The first indication of fire in the No.1 engine nacelle occurred shortly after Event 2 in the fan failure sequence when there was evidence from both the FDR and CVR that a fire warning had been triggered. Although the fire warning was not related to the No.1 engine by the FDR evidence, it was by the subsequent conversation between the commander and first officer recorded on the CVR.
The examination of wreckage on the accident site indicated that only the No.1 engine and nacelle had evidence of fire damage and the more detailed examination revealed that there was no evidence of fire elsewhere on the aircraft. Examination of the No.1 engine and nacelle had revealed that there were three separate and distinct fires which had occurred on this power plant. Of these, two were in zones monitored by fire detection loops, one on the outboard side of the fan case and the other on the underside of the combustion case. Neither had evidence of the presence of any forcing air draught to show that they had been burning in flight. The third fire had been on the outside of the nacelle, on the outboard side of the reverser duct, remote from any detectors. This fire showed the characteristics of having been slipstream driven and must have been burning whilst the aircraft was in flight.
Since the fire on the outboard side of the reverser duct was in an unmonitored zone, it is highly unlikely that it could have triggered the warning. Thus, since the warning had been triggered in flight, there must have been fire or very hot gases present in one of the other two zones. The fire on the underside of the combustion case, although in a monitored zone, had been very minor and showed the characteristics of a restricted fuel ground fire. It was also seen that this fire centred on a fractured fuel nozzle fitting which had clear evidence of having been damaged during the ground impact sequence, indicating that this fire was entirely post impact. Thus the fire warning must have been triggered by the fan case detectors.
The fan cowling itself is a nominally enclosed space, the only venting being via the cowl drain at the bottom, just ahead of the cowling firewall. Thus, even if the intake duct were breached by fan debris, it is unlikely that there was a fast airstream through the fan cowl in flight. The ground fire which had affected the fan case after the aircraft had come to rest had consumed a large proportion of the forward cowling and had overlaid any evidence, on the outside of the fan case, of any fire which might have occurred in flight. The reconstruction of the cowling, however, showed that some areas of the forward cowl which had broken free at impact had not been involved in the ground fire. By positioning these pieces it was possible to show that a fire had been present in the fan cowl before it had been broken up by impact and that fire appeared to have been present on the outboard side of the fan case and above the level of the MEC.
The investigation of the two other engines involved in fan blade fatigue failures showed that fuel pipe unions and the seal between the HP fuel pump and the MEC were susceptible to being loosened to the point of leaking if subjected to severe vibration. Furthermore, the trials performed to try to establish the characteristics of such leaks showed that atomised fuel spraying could occur as the result of unions being loosened to the extent found with the fuel pressures to be expected whilst the engine was running.
The exposure to vibration of the 3 engines, which suffered fan blade fractures, was compared. It was observed that the No.2 engine from MG had suffered a brief initial period of very high vibration at a high power setting, under similar conditions to that seen on the No.1 engine of ME but of about half the duration. Thereafter the engine was throttled back to a flight idle setting for a similar duration to that experienced by the ME engine during the descent. During the approach to East Midlands, the ME engine experienced two exposures to higher vibration levels as a result of engine power being increased. Thus the No.1 engine of ME had experienced greater exposure to vibration than the No.2 engine of MG and was likely to have at least as much loosening of pipe unions. Consequently, since the fuel unions found on the MG engine were sufficiently loose to produce atomising spray leaks, it is probable that immediately before Event 2 , at the time of the second power increase, atomised fuel sprays were present in the fan cowl of the ME engine.
If such leaks were present with no ignition source, the free fuel would have run to the base of the cowling and escaped through the vent and drain apertures. The airstream around the outside of the nacelle flows upwards and outboard, the upwards component increasing with angle of attack. The evidence of fluid streaking running in this direction showed that a significant quantity of fuel and oil, which could also have been liberated by a similar union loosening process, had been present in flight. When the fluid on the outside of the nacelle reached the trailing edge of the cowling, it would have been drawn off as a highly combustible atomised mist by the slipstream.
The source of ignition for the fluid mist from the trailing edge could have come from either the flames resulting from compressor stalling passing down the fan stream or incandescent fan blade particles generated by the fan break-up at Event 2. The same potential sources could also have ignited fuel sprays within the fan cowl if the inlet duct had been breached. Although no evidence remained to demonstrate that it had been, the likelihood of a breach being made by fan break-up fragments is reasonably high, and no evidence was found of an alternative ignition source within the cowl.
Although the effects of the fire were restricted to the No.1 engine and nacelle, this was principally due to the fact that the airfield fire service was able to attack it with appropriate extinguishants before it had time to spread. Had it been a significantly longer time before fire fighting was possible, although there was very little fuel spillage from the aircraft, it is probable that a much greater loss of life would have resulted.
The likelihood of a post crash ground fire will be much greater if there has been fire on an aircraft in flight. Fuel or oil leakage from loose pipe unions is an ever-present hazard and will always increase the possibility of a fire in flight.
It must be accepted that the vibration levels experienced on the No.1 engine of ME were orders of magnitude greater than those normally present and, therefore, more likely to cause loosening of pipe unions. However, it is under such conditions that there is likely to be an increased risk of accident. Although the fitting of locking wire to pipe unions would not entirely prevent loosening of these unions, it would limit the degree of looseness and, consequently, the likelihood of an atomised spray with a higher susceptibility to ignition. The fuel and oil pipe unions on the fan case of the CFM 56, in line with current practice, are not generally wirelocked; the control air pipes are the only ones, in this zone, wirelocked as a matter of course.
Although wirelocking of pipe unions will not prevent leakage of combustible fluids completely, its benefits and shortcomings should be reviewed in relation to the potential reduction of fire hazards in vulnerable zones. It is therefore recommended that the potential for fuel and oil system leakage within the fan case area of high by-pass turbofan engines, during conditions of excessive vibration, should be reviewed by the engine manufacturers and the CAA with a view towards modifying such systems to minimise such leakage and the associated fire risk (Made 30 March 1990).
The other major factor which affected the post-crash fire was the lack of a major release of fuel in the impact. This was partly due to good fortune, in that the centre section fuel tank, which was ruptured, did not contain fuel for this flight and that the damage to the left wing-tip occurred outboard of that fuel tank. It was, however, more largely due to the continuing integrity of the wing fuel tanks further inboard, which did not rupture despite the separation of both main landing gear (MLG) legs and the almost complete separation of both engines.
The wreckage showed that both MLG legs separated entirely consistently with the crashworthiness features of their design, failing the fuse-pin bolts and leaving the rear wing spars intact. In the case of the engines, the structural failures occurred within the pylons themselves, leaving the fuse-pin bolts in place; the separations were, in this instance, benign and the forward wing spars were not disrupted.
The excerpts in paragraph 1.17.16 are from the applicable airworthiness code (BCAR Section D) and the current code (JAR-25). They concern fuel tank penetration and address the MLG failure mode case (JAR-25.721) and the rear-mounted engine case. However, they do not address, other than in very general terms, the case for wing-mounted podded engines such as on the Boeing 737-400 and similarly configured transport aircraft. It is recommended, therefore, that the CAA should review the existing Joint Airworthiness Requirements concerning fuel tank protection from the effects of main landing gear and engine detachment during ground impact and include specific design requirements to protect the fuel tank integrity of those designs of aircraft with wing-mounted engines. (Made 30 March 1990).