[Updated] Asiana Air 214 crash landing analysis: the B777 flew a high, fast and steep approach and almost stalled on final

Jul 07 2013 - 13 Comments

On Jul. 6, 2013, at around 11.28 LT, Asiana Air 214 Heavy (OZ214), a Boeing 777-200 (HL7742) operated by Asiana Airlines from Seoul to San Francisco International with 291 people on board, crash landed on runway 28L at KSFO.

As a consequence of the impact with the ground the tail separated along with several parts of the fuselage, undercarriage and engine before the main part of the airframe came to a rest next to the runway.

Two passengers died and 49 were seriously wounded.

AAR214

Based on the reports and analysis of the images of the wreckage, it seems that the 777 touched the ground short of runway threshold. Asiana representatives said the aircraft did not suffer any technical failure. However, the NTSB investigation will say what went wrong.

In the meanwhile, we can try to analyse the details already availabel.

The aircraft flew the approach as expected: under VMC (Visual Meteorological Conditions) with visual final (since the ILS was under maintenance according to NOTAM for KSFO).

The playback feature of Flightradar24.com for AAR214 shows the standard route for 28L. Still, by comparing the type of approach flown on the day of the incident and the one flown for instance on Jul. 4 (again with FR24 playback), you’ll notice that on Jul. 6, the final approach started a bit higher. This is not a big deal, unless the crew set a steep approach that, for one reason or another one, did not properly correct until the aircraft was a few feet above the water.

Indeed, the Vertical Speed (at least the one broadcast using ADS-B by the airplane and recorded by FR24) is higher than one would expect (-1,536 fpm at 1,000 ft). Few seconds later, as the aircraft is at 75 feet of altitude, the airspeed falls to 85 kts with a vertical speed of -768 fpm, meaning that the plane is almost falling from the sky.

Provided the data collected by FR24 is correct, it seems to prove that for whatever reason, the incident unfolded as follows:

1) the aircraft under almost perfect weather was radar vectored for the final approach

2) the flying pilot flew the final approach higher and faster than expected: you can compare the approach of the AAR214 with the one of the All Nippon B777 that preceeded it. Abeam Coyote point, the ANA8 (NH7) is at 1,500 ft, -896fpm of vertical speed and 167 kts of horizontal speed whereas the doomed AAR214 is at 2,175 ft, -1,152 fpm and 186 kts. In other words, the Asiana was higher, faster and descending steeper (spoilers, full flaps etc) to intercept the proper glide path.

3) Most probably for a human error (whose root cause may be attention focused elsewhere, another type of cabin problem, distraction, not enough experience in flying visual approaches with the type, etc.) the aircraft approached the runway with a steep angle, the pilots tried to correct it with a bit of nose up to make it to the runway and avoid the plane from touching down before the threshold (if not in the sea) and this caused the aircraft to strike its tail first and then fall on the runway.

4) All the incident unfolded in front of UAL 885. Investigator will probably get some interesting details not only from the crew and CVR (Cockpit Voice Recorder) but also from the pilots of the United 885, a B747 holding short for departure to Osaka.

The above development is coherent with the video just released by the CNN.

You can see the final seconds of the flight: the B777 looks slow on final with a significant nose up attitude. It is struggling to remain in the air enough to get into the runway.
Beware: the video shows the very last part of the flight, when the pilot flying the plane has already applied the correction (advancing throttles to gain speed).

Noteworthy is also how the aircraft pivoted after impacting the ground: it’s a miracle almost everyone survived the incident.

Update: it looks like the pilot flying was actually in training, had just 43 hours of experience with the B777 and had never landed at San Francisco before. These could be contributing factors to the incident.

Enhanced by Zemanta
  • muiris houston

    Nice analysis, thanks!

    This is probably a really facile question but thanks for answering it: if the ILS is inoperable on that particular runway at SFO and the weather deteriorates to a point where autoland is the safest option, how can this be achieved?

    Muiris

    • cencio4

      You may opt for another runway equipped with a working ILS, wait for the weather to improve or divert to an alternate airfield :)

  • tdepp

    Great point by point analysis.

  • GogogoStopSTOP

    When there are incidents like this, I always leave the mass media & go for the true understanding of veterans like your self.

    First reports: ‘struck a sea wall.’ There is no sea wall.

    Thanks for the straight skinny. Your acumen shows thru!

    • William Benjamin Abbott IV

      Call it what you want. but don’t let preconceptions get in your way. “Seawall” is as good as “Shoreline” or “Coast” to describe the pile of big rocks at the interface between San Francisco Bay and the landfill that the runway, taxiway and access roads are built on. KTVU, channel 2, same ones who broadcast the prank names for the cockpit crew a few days later, had their helo up and were broadcasting high quality, live, video within 30-45 minutes of the accident. You could clearly see the airplane had struck the edge of the land, whatever you call it, and a trail of junk led from that point to the runway boundary. Their local talking head was completely uninformed, but the pictures couldn’t have been clearer. There was even something visible in the water right there at the edge. Later, one of their talking heads who was a private pilot had a lot of insightful comments about approach and landing. Of course, nobody imagined that the a plane with 3 pilots in the cockpit would piddle along at flight idle with nobody watching airspeed.

  • Zed

    Now I get it. It’s even repeatable in simulation. When you start high and fast, the PAPI always shows that the approach is high. You fixate on the PAPI because the ILS doesn’t work. You try to get down to the glideslope so the PAPI indicates a correct approach, then you’ll add a bit of power, flair and land. But when you cross the glideslope your too low, too slow and too close to compensate. You get behind the aircraft. How many visual approaches have the pilots flown in this type of aircraft? The copilot should have been watching airspeed and altitude and calling it out to the pilot, no mention of this was made by the NTSB. Why did this crew fall into this trap? Too much reliance and training on automated systems, not enough training on airmanship. These are my opinions and observations, we’ll see what the NTSB conclusions are later. Thanks very much for this post, enlightening!

    • Yuri Falconfire

      Supposedly this was the pilots first time landing a 777 ever from what I am reading elsewhere.

  • Nalliah Thayabharan

    The Boeing 777 crashed in favorable weather — partly cloudy skies and light wind.The twin-engine Boeing 777-200 is one of the most popular and safest long-range planes, often used for flights of 12 hours or more, from one continent to another.
    The Asiana Flight 214 was arriving on Runway 28 which originated from Shanghai, China, left Seoul’s Incheon International Airport 10 hours and 23 minutes before its crash landing. The 7 years old old, Pratt & Whitney PW4090-powered aircraft entered final approach at an abnormally low speed of 98 knots per hour (181 km/h) – well below the target landing speed of 137 knots per hour (253 km/h).
    The two pilots on Flight 214 were Lee Jung Min, 49, a graduate of the Korea Aerospace University who joined Asiana in 1996, and Lee Kang Kuk, 46, who started his career at the airline in 1994 and got his pilot’s license in 2001. Lee Jung Min has flown a total of 12,387 hours with 3,220 hours on a Boeing 777, while Lee Kang Kuk has flown a total of 9,793 hours, with only 43 hours of which were on a Boeing 777. The pilot did not make a distress call before landing. He came in short of the runway, tried to correct buy pulling up, got the nose too high and hit tail first. The plane crashed hard to the ground, smashed the landing gear and spun around. The fuel tanks ruptured, but being at the end of the flight did not contain a lot of fuel and the resulting fire was “relatively” small and nearly everyone was able to get out in good shape.
    The San Francisco Airport had intentionally disabled ground-based landing guidance system at the time of the crash. The aircraft, registered as HL7742, was delivered on March 07, 2006, and had accumulated 35,700 hrs on 5,185 cycles as of March 31, 2013.
    Passengers Ye Mengyuan and Wang Linjia, both 16 years old from Jiangshan in eastern China may have been run over and killed by a rescue vehicle. One child and five adults remained in critical condition. A total of 307 people were on board, 291 passengers and 16 crew members. The passengers included 77 Koreans, 141Chinese, 61 Americans, 3 Indians, 1 Japanese, 1 Vietnamese, and 7 of unknown origin.

  • Seashure

    It’s amazing anyone survived that let alone all but 2 of the passengers.

  • TaoKeo
  • Frank DeBenedetti

    Did anyone notice that the barometer reading at the airport was a couple of hundredths BELOW standard sea level barometric pressure that fateful morning? If the crew set the altimeter with the standard sea level barometric pressure (instead of the actual reading reported by the airport) that would explain the 20 feet or so of elevation it seems the crew thought they still had to do a missed approach reported only 1.5 seconds before impact. What was the setting used by the crew? It might also explain the failure of the aircraft to adjust the PAPI approach until the aircraft was in the water… And why else would pilots with 10,000 plus hours of experience pull back on the yoke on final with virtually no remaining altitude, knowing they are at the verge of stalling already…since that GUARANTEES further reduction in airspeed and no altitude to recover? And does ANYONE know if they actually tried to reduce the flaps to 20 (from 30) as they added power at 7 seconds until impact as recommended for a missed approach go-around?

  • martin jones

    I heard live control tower saying to Asian Air 214 890 seconds before crash the “emergency equipment is rolling”. Why?