When do we use autoland?

[Will]

Slightly off topic, what percentage of real-world landings use autoland?



Edit HH: Thread splitted from an autothrottle thread.

[Britjet]

In my airline. Hardly any.
Long-haul pilots need their recency. And if you do a "practice" auto land without Cat3 protection in force at the airfield (eg if there is traffic within the ILS protection area) and the aircraft does a dodgy autoland, it becomes a can of worms as to whether it is a viable defect.
Some airlines, I think, do more autolands.

Peter

[Avi]

I read somewhere an autoland is done once a month (in good weather) just to make sure it works well. This is probably a company policy.

Cheers,

[John H Watson]

My old airline had a 45 day max period between autolands.

And if you do a "practice" auto land without Cat3 protection in force at the airfield (eg if there is traffic within the ILS protection area) and the aircraft does a dodgy autoland, it becomes a can of worms as to whether it is a viable defect.

It's interesting to note that the flight crew's judgement of the cause of the dodgy autoland can influence the subsequent procedures which are followed by Engineering. If the pilot thinks that traffic or a late localiser capture caused the problem, and there are no recorded faults in the Central Maintenance Computer (CMC), then there is no necessity to carry out autoland ground tests. However, the A/P will be downgraded to Cat II until a successful autoland can be carried out by the flight crew. After the crew report a successful autoland, the A/P can be upgraded to Cat III.
If the flight crew cannot be sure, and there are no CMC recorded faults, then the ground tests must be carried out. These are not terribly onerous  After they have been carried out and no problems found, as above, the A/P will be downgraded to Cat II until a successful autoland can be carried out.

This should only cause a problem if the next flights are expecting some seriously foggy weather landings or the pilots need the next landings to remain current.

Only if a defect is found and rectification is carried out according to the Boeing Maintenance Manual can the aircraft be recertified to Cat III.

[Will]

Thanks John, that's interesting.

[tango4]

Hi Will,

I guess if you want to know how many autolands are done you would have to take into account the weather of the airfields you fly to.
I don't remember the exact figure, but at CDG we are less than 5% of the time in LVP on average.

Regarding the autoland made for training or aircraft certification, it is quite a problem. As Peter mentioned, you are not guaranteed CAT3 protection all the time, and it can even be dangerous if the pilot is not fully aware that he is not protected.

For that reason, CAT3 for training are only allowed in good weather at CDG (Vis more than 5Km, Ceiling more than 4000ft), and forbidden during rush hour. But sometimes, it is rather ambiguous, because some ATC will allow CAT3 for training, just saying you won't have separation. In some cases, pilots did not fully understand it meant that the signal would not be protected.
And the other problem is that if the field is not in LVP, you cannot guarantee "absolute" protection. What we do is we "lock" the ILS, meaning that it will lit a red light in the ILS "tech building" to let maintenance people know that they should not be doing any work on the ILS (for some maintenance, it can be done with the ILS operational as long as you operate CAT 1 only). And we can space the aircraft "as in LVP" to protect the sentive area from aircraft (but it means some aircraft will have to wait a bit, which is why it is forbidden during rush hours). But there is another intruder category: vehicles. During LVP, most vehicles are not allowed on the taxiways, and the ones that are should contact ground frequency (otherwise they just listen). AND NO VEHICLES ARE ALLOWED INSIDE THE CAT3 PROTECTION AREA. But if the field is not in LVP, as vehicle are not in contact, they are allowed to move freely between CAT1 and CAT3 protection, meaning you are not 100% protected.

Another critical moment is when you "lift" LVP and resume CAT1 operations on the field. You have to be EXTREMELY cautious to coordinate with the pilots who will be the last one to make a CAT3.

I don't know if you are aware of that incident:
http://avherald.com/h?article=445873f3
But if I recall correctly, the pilot failed to advise ATC that he was making a CAT3 autoland, with a pretty bad weather, but airfield NOT IN LVP. So ATC did not apply LVP separations, and well, just have a look at the last picture. I guess it is self explanatory !

Charles

[Hardy Heinlin]

Hi Charles,

re that B773 incident you quoted: I'm wondering why the 777 AFDS didn't use an IRS path during this temporary, massive radio interference. Or maybe it did, and the interference occured slowly, gradually within half a minute or so, during the takeoff roll of that Jumbolino, reaching the maximum error when the Jumbolino overflew the localizer antenna.


|-|ardy

[tango4]

Well, as you say, we are taught that there are two big categories of LLZ interference:
*The big, fast ones, where the aircraft will act as you say (replacing the LLZ signal with Inertial data), because in that case, the aircraft is able to detect that the LLZ signal is not valid, and compensate for it.
In that case, I heard that some autopilots would disconnect instead, but I don't know on which aircraft, or if it is even true.
This type of deviation, although potentially dangerous, is not the worst, because you have a chance of detecting the deviation yourself, or your aircraft has that same chance, and will possibly compensate with inertial data, or you can initiate a go around.

*Now the slow deviations:
Those are the nastiest, because you have no real way of seeing them. And in that case, for the autopilot, the signal can look perfect, and the aircraft is going to follow it. The only way you could see it I guess is if the deviation is SLOW, but rather IMPORTANT: in that case, you would perhaps see that you are strangely misaligned on your ND. The problem is that usually, those slow deviations are also not huge, so they are basically impossible to detect, and bring you right in the grass !

Now, regarding, the Munich incident:
The problem is that when you have an object infringing the safety area, you cannot know for sure what type of deviation it is going to generate. (I'd guess that an aircraft overflying would probably generate oscillations though, as you say. From what i heard, slow deviations are more frequently caused by a big stopped object: for example the tail of the A380 makes a fantastic mirror for the waves, and creates a second reflected signal, which is I think the cause of those slow deviations).

Here is a quote from the article I linked in my previous post:
The BFU reported that the flight data recorder showed all three ILS receivers recorded localizer signal deviations in both directions about 14 seconds before the autopilot changed into roll out mode, about 6 seconds prior to the roll out mode the localizer deviation increasingly showed a deviation of the aircraft to the right of the extended center line. The aircraft began to roll left in response to autopilot inputs to minimize the localizer deviation and reached 3.5 degrees bank angle to the left when the autopilot changed to roll out.

Hard to tell for sure, but perhaps there was a mix of deviations.
First some oscillations, compensated for by the inertial data. And then slower ones, unavoidable except by disconnecting the A/P. Or perhaps this inertial compensation is inhibited at some point or under certain conditions ? I don't know how you coded this in PSX, but I guess there are criteria to determine it is an oscillation and switch to inertial track: as you know, I mentioned two types of deviations, but in fact there is an infinite number of possibilities. And somehow in the A/P logic, I guess you have to set a limit where you say: that type of deviation I will compensate, and that type, I won't because I'm not sure it is a deviation.
Don't know if I'm really clear here !

Charles

[Magoo]

Not all autolands are about LVP's, if a runway is autoland approved by the company then we can use autoland even if not LVP at the airport. Of course be ready to take over if things turn to s....

[tango4]

I know some airlines have that policy.
The fact is we had some incidents were the field was not in LVP and pilots encountered deviations. The problem is not deviations per se, but the fact that they put in the ASR that they did not expect any, so it was really not clear in their mind what to expect (meaning that even though cleared for cat3 training it is not the same as LVP).
And if sh** happens without being 100% to take over very quickly, well it looks like it gives Singapore at Munich !

[Will]

Thanks all, very interesting replies. From the pilot's perspective, is an autoland more anxiety-producing than a manual landing? My assumption is yes, and if that's correct, it would make for an interesting distinction between real world ops and the world of simulated flight. There are plenty of amateur pilots who aren't good at manual landings and instead use autoland because they can rely on it, whereas in real life, pilots might feel less confidence in the automation and more confidence in their own manual landing skills. Any thoughts?

Also, if autoland is the more anxiety-producing situation, are there times when it's a relief to use it? Any combinations of weather and runway conditions in which real-world pilots feel more confidence in autoland than in their own abilities?

[DougSnow]

My airline, with a lot of types of Boeings, requires an auto land when TDZ RVR is 1000 or below.

[torrence]

Will's comment is thought provoking.

I concur that pure sim pilots like myself frequently use autoland as the 'easy' alternative to manual landings under many conditions.  I'm getting better at manual flying thanks to PSX, forum comments and especially the Britjet tutorials - they're much more fun now and I'm using it by choice more frequently, even though my best landings might still shake up real world passengers.  Of course this still leaves the reason a thoughtful sim pilot might want to master autoland - the old 'bad fish from the galley, both pilots disabled' scenario.  In this fantasy, even assuming all the radios are tuned properly and ATC is helping out, I know that autoland is the only mode I would have any hope of landing safely with (maybe), given that I've never had the yoke of a real 744 in my hands.  I'm sure that even the most expensive joysticks/throttles don't cut it.

For professional pilots I assume the answer is that it's stress producing but there are conditions of visibility etc that make this the only reasonable option, and therefore, per some of the discussion in this thread, you'd better have confidence in the system.  The point I guess for professionals using auto land under 0-0 visibility or similar is this is far from the public's perception of - 'you just push the auto button right, why do we need to pay you?'  It's a serious procedure needing a lot of attention and practice.

Cheers,
Torrence

[Magoo]

The A/P does a very good job at autoland, it just lands a bit longer than a manual one.
It's not more stressful than having the other guy landing, ie someone else is doing the work.
The adrenaline level goes up on an LVP approach though, but the A/C really works hard and always does a good job, I've never spoken to anyone who had to go around from an ILS that was badly handled by the A/P in LVP. Also, in rough weather, read typhoons, the A/P does a better job than any of us, but you tend to override the A/T a bit during the approach.