Aviation Investigation Report A98H0003
2.21 Fire Initiation
An evaluation of the available information indicates that the fire likely started within the confines of a relatively small area above the right rear cockpit ceiling just forward of the cockpit rear wall near STA 383. Although other potential areas were assessed, no other area was found that so comprehensively explained the initial indications of odour and the subsequent smoke and fire propagation. Support for the fire initiating and spreading from this localized area includes the following:
- The presence of electrical wires as potential ignition sources and easily ignited MPET-covered insulation blanket material;
- The known environmental conditions in the cockpit and cabin;
- The time frame in which the fire propagated from initial detection until the fire-related failures of various aircraft systems occurred;
- The air-flow patterns; and
- The fire and heat-damage patterns.
Within the localized area where the fire most likely started, a wire arcing event is the only plausible ignition source. There were several wire bundles containing hundreds of wires, including the four IFEN PSU cables and 16 American Wire Gauge (AWG) control wire, that passed through the localized area (see Section 184.108.40.206). It is most likely that the fire started from a wire arcing event that ignited the nearby MPET-covered insulation blankets. These MPET-covered insulation blankets are easily ignited and were prevalent in the area (see Section 1.16.8).
Of all the wires and cables that were located in the localized area of interest, the only arc-damaged wire that could be positioned in that area with relative accuracy was the 1-3791/1-3793 pair of IFEN PSU cables (see Section 220.127.116.11). Although it is possible that other wires from this localized area that were not recovered might also have arced, the only arcing event that is known to have occurred within that area is the forward arc on Exhibit 1-3791, located just forward of STA 383.
An assessment was completed to determine whether the forward arc on Exhibit 1-3791 could have been the result of fire damage, and therefore, would not have been involved with the lead arcing event. This possibility was considered unlikely. For this forward arc on Exhibit 1-3791 to be the result of fire damage, the fire would have to start from another unrelated arcing event within the localized area, and the fire would need to be sustained in the area for a sufficient time to melt and breach the ETFE wire insulation at the site of the forward arc on Exhibit 1-3791. To result in arcing at this site, the breached wire would have to be in contact with either grounded aircraft structure, or with a second wire of different electrical potential whose insulation was also breached by the fire. It is unlikely that a fire would selectively breach only one wire in Exhibit 1-3791 and breach at least one additional nearby wire to create conditions for an arc event to occur without also breaching the insulation on at least some of the other five wires in the 1-3791/1-3793 pair. It is unlikely that the insulation on these other five wires was breached at that location, as there were no arcs on these five wires at that location.
The forward portions of the 1-3790 and 1-3792 PSU cables, and the 16 AWG control wire, were not identified and may not have been recovered. Therefore, in the area where the forward arcing event occurred on the 1-3791 PSU cable, it is not known whether arcing occurred on either the 1-3790/1-3792 cable pair, or the 16 AWG control wire. However, as both of these PSU cables and the 16 AWG control wire subsequently arced at locations that were at least 50 cm (20 inches) farther aft, any arcing that might have occurred at the forward location did not trip the associated CB.
Given the number of unlikely circumstances and events that would be required, a scenario involving fire-related damage leading to the forward arc on Exhibit 1-3791 at the STA 383 location cannot be supported.
If the forward arc on Exhibit 1-3791 was not the result of fire-related damage, another potential scenario is that the arc occurred during the time of the lead arcing event; that is, it was associated with the lead arcing event, either alone, in combination with arcing on another wire or wires, or as collateral damage from an arcing event on another adjacent wire or wires. In any of these potential scenarios, the arcing was not sufficient to trip the associated CB.
The forward arc on Exhibit 1-3791 was assessed to determine whether it was, by itself, the lead arcing event that started the fire. For this arc to be the single lead arcing event, the wire would first have to be damaged, for example by chafing, at the location of the arc, to expose the conductor. The exposed conductor would then have to contact grounded aircraft structure, resulting in the arcing event. Although it was possible to position the cable segment (Exhibit 1-3791), and therefore the forward arc, relatively accurately, the extent and nature of the damage required interpretation of the damage patterns. This interpretation allowed for a small range of possible locations in the placement of the wires, as described in Section 18.104.22.168. At the forward end of the possible range, the arc was placed where it would be in contact with an aluminum wire support bracket. However, the chafing of any one wire by itself to this bracket would not result in an arc, as the bracket was isolated from the aircraft structure by a nylon stand-off and would not have provided an electrical path to ground.
There was aluminum found in one copper bead adhering to the wire strands slightly removed from the main arc site of the forward arc bead on Exhibit 1-3791. This suggests that the arc might have resulted from contact with aluminum. Arcing to the aluminum bracket would only be possible if there were two exposed conductors in contact with the bracket. This would provide an opportunity for arcing, as aluminum is a good conductor of electricity. Such a scenario would involve, for example, two phases of a PSU cable chafing separately against this same bracket until both of their conductors became exposed. This scenario could not be ruled out; however, there is no corroborating information to support it. The bracket was not identified in the wreckage. Neither were the two remaining PSU cables, the 16 AWG control wire or other aircraft wires from that area, that may have been involved.
Another potential lead arcing event scenario involving the forward arc on Exhibit 1-3791 would be that the arc occurred directly to another wire of a different electrical potential. This could be either an aircraft wire, or another IFEN wire. In either case, both wires would have to be damaged at the location of the arc, allowing their bare conductors to contact each other. Because the IFEN wires in the STA 383 area were routed separately and not along existing wire bundles, it is less likely that the IFEN wires would be in contact with aircraft wires within the localized area where the fire most likely started; therefore, the more likely candidate wires for this type of scenario would be the other wires in the bundle of four IFEN PSU cables and the 16 AWG control wire. It is known that the other wires in the 1-3791/1-3793 pair did not arc at that forward location. However, the wires from the other pair of PSU cables and the 16 AWG control wire from that area were not identified. Therefore, neither aircraft wires nor other IFEN wires can be ruled out as potentially being involved in such a scenario.
Damage to two or more wires in a wire bundle can be caused by chafing contact with the aircraft structure, by inadvertent damage occurring during installation or subsequent maintenance, or by the presence of swarf, such as a metal shaving, that could cut through the insulation on both wires, exposing their conductors. A metal shaving could also act as a conductor. If any of those events occurred, the subsequent arcing that took place on all of the PSU cables and 16 AWG control wire confirms that any arcing on these wires near STA 383 did not trip the associated CB.
An assessment was made to determine whether the forward arcing damage on Exhibit 1-3791 could have resulted from collateral damage; that is, damage from an arcing event involving other wires in the immediate vicinity that was of sufficient magnitude to breach the insulation on at least two other wires, including Exhibit 1-3791. For this to occur, the lead-event wires would have to be in very close proximity to the forward arc on Exhibit 1-3791. An arcing event of sufficient magnitude to damage other wires would likely have tripped the associated CBs. None of the IFEN CBs tripped at the time of the lead arcing event (subsequent arcing occurred on all of the PSU cables and the 16 AWG control wire); therefore, if such an arcing event occurred, it did not involve IFEN wires. Such a lead-arcing event would have to involve aircraft wires, but not result in any electrical anomalies that would be apparent to the pilots and not be recorded on the FDR. Although the possibility of a scenario involving collateral damage to Exhibit 1-3791 could not be ruled out, it appears unlikely that such a scenario occurred.
No determination could be made regarding how the insulation at the forward arc location on Exhibit 1-3791 was initially breached, or what that wire came into contact with, such as structure or another wire, to cause the arc. Although the available information indicates that the forward arcing event on Exhibit 1-3791 occurred during the time of the fire-initiating event, and in the area where the fire most likely originated, it cannot be concluded that the forward arc on Exhibit 1-3791 was the lead arcing event. It appears likely that at least one other wire was involved in the lead arcing event; however, it could not be determined whether this was an IFEN wire or wires, one or more aircraft wires, or some combination of both.
An arcing event or events provided an ignition source for the fire; however, this arcing would not have resulted in a threat to the aircraft had there not been material nearby that could easily be ignited by such an ignition source. The presence of significant amounts of flammable materials allowed the fire to spread and intensify rapidly, which ultimately led to the loss of control of the aircraft.
- Date modified :
- Date modified: