Diabetes mellitus is a disease with a wide spectrum of severity and many potential complications if inadequately treated. Historically, diabetic pilots have been permanently disqualified from flying duties. This policy was based on the increased risk of sudden incapacitation in-flight due to hypoglycaemia and cardiovascular disease in diabetics. In recent decades, a shift in worldwide aeromedical policy has occurred. This has resulted in diabetic pilots in several countries, including the United States and Australia, being granted limited flying certification. These pilots are required to satisfy a number of stringent medical criteria to achieve this certification. Aeromedical policy must be based on an appropriate risk management strategy, taking into account all relevant issues. Australian guidelines for the certification of diabetic pilots are designed to limit certification to all but the most well-controlled, motivated, and well supervised diabetic with no disease-related complications.

Boeing 747-438, VH-OJU

EXECUTIVE SUMMARY

An Australian registered Boeing 747-438 aircraft operating a regular passenger transport flight sustained the failure of an engine shortly after take-off from Los Angeles, USA. The engine was subsequently shutdown and the aircraft returned for an uneventful landing.

The failed engine was a Rolls Royce RB211-524G2-T model. Preliminary inspection by the operator's maintenance personnel found evidence of extensive internal mechanical damage within the high-pressure compressor section of the engine and as a result, returned the engine to Australia for inspection and overhaul.

The Australian Transport Safety Bureau examined the engine following its disassembly into primary modules. The engine had failed as a result of the liberation of a single blade from the first-stage high-pressure compressor section. That failure subsequently precipitated a titanium metal fire within the compressor, extensively damaging the following stages and rendering the engine inoperative.

The engine manufacturer has attributed three previous failures of RB211 high-pressure compressors to the loss of blades from the first-stage rotor. The blade losses were all associated with fatigue cracking of the dovetail root connection. The manufacturer identified uneven centrifugal loads on the blade roots as a significant factor in the development of blade cracking; possibly exacerbated by 'patchy' root friction and minor mechanical imperfections in the critical blade root transition region.

Evidence from the current investigation indicated the nature of the failure to be very similar to the previously reported events.

During a routine departure from Melbourne aerodrome, the inner left main landing gear wheel separated from a B727 aircraft (registration VH-TXH) and was later recovered near the aerodrome perimeter fence. Preliminary examination showed that the outer axle bearing had failed, allowing the wheel hub to move axially outward over the retaining nut and off the stub axle (figure 1).

There has been a long standing interest among various transport safety organisations, researchers, other stakeholders and the public concerning the relative safety of various transport modes. Questions are often posed along the lines: on average, is travel in a light aircraft safer than a typical journey in a private car? or - what is the safety difference between motorcycle riding and driving a car? Consideration of relative transport safety risks also has potentially important policy implications, particularly where contingent resource allocation or risk management decisions are involved. In an attempt to address the core substance of these questions this Discussion Paper explores some of the main issues associated with the development of comparative safety measures across various transport modes and compares the results available from Australian and other studies.

Boeing Commercial Aircraft Group, 747-436, G-BNLK

1. FACTUAL INFORMATION

1.1. Examination brief

The disassembled components of an electric air-cooling fan (figure 1) were received by the ATSB Technical Analysis unit for examination and analysis of the damaged fan impeller. The fan unit had been fitted to a Boeing 747 aircraft (registration G-BNLK) to provide forced air circulation for a forward galley chiller unit. During the early stages of a flight on 10 August 2002, a small fire developed in the forward cargo compartment adjacent to this unit (figure 2). Physical and recorded evidence suggested the fire had initiated from electrical arcing that was a result of a wiring short-circuit near the fan terminal housing.

FIGURE 1: Cooling fan assembly. Inlet is at top.	FIGURE 2: Cooling fan location (arrowed) behind cargo hold wall.

1.2. Samples received

Data plates affixed to the fan housing (figures 3 & 4) identified the unit as a three-phase unit (part number 73259E, serial number 3676), manufactured by Sunstrand (San Diego, California) in 1994. The fan was an axial flow design, with a single bell-shaped impeller manufactured from a moulded resin material. The motor and fan outlet guide vane assembly shared an integral housing which also carried the (damaged) electrical terminal housing. The motor was a brushless (induction) design, with the armature supported on sealed rolling element bearings. Surrounding the impeller was an aluminium shroud, which formed the fan intake and also provided for the mounting and support of the unit and its associated ducting.

FIGURE 3 & 4: Data and wiring information plates affixed to the fan housing.

1.3. Visual examination

1.3.1. Impeller

The impeller unit had been effectively 'cobbed', with all eleven blades fractured at or immediately adjacent to the impeller hub (figure 5). The uneven, irregular nature of the fractures suggested the failure occurred as a cascading fragmentation event, with multiple sections of blade breaking away and striking others, causing further break-up. A study of all fracture surfaces failed to identify any evidence of pre-existing defects or cracking that may have precipitated the initial blade failure, nor was any indication found of unusual hard-object impact damage that may have suggested foreign object ingestion. There was some evidence however that suggested early damage to the impeller blade forward corners - many blades showed breakage of a curved lip of material from the corners in a manner that suggested possible contact with the fan shroud (figure 6).

The inside surfaces of the impeller body carried a heavy coating of a powdery brown dust (figure 7), which was also evident inside the armature core and around the end of the motor housing. This material was loose and easily wiped away by hand and a sample was taken for later qualitative chemical analysis.

FIGURE 5: 'Cobbing' of all the blades from the impeller body.	FIGURE 6: Scalloped fracture form at the forward edge of several impeller blades.
FIGURE 7: Dusty brown deposit on inside of impeller body.

1.3.2. Fan shroud

Produced as a machined die-casting from an aluminium alloy, the fan shroud encased the full impeller length and showed no evidence of having failed to fully contain the fractured impeller blades. Around the blade tip path, the shroud internal surfaces showed several circumferential wear bands that indicated significant tip interference (figure 8). These were most clearly defined at the forward and rear limits of the blade path. Random indentation and scratching damage was noted around the shroud 'throat' region - this was consistent with the effect of multiple fragment impacts produced by the failing impeller blades. No specific evidence of hard-object impact damage was found across the blade path. Dark, waxy stains produced by an unidentified liquid draining through the fan shroud were noted at the low-point of the assembly (figure 9) and a series of tide-marks were formed on the front face of the shroud as the liquid had accumulated and later drained away.

FIGURE 8: Fan shroud showing prominent evidence of rotational contact.	FIGURE 9: Fan shroud with stains from liquid draining through the assembly.

1.3.3. Fan motor and housing

Being a brushless three-phase induction motor, the unit was comparatively simple in construction, with a star-wound stator containing a compact cage-type armature. Internal inspection found evidence of rub and erosion of the iron stator former over the innermost twenty millimetres of the stator length (figure 10), however the armature did not reflect this and showed no evidence of operational damage (figure 11). Several areas where material had been removed by bevelling the core corners were evidence that the armature had been re-balanced at some time following original manufacture (figure 12). Both armature shaft bearings rotated smoothly by hand and showed no notable indications of distress or abnormal operation. The rear armature bearing housing (figure 13) contained particulate debris that appeared to be the fine brown dust mixed with lubricant lost from the bearing unit. The contact points of the leaf spring and the housing bottom showed light fretting damage (figure 14), with some evidence of rotation also noted.

FIGURE 10: Motor stator inside surface showing clear evidence of rotational contact against the armature.	FIGURE 11: Motor armature showing no evidence of contact with the stator.
FIGURE 12: Motor armature with material removed for balancing purposes.	FIGURE 13: Motor end bearing housing with contaminants found.
FIGURE 14: Area of fretting damage found between the seating spring and housing base.	FIGURE 15: Electrical arcing damage and associated metal loss.

The fan motor electrical supply was routed behind the outlet guide vanes and terminated in an external housing fitted with an eight-pole connector. A localised area of heavy electrical arcing was clearly evident adjacent to the connector and had resulted in the melting and loss of around ten to fifteen millimetres of material in a v-shaped notch from both the front and side faces of the connector housing (figure 15). The metal loss had exposed some of the internal wiring and had produced an appreciable level of heating around the contact area, as evident from the tinting of the surrounding surfaces.

1.4. Dust analysis

The sample of brown dust recovered from the underside of the impeller body was qualitatively analysed using energy-dispersive x-ray spectroscopy techniques under the scanning electron microscope. The results of this test work (figure 16) showed the material to be comprised primarily of an iron-oxide compound, with traces of chromium, aluminium and silicon.

FIGURE 16: EDS spectra for the brown dust found beneath the impeller. Strong Fe and O peaks.

2. ANALYSIS

2.1. Impeller failure

All of the failed impeller blades exhibited brittle overload fractures. No evidence of material flaws, cracks or other pre-existing damage was shown by the blade stubs, however this was not conclusive, given the opportunity was not available to examine all of the blade remnants. Resin materials such as that used to manufacture the impeller have an inherent lack of ductility and as such, are prone to cracking and fracture under impact or elevated stress conditions. While no direct evidence was found to suggest the ingestion of foreign object/s, the damage sustained was not unlike that expected from such an event, and thus this possibility cannot be discounted.

2.2. Motor damage

The abrasion exhibited by the inner sections of the stator assembly was believed to be evidence of armature contact, presumably caused by the failure and collapse of the end bearing unit. The dust under the fan impeller was believed to be an accumulation of wear products from this contact - both the stator former and the armature core were ferrous alloys, while the remainder of the motor and fan assembly was predominantly aluminium based materials. Given that the installed armature showed no indication of stator contact and the end bearing was sound, it was concluded that the stator damage was a product of a previous failure that had been repaired by replacement of the armature and the re-use of the remaining components, including the impeller.

The collapse and failure of an armature shaft bearing and the shaft misalignment that results would be expected to alter the impeller - shroud clearances, with a risk of contact between the shroud and the impeller blade tips if the misalignment became severe enough. Impeller blade tip contact, if it did not produce immediate blade breakage, may produce latent cracking damage that could lead to later blade failures if not detected. While there was no evidence found to suggest that pre-existing damage of this nature existed, the possibility remained that this damage had been sustained given the evidence of the rear bearing failure and the contact marks inside the fan shroud.

2.3. Vibration

Rotary equipment such as the cooling fan relies upon accurate dynamic balancing to minimise the vibration induced during operation. Events such as the impeller failure will disrupt the balance of the assembly and can lead to significantly increased vibration levels. Wiring or piping that is installed against or in contact with the vibrating equipment may sustain fretting or erosion damage if the external protection or insulation has not safeguarded against this event.

3. CONCLUSIONS

3.1. Findings

1. The cooling fan impeller had sustained gross breakage of all blades at or adjacent to the body of the impeller.
2. All fractures were brittle in nature and showed no evidence of pre-existing defects.
3. No evidence of foreign object damage was found.
4. The fracture profiles of some blades suggested preferential tip breakage before the complete blade failure.
5. The fan motor showed evidence of a previous rear bearing failure that had produced contact and wear between the armature and the stator. The armature had subsequently been replaced, however the stator and other motor components remained in service.
6. The previous bearing failure may have allowed operating contact between the impeller blade tips and the fan shroud. This contact may have produced blade cracking which predisposed the impeller to failure in the manner observed.
7. After failure, the fan impeller would have presented a significant unbalanced load to the fan.

Boeing Commercial Aircraft Group, 747-436, G-BNLD

1. FACTUAL INFORMATION

1.1. History of the flight

On the evening of March 1, 2002, Boeing 747-436 aircraft G-BNLD sustained the failure of the number-3 (right inboard) engine during a scheduled regular passenger transport flight from Sydney to Bangkok. The flight crew experienced vibrations and received an ENG 3 REVERSER engine indicating and crew alerting system (EICAS) message. The crew shut down the number-3 engine and completed checklist items before returning the aircraft to Sydney.

An initial engineering examination found that a fan blade from the number-3 engine had failed and that debris had punctured the engine cowl, the right wing leading and trailing edge flaps and the fuselage; damaging a structural member above the wing root area. The inspection found fractured fasteners and other components beneath the fan cowls and damage to the structure associated with the thrust reverser assembly. Debris from the number-3 engine was also found embedded within the intake cowl of the adjacent number-4 engine.

Recorded voice data, such as from cockpit voice recorders (CVRs) or air traffic control tapes, can be an important source of evidence for accident investigation, as well as for human factors research. However, most of the analysis has been based on subjective interpretation rather than the use of systematic methods, particularly when dealing with the analysis of crew interactions. Conversation analysis uses highly detailed and revealing transcriptions of recorded voice (or video) data that can allow deeper analyses of how people interact. When analysing recorded voice data, and especially for understanding instances of human error, often a great deal rests on investigators' or analysts' interpretations of what a pilot said, or what was meant by what was said, or how talk was understood, or how the mood in the cockpit or the pilots' working relationship could best be described. Conversation analysis can be a tool for making such interpretations.

The ATSB Annual Review documents ATSB's achievements and safety activities from 1 July 2004 to 30 June 2005 and outlines its business planning for 2005-2006.

Executive Directors message

In 2004-05 the ATSB benefited considerably from the funding boost for aviation investigations and aviation database replacement that was provided in the May 2004 Federal Budget. During the year, the Bureau recruited and commenced training 12 extra aviation safety investigators, instigated 109 aviation occurrence investigations and released 98 aviation investigation reports, up from around 60 in recent years. High profile aviation safety investigation reports released in 2004-05 included reports on investigations into a fatal Cessna C404 aircraft accident at Jandakot Airport WA, a fatal Emergency Medical Services (EMS) helicopter accident near Mackay, Qld and a Boeing 737 terrain proximity warning near Canberra. The Bureau also released nine aviation safety research reports.

During 2004-05, the ATSB issued 19 aviation safety recommendations and two safety advisory notices and successfully negotiated valuable safety actions by regulators, operators, manufacturers and other safety stakeholders. For example, following ATSB recommendations from the investigation into the EMS helicopter accident, the Queensland Emergency Services Department is improving standards and support for Community Helicopter Providers including requirements for night Visual Flight Rules (VFR) flights and the helicopter operator is requiring and training all pilots to Command Instrument Rating standard. The Bureau also completed Stage 1 of the Safety Investigation Information Management System (SIIMS) aviation database replacement project, which involved developing ATSB user requirements and the trial of software tools to support the improved management of safety investigations.

The ATSBs marine achievements in 2004-05 included 11 marine investigation reports including on the grounding of the cruise liner Astor, and the fatal collision between the bulk carrier Asian Nova and the fishing vessel Sassenach. The ATSB also undertook an extensive education campaign within the fishing industry on commercial fishing vessel safety.

In November 2004 the ATSB established an Adelaide regional office as a base for its rail team leader and two other rail investigators. Two rail investigators are based in the ATSBs Brisbane office and one in the Canberra central office. During 2004-05, the Bureaus rail safety investigation team initiated seven investigations on the Defined Interstate Rail Network (DIRN) under the Transport Safety Investigations Act 2003 (TSI Act) and released the first TSI Act rail investigation reports on freight train derailments at Ararat, Victoria and Bates, South Australia. The ATSB also completed a rail investigation report under Victorian legislation into the derailment and subsequent collision at Chiltern between a freight train and a passenger train.

On road safety, the ATSB is coordinating the Australian Government involvement, in partnership with the NSW and Victorian governments and private sector organisations, in a large-scale trial of a best-practice education programme for novice drivers. During 2004-05, the Bureau released 32 road safety research and statistical reports, including a research report on road texture and crash risk and a survey of community attitudes. Steady progress was also made with road safety jurisdictions and stakeholders, but with substantial challenges remaining to meet or better the 2010 target of no more than 5.6 road deaths per 100,000 population.

The October 2004 International Civil Aviation Organization (ICAO) report on ICAOs May/June 2004 audit of the ATSB expressed high satisfaction with Australia's legislative, organisational and training framework for aviation safety investigation and the professional and efficient conduct of the ATSB investigations reviewed in detail. As expected, the audit team did make a number of recommendations for improvement against which the ATSB has undertaken corrective actions.

Following the 15-fatality Metroliner aircraft accident near Lockhart River, Queensland, in May 2005, the Bureau commenced its major investigation into the causes of this tragedy and released a preliminary factual report in June 2005. In November 2004 the Queensland Government asked the ATSB to chair a joint Queensland Transport investigation into the derailment of the Cairns Tilt Train north of Bundaberg, which injured a significant number of the 157 passengers and crew. The Queensland Minister for Transport released an interim report on this derailment on 16 February 2005.

The Bureau is continuing its commitment to training its investigators through the Diploma of Transport Safety Investigation course. In 2004-05, fifteen staff completed the TSI Diploma and a further 30 are progressing through the course.

During the year the Bureau signed a number of Memoranda of Understanding (MOUs) including with the Civil Aviation Safety Authority (CASA), the Australian Maritime Safety Authority (AMSA), the Victorian Rail Safety Regulator and with transport safety bodies in Indonesia and South Korea.

The ATSB continued to release all of its significant safety outputs to the public. Hits on the ATSB website www.atsb.gov.au amounted to around sixteen million.

A number of valued staff members retired during the year or prior to publication of this Review. I note in particular the contribution of Jon Henchy in Transport Safety Statistics. From early June 2005 I was off-line assisting the Rt Hon Sir John Wheeler with a review of Airport Security and Policing and Joe Motha has acted as ATSB Executive Director in addition to his normal duties with great professionalism and dedication which I wish to acknowledge with thanks.

I am grateful to the former Deputy Prime Minister and Minister for Transport and Regional Services, the Hon. John Anderson MP, the Minister for Local Government, Territories and Roads, the Hon. Jim Lloyd MP, and to the Secretaries of the Department of Transport and Regional Services, Mr Ken Matthews AO and Mr Mike Taylor for their support throughout the year. We look forward to working with the new Minister for Transport and Regional Services, the Hon. Warren Truss MP in 200506. The ATSB was again grateful for the bipartisan support it received for its safety work. The ATSBs ongoing effectiveness as the Australian Governments primary transport safety investigator remains reliant on both the perceptions and reality of its independence, fairness and professionalism.

Kym Bills

In 2004, there were two wirestrike accidents in New South Wales involving helicopters undertaking locust control operations. The first accident occurred in October 2004 near Forbes and resulted in minor injuries to one occupant and extensive damage to the helicopter. The second accident occurred in November 2004 near Dunedoo and resulted in the death of two occupants. A third occupant was seriously injured and there was extensive damage to the helicopter. A third accident, near Mudgee in November 2004, involved a helicopter that was being used for locust control, although the helicopter was not involved in locust control activities at the time of the accident.

The Australian Transport Safety Bureau (ATSB) began formal investigations into all three accidents and a research investigation into the practices used by Government organisations to contract aerial operators for locust control in order to identify issues that may enhance future aviation safety.

This report describes the results of a national survey of transport companies in Australia. The aim was to survey companies about knowledge and awareness of fatigue, about work-rest scheduling practices and about the factors which underlie the way schedules are organised. The survey was designed to provide complimentary information to that obtained in a national survey of drivers undertaken at the same time and reported elsewhere. Telephone interviews with 200 companies carrying freight over distances greater than 300km were undertaken, covering all regulated mainland states of Australia, and the Northern Territory. Companies were selected randomly from the telephone directory. The Northern Territory was included because it provided a comparison with an unregulated state. A middle management staff member, familiar with line haul operations was interviewed from each company. This report presents the main descriptive data obtained in the survey and provides an overview of views, knowledge and practices with respect to fatigue management. Key comparisons were drawn with the data obtained from the driver survey undertaken at the same time.

One of the key findings of this report was that there is a lag between increased awareness of fatigue and changes in operational practice. The majority of companies reported that awareness of fatigue had increased, both for themselves and their company, as well as for the industry at large over the last 5 years. However, from the results it seems that this increased awareness does not guarantee better management of the problem. Only half of the companies surveyed reported that they believed that fatigue was well managed in the industry and one fifth reported that it is badly managed. Even so, this is more optimistic compared with the verdict of drivers, half of whom reported that fatigue is badly managed in the industry.

Further evidence of the lag between increased awareness about fatigue in general and companies actually coming to grips with better management of the problem came from views of causes of and strategies to manage fatigue. Virtually all companies endorsed the significance of sleep and recovery before and during trips, and the contribution of long hours. However, other key contributors to fatigue were grossly underestimated. Company representatives failed to report the significance of night work as a prime contributor to fatigue and consolidated night sleep as prime strategy for reducing fatigue. Similarly, there was lack of recognition by companies of the substantial contribution of non-driving work, particularly loading and unloading, to the overall burden on drivers, and accordingly lack of endorsement of limits for such work as a fatigue management strategy. This picture is in sharp contrast to that presented by drivers, where awareness of the key contributors and likely effective strategies was much more in line with current knowledge.

It is hardly surprising that fatigue has become a more prominent feature of companies risk management agenda. There have been a number of high-profile initiatives in safety promotions and legislative directions over the last decade all aiming to focus industry attention on better management of driver fatigue. The results of the survey highlight that increased awareness does not immediately translate into increased knowledge and operational changes.

This systemic inertia was also evident in the persistence of industry perceptions that the freight task needs to be maximally responsive to the demands of customers and freight forwarders, often described as the chain of responsibility. In fact, the picture presented by the companies themselves was rather different. The majority of companies reported that they have considerable control over schedules, with only a minority reporting that their work was mostly irregular. Strict estimated times of arrival were uncommon, and trip times were mostly based on company and driver estimates, according to the companies surveyed. In other words, companies appear to have potential for far greater control over their schedules than is recognised or exercised.

The study provided some evidence that better attitudes to fatigue were associated with company practices that were more likely to manage fatigue effectively. For example, more aware companies were more likely to monitor fatigue and were more likely to change their schedules to accommodate driver fatigue. In contrast, companies who relied on the industry in general for management of fatigue and/or in the working hours regulations were less likely to be paying attention to the problem, were less likely to monitor fatigue and were more likely to change schedules to suit customer demands rather than for driver fatigue. They also used fewer management strategies and were less likely to otherwise restrict hours. These findings suggest that while attitudes do not seem to have a dramatic effect on practice, education and information for companies is a useful strategy for actively involving companies in better management of fatigue and for overcoming complacency about the driver fatigue problem.

Fatigue management strategies reported by companies surveyed focused on limitations of daily and weekly hours of service. Not surprisingly, there was less intervention and active management of fatigue for non-employee drivers. Active fatigue management strategies, monitoring of fatigue, or even formal policies for fatigue management for sub-contractor and independent drivers were reported by only a small minority of companies. Yet half of the companies surveyed reported that they hire these types of drivers. In many cases fatigue management for non-employee drivers is likely to become, by default, the responsibility of the individual driver. This is a serious problem because effective fatigue management is unlikely to emerge without not only company co-operation, but also active and formal company collaboration.

Surprisingly few differences were evident between companies of different sizes. Obvious and predictable structural differences were reported, for example greater reliance of smaller companies on non-employee drivers. Also predictably, formal policies and technical monitoring approaches were less common, reflecting the resource intensive nature of these strategies. However, little impact was seen of company size on the attitudes to fatigue and scheduling practices reported by companies surveyed. This pattern of findings suggests that the translation of fatigue awareness into operational practices is universally slow and is not just a feature of some segments of the industry having reduced access to information and so forth.

Overall, this survey suggests that there is considerable scope for improving understanding and management of fatigue in the industry. Companies do not seem to be doing all that could be done to improve management of fatigue. Partly, this seems to reflect a lack of understanding about the phenomenon. There was poor understanding among line haul managers of how driver fatigue develops, the key role played by time of day and the contribution of total burden of work, not just driving. There needs to be greater

understanding in the industry that the problem requires a more sophisticated approach than simply restricting hours of driving. Education and information for companies is likely to be a useful strategy to alert companies to the most appropriate practices and to overcome complacency about the problem. The survey revealed that approaches affecting global attitudes, general increases in awareness and so forth, have had little impact on practices. On the other hand, they are likely to have been important for raising the profile of the problem in the industry and laying the groundwork for more targeted information and education. Indeed, it is hard to imagine how transport operators could develop the most effective interventions for their particular freight task, as demanded by Fatigue Management Programs, without being better informed. From the results of this survey, improved understanding of fatigue and its characteristics among transport managers must be seen as an immediate priority.