Posted on 07/08/2013 7:37:00 PM PDT by justlurking
I used the data for Asiana Flight 214 from FlightAware to create a series of graphs to help Freepers understand how the approach to landing deviated from the expected behavior.
This data is collected from the FAA computer systems. The location data is either calculated by the FAA radar or reported by the aircraft transponder. The altitude is reported by the aircraft transponder. The ground speed and vertical descent rate is most likely reported by the aircraft's Mode S transponder, but it's possible these were computed by the radar system.
This first graph is a plot of the latitude/longitude in the last 15 minutes of flight. Each blue dot is a reported position, the blue line connects them. The red line is runway 28L. North is at the top of the graph. The aircraft approaches from the northwest, crosses over the runway and makes a descending left turn to final:
Pilots will recognize this as a modified crosswind VFR approach to land. For non-pilots: the classic crosswind entry to the landing pattern is:
The aircraft was descending through this entire maneuver. The track didn't follow the crosswind pattern exactly, but the pilot would have been following ATC instructions. My point is this is a pretty standard maneuver, and you shouldn't think it is unusual or difficult.
The next three graphs are are stacked together, so that you can see how the different metrics change together. All start 14 nautical miles from the runway threshold when the aircraft is established on the final approach, and end about 1,000 feet from the threshold. This is the last data point that I believe to be valid, as the final data point in the FlightAware data appears to be post-impact.
Note that the altitude is reported in increments of 100 feet. So, there will be a bit of error introduced in the individual points, but the overall trend will be accurate.
The first graph is altitude. The blue line is the aircraft, and the red line is the 2.85 degree glide slope that would be projected by the PAPI lights on 28L, if they were in operation. There are conflicting reports whether the PAPI was in operation, but either way it is still a way to project where the aircraft should have been on the glideslope.
The second graph is ground speed, in MPH. Note this isn't the same as airspeed: if there was a 5 MPH headwind, add another 5 MPH to the ground speed to yield the airspeed. The red line depicts the 157 MPH landing speed typically used by the 777. Stall speed depends on the flap configuration and gross weight: I'll defer to someone more knowledgeable to provide that information.
The third graph is vertical descent rate, in feet per minute. A constant descent rate for this approach would have been about 700 FPM.
From these graphs, you can see several things:
I'll repeat that this track log ends about 1000 feet from the runway threshold. Passengers and observers have reported the aircraft was too low, but that was apparently at the very end of the final approach.
If this data is accurate, the aircraft was too low and too slow because it was too high earlier in the approach and the pilot over-corrected.
Die Hard
Change the reading to the altimeter and the pilot won’t be able to recover at the last minute
CC
Sounds like he screwed the whole approach up, starting 12 miles away.
Ping for your info, review, and commentary.
Should have “gone around” at 3nm ?
Over-compensated for the altitude and airspeed, and followed later with an “oh sh!t” moment, probably. That second increase in descent rate happens at the same time as a sharp drop in airspeed. Is that possible? And then finally it looks like a serious pulling up on the stick.
It was still recoverable at that point. But, the closer they got to the runway, recovery became more and more difficult.
In a smaller aircraft, it's easier to correct. But, a 777 has a lot of inertia, so you must anticipate, rather than react.
Compare a jet ski vs. an oil tanker. The difference between a Boeing 777 and a Cessna 150 isn't as large, but the principle is the same.
Excellent presentation of the pertinent data to show a story as old as the Wright Brothers: running out of airspeed and altitude combine to become non-optimal career disintegration factors.
I would be interested in how far he was behind a heavy and what his flight profile was in relation to the aircraft he was following.
If he was slow and if he allowed the aircraft to get below the flight path of the heavy ahead, wake turbulence could have been a factor.
I am surprised that a cockpit full of highly experienced pros got caught in this trap. Too high too close is one of the oldest traps.
Instrument pilots know very well that if the aircraft is not in proper configuration four miles from touchdown: Power settings, aircraft config, speed and rate of descent stabilized, the landing will not be a good one.
Much has been made of the fact that the ILS was out, as was the VASI.
I quit flying quite a few years ago, so I no longer subscribe to Jepps, but I suspect that LAX has a GPS approach. Even if it does not, a prudent pilot with GPS or an inertial nav system would set up the touchdown point in the GPS and have from that a “glideslope” that would allow him to set up an approach that was proper.
That would have made landing at an unfamiliar airport much easier.
In other words, xx miles out at xxxx altitude would require a rate of descent of xxx.
Depending on aircraft type and how “slick” it is, one would pick a point 3 to 5 miles outside of the point at which he wished to start the final descent and at that outside point slow the aircraft, set approach flaps and power and have all check lists completed.
In a recip engine aircraft, which is all I ever flew, one would only have to pop the gear and slide down the proper path.
He was already at stall with nose high trying to maintain altitude.
W the throttles, instead of giving it forward yoke the engines kicked the nose up further and he stalled out!
Actually he should have executed a missed at least 4 miles out when he couldn't even come close to captuering the glide slope.
I used to fly a single into LAX twice a week and they always had me keep my approach speed to 150k minimum until crossing the numbers.
Remember that is ground speed, not air speed. Now, consider what happens if the aircraft is maintaining roughly constant velocity, but that velocity vector suddenly points down at a steeper angle, rather than forward and down. As if someone were diving for the glide slope... Vertical decent rate goes up while horizontal ground speed decreases. The aircraft was traveling more down than forward. Or at least the proportions of forward and down changed significantly. Dirtied up though the aircraft wouldn't have gained much actual airspeed to trade back for altitude later. Just scrubbing off energy...
Yes, if they reduced power and raised the nose.
Fortunately, the new digital flight controllers record all that information, and the NTSB will be able to recreate every action (and the change in aircraft attitude/speed) during the approach.
And it will be a lot more accurate than this data.
So you think that second dip in vertical speed was the actual stall?
There is no question the glideslope was out of service. The localizer was operating.
However, there is some dispute whether the PAPI was operational. The posted NOTAM is a few hours after the crash.
I went through a thread about flight 214 on a professional pilot's forum, and one pilot reported the PAPI was working the night before the crash, and another reported it was working only a few hours before the crash. But, it had been out of service earlier, which I think is confusing the issue.
what was the data entered re: elevation of SFO. I remember back in 2004/5 a 767 plowed into the ground b/c the elevation of destination was entered 40’ or so above actual. Fly by wire is good as long as data entered is accurate
As bad as it was, at least this pilot MADE it to the runway! I recall a commercial jet in the late 60s or early 70s that ended up landing in SFO bay about 1/4 mile short of the same runway that the Asiana Airways jet smacked.
The earlier jet spent a week or two in the water before a barge was brought in to lift it out and take it the last 1/4 mile to the asphalt!
Way past bedtime for a look-see...will ping tomorrow, PM. Thanks much.
Friends and I have been chatting this last couple of days. With all the pilots on the deck nobody said “too low” until the last minute when nothing could be done? Even going max power the plane doesn’t accelerate like the old prop jobs did. There’s sink rate involved and inertia to overcome before anything happens. A couple seconds to recover???? Good Grief!!! What did they say? “Gee that’s pretty water down there and oh my, look how close we are to it - hey lookit! A shark/whale/tsunami debris”
He was already in trouble and trying to save it from a stall is my guess.
At that point, he’s lost it and should execute a mised and try another approach.
He doesn’t have to ask for a missed, it’s his decision and all he has to do is announce it and execute it.
It is still mind boggling to me how such experienced pilots could eff up the approach that badly in really good weather without realizing it early on and going around safely to try again. But there is probably more to the story than we know at this time.
I was in an airliner years ago that had to go around twice before nailing it the third time, but this was in very poor visibility. I was sure glad they finally nailed it because I did not want to go around again....the nerves were starting to build and the in flight bourbon had worn off.
I've never flown anything larger than a C172, except in a simulator.
However, I believe ALPAPilot has actually flown the 777. I would be interested at what point he would have done so.
Great point!! I am also curious why the tower crew didn't wave him off and send him around? At some point, I would hope that someone was looking out the window and should have noticed that his descent rate was a little high and his forward speed should have appeared to be off.
I'm not trying to pin this on the tower crew, I know they're busy. But, when it comes to flight safety, it's a team effort - flight crew+tower crew+ATC. They ALL play a part!!
Only problem is that most aircraft do not drop the tail when stalling. I think the pilot added power and pulled up the nose to do a go-around, and when the tail dropped as the nose went up, it hit the seawall.
I think this was a cultural issue. It has happened before on a South Korean airline.
The captain was new to the 777. But, he was the pilot in command. The junior officer may have been afraid to correct him until it was too late.
CC
I just wondered if you had seen the visual footage of the plane crashing.
W the throttles, instead of giving it forward yoke the engines kicked the nose up further and he stalled out!
That would happen with a prop driven aircraft but in this case I question whether the engines had developed enough power at that point to cause a pitch up.
As for 150K to the numbers, that is fine on an airport you have used many many times but sure poison on an unfamiliar airport that is marginal.
Obviously 10 or 12,000 ft. runways are not marginal for a light aircraft. With that kind of runway, I always landed well down the runway in order to stay above the path of heavy aircraft ahead of me so as to avoid wake turbulence.
But marginal runway length, strange airport, I flew with the same seriousness as a 200-1/2 “ragged”.
This guy should have been as serious.
As it is, he had a career changing event.
I’ve taken several CRM classes (Crew Resource Management) where you learn job sharing. These crews are more intune I think than that. Maybe years ago but in this case even though the flying pilot (PIC) had the controls and even though he had over 8000 hours flying and even though he’d flown SFO before it was the duty of the non-flying pilot (FO) to monitor things and let the PIC know how he was doing. I heard Rush down here this very early morning and he had a retired 767 pilot on and it was a very eye opening expose of that cruise to landing drama. Plus I’ve seen flight profiles of that landing plus other landings this guy had done and they were scary. And - he had only 43 hours actual flight time plus sim time. He did a lot of sim landings into SFO. I’ve flown the 777/767 flight sim at Boeing (now called Flight International) - it can really be a sweat job. The guy running the sim can give you all sorts of nasty scenarios and you’d better ace them all. You’ve only got one shot at a landing.
I have, but I believe that video starts right as this track log ends.
It's really low and slow at that point.
CC
Bookmark
The crew wouldn’t warn the pilot (cultural), and going around would have been a loss of face (same).
We often hear about how having the “gotta-goes” can kill, this guy had the “gotta-lands”.
And so he flew a perfectly good airplane into the ground.
We have all heard really funny exchanges between ATC and pilots.
One I remember reading about was between a large jet and a controller in Spain, as I recall.
The pilot stated that he could not get down in the distance to the airport. The controller’s reply...”Slow it down enough and it will come down.”
Which reminded me of the many times I was kept too high too long...So I wonder when the pilot called the airport in sight and was cleared for a visual...or if ATC kept him high too long, thus adding to his inexperience in the 777?
The tapes between the aircraft and approach control...as well as the center, would be interesting.
A high, fast start is never good, the weight of a 777 and you have big problems.
When I say capture a glide slope, I mean a computed altitude to begin a normal approach. You could compute one or use the tip over point off the ILS approach. I realize the ILS was OTS.
CC
Then I take it angle of attack can do all that at low speeds, like during landing. So it’s looking to me like they tried to reduce speed by pulling up the nose, but then noticed they were dropping to fast. Which caused them to gun the throttle. But the airspeed continued to drop, maybe because the nose was too high. My shot in the dark at best. Time will tell.
The NTSB spokesperson said he was set up for a 17 mile straight in, and cleared for the visual at something like 12 or 14 miles. NTSB is now releasing that the airplane was going 30 kts below the target airspeed. That is all kinds of screwed up.
43 hours in the 777 with no landings at SFO. The instructor captain’s first training trip.
I don’t intend to be rude, but only to make a point.
It matters not what they did during the last minute of flight.
It is what they did not do 10 miles out that crashed that airplane.
Think of an airplane as like a slide rule. You put certain numbers in and you get certain numbers out of it.
No matter how hard you wish it, the numbers always come out the same.
Nothing they did or could have done within the last minute would have saved them.
10 miles out they should have been within an acceptable envelope...altitude, airspeed, configuration.
Five miles out, the numbers should have been exact. Airspeed, power settings, aircraft config...all exact to give a known rate of decent at bug speed and correct power settings that assure the ability to execute a missed approach at any point before actual touchdown.
I know nothing about a 777, but I am pretty certain that this crew was operating below Vmc and eventually ended up below every V.. Obviously if it quit flying.
OK.
Thanks.
So the next question would be at what altitude was he at the 17 mile mark and why was he at that altitude.
If ATC kept him high, they might have a part in this.
At 17 miles, what did the pilot do? Drop the nose and end up too hot? Drop approach flaps, gear, stabilize speed at an approach speed? Obviously not the latter.
Why he was high at the 17 mile mark is probably one of the serious keys to the cause.
And it could be that he did not start his descent at that point. Cleared for a visual means that all altitudes from that point to the ground belong to that aircraft, so he can dump it then or wait 5 or 10 miles to start down.
So in the end, no matter what, it is pilot error.
No you’re not rude, but it seems obvious that they tried to do something to correct their already incorrect approach. Look at the sudden maneuvers and that appear to get increasingly desperate toward the end. Is this a normal approach? There seems to be indications that it wasn’t. The question seems to be were they trying to correct their approach or didn’t even know it was off? And if so, what did they do wrong?
Like I said...what they did wrong was in not having the aircraft properly set up 10 miles out.
Look at it like this.
You shoot yourself in a vital area. Should you call 911, your family, your doctor or put a band aid on it?
You die.
What did you do wrong? 911? Band aid?
None of the above.
Correct answer....should not have shot yourself.
By the way, the distances I use in my examples are not necessarily the correct distances since I have no idea how long it takes to slow a 777 down.
In the light twin I flew, gear down speed could be pretty high.
I could keep the speed up to 3 miles from the marker, providing there was no significant tail wind at that altitude, drop the lights, drop the gear, set approach flaps, set the props and hit the proper speed for the glide slope.
A large jet would no doubt need a lot more room...
So my point was to the proper method, not the proper distances as they would vary according to type of aircraft.
The point being that good landings are already decided 3 or so miles outside the marker providing there is no total incompetence or mechanical problem.
Not a pilot, but from what I observed and heard from witnesses, it looks to me that at the last second they tried to gain altitude and pulled the nose up causing the tail to hit the break wall. This major impact of the tail caused the fuselage to slam onto the runway forcing the left and right landing gears through the wings disconnecting them from the plane. Without the gear the engines were on the ground and torn from their mounts.
Had they just landed the plane, albeit well before the landing zone, but paved, it would likely have been just a hard landing. It was too late for a go around, they should have realized that, but I guess they should have realized they were too high, causing too fast a decent rate, and too slow an air speed and didn’t until too late.
Multiple-pilot error.
Nearly all transitions between models for pilots require training and checkout in the new model, even if it is a step down in equipment. And then most airlines additionally require 'route checks' in the equipment, usually when in revenue service with a check pilot/trainer.
The PF did have only 43 hrs on the 777, with no landings in it at SFO, but had nearly 10,000 hrs total, and many landings at SFO in other types.
The PNF was experienced in type and route, but this was his first rodeo as a trainer/check pilot. After 35 years as an ATC, and dozens if not hundreds of trainees, it can be a fine line between too little or too much input. Obviously this trainer was paying way too much attention to the trainee and way too little to his primary duty....callouts, scans, checks.
So, while the PF may appear at first blush to be the bad guy, I would put the trainer/check pilot at the head of the bad guy list. More on my next post.
You can see the altitude, course, and ground speed for the entire track at FlightAware, starting shortly after departure, all the way across the Pacific. It's the URL link for this article, and the source of my data:
http://flightaware.com/live/flight/AAR214/history/20130706/0730Z/RKSI/KSFO/tracklog
At 17 NM out, the aircraft was still turning to final, descending through 6,000 feet at a ground speed just below 250 knots.
In the three graphs I posted, I started at 14 NM out, because that's when the aircraft was within a few degrees of the final approach course. At that point, the aircraft was only about 400 feet above the 2.85 degree glideslope, or 100 feet above the 3.0 degree glideslope.
After that point, the descent rate was almost right on at 800 FPM, but declined to as low as 200 FPM. And the airspeed remained above 200 knots. It wasn't long before the aircraft was almost 1000 feet above the glideslope, and still over 200 knots.
Providing there is no total incompetence or mechanical problem. Why didn’t they abort the landing, come around for another pass? They either didn’t know they were on a bad approach, or someone thought they could correct it. But it doesn’t seem like mechanical failure yet.
Never got to LAX. All of my flying was from Denver to the east so I have no idea what that end of the runway looked like.
But you are probably right that if the only thing that hit the dike was the tail section, that if the aircraft had been in a normal landing attitude it would have cleared the dike and landed in the over run area which probably would have resulted in a hard landing and wiped out the gear and still done a lot of damage, but maybe not as bad.
However, the view from the cockpit might have been more alarming so there could have been a valid reason for attempting a go-around.
But there is no valid reason for the stupid airspeeds.
All milti-engine aircraft must be operated at speeds that enable them to execute a missed approach even in the event of an engine failure.
Airline pilots will be furloughed or worse for violating that rule.
Many years ago, in the days of Eastern, a guy I knew who was a captain with Eastern flew a route that had a stop in his/our home town.
He thought it great fun to impress us with how he could land and turn off at the first taxiway, and on the favored runway that was a feat in a Bonanza.
He was doing it in a 737.
Eastern caught him at it and he got a long long furlough.
The stupid thing about that was that the terminal was at the far end of the same runway and it was actually an operational advantage to roll all the way down the runway.
He later killed himself and the son of a friend of mine, showing everyone how good he was.
But that is another story for some other time.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.