In flight break up

Citation

(2002), "In flight break up", Aircraft Engineering and Aerospace Technology, Vol. 74 No. 5. https://doi.org/10.1108/aeat.2002.12774eab.011

In flight break up

Keywords: Safety, Aircraft

A Rockwell Commander 114 piston-engine aircraft was parked in a hangar on the fouthere side of Biggin Hill airfield. At 09.00h on the day of the accident the pilot telephoned the hangar manager inquiring as to the aircraft's fuel state. He was told that the tanks were full but the aircraft was being flown earlier that day by another member of the syndicate to which the pilot belonged. The pilot asked that the aircraft should not be refueled upon its return as he intended carrying three passengers and did not want full fuel on board. The aircraft landed from its first flight that day at 11.45h with its tanks 3/4 full. The pilot operating the first flight subsequently reported that the aircraft was fully serviceable with all systems operating normally.

The aircraft took off from Biggin Hill at 12.47h and arrived at Shoreham without incident at 13.12h. The pilot and his three passengers left the aircraft parked close to the terminal and entered the airfield restaurant for light refreshments. Having eaten and carried out some pro-flight planning the pilot and passengers boarded the aircraft for the return flight. The passenger seated in the front right seat next to the pilot, was also a qualified pilot who had flown with him on previous occasions on flights to the Continent, North Africa and the Middle East.

The aircraft took off from Shoreham at 14.07h to return to Biggin Hill. The pilot was in contact by radio with the Shoreham. Approach controller during departure but 15min into the flight he requested to change frequency. There was no record of him subsequently having made contact with any other air traffic unit. The progress of the flight from Shoreham to Biggin Hill was recorded by the National Air Traffic Service (NATS) radar at Pease Pottage several miles to the south of Gatwick. Only the primary radar returns were recorded as the aircraft's Mode C (height recording) transponder transmissions were not received.

The aircraft was first detected by radar at 14.07h, as it took off from Runway 03 at Shoreham. The aircraft turned right after departure, flying towards Brighton and crossing the coast by Brighton Marine. It flew approximately 0.5km from the coast until it turned inland close to the Seaford (SFD) VOR. The aircraft then followed a track of approximately 345 passing just to the east of the village of Newick (5km southeast of Haywards Heath) before turning onto a track of 340 towards Sharpthorne. It approached the southern edge of the Gatwick Control Zone (CTR) (controlled airspace extending from the surface to 2,500 feet amsl) just north of Sharpthorne and was recorded as entering a tight right turn before the radar return disappeared. The last recorded position was at 1424.53h, directly above the Bluebell Railway line, 2km north of Sharpthorne.

There were several eye witnesses to the aircraft's progress just prior to the accident. One witness, positioned east of Sharpthorne, saw the aircraft flying normally on a northerly track with its wings level and the engine sounding normal. The witness, an experienced pilot himself, estimated the aircraft to be at a height of 500800 feet agl (9001300 feet amsl). He heard the engine noise then increase for 510sec before it stopped abruptly. He described the weather at his location as being fine but with a large cumulus cloud in the area. Moments after the engine noise ceased visibility on the ground reduced to 200m in snow that continued to fell for a further 510min.

One witness saw the aircraft "enter or fly behind the cloud". As it did so the engine noise started to rise as if the aircraft had started a dive without throttling back. A further witness was also aware of changes in engine noise whilst the aircraft was in the vicinity. As the noise became particularly loud he saw the aircraft 800m away from him coming out of a shallow 10 descent, still under power at or below the tree line. He reported that it appeared to attempt a level right turn but both wings fluttered, vibrating unnaturally before there was the sound of mechanical failure. The aircraft continued its turn to straighten on a southerly heading now with the left wing barely attached. The left wing separated from the aircraft which then impacted with the ground. A witness close to the Buebell Railway line looked up to see the aircraft's left wing break away, the aircraft start spinning and spiral into the ground. Other witnesses described both wings as folding upwards and rearwards prior to impact. The accident was reported to the police at 14,25h. The other emergency services were alerted immediately and arrived on the scene within about 10min. The pilot and passenger however had all received fatal injuries in the impact. There were no reports received from the emergency personnel on the scene of any ice or snow adhering to the wreckage.

Examination and wing failure mode

The aircraft crashed in an open field with the wreckage spread over a distance of 400 feet in a northsouth direction. The fuselage heading on impact was easterly. The impact speed was assessed as high, with the aircraft in a steep nose down attitude as it made contact with the ground. Ground marks and relative degrees of damage indicated that the right wing, which had come to rest some 80 feet from the fuselage, was not attached to the fuselage at impact. It was apparent from the limited damage to the right wing that it had finally separated not long before the ground impact. Although, several witnesses had stated that the left wing detached in flight, the left wing had travelled with the fuselage to the point of impact even though its major structural attachments were broken.

The main spar of the right wing had separated near the root, with all breaks inboard of wing station (WS) 25 (right hand). The bottom spar cap had fractured at about WS 25 (right hand) in tensile overload with some twisting. The top spar cap joiner plate on the aircraft centreline had failed in upward bending, reaching an angle (dihedral) of 30400 before fracture. The spar web had collapsed on the left side of the fracture, and torn away on the right side. The left wing top and bottom spar caps had collapsed towards each other and rotated out of plane by about 20. There was evidence of compression budding of the right wing top spar cap at WS 26, but this spar cap had not completely collapsed. There was also some compression buckling in the left top spar cap at about WS 8, combined with some bending rearwards at WS 26. Metallurgical examination of the fractures showed that overload condition had caused the spar caps to break. There was no evidence of fatigue. Material hardness tests carried out on the spar caps and doublers found them to be of normal hardness. No evidence of control surface hinge over-travel was found on either wing, the fin, or the horizontal tail. Hinge cutouts were not enlarged, and there was no corresponding damage at the edges of the control surfaces. The assembly of the main spar truss "A" frame had been satisfactory with no pre-impact fretting or free play, and there was no evidence of any initiation of the wing break up from this area of the wing structure. None of the other wing attachments were loose or showed any evidence of relative movement or fretting.

Witness observations of the aircraft during the later stages of its descent raised concerns over the wing break up such that, a proposed safety recommendation was formulated to review the flutter substantiation of the aircraft. During the consultation phase, however, information that was not previously known was submitted by the manufacturer on previous flutter analyses and test results, and a recent review which they had commissioned through the FAA Designated Engineering Representative (DER). Using this data AAIB commissioned its own review, which concluded that the aircraft is free of wing flutter to speeds well in excess of the requirements. In addition the review showed that major damage to the wing spar and/or attachments was necessary before aeroelastic effects could arise, with divergence becoming the first likely effect, possibly only after substantial damage had been incurred.

The manufacturer also submitted data and photographs of the wing upbending destruction test carried out for certification. This data showed that the initial failure mode was a buckling instability failure of the top spar cap inboard of wing station WS 26.5. Elements of the buckling deformation were in a downward sense at about WS 10 and WS 26, with a reversed, upward clement interposed. There was also a degree of rearwards deformation. The test was halted before further failures occurred.

The similarities between the wing failure and the wing destruction test carried out for certification purposes are strong and show that the observed pattern of damage can be explained by a simple recovery manoeuvre, albeit involving high "g" levels. The calculations show that the height loss in the descent would allow sufficient speed build up for such a recovery to cause structural failure in the manner of the wing destruction test. This, combined with the flutter investigations carried out by the manufacturer, the FAA DER and the AAIB, lead to the conclusion that the in flight break up was caused due to a recovery manoeuvre as the pilot attempted to pull up to avoid impact with the ground. The observations of the witnesses concerning unusual behaviour of the structure were made after major structural damage had been sustained in the pull up manoeuvre. In the light of the work described above, it is considered that the intent of its previously formulated safety recommendation to review the aeroelastic behaviour of the aircraft has been met, and that the aircraft has been shown to conform with the relevant requirements.

ReferenceAAIB Bulletin 4/2002