Description

On 18 December 2018, a Boeing 787-9 (ET-AUP) being operated by Ethiopian Airlines on a scheduled international passenger flight from Oslo to Addis Ababa as ETH175 and due to make an intermediate stop at Stockholm was taxiing into a remote stand to be de-iced prior to departure when it collided, in normal night visibility, with a lighting mast. The lighting mast was damaged to the extent that it required replacement and the collision caused significant damage to the right outer wing of the aircraft but there was no resultant fuel leak.

The accident scene once the 787 had stopped about 4 metres after the collision occurred. [Reproduced from the Official Report]

Investigation

The Accident Investigation Board Norway (AIBN) was not notified of the event, but after learning of it from media sources the following day, an Investigation was immediately commenced. The delay in reporting the accident and the failure to electrically isolate the two-hour CVR after it had occurred meant that its data were lost. However, relevant recorded ATC ground radar and communications data and CCTV pictures were available. It was deemed unnecessary to download data from the FDR.

It was noted that the 44 year-old Captain had a total of 20,137 hours flying experience which included 2,697 hours on type and the 28 year-old First Officer had a total of 2,868 hours flying experience which included 1,215 hours on type. A third pilot, who was “undergoing line training”, was present on the flight deck for the departure. The operator’s aircraft type was the one normally used for this service.

The Ground de-icing system at Oslo

The airport provision of ground de-icing was based on the availability of three dedicated de-icing platforms. At the time of the accident, runway 19L was active for departures and so the ‘B-North’ platform, situated at the north-eastern end of the airport, was in use. De-icing services provided by various suppliers were coordinated by SAS Ground Handling (SGH). The SGH office, from which the “de-icing coordinators” involved were working, was located next to a different de-icing platform ‘A-South’, situated at the south-western end of the airport (see the illustration below). Two de-icing coordinators were on duty in that office at the time. To manage the provision of de-icing services, they relied on information from ATC ground radar, the traffic lists generated by the A-CDM and monitors displaying CCTV pictures from cameras at the de-icing platform(s) in use. They also used a proprietary computer software de-icing management system called ‘Wise’ to assign a de-icing platform and a de-icing truck to each aircraft. It was learned from the two de-icing coordinators involved that it was “common practice” to submit an electronic request to the allocated de-icing personnel using the ‘Wise’ system only after an aircraft had arrived on its designated de-icing stand. This request includes the aircraft flight number, its category, the de-icing stand it is using, the type of de-icing fluid and the extent of de-icing required.

It was noted that aircraft were divided into different categories for de-icing purposes designated by letters in alphabet sequence beginning at ‘A’ based on their wingspan, a system which assigned the Boeing 787 to category ‘E’. At the B-North platform, only the central stand of each sub-set of three - stands 992, 995 and 998 (see the illustration below where stands 991 and 992 out of the sub set 991, 992 and 993 are ringed in red) could be used by aircraft of Category ‘D’ or greater. The de-icing platforms were not within the designated airport manoeuvring area so aircraft on them were not subject to clearance or any other type of ATC service and temporarily under the control of the de-icing service providers.

The three available de-icing platforms. [Reproduced from the Official Report]

What Happened

It was established that prior to the aircraft pushing back from its parking gate, and about 30 minutes before it arrived at the de-icing platform, the crew had used datalink to make a request for de-icing. Following the subsequent pushback, a clearance had been obtained from GND to taxi using taxiways ‘Z’, ‘V’ and ‘S’ towards the B-North de-icing platform. As it reached there, the GND controller instructed the crew to contact ‘De-icing’ on 121.855 MHz whilst continuing to listen out on GND.

Prior to the aircraft calling on their frequency, the two de-icing coordinators reported having discussed the current traffic situation, indicating that their assessment of their workload at that time was “low to moderate”. They reported having agreed that the 787 and a following Atlas Air Boeing 747 should be assigned the ‘centre-of-sub-set’ de-icing stands 992 and 995 respectively, the 747 entering 995 only after a KLM Cityhopper Embraer 190 had vacated stand 994. However, on contacting the De Icing frequency as instructed, the 787 crew were told by De Icing Coordinator 2 to taxi “proceed to stand 991” and call when ready for de icing to commence which was read back correctly. The illustration below shows the situation as the 787 enters stand 991 whilst the Atlas 747 allocated stand 995 waits for the Embraer 190 to vacate stand 994.

CCTV recordings showed that the 787 taxied as instructed at a slow speed. The aircraft Captain, who was taxiing the aircraft, subsequently stated that he had seen the light mast to the right of the aircraft and assessed that although clearance from it was “marginal”, he had been “confident the airport had control of the situation”. Nevertheless, he did ask his First Officer if he could see the distance between the wing tip and the light mast but, as the 787 wing tip is not visible from the seated pilot position, he had only limited ability to estimate the likely clearance. The Captain reduced the taxi speed further for the last few metres prior to the collision but said that no signals had been seen from the ground crew to indicate that they were on a collision course with the light mast.

The B-North de-icing platform showing aircraft locations as the 787 approached the mast. [Based on ground radar data and reproduced from the Official Report]

Prior to the 787 being instructed to enter stand 991, two SGH de-icing vehicles with an operator in each were parked at the forward limit of the stand. These vehicles, unlike some newer ones also in service were not fitted with a VHF radio which would have enabled either to communicate directly with aircraft flight crew if necessary. The driver of one of these vehicles saw that the 787 was being directed onto stand 991 and was aware that this stand could not be used by an aircraft of this size. He therefore called the de-icing coordinators directly to tell them that the aircraft had to be stopped and five seconds later, one of the coordinators made a less than clear transmission on the de icing frequency “Etian seven one five you are going to wrong … stop stop Etian stop stop stop”. Eight seconds after this transmission began, the right-hand wing of the 787 collided with the light mast. Following the collision, the aircraft continued moving slowly forward for another four seconds before coming to a stop approximately four metres further on with the nose of the aircraft turned slightly to the right and the nose landing gear 0.75 metre off the stand 991 centreline.

Once stopped with the park brake applied, the 787 crew, unaware of the collision, contacted the de-icing coordinator to inform them of the type of de-icing they required and were then told of the collision. The Captain stated that although he had noticed the nose of the aircraft pulling right as it came to a stop, he had thought that the cause of this had been the aircraft skidding slightly. In the absence of any evidence of fire, the passengers and crew remained on board. The first RFFS vehicle to respond reached the scene after 10 minute and all 69 occupants were subsequently disembarked via mobile air stairs. Damage to the wing occurred at a position adjacent to the front of the surge fuel tank which, due to the low fuel load carried (9 tonnes against a capacity of 107.5 tonnes), was not punctured, although it would have contained fuel vapour.

Damage to the leading edge of the right wing. [Reproduced from the Official Report

The Immediate Cause of the accident was determined as the de-icing coordinator’s instruction to taxi onto a different de-icing stand (991) to the one he and his colleague had previously planned (992).

Other Findings of the Investigation included, but were not limited to, the following:

• The absence from Norway AIP or any NOTAM information about what de-icing stand could be used by the various aircraft categories.
• The fact that the aircraft flight crew had no other access to information about which de-icing stands were authorised for their type of aircraft at Oslo.
• The absence of any markings, lighting, signage or other technical barriers to indicate to the crew that they had been assigned the wrong de-icing stand.
• The de-icing coordinators did not have adequate technical aids to ensure that they assigned the correct de-icing stand based on aircraft category.

It was observed that some type of anti-collision aid, such as a wingtip camera, would have clearly indicated to the crew that there was insufficient clearance between the wing and the light mast and prevented a collision.

It was also noted that AIP entries for other Norwegian Airports operated by Avinor do not provide any information about the maximum wingspan allowed at de-icing platforms.

Safety Action taken as a result of the accident whilst the Investigation was in progress was noted as having included:

• SAS Ground Handling (SGH) has:
  • required that for all aircraft larger than ‘code C’ a marshaller shall be present to guide the aircraft onto the correct de-icing stand.
  • stipulated that there must always be two de-icing coordinators on duty and that in the event that the A-North and B-North platforms are in simultaneous use, this should be increased to three.
  • for the first time (a matter of concern to the Investigation) issued a SOP for the allocation of de-icing stands to individual aircraft by de-icing co-ordinators.
• Airport Operator Avinor has initiated a revision to the AIP Norway to add information about the maximum wingspan permitted on each de-icing stand and added ground markings to the lead-in lines to all de-icing stands.

Three Safety Recommendations were made as a result of the Investigation as follows:

• that All Airport Operators of Norwegian Airports review and issue information in AIP Norway for all other airports in Norway, stating limitations relating to what de-icing stands/areas the various aircraft categories can use. [SL No. 2020/12T]
• that Avinor, the Oslo Airport operator, in consultation with the relevant handling operators, find suitable technical solutions for Gardermoen to prevent aircraft from being directed to the wrong de-icing stand. [SL No. 2020/13T]
• that the European Union Aviation Safety Agency (EASA) consider requiring large aircraft to be equipped with anti-collision aids for use during taxi. [SL No. 2020/14T]

The Final Report was completed on 29 June 2020.

Related Articles

• Ground Collision
• Ground Operations
• Wing Tip Clearance Hazard
• Aircraft Ground De/Anti-Icing
• ATC Radio Use by Airside Vehicles
• Taxiway Surface Markings and Signs
• Safety Assessment of Airport Collaborative Decision Making (ACDM)