Revised Occurrence Brief, 8 December 1999
The pilot of Kawasaki 47G3B-KH4 helicopter, registered VH-KEB, reported that following a normal take-off, he heard a loud cracking sound and felt some restriction through the cyclic control. The pilot placed the helicopter in an autorotation descent from about 50 ft above a tidal estuary and landed in about 2 ft of water. Following contact with the water, the helicopter nosed over and rolled through 360 degrees, coming to rest upright. The pilot exited the substantially damaged helicopter unaided.
The damage to the helicopter was consistent with impact with the water and riverbed, except for damage to the engine-cooling fan assembly, which was consistent with pre-impact failure. Several blades had fractured and separated from the fan assembly. It was also evident that the collective control rod located in front of the cooling fan had been damaged by the impact of a released blade. The fan was identified by the serial number A34-08516 etched onto the inner edge of the cooling fan disc.
Analysis of the cooling-fan fractures established that a two-bladed segment of the fan separated from the assembly as a result of the initiation and growth of fatigue cracks. A single blade separated as a result of the development of an excessive stress in the region of the blade root; this fracture was considered to be secondary to the separation of the two-bladed segment by fatigue-crack growth.
The cooling fan, s/n A34-08516, was physically consistent with the fan, part number 47-661-029-7, specified for this helicopter. However, significant differences in the surface and assembly of the fan fitted to KEB were evident when compared to the "as manufactured" condition of engine-cooling fans and the assembly requirements of the helicopter manufacturer. The fan disc surface had been shot-peened. Fans manufactured by Kawasaki Heavy Industries (Kawasaki) were not shot-peened, and the Bell Helicopter Company did not specify shot peening as a surface treatment for engine-cooling fans. The surface of the fan had been abraded in areas around the boltholes. The resistance of a component to fatigue is sensitive to surface finish. Abrasion of a shot peened surface would be expected to reduce the resistance to fatigue.
Significant factors that affect fan assembly stiffness, and consequently fatigue, are the establishment of the required (designed) clamping force in each fan assembly bolt and the retention of the required clamping force during operation. The tightening instructions included in the Kawasaki service bulletin KSB-BELL-305 did not indicate whether the final torque included, or should be adjusted for, any frictional torque effects. The frictional torque of nut self-locking features will affect clamping forces established by the applied torque method. Clamping forces may be reduced during operation by the deformation of materials included in the bolted joint. In addition, the fan had been painted, and deformation of the paint film under the heads of the bolts had occurred. In a letter dated 18 September 1998, Kawasaki stated that it did not recommend the painting of cooling fans part numbers 47-661-029-2 and -7 for Kawasaki/Bell model 47G3B-KH4 helicopters. The investigation was unable to find such a reference in the manufacturer's manuals.
The loud cracking noise and the restriction in the cyclic control reported by the pilot may have resulted from the fatigue failure of the engine-cooling fan. The investigation was unable to determine if the damage to the control rod would have been sufficient to adversely affect the flight controls.
Although it is likely that the specified torque was previously applied to the fan retaining bolts, the required clamping force may not have been achieved due to frictional torque effects and the painted surface of the fan.
The difference in the surface finish from that specified by the manufacturer would be expected to have a detrimental effect on the resistance of the fan to fatigue failure. Differences in assembly practices from those specified by the manufacturer may alter the magnitude of the alternating stresses created in the fan. A determination of which difference was the most significant, or whether failure resulted from the combination of the effects of several differences, would require further testing and analysis by the manufacturer.
The Bureau of Air Safety Investigation (BASI) released occurrence brief 9501019 on 2 February 1996. On 27 April 1999, following receipt of new information from the helicopter manufacturer concerning the identity of the part number of the fan fitted to KEB, a revised occurrence brief was issued by BASI.
The owner of KEB drew attention to a number of potential safety deficiencies in the maintenance and inspection of Kawasaki 47G3B-KH4 helicopter engine-cooling fans.
In response to the issues raised, and in recognition of the hazard to flight safety created by engine cooling-fan blade fracture, the Australian Transport Safety Bureau (formed on 1 July 1999, and which integrated the Bureau of Air Safety Investigation in a multi-modal agency) then undertook an analysis of the failure of Kawasaki 47G3B-KH4 engine cooling-fans. This analysis did not consider operational or piloting aspects, but was confined to the analysis of the cooling fan. As a result of technical analysis of this and other engine-cooling fan blade failures, the following recommendations are made:
The Australian Transport Safety Bureau recommends that the Civil Aviation Safety Authority alert all aircraft maintenance engineers, especially those conducting non-destructive inspections, to the detrimental effects that may be created in critical components by the abrasion of surfaces during general cleaning or cleaning in preparation for nondestructive inspection.
The Australian Transport Safety Bureau recommends that the Civil Aviation Safety Authority alert helicopter maintenance organisations and all aircraft maintenance engineers to the potentially detrimental effects of painting engine cooling fans.
The Australian Transport Safety Bureau recommends that Kawasaki Heavy Industries clarify the bolt-tightening instructions contained in Service Bulletin 305 to establish if an allowance should be made for the frictional torque effects of self-locking nuts.
|Date:||06 April 1995||Investigation status:||Completed|
|Time:||1158 hours EST|
|State:||New South Wales|
|Release date:||22 December 1999||Occurrence category:||Accident|
|Report status:||Final||Highest injury level:||None|
|Aircraft manufacturer||Kawasaki Heavy Industries|
|Type of operation||Charter|
|Damage to aircraft||Substantial|
|Departure point||Moonee, NSW|
|Departure time||1158 hours EST|
|Destination||Coffs Harbour, NSW|
|Role||Class of licence||Hours on type||Hours total|