On 8 July 2005, a Boeing 737-400 (G-DOCT) being operated by British Airways on a scheduled domestic passenger flight from Aberdeen to London Heathrow was damaged during departure in daylight and normal ground visibility when jet efflux broke up the surface of the blast pad adjacent to the beginning of the paved surface of Runway 16 causing significant damage to the tailplane and elevator. The crew did not experience any unusual in flight effects as a result and only became aware of aircraft damage when advised after flight that it had been discovered during a routine engineering inspection.
An Investigation was carried out by the UK AAIB. Flight Data Recorder (FDR) data was successfully downloaded; data from the 30 minute Cockpit Voice Recorder (CVR) had been overwritten during the subsequent flight.
The pilots of the next aircraft in line for departure saw "two large sections of asphalt, the largest section estimated to be 2 m by 3 m, slowly lift and disintegrate as the aircraft started its takeoff roll" and reported this to ATC. Having heard this, the Commander of the departed aircraft asked if the observing crew had seen any damage to his aircraft and were told that they had not. In the light of this information, and in the absence of any evidence of abnormal handling, he decided to continue the flight. No evidence which might have suggested damage was detected during the remainder of the flight and it was only after a routine engineering inspection at the destination after the crew had left the aircraft that significant damage to left tailplane and left elevator, as illustrated below, was discovered.
At Aberdeen, pieces of the damaged elevator were found in the grass area behind the Runway 16 threshold close to the extended centreline with the most distant being 132 metres from the threshold. Most of the damage to the blast pad (also sometimes referred to as an erosion strip), which was an 8.4 metre long asphalt surfaced area extending well beyond both edges of the 45 metre-wide runway, consisted of the complete break up of the material comprising its 13 metre-wide central section. Most of the asphalt was found in the grass beyond the blast pad, some as far as 20 metres away, with the rest piled up loose in its damaged area or on the adjacent runway surface. The largest piece found within the area of the damaged blast pad was approximately 1.8 metres by 1.5 metres, 6 cm thick and weighed approximately 340 kg. The surface below the missing asphalt was found to consist of "stones and dirt with almost no bitumen residue". Some of the stones from this surface were found on the runway. It was noted that most of the bitumen sealing between the end of the runway and the blast pad had detached with the asphalt. An abbreviated form of runway inspection was found to have been carried out two minutes prior to the departure under investigation and a full runway inspection some 6/7 hours earlier during the hours of darkness.
View of damage to the left tailplane (a 2.4 metre-long gash) and to the left elevator from below (reproduced from the Official Report)
Damage to left elevator from above showing separation of the outboard section including the loss of a section approximately 0.9m x 0.6m (reproduced from the Official Report)
Using FDR data, a reconstruction of the ground track used by the aircraft was made. This indicated that, having determined that a full length take off with high thrust applied prior to brake release was appropriate for the take off, the aircraft commander had taxied the aircraft onto the runway by taking the main landing gear close to the end of the runway before completing the 125º right turn necessary to get the aircraft lined up on the runway centreline. Left engine thrust up to 40% N1 was used to achieve the right turn, at which time that engine would probably have been over the blast pad. Once on the centreline, the aircraft had held position with the brakes applied for a minute. Prior to brake release, engine thrust was in line with normal procedures to "approximately 40% N1" to allow engine stabilisation and then, after three seconds, the increase in thrust was resumed, reaching 95% after 14 seconds, and 100% 2 seconds prior to brake release.
The history of the blast pad was reviewed and it was found that it had "probably" been constructed during the early 1990's with the central section - part of which had disintegrated in the investigated event - having been subsequently re-surfaced at an unrecorded date, possibly in 1992 after an overrun in the 34 direction. The depth of the damaged asphalt was found to have varied between 4.5cm and 6.5 cm. The blast pad was not marked in any way which might have reminded pilots that it was not intended to take the weight of any aircraft permitted to use the runway. The applicable UK CAA regulatory guidance on 'Licensing of Aerodromes' CAP 168 9th edition, was found not to include any guidance or reference to blast pads or erosion strips. ICAO Annex 14 was also found to lack any such references, although it was noted that the need for blast pads to be able to "accommodate an occasional passage of the critical aircraft for runway pavement design" was noted to appear in the International Civil Aviation Organisation (ICAO) Aerodrome Design Manual Doc 9137. This Manual also recommended that the minimum surface thickness for blast pads associated with a runway such as the one involved should be 7.5cm and that they should be 60 metres long. Federal Aviation Administration (FAA) guidance on blast pad capability and asphalt surface thickness were noted as similar to those of ICAO. The Airport Operator's own 'Airport Planning Standards' document was found to state that blast pads should be provided for runways used by jet aircraft and should be 30 metres long but that "there are no particular strength requirements" other than that "the surface should be sealed to prevent flying debris" and that "it needs to be able to support the passage of airport vehicles".
The fact that the aircraft commander had deemed it necessary not to follow the marked taxiway lead on line onto the runway centreline prompted the Investigation to examine the basis on which such lines are marked. It was found that the applicable UK CAA publication on visual aids, CAP 637, which is based on ICAO Annex 14 Chapter 5, stated that lead on/off lines should ensure safe clearance for the largest aircraft permitted to use a runway provided the 'cockpit' of that aircraft is kept on the centreline although there is no mention of any requirement for pilots to follow such lines.
The effect on take off performance in terms of the declared TORA/TODA/ASDA and the reduction in distance attributable to lead on lines was reviewed noting that the transition which had been made onto runway 16 had required "slightly more than a 90º turn". The Operator's performance calculations for the aircraft type were found to be based on alignment distances provided by Boeing which ranged from 10 metres for a 90º turn to 18 metres for 180º turn based on a minimum 'safety margin’ of 3 metres between the edge of the runway threshold and the most proximate main landing gear wheel. The Investigation calculated that the minimum attainable alignment distance required to position the aircraft main landing gear approximately 10.5 metres from the threshold would have required the outer left main wheel to run along the edge of the threshold before the right turn onto the centreline was commenced. In this position, the aircraft tailplane would have been directly over the blast pad. It was also calculated that if the aircraft had followed the marked lead-on line, "its main wheels would have been about 66 metres from the threshold" by the time it was aligned with the centreline.
The Operator was found to have published information to its pilots on TORA to the effect that "alignment distances are incorporated into the takeoff performance calculation" although this was found not to specify the exact distance from the beginning of the TORA from which it was assumed that an aircraft would start its takeoff run from. It was noted by observation that aircraft routinely did not follow the lead on line when entering runway 16 - which had a TORA/ASDA of 1829 metres.
The applicable Operations Manual guidance was found to state that "The rolling take off procedure is recommended for setting takeoff thrust" and that such a procedure "expedites takeoff and reduces risk of foreign object damage". No other techniques were described in the relevant Operations or Training Manual but it appeared that the practice of standing starts (full thrust applied against the brakes) when departing from 'limiting runways' was, in practice, widespread. The Boeing FCTM was found to give slightly more detail in respect of the procedure for applying takeoff thrust and to include the following remarks:
- Flight test and analysis prove that the change in take-off roll distance due to the rolling takeoff procedure is negligible when compared to a standing take-off.
- Brakes are not normally held with thrust above idle unless a static run-up is required in icing conditions. A standing take-off procedure may be accomplished by holding the brakes until the engines are stabilised, then release the brakes and promptly advance the thrust levers to take-off thrust
It was noted that the aircraft commander had "believed, in the absence of any information to the contrary, that the performance restrictions imposed on the aircraft’s takeoff were due to runway length" whereas "in the event, the restriction was actually due to obstacle clearance requirements during the climb out".
The potential effect of jet efflux on the integrity of an asphalt surface was considered and it was estimated that at 90% N1, the speed of the efflux over the ground directly below the tailplane and in line with the engines would have reached a maximum of approximately 190 knots. It was also found that "the static pressure of the air within the jet exhaust directly below the tailplane at ground level was equal to the ambient static pressure" so that the passage of the efflux was not generating any suction.
A further study was made of the potential suction effects of air intake to the left engine as the aircraft had been taxied into position using 40% N1 on the left engine to achieve the right turn onto the centreline as the engine was likely to have been passing over the central part of the blast pad. It was found that in calm conditions, the static pressure on the ground ahead of the engine intake was the same as ambient pressure but that in the presence of a 5 knot crosswind component, it was likely that a vortex generated in front of the inlet would have applied a 0.2 psi suction force to the ground. The prevailing surface wind velocity at the time was recorded as 140º (M) at 7 knots which equated to a crosswind component of approximately 6 knots. It was noted that the 0.2 psi suction was slightly more than the equivalent measure derived from weight per surface area for a 6cm asphalt depth so that, provided any adhesive force there might have been between the base of the asphalt layer and the material beneath were ignored, the applied suction might have been sufficient to start to lift the failed asphalt surface.
Previous events in the ICAO "Accident Database" in which jet blast damage to runway surfaces during take off had been accompanied by damage to the aircraft which had been the source of the damage were reviewed by the Investigation but it was found that no formal Accident Investigation in accordance with the principles of Annex 13 had taken place and that as a result, "very little information is available about what caused the asphalt surfaces to delaminate". However, the publication of an ANSV Report into an accident to an Airbus A320 at Treviso in 2002 was noted - in this case, tailplane damage not dissimilar to that sustained by G-DOCT occurred but significant damage was also caused to the one of the hydraulic systems and the corresponding warning led to a rejected take off.
Conclusions arrived at on the basis of the findings of the Investigation included the following:
- The damage to the tailplane would have had minimal aerodynamic effect, but the elevator was missing a section almost 1 metre long and this would have reduced the elevator’s effectiveness.
- Further elevator surface loss could have prevented rotation and resulted in an aborted takeoff beyond V1 speed and a potential runway overrun. A more severe outcome could have resulted if the elevator’s structure had been compromised to the point where the aerodynamic loads in flight caused further elevator damage and possible separation. The change in the elevator’s aerodynamic and mass properties could also have made the elevator more susceptible to flutter.
- The cumulative effect of occasional aircraft taxiing loads may have weakened the blast pad surface making it more vulnerable to the effects of jet blast and/or to engine intake suction effects
- A blast pad should be designed in a similar way to a stopway (which may function as one) but if they are not as strong as a stopway then a clearly evident marking scheme must identify them.
- There is no requirement to alter the takeoff technique in order to achieve a safe departure provided all specified performance criteria are satisfied. Setting and then briefly holding full thrust against the brakes may have contributed to break up of the affected area of the blast pad whereas a rolling take off would have resulted in the tail of the aircraft being clear of it before a potentially damaging level of thrust had been applied.
- The provision of lead-on lines does not take account of the provision of runway performance data for take off and additional guidance is necessary.
Safety Action including the following was noted as having been taken during the course of the Investigation:
- the UK CAA reminded all licensed aerodromes of their responsibility to identify and clearly mark pavement ancillary to runways such as blast pads which may not be able to take the weight of the heaviest permitted aircraft and stated that any replacement blast pads should adopt the thickness recommended in the ICAO Aerodrome Design Manual. It also carried out a survey of blast pads, turn pads and other similar surfaces at UK aerodromes and found eight where "closer attention is going to be paid and potential redesigns considered".
- the Aberdeen Airport Operator (BAA) installed a completely new blast pad at both ends of the runway with a 10cm depth and capable of accommodating a Boeing 767, the most critical aircraft. All other UK airports also operated by BAA have been reviewed and it has been found that no remedial action in respect of blast pads is required.
- British Airways has provided its 737 pilots with additional clarification on the use of recommended take off procedures in line with the corresponding content of the Boeing 737 FCTM.
Ten Safety Recommendations were made as a result of the Investigation as follows:
- that the International Civil Aviation Organisation (ICAO) should consider amending Annex 14 to include requirements for paved blast pads that will ensure that they cannot be damaged by the engine inlet suction, the engine jet blast or the taxiing loads of the most critical aircraft. [2007-023]
- that the International Civil Aviation Organisation (ICAO) should review the requirements of Annex 14 to ensure that runway surfaces, stopways and other adjacent areas susceptible to high-power jet blast cannot be damaged by the engine inlet suction or the engine jet blast of the most critical aircraft. [2007-024]
- that the Civil Aviation Authority (CAA) should consider amending Civil Air Publication (CAP) 168 to include design requirements for paved blast pads that will ensure that they cannot be damaged by the engine inlet suction, the engine jet blast or the taxiing loads of the most critical aircraft. [2007-025]
- that the Civil Aviation Authority (CAA) should ensure that paved blast pad surfaces, stopways and turnpads at all licensed UK airports are constructed such that they cannot be damaged by the engine inlet suction, the engine jet blast or the taxiing loads of the most critical aircraft. [2007-026]
- that the International Civil Aviation Organisation (ICAO) should establish standardised markings for paved blast pads and amend Annex 14 accordingly. [2007-027]
- that the Civil Aviation Authority (CAA) should, in consultation with the International Civil Aviation Organisation (ICAO), establish standardised markings for paved blast pads and amend Civil Air Publications (CAPs) 168 and 637 accordingly. [2007-028]
- that British Airways should review the training of takeoff techniques across all fleets to ensure that it is consistent with the operator’s intended procedures. [2007-029]
- that British Airways should incorporate information on appropriate takeoff techniques in relevant flight crew documentation for all fleets. [2007-030]
- that the Civil Aviation Authority should review the implementation of current performance requirements for ‘Performance A’ aeroplanes, to ensure that they adequately reflect desired line-up techniques, in particular following ground markings provided for taxi guidance. [2007-031]
The Final Report was published in June 2007.