I’m not sure that it serves any useful purpose trying to pin the tail on the donkey in seeking a single cause to this accident. Perhaps we should move away from the use of human error and towards the less pointed human factors. The primary cause of this accident appears to be a mechanical failure; to establish all secondary causes requires that we look back and try to understand why the previous incident did not ‘raise the hackles’ and why it was felt that there was ‘room to manoeuvre’ after the first stud failure last year.
It is clear that the precursors to this accident had already been seen; on July 3rd last year malabo hinted at a transmission problem in Australia; Heli-kiwi then indicated a failure of oil filter studs; Malabo and Heli-kiwi traded quips until, after being pressed by NorthSeaTiger, Heli-kiwi gave us the essential information
Day before yesterday an S92 was onroute to Broome on a return flight from an offshore rig when a main Txmsm low oil pressure was noted, shortly afterwards remaining oil pressure ceased completely and a fairly rapid decent was carried out. From time of intial low px indication at 6000ft to touchdown was 8 minutes. The oil filter housing which apparently has 3 studs fastening it to the txmsn was hanging on by one which hadn't broken and most of the oil had been pumped out.
Following this report, the thread meandered on but was more engaged by the incident of a pump failure than the fact that we had seen an extremely serious incident caused by a major failure of the MGB assembly.
Within hours, it was known that all MGB oil had been lost through the filter bowl assembly. Shortly after that, it was also known that this was caused by the failure of the titanium studs. As was pointed out in an earlier post, we now had a failure mode which had not been considered during certification. In fact it must have been known quite soon after the inspection that, without a change of material for the studs, another failure was foreseeable with a likelihood of reasonably probable. Such a conclusion (if recognised, and it is not clear that it was) should have triggered a hazard analysis which should have impacted, in turn, the manufacturer, the regulator and the operator.
The manufacturer and the regulator should have conducted an assessment of whether the aircraft should be grounded until the titanium studs were replaced, against the likelihood of another failure. The fact that the aircraft was not grounded and, in the letter to operators in January, it was still not felt necessary permitting, instead, another 1250hrs or 12 months for replacement is not easily understood (in hindsight). Even with a 30 minute run-dry capability this would appear too relaxed. However, having made those decisions, all operators should have been brought into the decision making process because it was them, and their customers, who were now to take the front-line risk.
At the very least, the results of the investigation of the Broome incident should have been promulgated to all operators so that they could educate/inform their pilots, so that the likelihood and consequence of such a failure would be foremost in the minds. More, the whole notion of a 30 minute run-dry time for gearbox should have been debunked and the necessity, regardless of environment, of a ‘land immediately’ in any case of a loss of oil pressure reiterated. In view of the fact that data had been collected during the Broome incident, it would also have been possible to put the sequence of events and cockpit indications into a file, and distributed for the purpose of demonstrating the symptoms and indications.
This leads to the questions:
(1) Was the probability of another stud failure established?
(2) Was the probability of this event put to the regulator?
(3) Did the regulator accept that nothing need be done until the new stud had been sourced?
(4) When the new stud was sourced, was the rectification interval accepted by the regulator?
(5) Were the operators informed of the Broome event and the probability of another until such time as the studs had been replaced?
(6) If the operators were made aware of the situation, did they risk assess their procedures with the aim of minimising the consequences of a failure?
(7) were the operational and training departments aware of the probability of failures and did they place appropriate focus on the required actions?
The main difference between the first incident and the Cougar accident is the action of the crew following the manifestation of the same emergency. We can only speculate on why there were differences in reaction but it is probable that it was due to the operational environment: in Australia, the aircraft had coasted in and was over land; in Canada, the aircraft was over Sea State 5 and a water temperature of just above zero. Regardless of this, had the sequence of actions shown above been put into place, this accident might have been another (however, more serious) incident.
Anyone who reads this (and the S92) thread now knows that ‘land immediately’ provides no leeway. It should never have been in doubt; it might not have been if we had learnt the lessons from the Broome incident and paid less heed to those for whom such rules/procedures are seen as a challenge to their intellect.