enough to be pilot flying. He estimated he had done around 2 landings per month in the last 20 years of flying part 121, so about 500 landings in that part of his career. I had just come of flying ACMI for a European regional, doing 10 days on 5 days off, 6 legs per day, and in 8 months had about 500 landings as PF that year. He also had 8000 hours "dozing for dollars" (sleeping in the crew rest quarters as PIC). While I was commuting from the us to the EU I definitely flew more miles as a passenger than as a pilot....

[vilas]

Originally Posted by hans brinker

About a decade ago, when I flew corporate, sat at the bar with a retired legacy US legacy carrier pilot. He had always been able to do long haul, but most not senior enough to be pilot flying. He estimated he had done around 2 landings per month in the last 20 years of flying part 121, so about 500 landings in that part of his career. I had just come of flying ACMI for a European regional, doing 10 days on 5 days off, 6 legs per day, and in 8 months had about 500 landings as PF that year. He also had 8000 hours "dozing for dollars" (sleeping in the crew rest quarters as PIC). While I was commuting from the us to the EU I definitely flew more miles as a passenger than as a pilot....

So how do they maintain their skill? Everyone has to manage within what is available. Long haul guys have to make their simulator sessions count because online situation is not going to change. The only advantage they have is they land at international airports which are generally not critical infrastructure wise. Also doing more manual landings is the corporate sector safer than airlines? What does the record suggest?
https://www.google.com/amp/s/qz.com/work/1169517/executives-are-dying-in-private-planes-while-commercial-flights-have-never-been-safer/amp/

[Bueno Hombre]

https://futureairlinepilot.********....-there-for.htm

[Uplinker]

Originally Posted by Centaurus

(quote):
The following statement is from an earlier Pprune contributor. I don't know the date of the original post or even which forum or the author, but in my book it should be kept in every pilots navigation bag and shoved in the face of any pilot (first officer or captain) that baulks at a reasonable request to switch off the flight director and practice hand flying.

"Raw data as a competence demonstration is intended to check a fundamental ability to fly some procedure at the lowest level of instrument display required to be provided in the aircraft.
Thus, no flight suggester, no thrust management slave, no 'noughts and ones' translating input data and no stress-relieving flight control manipulator. Just that most elusive of skills, manual flight on basic instruments." (Unquote)

Perhaps the operative words being "a fundamental ability to fly."

The contributor you quote is saying that we should all be able to fly an aircraft with very basic controls - not even FBW, which is fair enough - yes we should, and when it all goes wrong that is what we have to do.

But do we drive our cars holding a paper map or remember the route like a taxi driver, or do we use Sat Nav? Do we go down to the local library to look up information, or do we Google it? Do we use a fully manual film camera and a hand held light meter, setting the shutter speed and f-stop, or do we use an automatic digital camera? Do we keep a little black book of all our contacts' phone numbers or do we use our smart phone?

Technology has moved on, and there are better, more reliable and more efficient ways of control, than there were in the 1950's. (Even cross-channel ferries use autopilots, and they are moving quite slowly!)

[alf5071h]

Uplinker, "… be able to fly an aircraft with very basic controls …"
An alternative view is that pilots should first be able to interpret the situation from basic (unaided) instrumentation, enabling awareness and understanding before acting through the controls.

http://understandingaf447.com/extras/Gillenstudy.pdf

[Centaurus]

Technology has moved on, and there are better, more reliable and more efficient ways of control, than there were in the 1950's

A true statement. However there is one big difference between the examples you quote and the ability of the pilot to fall back seamlessly on raw data manual flying. That is lack of basic instrument flying skill has the potential to kill you and your innocent trusting passengers. Using an obsolete camera or a road map without GPS won't.

[vilas]

In modern automated aircraft and the environment like RVSM, CAT II/III the opportunities to keep safe levels of raw data flying are limited. Here the Airbus flight path stable concept helps because it doesn't demand the same levels of skills as say a 737. You just make what ever pitch and bank changes you require and the aircraft stays there. It trims itself and doesn't deviate due to speed or thrust changes. It's just a question of not loosing the scan. Only degraded mode like direct law which shouldn't normally extend beyond 3 or four minutes before landing you need to fly like conventional aircraft. And that in any case is practiced only in the simulator. Even engine failure after takeoff is automatically partly assisted leaving the pilot to do very little. This cannot be removed even in the sim for practice nor is it required. So the raw skill requirements are much less than ancient times, it cannot be denied nor is anyone unhappy about it.

[Uplinker]

Yes, I think there are two elements to this.

One is being able to fly manually, the other is being able to navigate accurately using only raw data.

Being able to fly manually means having the proven ability to direct the attitude, path and speed of an aircraft in the air mass, by reference to the primary flight display - pitch, bank, speed, Alt and V/S - and no flight director. Then trim it to maintain attitude with hands and feet off the controls; and manually adjust thrust to set the required speed. Then do it again following config or speed changes. This is a most fundamental requirement and ALL pilots should be able to do this to keep their licence.

The second element is navigation with reference to raw data. This is more of a grey area in my mind. If you navigate with raw data on a daily basis, you become very accomplished and accurate - I am sure that many of us flew our first commercial contracts in raw data aircraft, so this became second nature to us. However, some raw data displays and presentations leave a lot to be desired. Tracking an NDB in a PA 28 can be hard work; doing the same in a Dash 8 - in which you can overlay the NDB course bug onto the magnetic heading rose, which is automatically synced to magnetic north - is much easier. But is that "cheating"? Having an FMGS do it and drive flight directors, is obviously much easier again, (and essential for some navigation protocols).

In modern flying, do we need pilots to be able to accurately track an NDB with raw data? If you have lost most of your Nav systems in a modern aircraft AND have to land only via a raw data NDB procedure - displayed on the tiny back-up instrument(s) - it is a very very bad day - normally systems can be reset, or there will be at least a VOR or a Cat l ILS approach or radar vectors available, or within easy reach.
.

[FullWings]

In modern flying, do we need pilots to be able to accurately track an NDB with raw data? If you have lost most of your Nav systems in a modern aircraft AND have to land only via a raw data NDB procedure - displayed on the tiny back-up instrument(s) - it is a very very bad day - normally systems can be reset, or there will be at least a VOR or a Cat l ILS approach or radar vectors available, or within easy reach.
.

I agree.

In my experience of modern highly redundant aircraft systems, you can pile on the faults and the overall experience slowly degrades but is essentially not that much different to normal, until the last failure when everything goes blank. Given that “raw data” on EFIS goes through a lot of processing before being synthetically presented, problems that are significant enough to cause serious navigational issues will likely leave you with little to go on as MMRs have combined a lot of independent units into one. On the aircraft I fly at the moment, you need a CDU to tune navaids manually, so if it becomes unavailable, that’s it...

[KayPam]

If it is of interest for some people, I measured the stick forces on the airbus sidestick.
The results are : 17Nm in pitch, 10Nm laterally, both for full deflection.
It means that full deflection measured just two centimeters below the red button (where one could typically apply force) would give 8kg(force) for full pitch deflection and 4.8kg for full roll deflection.

Edit : and going from idle reverse to forward idle requires about 1.5kgf per thrust lever !

[KayPam]

Originally Posted by Uplinker

The contributor you quote is saying that we should all be able to fly an aircraft with very basic controls - not even FBW, which is fair enough - yes we should, and when it all goes wrong that is what we have to do.

But do we drive our cars holding a paper map or remember the route like a taxi driver, or do we use Sat Nav? Do we go down to the local library to look up information, or do we Google it? Do we use a fully manual film camera and a hand held light meter, setting the shutter speed and f-stop, or do we use an automatic digital camera? Do we keep a little black book of all our contacts' phone numbers or do we use our smart phone?

Technology has moved on, and there are better, more reliable and more efficient ways of control, than there were in the 1950's. (Even cross-channel ferries use autopilots, and they are moving quite slowly!)

Automation dependency is also a thing in daily life.
My girlfriend no later than yesterday typed an address and was willing to embark on a bus, until I told her that the address that she typed was completely wrong. She disregarded too quickly the map showing the itinerary and only looked at the resulting bus number and waiting time. At a time when we are under curfew here, she could have finished away from home at "shutdown time", and risked a fine.
At the first times of car GPS, we heard stories of people going through very narrow roads and ending up stucked or worse in the middle of nowhere.
A beginner photographer but who has understood the concepts can make significantly better pictures than the automatic mode in many cases.

However, it's clear that all this won't make it to news headlines.

Depending on automation is OK if it is not a matter of life and death.
It can be OK if it can be guaranteed that the automation will be available and functional in 99.999..% of cases.

But in aviation there are two problems :
We are commonly required to be able to fly without automation. 100% of flights end with a manual landing, and a significant proportion can require raw data visual approach, or raw data sidestep, or this kind of maneuver.
It has been demonstrated that using automation reduces the instrument scanning, and hence reduces the performance in recognizing situations where the aircraft does not do as wanted (and why this situations happen does not matter, the pilot is required to correct them whether they are caused by automation malfunction, or misuse of the automation)

Originally Posted by vilas

In modern automated aircraft and the environment like RVSM, CAT II/III the opportunities to keep safe levels of raw data flying are limited. Here the Airbus flight path stable concept helps because it doesn't demand the same levels of skills as say a 737. You just make what ever pitch and bank changes you require and the aircraft stays there. It trims itself and doesn't deviate due to speed or thrust changes. It's just a question of not loosing the scan. Only degraded mode like direct law which shouldn't normally extend beyond 3 or four minutes before landing you need to fly like conventional aircraft. And that in any case is practiced only in the simulator. Even engine failure after takeoff is automatically partly assisted leaving the pilot to do very little. This cannot be removed even in the sim for practice nor is it required. So the raw skill requirements are much less than ancient times, it cannot be denied nor is anyone unhappy about it.

I try to fly direct law when flying the sim, both because the 5 minutes I did during type rating were very funny, and also because since it is more demanding, it is excellent training for manual flying in normal law as well.
But we rarely have the time, instructors tend to be reluctant to train us for something that could never happen to us.

[pineteam]

Originally Posted by KayPam

But we rarely have the time, instructors tend to be reluctant to train us for something that could never happen to us.

Direct law in the real aircraft does happen. Even so it's a rare event. Happens in my outfit in 2017. I talked about it to the captain who had the chance to fly direct law on that occasion. They had dual IR fault. He said direct law is a lot easier to handle in the real aircraft than the sim. Not a surprise here. Direct law in the sim is way too sensitive. I quite enjoy it tho.

[sheppey]

instructors tend to be reluctant to train us for something that could never happen to us.

"Training" implies the ability to demonstrate a sequence where necessary. After all isn't that what the most junior flying instructor is required to do in a Cessna? It is indeed a rare simulator instructor who has the confidence and technical skill to confidently demonstrate a hand flown IMC sequence. Engine failure at V1 for example. Jet Upset is another. It is easier to gabble from the jump seat.

That is entirely understandable as it would too embarrassing if they stuffed up a demonstration. They might be able to talk the talk but very few can walk the walk in front of the student. . That is precisely the reason why check captains who are simulator instructors avoid demonstrating; even though a pilot under training in the simulator can often benefit from first observing a good demonstration before having a go himself. Just like our junior flying instructor and his student pilot in their Cessna.

It should be relatively easy to make time in a simulator session if you have the mind to. All you need is a modicum of common sense and avoid long drawn out briefings when the time should be spent flying the simulator.

Re the highlighted quote. Instructors are not necessarily reluctant to train us for something that could never happen. That is just an excuse, It is just they know they are not competent to demonstrate; despite their paper qualifications as examiner/check captain or what have you.

[vilas]

Originally Posted by pineteam

Direct law in the real aircraft does happen. Even so it's a rare event. Happens in my outfit in 2017. I talked about it to the captain who had the chance to fly direct law on that occasion. They had dual IR fault. He said direct law is a lot easier to handle in the real aircraft than the sim. Not a surprise here. Direct law in the sim is way too sensitive. I quite enjoy it tho.

Direct law experience is very very rare like winning a lottery. How it feels in real aircraft only those winners know. But other mortals can only say It feels different from normal law. It may be because since you directly move the surfaces unrelated to g, the stick movements to get a certain pitch change are different than what we are used to in normal law. Also because now it's like conventional aircraft that need trimming but without the aerodynamic feel. Direct law actually is only a flare mode just to give conventional handling during flare and touchdown and not meant for large manoevering because it always has gear down. So there's nothing much to practice except trimming.

[FlyingStone]

Originally Posted by KayPam

100% of flights end with a manual landing, and a significant proportion can require raw data visual approach, or raw data sidestep, or this kind of maneuver.

I don't know where you fly, but there's a fair amout of autolands done during the year due to weather, so you can shed couple of percent of the 100% figure.

Significant proportion requiring raw data visual approach or sidestep? How many flights in the last 12 months would you be unable to complete (as in would have to divert) without doing a visual approach or a sidestep?

[gums]

Thanks, Vilas and Pine.

One thing about the real plane in flight with the "direct" law is that is itn close to what we had 50 years ago with pure hydraulic controls that did not have a lotta electronic or other inputs , well, maybe some dampers. So we basically replaced pulleys, cables, pushrods, torque tubes and such with hydraulic pressure and actuators that moved the elevators, ailerons and rudder.

Up to me I would require a few training moments every year in the real plane with degraded computer modes - maybe not pure, direct law at the beginning, but at least feeling how the plane reacts without all the protections, gains and smoothing and...

[giggitygiggity]

Originally Posted by gums

Up to me I would require a few training moments every year in the real plane with degraded computer modes - maybe not pure, direct law at the beginning, but at least feeling how the plane reacts without all the protections, gains and smoothing and...

Don't we already do that? We always have some element of direct and alternate law flight at some point in the year. At the legal minimum, I imagine it is required at least once every 3 years during the ATQP cycles. The move to EBT should demonstrate when extended modules are required to to a reduction in proficiency.

[vilas]

It is indeed a rare simulator instructor who has the confidence and technical skill to confidently demonstrate a hand flown IMC sequence. Engine failure at V1 for example.

I don't know which aircraft you fly but engine failure at V1 in Airbus as compared to even other non FBW airbuses is a joke. Even if one flies EFATO once in a year it can be done in AB FBW within a few degrees of deviation. So anyone who cannot demonstrate EFATO with competence how does he become an instructor is mystery to me. Direct law is no big deal either. Actually if you can't fly Airbus well Bernard Ziegler will be offended. In any case you won't be able fly any other aircraft well either.