JS blocked

Database Error Complicit In Turkish Airlines Landing Accident


When I was control systems engineer-in-training at NASA in the early 1980s, my boss, a very wise man named James Donohue who cut his teeth working the X-24 lifting body program for Martin Marietta, always warned me of the dangers of GIGO – Garbage In, Garbage Out. 

In our case, it related to the computer simulations underway to develop the attitude control system for what would later become the Cosmic Background Explorer, a NASA Goddard satellite mission that would change the way scientists analyze the Big Bang, and because of that, win the Nobel Prize in Physics for its two principal investigators, John Mather and George Smoot.

Donohue’s issue: If you didn’t properly model the spacecraft and its environment and use the correct inputs to the simulation (mass, moments of inertia, CGs, etc), your results were likely to be bogus, and potentially dangerous. COBE’s success was proof of a GIGO-free development, at least when it counted.

I was reminded of Donohue earlier this week when I read a final accident report where GIGO was a factor – although it shouldn’t have been – in the runway excursion by a Turkish Airlines A330-300 at Kathmandu in March. No one was killed but a very beautiful and expensive aircraft remains “prematurely destroyed” on the tarmac there, according to Aviation Week’s databases. 

GIGO shouldn’t have been a factor because if the pilots had followed standard operating procedures (SOPs), they would have performed a go-around at 360 ft. above the ground due to near-zero visibility. Instead the fateful approach ended with a heavily damaged aircraft off to the left side of the runway in the grass. Passengers and crew exited via escape slides and there were no serious injuries.

Turns out the pilots let the autopilot continue the approach until 14 ft. above the ground, where presumably they finally saw the runway and attempted to flare the aircraft a bit too late – it hit the runway with a vertical acceleration of 2.7G. A hard landing on centerline is one thing, but this A330’s nose wheel was offset 85 ft. to the left of the centerline….. thanks to GIGO on the part of the Civil Aviation Authority of Nepal (CAAN).

CAAN provided the coordinates for the runway end that were used in the navigation database that the flight management system and autopilot used to fly the approach. Some other errors were thrown into the mix, but what put the A330 far off centerline -- and pretty much in the exact location of the faulty runway end coordinates, thanks the the advanced navigation and automation systems on modern aircraft like the A330 -- were some very important missing digits in the degrees, minutes and seconds marking the latitude and longitude of the runway end. Missing were the three digits to the right of the decimal point of the seconds field.

Who’s to blame? CAAN was obviously complicit for the error, but Turkish was also required to “validate” this type of RNP authorization-required approach. 

In the aftermath, GIGO in the database at least for this one approach *should* be history. The Nepalese accident investigation authority recommended that CAAN ensure that the raw navigation data is correct (as required by EASA regulations) and that Turkish Airlines ensure the correctness of the navigation data are uploaded into the FMS navigation database (as required by EASA regulations). Turkish also said it had deployed “a new procedure” to check the validity of the FMS database for the approaches to keep GIGO out of the cockpit.

Read the final Turkish Airlines accident report

Please or Register to post comments.

What's Things With Wings?

Aviation Week's civil aviation blog

Blog Archive
We use cookies to improve your website experience. To learn about our use of cookies and how you can manage your cookie settings, please see our Cookie Policy. By continuing to use the website, you consent to our use of cookies.