French investigators have detailed the extensive efforts to hunt for engine fragments which fell from an Air France Airbus A380 over southern Greenland nearly 20 months ago.
As the aircraft headed for Los Angeles on 30 September 2017 it lost the entire fan disc, inlet and casing from its outboard starboard Engine Alliance GP7200 powerplant, and diverted to Goose Bay in Canada.
“Quite early in the investigation, it was established that the recovery of the missing parts, especially of the fan hub fragments, was the key to supporting the investigation of the cause of the engine failure,” says French investigation authority BEA.
But the search has so far failed to locate the crucial components from the powerplant.
Flight-data recorder information enabled the probe to identify the time at which the incident occurred, and to demarcate a search area for the fragments.
“This area turned out to be a wasteland covered with ice,” says BEA. The remote zone is situated some 150km south-east of Paamiut on the west coast of Greenland. It features an 800m-thick ice sheet and is exposed to cold and harsh weather conditions – as well as limited daylight – for much of the year.
Investigators drew up a primary search area of 115km² and a secondary surrounding region of 200km². While inertial reference system data provided initial co-ordinates for the engine failure, this was corrected to a position 500m north and slightly east, through the use of GPS data, to account for an offset of the inertial reference system relative to runway locations.
Thirty sections of debris retrieved early included pieces of fan blade, the spinner, pylon and inlet lip. But while the fan hub section, weighing 220kg, was considered a priority, it was not spotted during helicopter flights over the area in the first half of October.
“No primary radar was available in that region which would have helped to determine the trajectories of the ejected parts,” BEA points out.
Snowfall quickly covered the area, burying any parts which remained on the ground, forcing the search team to explore alternative methods to visual detection.
BEA says all recovered parts were relatively light, and found upstream from the event point, while the search area for the heavier fan hub was defined to be downstream.
“This did not lead to the definition of the search area being reconsidered,” it adds.
High-resolution civil-military imaging satellites from the Pleiades constellation were recruited but were unable to obtain unobscured photographs until 11 and 14 October, two weeks after the incident, and no debris was detected.
“Observation of the damage on the engine led the team to consider a fan hub failure as a probable scenario,” says BEA, adding that fracture surfaces from hub parts still attached to the engine were “consistent” with overload failure.
Investigators believe the fan disc probably separated into two large fragments, and one smaller one, as the aircraft travelled at 483kt on a true heading of 243° while cruising at 37,076ft. The fan hub was rotating at 2,638rpm at the time.
Ballistic considerations taking into account the fragment weights and wind speeds, as well as snow effects, led to team to estimate that the fragments would be buried to a depth of 2-3m by spring 2018, after the first winter season.
Airbus calculated hundreds of possible impact points using various ejection velocities and radial trajectories, aided by joint venture ArianeGroup. The US National Transportation Safety Board also carried out its own ballistic calculations.
Overlapping the Airbus and NTSB calculations enabled a high-probability search area to be defined – a 12km² box, the southern 7km² of which is the “most promising”, says BEA.
Once a piece of debris is covered by snow it will not reappear on the surface, even after summer melting, owing to the dynamics of the ice system in the area, the inquiry states: “Snow will accumulate above the part with time, meaning that the longer it lies on site, the deeper under the snow it will be.”
Several techniques were considered for the second phase of the search. These included radar imagery, ground-penetrating radar, electromagnetic detection systems, and synthetic-aperture radar.
Investigators have concentrated particularly on the latter of these options, using an airborne synthetic-aperture radar system designated SETHI. This system features X-band, L-band and UHF radar carried in underwing pods on a Dassault Falcon 20.
BEA asked French aerospace laboratory ONERA to evaluate SETHI’s ability to detect the fan hub fragments, conducting preliminary tests on similar pieces placed on the taxiway at Nimes-Garons airport in November-December 2017.
ONERA was “quite confident” that the radar would penetrate snow at a sufficient depth to detect the fan hub fragments, says BEA, so no snow-penetration flight tests were performed.
The Falcon was stationed at Kangerlussuaq with Narsarsuaq as a secondary airport, and initial tests indicated good detection capabilities.
BEA says the first measurement flights over the search area were carried out on 6-8 April 2018, during which several candidate signals were identified – requiring work to distinguish true and false results.
A second set of flights took place on 10-11 April to cover areas previously unchecked, and to reconfigure the radar to improve detection within crevasses.
X-band radar did not penetrate the snow as much as expected, but L-band radars penetrated into the ice sheet. Large numbers of crevasses in the search field resulted in various algorithms being created to eliminate unreliable detection.
BEA says a third, and final, set of flights was carried out on 14 April.
Six particular targets were identified for further examination by ground teams – examination which had to be conducted carefully, given the terrain and other environmental conditions – but BEA says none of the six produced positive results.
This ultimately led the search to evolve into a systematic effort using ground-penetrating radar but BEA says this proved to be an “inadequate” sensor.
“Even with 430km of search tracks analysed, it could not be certified that if the part had laid under those tracks, it would have been detected,” it adds.
BEA says that, despite the substantial work, the fan hub fragments from the A380 were “not reliably detected” by the end of June 2018.
It says the synthetic-aperture radar is experimental and that its use over the ice sheet to detect buried parts amounted to a “premiere”.
“However, due to higher background scatter noise levels and lower penetration of X-band radar waves in the search zone than expected, no high confidence target was detected,” says BEA, and only moderate-confidence targets could be explored further.
But it states that ONERA has continued to process the radar images acquired, using greater computation facilities than those available on-site in Greenland. BEA adds: “The team considered it was necessary to test new sensors with a larger swath, and a more certain response, for potential new ground searches, later on in 2019.”