APU

The aircraft's Auxiliary Power Unit (APU) has received much of the blame for the accident, as it is located in the tail and was in operation when the accident occurred. An APU is a turbine engine that drives an alternating current generator. When the fuel supply is shut off, the APU stops within a few seconds.

The accident investigation board's conclusion is built on the premise that the use of the APU on this flight was a decisive factor. It is not necessarily important to know exactly when the APU was stopped, but it is so central to the accident investigation board's causal hypothesis that there should be no remaining open questions surrounding it. There are contradictory findings regarding when in the sequence of events the APU was stopped, and the accident investigation board itself expresses doubt as to whether the APU could have been the initial cause of the damage to the tail section.

«During normal operation, the APU turbine rotates at over 40,000 RPM. If the friction led to vibration, it would have been high-frequency. It therefore seems improbable that the APU, with its defective forward mount, would be the initial cause of the large alternating loads to which the rest of the tail has been subjected.» (The Accident Report, 1993, p. 99)

«The forces to which the control surfaces were subjected overloaded the linkages between the cockpit and the control surfaces, causing them to fail.

The crew probably knew at this point that something was wrong in the tail, and activated the fire extinguisher for the APU. This caused the fuel supply to the APU to be shut off, but had no influence on the subsequent course of events.» (The Accident Report, 1993, p. 107)

«In the unit's turbine section, fragments of melted plastic were found. The plastic fragments were identified using IR spectroscopy and had the same composition as materials used in the cabin interior.» (The Accident Report, 1993, p. 68)

The accident investigation board's hypothesis is that the APU was stopped during the period 34–36 minutes FDR time, when the crew was in the process of losing control of the aircraft. This does not, however, reconcile with the fact that an APU takes only a few seconds to stop. If burnt plastic from the cabin interior was found in the APU turbine, it seems more probable that the APU was in operation when the aircraft broke apart. Alternatively, it may mean that the APU was stopped only seconds before the fuselage disintegrated.

Stopping an APU that was powering one of the aircraft's two alternating current systems seems unlikely if the crew was attempting to manage a power problem. If one is in the process of losing control of the aircraft, a pilot would naturally focus on the aircraft's primary flight controls and engine settings. In that case, it may mean that the pilots made no attempt to stop the APU, which means it was in operation when the aircraft disintegrated. This is consistent with the finding of melted plastic in the APU's turbine.

«The squib, which is intended to trigger the fire extinguisher for the APU, was found in a fired state. The charge is fired electrically. Tests have shown that an applied voltage of approximately 15 V DC is sufficient to fire such a charge.»
(The Accident Report, 1993, p. 69)

The accident investigation board concludes that the fire extinguisher for the APU had been triggered, but does not explain how it could have been triggered without the aircraft having power on either the alternating current system or the direct current system. It would therefore have been natural to question whether the fire extinguishing mechanism could in fact have been triggered by the crew. The absence of electrical power and the finding of burnt plastic from the aircraft's interior weaken the accident investigation board's hypothesis that the crew activated the fire extinguisher.

«The technical findings show that the APU, after the forward mount had failed, remained in place and swung about the lateral axis. A large number of impact marks on the forward mount demonstrate this. Movement about the lateral axis was possible due to elasticity in the mounts that were still intact. "These oscillations must necessarily have affected the properties of other elements in the tail section."» (The Accident Report, 1993, p. 100)

«It is probable that the fact that the APU vibrated about the lateral axis is related to the oscillations in the elevator, although it cannot be established with certainty which of these two vibration sources initiated the reaction in the other.» (The Accident Report, 1993, p. 107)

It would have been interesting to know whether the impact marks occurred on the accident flight, or whether they developed over time when the aircraft flew in turbulence with the APU switched off. Gyroscopic precession is a physical phenomenon that causes a rotating object to resist changes in its position—in the same way that a spinning top remains upright as long as it is spinning. When flying straight ahead in stable flight, this effect will stabilise the APU about two axes while it is in operation. Given that the APU's shaft is mounted along the aircraft's longitudinal axis, this means the APU will resist movements about the lateral axis and the vertical axis. Only rotation about the longitudinal axis would be able to occur without the influence of gyroscopic precession. Without any further explanation of what is meant by the APU "remaining in place and swinging about the lateral axis," it is worth questioning whether this is even possible while the APU is in operation.