Turkish A330 incident, Kathmandu

[Chronus]

According to a Turkish daily great difficulties are being experienced in moving the aircraft. For recovery photos see:

Nepal'de pistten ç?kan Göbeklitepe uça?? imkans?zl?klar içinde kurtar?lmaya çal???l?yor - Milliyet Haber

The aircraft`s name painted below the cockpit, translated from Turkish reads:

Belly Hill

Rather an ironic coincidence.

[silvertate]

B777

Turkish Airlines has the approval for this approach , the aircraft and the crew were certified and they had the necessary sim training for this app to ktm

As a matter of interest, does the A330 give adequate error warnings for this type of approach? Our more steam-driven type does not adequately inform you if it has not gone to RNP 0.3 mode; it gives little warning of deviating from lateral track; and gives no gives warnings for deviating from vertical track (just the path indicators). At least the ILS will shout 'glideslope' at you.


In a similar fashion, our steam-driven FMC gives very little indication that it has lost GPS signal, or is getting GPS ghosting from mountains etc: Our current FMC manual gives no error messages for RAIM monitoring of the GPS. Some warnings may be in the box, for all I know, but they are not in the current manual. And I have seen no method of verifying "RAIM OK" before making an approach. And since there is no WAAS in Turkey (and most of Europe), you are entirely reliant on RAIM checking the satellites correctly - especially in somewhere like KTM or Turkey where signal ghosting from high ground may be problematic.

This is why our outfit is not certified for GNS approaches, and why crews should not be doing ad-hoc GNS approaches just because they are in the database. Someone said earlier that we should stop blaming crews? Who is blaming the crew? The problem is that various CAAs have started leaking out GNS approaches without informing anyone as to the requirements. And many companies have been slow to act with clear policies and instructions.

[cucuotto]

Both captain and FO Turkish nationals. Female captain I' m been told. Touched down out of runway at least with one mlg...possibly confused edge with centerline.All actors RNP approach certified but approach type seems to have no influence on event. I stand to be corrected.Info coming from friend working in THY.

[turker339]

As mentioned above, Turkish was authorized for the RNAV AR approach and the pilots that fly it were trained and signed off for it.

Apparently pilot announced a hold of 90 mins but told pax not to worry as they had more than enough fuel for it.

[aterpster]

Silvertate:

As a matter of interest, does the A330 give adequate error warnings for this type of approach?

Dual GNSS
Dual FMS
FMS most provide continuity, integrity, accuracy, and alerting.

Auto-flight
Auto-throttles

This is the case with all airframes certified by the O.E.M. for RNP AR. Additionally, for RNP of less than 0.3, and/or missed approach of less than 1.0, there must be at least one IRU or its equivalent.

[ATC Watcher]

cucuotto

Touched down out of runway at least with one mlg...possibly confused edge with centerline.All actors RNP approach certified but approach type seems to have no influence on event.

Well if your info is correct that confirm us that there was not 900m vis.

As to whethter the APP type had nothing to do , well. we'll see but one has to remember the definition of RNP 0.3. . the GPS is not designed to align you with centre line and to maintain you there , but just to align you with the runway with a precision of 0.3 of a NM orv 500 meters.

[Capn Bloggs]

Originally Posted by ATC Watcher

one has to remember the definition of RNP 0.3. . the GPS is not designed to align you with centre line and to maintain you there , but just to align you with the runway with a precision of 0.3 of a NM orv 500 meters.

In theory... but do you really think a system would be approved for RNP 0.3 (eg RNP Approach LNAV) if all it could do was put you within 0.3nm of the Centreline? That is "miles" off and you'd never get in on any min-vis approach. What about RNP-AR at RNP 0.1? "What are you complaining about? You're only 200m off the centreline!".

[ATC Watcher]

Cap'n. bloggs

I'm with you but that is what the system is certified to , due various factors among which is the integrity of the signals . That is why some build expensive ground augmentation system to get precision approaches . We agree 95 % of the time GPS will get you single meters precision , but in the other 5% you can be off mark and it can be sometimes quite a bit. Fact of GPS life.
I am not saying this is what happened here , no idea , like most of us , but something to keep in the back of one's mind when trying to disregard the approach type as a contributing factor.

[de facto]

JFI there are hundreds of below standard Expats CP's (I've been exposed to at least 50+ in the sims and Check flights) out there operating day in ant out! Ask yourself every day "Do you feel lucky" or are you really doing your job?

Were you working at THY as TRI/TRE?

[aterpster]

ATC Watcher:

Well if your info is correct that confirm us that there was not 900m vis.

Isn't the AB330 an Approach Category D airplane? I so, then minimums were DA of 4670 (352) and 1600 meters. Or, if the ALS were considered inoperative because of the temporary displaced threshold, then visibility minimum of 2000 meters.

As to whethter the APP type had nothing to do , well. we'll see but one has to remember the definition of RNP 0.3. . the GPS is not designed to align you with centre line and to maintain you there , but just to align you with the runway with a precision of 0.3 of a NM orv 500 meters.

Agreed. But, the airplane would only be off by .3 of a NM 5% of the time on a statistical basis. Usually, with autoflight connected the lateral displacement would be significantly less. Something that could easily be handled with visibility of 1600 meters.

[Stone_cold]

Actually ATP , the A330 is cat C/D depending ...not sure what weight variants TK was using .

Re your comments regarding "out of the box " RNP AR certified and ICAO . The aircraft are not necessarily certified out of the box , it depends on operator requirements and many operators have databases and charts for approaches that they are not approved for( in other parts of the world ), read ICAO . E.g. RNAV / CAT II/III ,Circling .

Seems you are not so critical of the Delta runway excursion which seems to be closer to your area of regulatory oversight and happened on the same day .

[Bus Driver Man]

Originally Posted by aterpster

Isn't the AB330 an Approach Category D airplane? I so, then minimums were DA of 4670 (352) and 1600 meters. Or, if the ALS were considered inoperative because of the temporary displaced threshold, then visibility minimum of 2000 meters.

The A330-200 is a Cat. C airplane. I assume that the A330-300 is Cat. C as well.

According to AVHerald, the visibility was 1000m at the time of the approach.
The minimum required visibility is 900m, according the AIP approach chart.
However, according the THY approach chart from their LIDO database, the minimum visibility is 1.5km, which means that THY is only allowed to fly the RNP AR approach with a higher required visibility.

[Algol]

Algol:" Were the crew local or expat? ".

Boeingrestricted: "Will your professional opinion depend on the answer to your question".

Many early contributors to this thread made derogatory references to pilot standards at THY. The drift was the Turks are all dangerous.
My question was meant to remind them that there are many expat pilots on the A330 at THY.

I was wondering how they'd react if the crew turned out to be wholly or partly Expat.

Hope that answers your question.

[silvertate]

Cpt Blogs

In theory... but do you really think a system would be approved for RNP 0.3 (eg RNP Approach LNAV) if all it could do was put you within 0.3nm of the Centreline? That is "miles" off and you'd never get in on any min-vis approach. What about RNP-AR at RNP 0.1? "What are you complaining about? You're only 200m off the centreline!".

On all the GNS approaches I have done (practice VFR approaches for training), the approach is a pseudo non-precision, and not an ILS replacement. The idea is to facilitate difficult approches in high terrain. So many of them bring you in with a large approach offset, terminating in a decision of about 1,000 ft or so.

If you have a 3 nm missed approach point, there is not much point in having lateral accuracy much greater than 0.3nm. The approaches are nice, down the valleys etc. But there are gotchas, like where do you go if the FMC loses the curving path half way down.

[silvertate]

It wasn't my point. As far as I read it, Capt Blogs was saying the system should be better than 0.3 nm rnp. I am just saying that for all the GNS approaches I have done, 0.3 nm was quite sufficient because of the offset and high missed approach point.

I note on page 2 of this thread, that the GNS approach at VNKT is actually straight in for the last 3 nm, with a decision of just over 300 ft. That is a max 15º lateral turn to achieve the threshold, if you are at the margins of the 0.3 nm error for some reason. Not too bad. I'm sure someone has worked out that that is a sufficient margin. But do bear in mind that the rnp is only a 95% probablity, so five times in every hundred approaches you could be off by more than that.

[silvertate]

Aterpster

Dual GNSS
Dual FMS
FMS must provide continuity, integrity, accuracy, and alerting.

Indeed, that's what ours says too. But as far as I am concerned a little message on the scratch-pad of the FMC or on the PFD, is not the same as a big red flag on the PFD or the GPWS shouting 'glideslope'.

And dual GPS receivers are of no benefit if there is ghosting or other signal interference, as both receivers will pick up the same error. GPS signals are not as strong as an ILS, and low-level GNS navigation is a whole new ball-game.

[aterpster]

The alert is supposed to be in your primary field of view. And, of course both receivers will pick up the same error, but the receivers will detect such and send it to the FMSes.

[aterpster]

Silver:

On all the GNS approaches I have done (practice VFR approaches for training), the approach is a pseudo non-precision, and not an ILS replacement. The idea is to facilitate difficult approches in high terrain. So many of them bring you in with a large approach offset, terminating in a decision of about 1,000 ft or so.

There cannot be any offset in an RNP AR approach that has RF legs. The final roll out point (FROP) must be aligned with the runway centerline.

If you have a 3 nm missed approach point, there is not much point in having lateral accuracy much greater than 0.3nm. The approaches are nice, down the valleys etc. But there are gotchas, like where do you go if the FMC loses the curving path half way down.

If the approach value is less than 0.3 and/or the missed approach is less than 1.0, then you must have independent position information that provides sufficiently accurate position information for extraction (typically one or more IRU). It's called extraction because you may get "unable RNP" well prior to the MAP.

"Dual string" is the term the avionics engineers use. The VNKT approach is dual string because of the missed approach requirements.

[Capn Bloggs]

Silvertate, I don't want to be harsh but it seems that you and I have been on different GPS-Approach planets for over a decade. I have done dozens and dozens of these over a decade or more using three different systems and the navigation accuracy is "on the centreline", and they were not RNP-AR systems.

I am just saying that for all the GNS approaches I have done, 0.3 nm was quite sufficient because of the offset and high missed approach point.

Most of ours (RNP APCH LNAV) have MDAs of around 450-500ft. Very few are offset. 0.3nm is not "quite sufficient" to be able to get in if you pop out at the MDA.

I note on page 2 of this thread, that the GNS approach at VNKT is actually straight in for the last 3 nm, with a decision of just over 300 ft. That is a max 15º lateral turn to achieve the threshold, if you are at the margins of the 0.3 nm error for some reason. Not too bad.

Are you serious?? 0.3nm offset from the centreline "at just over 300ft". I defy anybody to safely get a medium/big jet onto the runway from that far off the centreline. Have you actually tried it? I am sometimes 0.1nm off and it is very hard work to get onto the runway. BTW, there is no offset on the approach posted on page 2. The runway QFU is 022° and the approach track is also 022°.

But do bear in mind that the rnp is only a 95% probablity, so five times in every hundred approaches you could be off by more than that.

That's nice to know; we and I assume hundreds of other operators around the world are using dual-GPS sensors for sole-means approaches. No other navaids, and no alternates. If a VOR, NDB or ILS approach had such bad "reliability" it would never be certified.

[ATC Watcher]

capn Bloggs

That's nice to know; we and I assume hundreds of other operators around the world are using dual-GPS sensors for sole-means approaches. No other navaids, and no alternates. If a VOR, NDB or ILS approach had such bad "reliability" it would never be certified.

What Silvertate said is quite correct . It is 95% of the time for RNAV and no, VOR DME or ILS Approaches performances are not the same because they are fixed based and their exact position is known 100% of the time..GPS stand alone APP is no ILS substitute . The 5% uncertainty are still there and will remain there as long as GPS ( as opposed to EGNOS , etc..) is used.
I am very surprised this fact is not known to Pilots operating the sytem.


Just a reminder for the younger generation :
When RNAV/RNP was designed RNP1 was the lowest. RNP 5 the norm aimed at. . "Nav capabilities based on sensors able to calculate your position within 5 ( or 1) NM radius 95% of the time" .
This was designed initially to reduce lateral separation on non-radar enviroment, and allow the design of closely parallel routes in radar airspace.

RNP 0.3 was intially never designed for precision APP but for helicopters in Terminal aereas. How it progressed to what it is now is due to the pressure ( lobby) of operators ( Alaska Air was the first if I remember correctly to test it in remote places in Alaska where no ground aids were available)
Nobody at the time expected the thing to be used in KTM by a non-based airline in bad weather /visibility. ( no pun meant against TK, just to illustrate a foreign crew using a system in a difficult airport a few times a year and again , not suggesting this is the cause of this accident.)
RNP 0.3 may be used in KTM in 2015 with 900m RVR successfully by many, but it is still a GPS and 95%, plus a having a USAF general somewere having access to a button that can further degrade accuracy (SA). But this is probably included in the 5%.