Damaged on Departure

Ground incidents place air safety at risk

Setting aside failure to de-ice and engine failure on takeoff, aviation is replete with stories of aircraft damaged on the ground before departure (or otherwise made unairworthy) that nevertheless got to where they were going. In some incidents, such as the one described below, crews were blissfully but dangerously unaware of any pre-departure damage or unserviceability.

747 hits de-ice truck and departs

The right wing of a Singapore International Airlines (SIA) B747-400 struck a parked and unattended de-icing vehicle during taxiing at night in thick fog on the Romeo taxiway at Paris Charles de Gaulle (CDG) before its Jan. 18, 2003, takeoff from runway 26 right. The crew did not notice the collision with the vehicle and undertook the cargo flight to Dubai without further incident. Some debris (leading-edge wing slats) was found in Paris. At Dubai, some further debris (LE slats) was found on the ground and damage was observed on the airplane. The collision had occurred in the Sierra Echo 1 de-icing zone, a holding area just short of the runway 26R threshold.

The flight data recorder (FDR) was read out. The data showed that the heading during taxiing on the taxiway was constant and that there was no noticeable change of direction, either before or after the collision. The recording did not provide any further information relevant to the investigation. In addition, ground radar data and on-site observations after the collision showed the approximate position of the vehicles on the de-icing zone a few seconds before the collision. The imagery shows that the de-icing vehicle parked to the southeast was positioned differently than the three others on that side; they were in a more withdrawn position.

The hazards of icing, de-icing and not de-icing

The captain said that the vehicles in SE1 stationed to the right of the airplane seemed to him to be perhaps a little closer than those on the left but, since the latter were at a good distance, he supposed that those on the right were as well. From his side, the co-pilot stated that at that moment he could see the right wingtip and the airplane’s navigation lights. A vehicle was slightly visible but he could not make out the exact shape and size. He was, however, able to check that the clearance available appeared adequate and, in particular, that the flashing light was clearly lower than the airplane’s wing. A little later, he noticed the outline of another vehicle to the right that seemed to be a little farther forward than the first one. He continued to taxi slowly and he was sure that the wingtip was higher than that of the vehicle’s flashing light. There was no noise and no vibration at the point of collision. He’d considered, logically but incorrectly, that this light marked the highest point on the vehicle.

The de-icing supervisor had supervised the movement of the de-icing vehicles but did not check that they were correctly positioned or configured. Because of the darkness and the fog, it was impossible for him to determine, from his control center, that the de-icing vehicle was badly positioned. The vehicle driver positioned the de-icing vehicle perpendicular to the Romeo taxiway on the eastern edge of the tarred SE1 zone. Fog and darkness also prevented the copilot from ascertaining that the vehicle was badly positioned in relation to the right wingtip (55 degrees to his right rear). A case where vehicles are left positioned in the absence of operators is not specifically referred to in the airfield’s de-icing procedures documentation.

Prefect’s order N degrees 00-2988 of Aug. 8, 2000, applicable at CDG, stipulates in Article 13 that “no vehicle shall be left unattended on the movement area.” However, everyday practice had meant that, when the de-icing zones were deactivated, the de-icing vehicles were moved back behind the white lines marked on the ground, so as to allow aircraft to pass and/or be positioned for de-icing.

The painted lines are not mentioned. Equally, there was no instruction about the “at rest” position of the vehicle’s spray control cabin when the trucks are left in position (in the absence of the operators).

Collision was inevitable

Only one possibility remained for avoiding the collision. The ground controller had a radar screen that showed a spot, thus an obstacle, closer to the airplane’s route than the other comparable spots. There was, however, traffic on the aerodrome, despite the early hour, and the controller’s role was to manage the traffic, not to ensure that no collisions occurred with obstacles. Furthermore, as far as he was concerned, the only possible risk that could occur on a taxiway would be from moving objects, which he would normally be aware of.

Of course, “pressing on” with knowledge of a condition is another matter. Up to this point, we are assuming that a flight crew was unaware of any untoward departure event. However, the end product can have the same potential. But in the case of “pressing on,” the pilot is making a judgment call.

His responsibility with respect to the Minimum Equipment List (MEL) of allowably unserviceable items ceases when he applies power for takeoff. Clearly, with fuel as expensive as it is and the European Union ruling that inconvenienced passengers must have restitution, the pressure is on aircraft commanders to make difficult judgment calls. Dumping (or burning off) fuel and landing is usually far more disruptive than biting the bullet and plowing on.

Press-on-itis

We have seen the widely publicized British Airways cases this year (Los Angeles to Heathrow and Singapore to Heathrow on three engines, see ASW, Feb. 28, and ASW, March 14). However, it happens more than one ever hears about, particularly with the B747. In mid-November 2004, a European Aviation Air Charter B747 departing Reunion (East of Madagascar) had a bird-induced surge on takeoff, an engine shut-down and still continued across Africa for 10+ hours to Paris. It “redispatched itself” from abeam Rome after reviewing fuel, weather and serviceability – without company consultation. It then diverted to Nice but, hearing that comely siren Paris’s proximate call, from there “redispatched” direct to Paris’s Charles de Gaulle Airport. Redispatching is like “restarting the plot,” and is all about an ongoing review of fuel and other fudge-factors. A similar incident happened with a Corsair 747 flying from Paris (Orly) to Reunion on Oct. 29, 2004. A Swissair DC-10 made it all the way to Athens from Karachi (via Dubai and Kuwait) after an engine failure on departure in September 1978. In a more recent case, in March 2003 an Air Europa L-1011 captain tried to persuade his passengers that a noisy engine problem on departure had been resolved but put it to a vote whether or not they should press on to Havana. The passengers universally voted for a return to Milano. Actions may speak louder than words – but sometimes words just cannot speak louder than reactions. That philosophy cost Phuket Air much bad public relations when passengers mutinied rather than allow their B747 crew to take off with a bad fuel leak out of Sharjah for the UK on April 3 (see ASW, April 11).

When and if the Federal Aviation Administration‘s new ETOPS (extended range operations) rules come into effect, three- and four-engine airplanes will be severely constrained with respect to droning on to the planned destination after a failure (see ASW Dec. 15, 2003).

The Singapore Airlines crew was lucky in that the damage to their aircraft did not lead to disaster. This is not always the case, as shown by randomly selected examples of (shall we say) “Disadvantaged Departures”. Indeed, the Space Shuttle has been plagued with problems on takeoff, notably shed insulation, that has led to damage and, ultimately, the loss of Shuttle Columbia on reentry. The lesson is plain: the flying machine must be free of defects before departure. (For this case of de-ice truck induced damage, see http://www.bea-fr.org/docspa/2003/9v-h030118a/htm/9v-h030118a.html)