Description

On 10 February 2023, a PW4060-engined Boeing 767-300ER (N187-DN) being operated by Delta Air Lines on an international passenger flight from Edinburgh to New York JFK with an augmented crew in day IMC encountered problems with the right engine when the crew set climb power but the engine continued to function. It was decided that the engine control issue meant a diversion was appropriate and on arranging this to Prestwick, no urgency or emergency declaration was deemed necessary. As the diversion and overweight landing was completed, right engine vibration became apparent on approach and it was set to idle and shut down as the aircraft approached its allocated parking position. On arrival there, fuel was observed leaking from the right wing and the RFFS, in attendance for potential hot brakes, began collecting it whilst monitoring brake temperatures. The Captain ordered what amounted to a precautionary rapid disembarkation, although the airline company involved include a similar procedure in its OM. Once it began, the 211 passengers were all off the aircraft in just over five minutes.

Investigation

A Field Investigation was carried by the UK Air Accident Investigation Branch (AAIB). Relevant recorded data were recovered from the aircraft CVR, QAR and ACARS and a passenger video showing flames coming from under the wing in flight (which was widely circulated on social media) and TWR video and airport CCTV recordings were also available. Relevant engine management system data were subsequently downloaded at the operator’s maintenance facility.

It was noted that the 62 year-old Captain had a total of 21,000 hours flying experience which included 12,000 hours on type. Similar experience and age information for the First Officer and the relief pilot was not recorded.

What Happened 

With the First Officer acting as PF and the relief pilot occupying a supernumerary crew seat on the flight deck, the flight was taxiing towards runway 24 when two members of the cabin crew at the overwing exits recalled hearing a “rattling” sound “as though something was loose in the cargo compartment”. Although there was no corresponding sound on the CVR, several of the cabin crew reported having been aware of a continuous loud vibration from mid-cabin once takeoff thrust had been set. One of the overwing exit cabin crew reported this to the SCCM by interphone and the latter then called the flight crew but did not get a response.

Soon after the flight crew set climb thrust, the A/T disconnected and on the EICAS, the right engine EEC ‘alternate mode’ status appeared. The Captain responded by calling an “overtemp on the right side” (considered to have been a reference to a high EGT). The flight was cleared to climb to FL140 and the After Takeoff Checklist was completed. An attempt to reset the EEC using the applicable NNC (non-normal checklist) was not successful. The Purser was able to make contact with the flight crew and reported the vibration which had begun during the taxi out was now “quite bad” with some of the passengers appearing to be concerned. ATC were asked to confirm that the flight could stop the climb at Fl 100 and maintain 250 KIAS which was approved. The Captain could not detect any vibration on the flight deck and asked the relief pilot to contact company engineering to see if EEC alternate mode would affect their ETOPS clearance. He then asked them to go into the passenger cabin to check the vibration and unsuccessfully tried to re-engage the A/T at which point he noted that the right engine EGT was fluctuating. Having decided that the flight should not proceed as planned, he advised Edinburgh ATC “we don’t have an emergency at this time we just have a maintenance problem” and briefly took over as PF before re-designating the First Officer as PF.

Having realised that a longer runway than the one at Edinburgh would be preferable for an overweight landing, the Captain accepted an ATC suggestion that runway 30 at Prestwick (2,987 metres long) was suitable. The relief pilot returned and advised that he had not noticed “anything obviously abnormal”. ATC asked whether the flight needed any special handling at Prestwick and the Captain responded by requesting the airport RFFS to be in attendance because of the possibility of hot brakes.

He decided to perform the overweight landing himself and again took over from the First Officer. He then briefed for a 25° flap landing on the basis that the associated VREF was lower than for flaps 30. On a long right base leg at 6,000 feet QNH, the Captain, having become aware of some airframe vibration, decided to reduce thrust on the right engine which had the desired effect. Around the same time as ATC instructed the flight to turn right towards an intercept heading for a 21 nm final, a passenger videoed what appeared to be flames coming from under the right wing (see the illustration below). The Captain had decided to reduce the right engine to idle thrust and by the time the Purser was advised of the sighting, the flames had ceased and were not reported to the flight crew.

An image taken from the passenger video showing flames under the right wing. [Reproduced from the Official Report]

The single engine go around procedure was briefed and the Captain advised that although an autoland could not be performed with one engine shut down, as both were “still running”, he was going to use it. The landing was uneventful and the Captain requested the RFFS vehicles to follow them to the parking position. At this point, the pilots noticed that all primary indications for the right engine were showing zero and the Captain remarked that he thought the right engine had failed on touchdown before selecting its fuel off and calling for the ‘After Landing’ Checklist.

ATC then passed a message from the marshaller that “something was leaking from the aircraft’s right wing”. The Captain requested the stairs “real quick” and advised passengers to expect an “orderly deplaning”. The relief pilot left the aircraft to liaise with the RFFS crew who were monitoring brake temperatures with thermal imaging cameras. Meanwhile, the RFFS had spoken to the First Officer to advise that fuel was leaking behind the right engine and that they intended to put a water hose on it.

Once the stairs on door 1L were in place, the Captain made a PA to instruct the passengers disembark but leave their cabin baggage behind. This took just over five minutes and an engineer from a maintenance organisation at the airport then arrived and helped co-ordinate actions to stop the fuel leak. He recognised that the fuel was coming from a drain hole in the wing dry bay and concluding that emptying the right wing should stop the leak, he helped to arrange the transfer of fuel from the right wing tank to the centre tanks until the leak stopped.

Why It Happened

Initial examination of the engine found that all the debris from the fan-blade initiated failure had been contained within the engine. A strip down of the engine found that the high cycle fatigue failure of one of the No 2 High Pressure Turbine blades had caused five consecutive following ones to sustain non-directional fractures which were indicative of “rapid tensile overstress due to secondary impact” (see the illustration below). The initial blade fracture was attributed to fatigue cracks initiated from two Stress Corrosion Cracks on the exterior convex face of the blade. Similar cracks were also seen on the adjacent (non-damaged) blades with a detailed metallurgical examination of all such cracks finding evidence indicative of a sulphation attack. This showed that the blade failure had occurred in normal operation rather than because of a specific blade problem. The level of vibration created by the out of balance HPT 2 disk was found to have exceeded that observed in two previous similar events to the same aircraft type. 

The HPT 2 disk looking forward (left) and the failed blades close up (right). [Reproduced from the Official Report]

It was considered that a fuel drain tube found fractured within the right wing tank dry bay was likely to have failed as a result of the vibration caused by the engine imbalance created by the blade failure. This fracture had led to fuel escaping overboard via a drain hole and igniting in flight. The Investigation did not determine the exact vibration spectra to which the tube had been subjected, but it was concluded that the sustained high level of vibration from the engine had caused the drain tube to oscillate and fail due to fatigue. A detailed analysis of the tube fracture faces led to the conclusion that the fatigue cracks had initiated from intergranular cracking which was “deemed most likely to have come from accidental acid splash” which “may have occurred during the early stages of its manufacture”. Further work also led to the conclusion that the crack had started and fully fractured during the investigated 30 minute flight.

It was estimated that approximately 500kg of fuel had been lost overboard from the aircraft between the fracture of the drain and the completion of the fuel transfer once the aircraft was on stand. Approximate calculations of the flow rate suggested that the leak began soon after takeoff and continued until fuel transfer had been completed about 90 minutes after landing and that about a quarter of the fuel lost (125 kg) had been lost during the flight.

It was considered likely that the passenger video of flames coming from the rear of the wing was likely to have been fuel exiting the dry bay drain hole into the air flow and then being ignited in the hot jet efflux. The absence of any soot or heat damage on the engine cowlings was considered to indicate that the flames seen were behind the engine rather than from an engine fire. However once parked, the wind direction in relation to the aircraft orientation had resulted in fuel being blown rearwards towards the landing gear when the brake temperatures as monitored by the RFFS team were in excess of the fuel ignition temperature which meant there had been “significant potential for a fuel fire”.

As a result of the event under investigation, Boeing was noted to have begun a project to review the design of the drain tube for potential reliability improvements. Although at the time the Investigation was completed Boeing had concluded that there had been no increased risk of a catastrophic outcome from a failed drain tube, it was noted that the FAA Review of Boeing’s analysis was still awaited and so a corresponding Safety Recommendation was formulated.

The narrative Conclusion of the Investigation was as follows:

"During the early stages of the event flight a high-pressure turbine blade fractured through fatigue cracking. The fatigue crack was initiated by a possible combination of hot corrosion and pitting from external contamination. The detached blade caused damage to a further five blades resulting in an engine imbalance. A drain tube in the right fuel tank fractured, probably as a result of the vibrations transmitted from the out of balance engine. This resulted in fuel escaping from the right fuel tank out of a wing drain hole and igniting in flight. The flight crew landed the aircraft promptly at Prestwick. They were unaware of any flames, or the right engine running down, until after landing.

The fuel coming from the wing was noticed by the ground crew as the aircraft parked near the terminal buildings. The wind was blowing the fuel towards the hot brakes and there was a risk of a fuel fire. The passengers were rapidly disembarked, and actions were taken to contain and stop the fuel leak.

As a result of the vibration-driven fatigue fracture of the drain tube, the aircraft manufacturer has launched a project to identify potential reliability improvements to the design while the Federal Aviation Administration reviews the manufacturer’s finding that the risk of a catastrophic outcome from a failed drain tube is not elevated."

One Safety Recommendation was made on completion of the Investigation:

• that the Federal Aviation Administration requires the Boeing Airplane Company to demonstrate that following this serious incident, the design of the slat track housing drain tube on the Boeing 767 family of aircraft continues to comply with the certification requirements for large transport aircraft.

The Final Report was published on 18 July 2024.

Related Articles

• Contained Engine Failure
• Engine Failure During Takeoff - Multi-Engine Transport Category Jet Aircraft
• Engine Failure in Cruise
• Fire in the Air
• Ageing Aircraft - Structural Failure
• Overweight Landing