Far Eastern / Thai Airways TCAS incident

[armchairpilot94116]

Far Eastern Air Transport Flight EF306,
Boeing 757-200 / Thai International
Airways Flight TG659, Boeing 777-300
Near Collision at an Altitude of 34,000
Ft. and 99 NM South of Jeju Island,
Korea
On November 16, 2006
Factual Data Report
July 30,


http://www.asc.gov.tw/acd_files/258-c1contupload.pdf

++++++++++++++++++++++++++++++++++++++++++++++++


What can we learn from this?

[ATC Watcher]

What can we learn from this?

among others :

1: That having TCAS training once every 3 years in the sim is not enough.

2 : Following an RA with a rate of descent of 8000 ft a minute , then leveling off after that to get 2,5 G is going to get you into trouble.( 164 pages of it)

3: That the Taiwanese are very good at collecting data , but much less for learning from these and issuing ( en enforcing ) recommendations.

[armchairpilot94116]

02:05:07.6 02:05:05 UNS514T ROGER THANK YOU UNS514T

02:05:12.0 02:05:09 AI310 INCHEON GOOD MORNING AI310

02:05:15.0 02:05:12 ACC AI310 SQUAWK 4120

02:05:19.0 02:05:16 AI310 4120 AI310

02:05:21.2 02:05:18 ACC
CE5053 RADAR CONTACT CLEAR
DIRECT CHILE JULIET UNIFORM

02:05:25.6 02:05:23 CE505
DIRECT CHILE JULIET UNIFORM
CE5053

02:05:56.5 02:05:53 ACC
UNS514T CONTACT JEJU APPROACH
121.2 GOOD DAY

02:06:01.8 02:05:58 UNS514T 1212 UNS514T GOOD DAY

02:06:21.0 02:06:17 ACC
AAR363 RADAR SERVICE TERMINATED
CONTACT SANGHAI CONTROL 132
DECIMAL 4

02:06:26.1 02:06:22 AAR363 132 DECIMAL 4 AAR363

02:06:29.4 02:06:26 ACC AFFIRMATIVE

02:06:30.5 02:06:27 ACC KE887 CLIMB TO FLIGHT LEVEL 350

02:06:34.4 02:06:31 KE887 CLIMB FLIGHT LEVEL 350 KE887

02:06:50.8 02:06:45 ACC
EF 308 STOP.... AHA... IMMEDIATELY
CLEAR AND DESCEND

02:06:55.9 02:06:50 ACC
TG659 TURN RIGHT HEADING 270, 270
IMMEDIATELY

- 02:06:56 TG659 725, 270, TG659.

02:07:15.4 02:07:08 EF306
OKAY. INCHEON CONTROL EF306,
TCAS TCAS CLIMB

02:07:19.2 02:07:14 ACC ROGER, NOW DESCEND, DESCEND

02:07:22.3 02:07:17 EF306
NEGATIVE, NEGATIVE WE FOLLOW
TCAS

02:07:46.4 02:07:42 FEA306
INCHEON CONTROL FEA306, WE ARE
CLEARED TRAFFIC, MAINTAIN FLIGHT
LEVEL 310

02:07:53.6 02:07:49 ACC FEA306 ROGER, MAINTAIN 310

02:07:57.3 02:07:52 FEA306 ROGER, FEA306


02:07:59.1 02:07:55 ACC
THA659 MAINTAIN FLIGHT LEVEL 340,
NOW DIRECT ATOTI

02:08:05.3 02:08:00 THA659
DIRECT ATOTI, THA659. HOW COME
YOU LET THE THINGS LIKE THIS
OCCURRED. THA659.

02:08:11.2 02:08:06 ACC ROGER, THANK YOU.

02:08:12.8 02:08:08 TG659
HOW COME! WE LIKE TO KNOW THAT
YOU LET THE THINGS LIKE THIS
HAPPEN? WE HAVE TO DESCEND. I
HAD TCAS CLIMB.

02:08:20.7 02:08:17 ACC TG659, ROGER.

02:08:24.2 02:08:20 TG659
NO WE'D LIKE TO KNOW THAT HOW
COME YOU LET THING LIKE THIS
HAPPEN?

02:08:29.2 02:08:25 ACC OKAY, STAND BY

02:09:24.1 02:09:18 ACC
EF306, TURN RIGHT HEADING 070, 070
FOR DESCENDING

02:09:30.8 02:09:26 EF306
TURN RIGHT 070 FOR DESCENDING
EF306. WHAT'S THE PROBLEM. WE
HAVE SOME PERSONAL INJURY.

02:09:44.3 02:09:39 ACC
ROGER. EF306 NOW DESCEND TO
FLIGHT LEVEL 150

02:09:50.2 02:09:45 EF306
DESCEND TO FLIGHT LEVEL 150. WE
ARE TURN RIGHT HEADING 070, EF306.

02:10:34.6 02:10:31 EF306
CONFIRM HEADING 070, DESCEND
FLIGHT LEVEL 250

02:10:42.0 02:10:38 ACC
EF306, HEADING 070 DESCEND TO
FLIGHT LEVEL 150

02:10:46.5 02:10:41 EF306
DESCEND TO FLIGHT LEVEL 150,
EF306. HEADING 070.

02:11:33.2 02:11:28 EF306 INCHEON CONTROL DUE TO… AH

02:11:39.4 02:11:35 ACC SAY AGAIN CALL SIGN.

02:11:41.1 02:11:37 EF306 EF306, WE HAVE PERSONAL INJURY.
REQUEST FOR EMERGENCY LANDING
FOR JEJU AIRPORT.

02:11:53.5 02:11:49 ACC
EF306 ROGER. NOW DIRECT MARIN.
CLEAR DIRECT MARIN. SAY YOUR
INTENTION. SAY AGAIN.

02:11:59.7 02:11:52 EF306
OKAY, DIRECT TO MARIN. WE
REQUEST EMERGENCY LANDING FOR
JEJU AND WE NEED MEDICINE HELP

02:12:08.9 02:12:04 ACC
AH, 306 ROGER, STAND BY. DO YOU
HAVE ……

02:12:26.6 02:12:22 ACC
EF306, ROGER. DO YOU HAVE ANY
NEED? DO YOU NEED ANY?

02:12:31.8 02:12:27 EF306
YES WE NEED EMER...... MEDICINE
HELP AND EMERGENCY AMBULANCE
FOR HELP

02:12:41.5 02:12:37 ACC
OKAY, DO YOU NEED AMBULANCE
OKAY. I.......

02:12:44.7 02:12:40 EF306 WE HAVE PERSONAL INJURY.

02:12:48.2 02:12:44 ACC ROGER, PERSONAL INJURY.

02:12:50.1 02:12:47 EF306 AFFIRM

02:13:08.2 02:13:03 TG659 INCHEON TG659.

02:13:10.5 02:13:06 ACC-2 TG659, GO AHEAD.

02:13:12.2 02:13:08 TG659
IF YOU WANT TO KNOW THE REASON
WHY YOU LED THE AIRCRAFT WITH US
ANOTHER ONE NEAR MISS THAT WE
HAPPENED TCAS CLIMB.

02:13:22.8 02:13:18 ACC-2 TG659, ROGER.

02:13:24.9 02:13:20 TG659
NO, WE'D LIKE TO KNOW THE REASON.
NOT ROGER. WE'D LIKE TO KNOW THE
REASON WHY YOU LET THINGS LIKE
THAT HAPPEN.

02:13:29.8 02:13:25 ACC-2 OKAY, I KNOW. STAND BY.

02:13:36.4 02:13:30 ACC-2
TG659, WE HAVE TRAFFIC. SO TCAS
CLIMB. DO YOU UNDERSTAND?

02:13:44.5 02:13:40 TG659
TG659, WE HAD TCAS CLIMB. THE
OTHER TRAFFIC MAINTAINING WRONG
ALTITUDE.

02:15:49.0 02:15:43 ACC-2
EF306 CONTACT JEJU APPROACH
121.2. EF306 GOOD DAY.

02:15:54.2 02:15:49 EF306
INCHEON 12 … JEJU APPROACH 121.2
EF306

02:15:59.2 02:15:53 ACC-2 AFFIRMATIVE.

02:16:00.4 02:15:55 EF306 GOOD DAY


+++++++++++++++++++++++++++++++++++++++++++++

(unquote)

ATC advising against TCAS. Not helpful and similar situ has caused accidents in the past when pilots couldnt decide who to follow.

[Aviator_IT]

I am currently reading the report, but from the transcript, this is one of the scariest and dangerious conversations I have ever read.

A controller that does not comprehend english, makes a mistake, ignores questions in such a situation from pilots and above all, gives command when he knows that TCAS RA is active!

Thank God those pilots didn't listen to him.

I wonder what will happen to the controller.

[cwatters]

> Following an RA with a rate of descent of 8000 ft a minute , then leveling
> off after that to get 2,5 G is going to get you into trouble.( 164 pages of it)


Interesting reading the report, looks like not a lot of choice but to push down rather smartly...

"When the RA aural warning tone “descend, descend” was issued, I followed the TCAS red T-bar on the ADI and pushed down the aircraft smoothly. Afterward the TCAS issued an “increase descend” aural warning, at the same time I looked outsight from the left to the right and visual contacted that there was a flying object approaching rapidly in front us. So I pushed down the aircraft hard to avoid the traffic."

[gonebutnotforgotten]

Quite extraordinary, in 150 odd pages of accident report about an incident where the issue is gross over-reaction to an RA, there is not a single mention of what the appropriate responses should be. There is a piece of TCAS lore which should be understood by everyone, but is either not known, or ignored, by many. TCAS assumes that you will respond to changes in vertical speed at a delta g of 0.25 (and enhanced RAs, increase climb etc, at 0.35g). Neither is an enormous amount, but it is more than you experience at any other stage of flight (e.g. a 30 deg banked turn = about +0.15g), landing excluded. Right at the beginning of TCAS the FAA didn't help by describing the required manoeuvre as 'undetectable', and ignored completely any suggestion that they come clean, not only about the magnitude (which is certainly very detectable) but even more importantly, how to do it right. In the BA TCAS description and the JAA advisory material there are some simple rules of thumb to help. If you achieve the normal desired V/S change of 1500 ft/min in about 5 secs, then you will be close to the right acceleration (TCAS also assumes some thinking time); for the enhanced manoeuvres 2-3 secs would be appropriate. But that doesn't completely solve the problem as you need still some idea of how you are getting on during the manoeuvre.
Enter the crude estimate of pitch change needed to satisfy the demand (if you have the Boeing trapezoid in front of you you can see this directly, but it still helps to be prepared). Very roughly, the attitude change required to hit +/- 1500 ft/min is 1000/TAS; that translates into 2 degrees only at cruise speed increasing to 7 on the approach. 'Now', I hear you cry, 'you surely don't expect me to work this out in the heat of an RA, do you?' No I don't, I expect you to use the thinking time between the TA and RA to prepare yourself for whatever might happen next. If you can't do that in the typically more the 10secs available, you don't have the mental arithmetic skills to fly a modern jet. The rules of thumb work and should be known by every pilot. Well done to the Thai 569 crew who appear from the traces in the report on page 142 to have got it exactly right.

[ATC Watcher]

This report is list of facts of a bad scenario.
Everybody in this mess did not do what they were supposed to do.
1. The controller levels off an aircraft descending when passing through the intruder/s altitude, at the same altitude at that intruder ( 340) instead of just increasing his rate ( they had 11 NM at that point !).
2. The FAT crew report to ATC a TCAS climb when they have an Decend RA.
3.The controller then intervene and instruct the crew to Decend ( opposite to the RA !)
4. The FAT Capt/ PF overreacted and blew the VSI and the speed limits , (and got the pax injured ).
5. The Thai got a RA to climb but seem not to follow it . ( or only climbed 126 ft in 15 seconds )
TCAS training seems not to be working in that region ( But only there !)
Finally during the conflict, according the report, both aircraft had about 10 NM when TCAS Ras went off , had 2000 ft vertical when they were 5 NM apart and at their closest point of APP ( 1.1 NM ) they had 3000 ft separation .
When TCAS went off in fact the FAT was already 100 ft below the Thai.
So hardly a " near collision " in mathematical terms.
Also the FAT Cpt report that he increased the rate of descent to " visually" avoid the 777 does not match facts. The FAT F/O report is more fitting the facts described later in the report.

[armchairpilot94116]

The island nation of Taiwan has some 23 million people but less then 200 airliners. Thats less then a large American Airline. You can bet they investigate every notable incident. This incident was important because of serious injury to 4 persons. As well as as being potentially serious as a TCAS could be. All TCAS incidents should be looked into I would imagine. Finding out exactly what happened and why it happened would certainly benefit in avoiding similar. Better training all around sounds like a start?


This was a FActual Data Report , not a Final so therefore no conclusions yet.

[Brookmans Park]

A prime case where some ICAO Level 4 English would have made at least a small contribution to Safety

[Capt Pit Bull]

ATCWatcher

5. The Thai got a RA to climb but seem not to follow it . ( or only climbed 126 ft in 15 seconds )

Well, you'd expect their RA to soften, and with the other aircrafts large overreaction it would soften earlier.

Haven't got the tables in front of me, but IIRC TCAS is shooting for a miss of about 700 feet at that altitude. By the time they had started to pitch up, the other aircrafts predicted separation at CPA was probably increasing rapidly through 700', thereby allowing their RA to soften.

pb

[ATC Watcher]

Cat pit Bull :
Well, you'd expect their RA to soften,

Yes, I thought so as well, but when looking at the TG TCAS memory download table on the report pages 39/40 ( or 52 of the pdf) you can see that the RA started at 07.01 asking for 1500ft/min ( normal ) and went on for 11 seconds, always with 1500 . The A/P disengage recorder at 07.00 was probably for the heading change and not the RA.
They did pitch the nose up but did not really climb much.
In any case you are right, with the rate of descent of the FAT, the CPA conflict was soon over. If the TG reacted late, it could explain the small vertical deviation .

Anyway the TG is the least involved in this accident, I grant you that.

[songbird29]

After reading the near collision report I have a few observations:
1. Training. The declarations of the captain and the first officer differ. The captain reports “As I know, pilots in our company receive TCAS training at least once every three years”. The FO must have had a different training schedule: “As I remember, the company regulated the flight crew receiving the TCAS training once every half year.” The report contains data and statements only. More clarity would be required before one can draw conclusions that TCAS training seems not to be working in that region.
2. Confusion. The TCAS RA brings a lot of confusion and agitation, perhaps panic, both in the cockpit as well as at the radar screen.
a. When the controller observes the traffic conflict, he uses the wrong callsign (EF308 instead of EF306) and utters an unintelligible phrase (stop eh immediately clear and descend), which is interpreted as level off. When the controller receives the verbal message about the TCAS RA, he gives an instruction to the opposite.
b. The Pilot in Command overreacts the descent after an unclear visual contact, to 8000’ a minute.
c. The First Officer reports TCAS climb instead of descent.
The human reactions are understandable. More than criticising the controller and the pilots for being inarticulate and overreacting, it shows again that a lot of work and investment is required to get TCAS real safe.
3. Time lapse RA-ATC. Between the TCAS RA and the FO’s advice to the controller the time lapse is 13.8 seconds. That is a pretty good reaction time, far better than the average, but it is still more than could be achieved with an automatic message to ATC, connected to the RA instruction to the pilot.
4. Visual contact. The instructions in the manuals (both from the company and Boeing) to try and get visual contact does not enhance safety, but gives rise to further misunderstanding. The captain has stated that "When the RA aural warning tone ‘descend, descend’ was issued, I followed the TCAS red T-bar on the ADI and pushed down the aircraft smoothly. Afterward the TCAS issued an ‘increase descend’ aural warning, at the same time I looked outsight from the left to the right and visual contacted that there was a flying object approaching rapidly in front us. So I pushed down the aircraft hard to avoid the traffic. At that time I neither know what the flying object was nor identify the distance between us."
This incident obliges the air traffic community once again to look into some necessary refinements of TCAS, beyond its version 7.0. It will require initiative and it will cost money. But look at the alternative, pilots and pax, not to forget the controller, will not always be so lucky as EF306 and TG659. As a minimum the following is required:
1. An automatic TCAS-RA signal to the ground is urgently required so that the controller within 0.1 to 8 seconds knows that something is going on in which he should not intervene anymore. This is technically feasible through the transponder.
2. The manuals about getting visual contact should be reconsidered. Visual contact dates back to the piston engine era, it may still be useful close to ground but it is counterproductive for jets at cruising level.
3. International rules should be refined to indicate that controllers should cease to instruct aircraft after reception of the message that TCAS RA has instructed the pilot to take safety action. If the ATC system has made mistakes leading to the TCAS RA, they should be held resposible for these mistakes (in line with the Swiss verdict after Uberlingen, not so much the individual but the total system). But ATC should not be held responsible for what happens as a result of the RA.

As Armchairpilot has said:
ATC advising against TCAS. Not helpful and similar situ has caused accidents in the past when pilots couldnt decide who to follow
Songbird

[Spruit]

Hi,

Reading this thread is both concerning and enlightening at the same time.

I was wondering about TCAS and the Auto Pilot, when you receive a TCAS alert and the Aircraft is on AP, does the AP disengage or must you manually resume control to action the TCAS RA?

Many thanks in advance for any answers.

Spru!

[ATC Watcher]

Songbird :
More clarity would be required before one can draw conclusions that TCAS training seems not to be working in that region.

And then a few lines later you add :
The human reactions are understandable. More than criticising the controller and the pilots for being inarticulate and overreacting, it shows again that a lot of work and investment is required to get TCAS real safe.

I put that down to system knowledge and training. More work on the TCAS logic you are not going to have.

but it is still more than could be achieved with an automatic message to ATC, connected to the RA instruction to the pilot.

Do not use this case to promote RA downlink, it would not have changed anything in this case, like in most of the others accidents , including Ueberlingen.
This particular accident ( i.e. the pax injured) would not have been prevented by an RA downlink.

I fully agree with your comments on visual contact. The Visual rules have to be reviewed, and some form of EFR ( Electronic Flight rules ) have to be implemented. But this is an ICAO initiative , that will , at best , takes 20 years to set up and 10 more to implement worldwide.
We even do not agree yet on a universal data link 10 years after the adoption of FANS II !


and finally :
Spruit : does the AP disengage or must you manually resume control to action the TCAS RA?

TCAS is not coupled to the AP ( it if was a lot of our problems will be solved :E ) so yes, pilots have to manually disengage AP and manually fly the RA.

[Permafrost_ATPL]

Enter the crude estimate of pitch change needed to satisfy the demand (if you have the Boeing trapezoid in front of you you can see this directly, but it still helps to be prepared). Very roughly, the attitude change required to hit +/- 1500 ft/min is 1000/TAS; that translates into 2 degrees only at cruise speed increasing to 7 on the approach.

:suspect:

Follow the TCAS commands on your PFD, no more, no less, and keep it smooth. KISS principle is never more relevant than when the s*** hits the fan...

The manuals about getting visual contact should be reconsidered. Visual contact dates back to the piston engine era, it may still be useful close to ground but it is counterproductive for jets at cruising level.

Unless someone can show me evidence that TCAS RAs are not reliable, I'd have to say I agree with the above. At the kind of closing speeds involved, humans probably make a poor job of judging what would be the most effective way to resolve the conflict. We might even disregard the RA just because our eyes tell us so, ignoring that fact that we do not know what the pilot in the other aircraft is doing.

P

[songbird29]

The Visual rules have to be reviewed, and some form of EFR ( Electronic Flight rules ) have to be implemented. But this is an ICAO initiative , that will , at best , takes 20 years to set up and 10 more to implement worldwide.

Let's not be too pessimistic.

A change of visual rules may be a cumbersome process, but I was rather speaking of the Boeing manual and the airline manual. ICAO is not writing the Boeing manuals. It's them at Boeing who should not put so much weight on the visual contact following a TCAS RA. It's unrealistic and if it comes to the crunch, unsafe for jet aircraft at cruising level.

Furthermore, don't be too hard on ICAO, which is an assembly of States and influence groups. I would be in favour of reinforcing the ICAO machinery and power, so that the necessary processes can be speeded up.

[gonebutnotforgotten]

Permafrost
'Follow the TCAS commands on your PFD, no more, no less, and keep it smooth. KISS principle is never more relevant than when the s*** hits the fan...'
Well, yes, use whatever you have. But I am not aware of 'TCAS commands' as such yet; Boeing simply tells you where not to be pitchwise, and Airbus does absolutely nothing on the PFD, so that's one case at least it where helps to have a rule of thumb, and I just don't know about Embraer etc. In all cases you still need to know roughly the right rate to react in order to avoid depositing the pax on the ceiling or their knees, while still achieving worthwhile vertical separation. It helps to use the thinking time between the TA and the RA to think about the response needed, that is all. If you get an RA then just count slowly up to 5 as you carry out the manoeuvre, that paces you nicely. A Eurocontrol study program ('Emotion 7'?) showed pilots tended to fall into two distinct camps in responding to RAs, either too much or too little, with precious few in between getting it right.
The Far Eastern 757 would have had the Boeing 'lozenge' appear, but he apparently didn't know what to do with it, or how fast. But we should at least be grateful he went the right way (see many other threads about the dangers of deciding you know better which way to go).

GBNF

[Permafrost_ATPL]

Actually, the NG FCTM does use the words "pitch command" for TCAS. But interesting that Airbus does not give you anything. So are the Boeing indications completely arbitrary or are they based on specific closure rates? I can't find anything in Vol1, FCTM or AFM on how they are generated. If it's not arbitrary, I wonder why Airbus don't have the equivalent. It might not be perfect, but surely it's better than nothing?

Do you have a reference for that Eurocontrol study? I'm curious to see what was deemed "too little" when responding to RAs. Given that it's 25 seconds to point of closest approach, I would have thought that as long as you don't sit there doing nothing for 10 seconds, a smooth transition to commanded pitch should do. In fact the FCTM on the NG says "Properly executed, the RA maneuver is mild and does not require large or abrupt control movements."

Cheers,

P

[ATC Watcher]

Songbird :
Furthermore, don't be too hard on ICAO, which is an assembly of States and influence groups. I would be in favour of reinforcing the ICAO machinery and power, so that the necessary processes can be speeded up.


I will also be in favour of reinforcing ICAO's power, but the current General Assembly is not made of young progressive and dynamic States representatives !
At working group level, things go generally well, but later, the State reviews and the ANC votes !
I was involved in FANS II and looking back at its results 10 years later , I am somewhat bemused. Same story for the AWOP / MLS fiasco.

Back to EFR rules, NEXTGEN in the US and SESAR in Europe might get a welcome boost . Who knows ?

But for ACAS/TCAS , we might get with a lot of difficulties a version 7.1 to take care of the Adjust vertical speed RAs and the sense reversals, but in my opinion that will be the end of it after that. the FAA might even not mandate that change at all.

[gonebutnotforgotten]

Permafrost
I think Boeing's terminology is not the best - the prohibited area isn't a 'command' in the same way that say a Flight Director bar is (adjusting itself to what you are doing) but perhaps I am hairsplitting a bit. Nor is it arbitrary, it shows the pitch attitude necessary to achieve the v/s demanded by the TCAS RA. Nothing to do with closure rates as such, they just alter the timing of the issuance of any RA. I guess the algorithm is proprietary, but can't be a million miles away from the very unsophisticated rule of thumb. The reason why Airbus don't use it is rumoured to be because of patent issues... which is frankly baffling. I recall seeing a display essentially identical to the current Boeing one in a very early A320 development sim down in Toulouse in 1986 or 7 (it also included horizontal advisories...) long before Boeing/Honeywell offered it for glass cockpits.

Re slow and aggressive pilot responses - the study was in fact 'ACASA', part of Emotion 7 I think. I can't paste the relevant chart into this page for some reason, but you can find it on page 37 in the original report at http://eurocontrol.int/msa/gallery/content/public/documents/Safety/WP1.pdf. the data only respresent a handful of analysed flights, but the message is clear enough. This stuff was published in 2002, but I very much doubt things have changed a great deal since.

GBNF