Sequence of events
Shortly after selecting the landing gear up after departure on a
flight from Brisbane to Barcaldine on 27 November 2003, the crew of
the de Havilland Canada DHC-8-202 (Dash 8), registered VH-SDA,
observed the number-two hydraulic pump caution light illuminate,
followed by a zero hydraulic pressure indication. The number-two
hydraulic system powered the aircraft's roll spoilers, ground
spoilers, parking brakes, nose wheel steering and landing gear
extension and retraction. The crew diverted back to Brisbane and
extended the landing gear manually for the landing. The failure of
the hydraulic system pressure was traced to the failure of a
flexible hydraulic hose in the nose landing gear actuation system,
which allowed the loss of system pressure and hydraulic fluid.
The failed component was forwarded to the Australian Transport
Safety Bureau in Canberra for examination. The flexible hydraulic
hose, part number DSC252B4-0124, failed at 12,369 cycles since new
and had been installed since the aircraft's manufacture. When first
introduced into operation, the hose life limit was set at 15,000
cycles. However, the aircraft operator indicated that a company
replacement life limit of 13,000 flight cycles had been introduced
in response to a history of failure of the same or similar hoses in
the Dash 8 landing gear system.
Examination by the ATSB found the hose had failed by localised
rupture at the point of swaged connection to an end fitting.
Associated with the rupture was evidence of fatigue cracking and
breakage of the external reinforcing braid wires, with cracking
also found to a lesser degree on the opposite side of the
connection. There was no evidence suggesting that a manufacturing
or material defect had contributed to the hose failure. Assembly
diagrams showed that the hose failed at the point of maximum
flexure when the landing gear was extended or retracted.
Failure of the hose was attributed to the localised fatigue
cracking and breakdown of the external braided hose reinforcing
sheath and the subsequent rupture of the tubular core in the
absence of the support afforded by the sheath. In-service flexures
of the hose and pressure cycles and pulsations inherent in the
operation of the aircraft's hydraulic system were considered to be
likely contributory factors.
The hose failed 631 flight cycles before it was due for removal
in accordance the aircraft operator's maintenance schedule. Given
that the failure was a result of a fatigue cracking mechanism and
there has been a history of similar occurrences, it would be
expected that the identified cycles-to-failure for any such
component would lie within a distribution represented by lower and
upper limits ('outliers'). While most of the failures would occur
between those limits, there remains a probability that some
components may fail before the lower limit, or indeed after the
upper limit. The establishment of a service life limit for a
component based on the distribution takes into account the
criticality and acceptability of component failure and hence, the
necessary confidence level in defining a service limit for removal
of the item before failure.
The evidence to hand suggests that this flexible hydraulic hose
failure was a distribution 'outlier'. Given that the reliability of
the assembly is being managed by life limiting the susceptible
component, and the loss of a hydraulic system in this case was able
to be managed using existing aircraft flight manual procedures, no
Bureau recommendation has been made with respect to the maintenance
philosophy applied to that system.