Aircraft operator

As a result of this occurrence, the aircraft operator has advised that they have conducted a fleet wide inspection of all landing gear actuator locking devices to ensure they conform to the actuator manufacturer's specifications.

Aircraft Manufacturer

The aircraft manufacturer has advised that it intends to issue a 'mandatory' Service Bulletin, SB F406-56, which will instruct that only the correct landing gear actuator locking devices are to be fitted. The Service Bulletin will also require that strict compliance with the actuator manufacturer's requirements. In addition, the Nose Landing Gear maintenance requirements will be amended in the aircraft maintenance manual to emphasise the requirements of the actuator and control indication check.

Direction Generale de l'aviation Civile of France

The French Direction Generale de l'aviation Civile (DGAC) have advised the aircraft manufacturer that they intend to mandate the requirements of the aircraft manufacturer's Service Bulletin through the issue of an Airworthiness Directive.

Local safety action

Airservices Australia safety action

In the January issue of one of its newsletter to controllers, Airservices reminded controllers of the factors that contribute to runway incursions. Those factors included inadequate supervision of the manoeuvring areas of the airport, lack of adequate coordination between controllers in the tower, ambiguous clearances and instructions issued by controllers, incorrect read-back of clearances and instructions by pilots and vehicle drivers, and controllers not detecting the errors in those read-backs.

Airservices is investigating the International Civil Aviation Organization's concept of the Advanced Surface Movement Guidance and Control System for Sydney and other Australian airports. The concept includes consideration of improved surface movement radar, improved Aerodrome Mandatory Instruction Signs and Aerodrome Information Signs and the installation of stop-bar lighting. Stop bar lighting consists of a row of red unidirectional, in-pavement lights installed on the taxiway along the holding position marking the entrance to a runway. These initiatives could improve conflict detection for controllers, reduce the incidence of runway incursions and reduce the risk of a collision as a result of a runway incursion.

Overall, the system has the potential to improve the level of safety for operators at airports and improve controller situational awareness.

Sydney Airport Corporation Limited safety action

Since this occurrence, SACL has:

• formulated an updated Letter of Agreement with Airservices on the exchange of safety information;
• nominated single points of contact between SACL and Airservices to act as representatives for the distribution of safety related information;
• established a Runway Incursion Working Group with participation from Airservices, Qantas, Regional Express, Virgin Blue, Jetstar and Eastern Australia Airlines (involvement from other parties will be sought as required);
• audited all AIA's in accordance with the AVCH (October 2003, s6.4.1, p.46) and acted to ensure any recommendations made, as a result of those audits, were implemented;
• developed and issued updated category 2 testing examinations to all ADA issuing authorities which include guidelines for use by Approved Issuing Authorities when conducting the written tests;
• re-issued the category 2 airside driver's pocketbooks with advice to drivers to stop and wait for assistance if they become lost or disorientated while driving airside; and
• established an Airside Driving Forum co-facilitated with the NSW WorkCover that includes various airside users.

ATSB safety action

As a result of the investigation the Australian Transport Safety Bureau issues the following recommendations:

Recommendation R20040059

The Australian Transport Safety Bureau recommends that Sydney Airport Corporation Limited review the procedures used to ensure initial and ongoing driver competency and knowledge.

Recommendation R20040060

The Australian Transport Safety Bureau recommends Sydney Airport Corporation Limited ensures that Approved Issuing Authorities' driver training programs at Sydney Airport include a course of action that drivers can take should they find themselves lost or disorientated while driving airside.

¹ Civil Aviation Regulation 89 defined the Airport operator as 'in relation to a licensed aerodrome-the licence holder'.
² In accordance with the Airports Act 1996, s172, Airports (Control of On-Airports Activities) Regulations 1997.
³ Airside Vehicle Control Handbook, June 2003, p1.
⁴ Airside Vehicle Control Handbook, Sydney Airport, June 2003, p. 42.
⁵ Airside Vehicle Control Handbook, Sydney Airport, June 2003, p. 44.
⁶ A 'Runway Strip' is 'a defined area, including the runway (and stopway if provided), intended both to reduce the risk of damage to aircraft running off a runway and to protect aircraft flying over it during take-off, or landing operations. (ICAO) (Manual of Air Traffic Services effective 15 April 2004, Pt. 10, s.1, p. 10-18).
⁷ Drivers Pocketbook, Category 2, May 2002, p3.

On the morning of 15 August 2003, the pilot hired a Victa Ltd Airtourer, registered VH-MVP, to practice basic aerobatics. Prior to flying to the training area, the pilot was conducting a number of touch and go circuits. Following the fourth take-off and while climbing through about 200 ft, witnesses described hearing the aircraft engine surge then stop. Shortly after, witnesses saw the aircraft turn left. The turn steepened as witnesses lost sight of the aircraft behind trees and houses and a short time later they heard the aircraft impact the ground. The pilot was fatally injured in the accident.

The pilot was appropriately licensed and was reported by a family member to be fit and well and looking forward to the flight. The family member believed that the pilot was going to conduct circuits in another aircraft, not aerobatics in the Victa. However, the company flight details log sheet contained an entry for `aerobatics' and the pilot had discussed the aerobatic component of the flight with an instructor.

The Australian built, two seat, aerobatic rated, single-engine, low-wing aircraft was originally fitted with a 100 hp engine driving a fixed pitch propeller. However, the aircraft was later fitted with a 180 hp engine with a constant speed propeller.

Examination of the aircraft fuselage, flight controls, fuel system and engine including magnetos and spark plugs, by the Australian Transport Safety Bureau (ATSB) investigation team, provided no explanation for the sudden loss of power observed by witnesses. During the onsite phase, approximately 6.8L of fuel was removed from the aircraft fuel tank, which was still securely attached and not deformed. This included approximately 1L that drained from the tank as it was removed from the wreckage. The fuel was free of visible contaminants and was the correct colour. No other fuel was recovered from the site, nor was fuel staining or odour evident on the fuselage or the ground immediately below the wreckage. There was no evidence of excessive fuel consumption. Emergency workers and witnesses, who arrived at the accident site within minutes of the accident, could not recall a fuel smell. The ATSB did not receive any fuel related incident reports from other aircraft that had refuelled from the same source following the accident.

The fuel system consisted of a 130L (total capacity) rubber bladder fuel tank located in the fuselage. Fuel addition and manual contents checking was via an angled filler tube into the bladder tank through the side of the fuselage. Fuel was supplied to the engine via an engine-driven fuel pump and carburettor, with an electric boost pump as backup. Fuel quantity is checked via an electric fuel gauge and a flexible dipstick graduated in imperial gallons. The dipstick is made up of a number of five imperial gallon graduated segments, held taut by a chord under tension. Dipping of the fuel contents required a pilot to depress the button on the top of the dipstick to relax the tension on the segments and allow the flexible dipstick to travel down the angled fuel filler tube. After traversing the angled section of the filler tube, pressure on the button was required to be relaxed, thereby re-tensioning the segments prior to the end of the dipstick contacting the bottom of the tank. The procedure required the dipstick to be under tension prior to contacting the bottom of the tank and that the dipstick did not rest on a fold or ripple in the bladder. Failure to do this may result in an erroneous reading of the tank's content.

The approved aircraft flight manual stated that 1.3L of fuel was unusable. Although not stated, this is assumed to be for level flight. Calculations by the investigation indicated that about 6.2L of fuel would be unusable while the aircraft had a 5 degree nose up attitude in a climb. This would increase to about 12.4L at 10 degrees of nose up attitude.

Examination of the tank sender unit, a wire-wound wiper type, showed wiper shaft bearing surface wear. This would have allowed lateral movement of the wiper arm resulting in intermittent contact with the wire wound former and is likely to have caused the fuel gauge to display intermittent readings to the pilot. However, due to impact damage of the sender unit, it was impossible to determine the extent of the intermittent readings.

Three small pin size holes were located in the rubber bladder. A test indicated that 250 ml could have leaked from the tank from the time of the last refuelling to the time of the accident. The holes were not collocated and were not in the vicinity of the dipstick. It could not be determined if the holes were due to impact damage.

The tank had been filled on the day before the accident, prior to completing a 2.2 hour cross country flight. On the day of the accident, the pilot checked the tank contents and informed an instructor who assisted her to push the aircraft from the hangar that it held 15 imperial gallons (68L). However, presumably in error, the pilot entered 75L (16.5 imperial gallons) in the company flight details log sheet. The instructor did not see the pilot dip the tank or check the fuel gauge.

The investigation examined the aircraft engine manufacturer's fuel consumption tables and company flight details log sheet to establish the fuel consumption rate. Based on the record of total hours and fuel consumed, for the previous two weeks, the aircraft had consumed about 37.6L per hour for all modes of flying. Therefore, the tank should have held about 47.3L after the previous day's cross-country flight. If the tank contained 75L prior to the accident flight, as entered in the log by the pilot, the aircraft consumption rate would have been about 25L per hour. That fuel consumption rate was unlikely to be achieved during the cross-country flight. The chief flying instructor commented that the company instructed students to plan using a fuel consumption rate of 40L per hour.

An instructor reported that the aircraft had completed three circuits prior to the accident, which would equate to roughly 35 minutes of taxi and flying time. Allowing for an average fuel burn of 37.6L per hour, there should have been about 25L remaining at the time of the accident. The ATSB fuel consumption calculations for some individual flights ranged from 35L to 50L for circuits, aerobatics and cross-country flights. However, the log was incomplete, so it was not possible to derive a fuel consumption per flight for some of the previous flights. The reason for the discrepancy between what should have been in the tank and what was recovered at the accident site could not be determined.

The investigation examined a similar model aircraft and its fuel system. It was found that it is relatively easy to have the bottom (five imperial gallon) segment of the dipstick bend sideways, when the dipstick contacts the bottom of the tank. This will occur if the segments are not tensioned by releasing pressure on the tension button prior to the dipstick contacting the bottom of the tank. This results in the dipstick over-reading by about 4 to 5 imperial gallons (18.2L to 22.7L) and could have led the pilot to believe that there was adequate fuel for the flight.

Based on the examination of the recovered engine components, the witness reports of engine surging just prior to the engine failure, the lack of fuel odour at the accident site and the company fuel details log sheet, it is probable that there was insufficient fuel to complete the flight, either due to fuel exhaustion or starvation. It is possible that the pilot's operation of the dipstick provided an erroneous reading, which led her to believe that the aircraft's tank contained more fuel than it actually did. Additionally, the intermittent fuel gauge reading may have meant that she was not able to check the fuel quantity by a secondary means, other than the log entry for the previous day's flying.

In a take-off climb attitude of between 5 to 10 degrees, the fuel pickup point in the tank was probably unported, interrupting the fuel flow to the engine thus causing it to lose power. The flight path of the aircraft after the loss of engine power indicates that the pilot may have been attempting to turn the aircraft back to the runway at a low height and lost control with insufficient height to effect recovery.

On 21 August 2003, during approach to Bankstown aerodrome, the left engine of a Piper PA31-350 (Chieftain), registered VH-OCF, malfunctioned. The pilot first became aware of the malfunction through a series of 'explosions' followed by severe vibration and the observation of flames. The pilot shut down the engine immediately and feathered the propeller, at which stage the fire extinguished.

Subsequent examination of the engine revealed that the number-6 connecting rod little end had fractured and separated from the piston. The separated end of the connecting rod had collided with the underside of the number-6 piston, driving it into the cylinder head and fracturing the cylinder attachment fasteners. The force of the collision and the flailing of the connecting rod fractured the camshaft and extensively fractured the crankcase. Both upper engine mounts separated from the crankcase during that sequence.

Number 6 connecting rod little end housing fracture

Examination of the fracture surfaces in the number 6 connecting rod little end housing revealed that fracture initiated from a region of fatigue cracking. Fatigue cracking initiated on the inner surface of the housing. Examination of both the housing inner surface and the piston pin revealed that galling had occurred between the pin and the housing inner surface. Galling is a term used to describe surface damage created by adhesive wear. In this process sliding contact between two surfaces results in localised welding, fracture of localised welds and transfer of material from one surface to the other.

The surface damage created by galling lowers the fatigue resistance of a component, making the initiation of fatigue cracks more likely under normal loading. For galling to occur between the piston pin and the inner surface of the housing, the bronze bush normally fitted to the housing must not be present. It is evident in this case that the bronze bush had been destroyed during engine operation.

An examination of other connecting rods from the engine revealed that the bushes were in various states of destruction. Subsequent examination of a connecting rod from the right engine of the aircraft (following engine overhaul at a time after the occurrence involving the left engine) revealed that the little end bush was being destroyed progressively.

The issue of little end bush destruction is being addressed in detail in a comprehensive ATSB technical investigation (BO/200305443), titled Aircraft Reciprocating Engine Structural Failure - An Analysis of Failure in a Complex System.

Local safety action

Aircraft operator

As a result of the contractor's requirements, the operator reported that future flights for the contractor would be operated by two pilots.

Australian Transport Safety Bureau

Fuel exhaustion and starvation accidents accounted for over 6 per cent of all accidents between 1991 and 2000 and the rate remains relatively constant.

In December 2002 the ATSB published a research paper titled `Australian Aviation Accidents Involving Fuel Exhaustion and Starvation'. It is available on the ATSB's website www.atsb.gov.au, or from the Bureau on request.

On 2 July 2003, the Boeing 747-438 aircraft, registered VH-OJU, operating on a scheduled flight from Singapore, arrived at Sydney at 0511 Eastern Standard Time, during the airport's curfew period. There was a tailwind of around 12 knots when the aircraft landed. The pilot flying selected auto brake setting three and idle reverse thrust in accordance with the curfew requirement. However, during the landing roll the reverse thrust was inadvertently de-selected.

On arrival at the terminal, the pilot in command (PIC) observed a BRAKE TEMP advisory message and notified the ground engineers. At that point, a fire ignited on the right wing landing gear. The flight crew were advised and the PIC ordered an evacuation of the aircraft. On receiving the evacuation announcement, the cabin crew commenced the evacuation drill deploying the aircraft's escape slides. The upper deck left (UDL) door and doors 2 left (L2) and 4 right (R4) escape slides, did not deploy. During the evacuation, the over-wing slide at door right 3 (R3) deflated while in use. As a result of the evacuation, one flight crew member and three passengers were seriously injured. Some passengers evacuated down the slides with their cabin baggage.

During the accident, an additional two brake fires ignited on the right body landing gear, one of which was extinguished by the Aerodrome Rescue and Fire Fighting Service (ARFFS). A subsequent inspection found that the aircraft's landing gear contained an excessive amount of grease with the presence of inappropriate grease on all of the landing gear axles. The three brake units that had caught fire were found to be serviceable but in a worn condition.

The investigation determined that slide R3 did not have any pre-existing defects that contributed to its failure. The nature of the failure was found to be overload of the fabric fibres during the evacuation. The inappropriate grease found on the landing gear axles was general purpose grease used on other components of the landing gear. The time and point of its application to the aircraft axles could not be determined.

The investigation found deficiencies in the operator's maintenance, flight crew and cabin crew procedures. As a result, the operator has issued maintenance memos to its engineering staff clarifying aircraft landing gear lubrication procedures, amended its Aircrew Emergency Procedures Manual, and reviewed cabin crew and flight crew emergency procedures.

As a result of this investigation, the ATSB is issuing safety recommendations to the operator and the Civil Aviation Safety Authority concerning the use of over-wing slides during known brake fires.

Related Documents: | Media Release |

(Cessna 404, VH-ANV Jandakot WA, 11 August 2003)

Introduction

At 0735 UTC on 11 August 2003, VH-ANV was cleared on a MANTL 1 departure from runway 24R at Jandakot airport. Onboard were the pilot and five passengers. The aircraft called ready and was cleared to climb to 3,000 feet. The aircraft rotated and the tower staff noticed a sound similar to an asymmetric operation. The aircraft was turned left and subsequently impacted the ground to the southeast of the tower near the NDB site. This Technical Analysis Investigation report should be read in conjunction with ATSB report BO/200303579.

Sequence of events

On 22 June 2003, a Cessna Aircraft Company 172M, registered VH-TUR, drifted to the right shortly after take-off from runway 35 at Wedderburn airfield in NSW and impacted the ground to the north-east of the airfield. The aircraft was destroyed, and the four occupants were fatally injured.

The pilot held a valid Private Pilot Licence (aeroplane) and current class 2 medical certificate. There was no evidence that any physiological or psychological factors had affected the pilot's performance.

A witness at the airfield videoed the aircraft as it took off. Examination of the video revealed that the aircraft became airborne after a take-off roll of about 500 m, with 10 degrees of wing flap extended. As it climbed, the aircraft drifted to the right and entered a right-wing-low sideslip with a nose-up attitude. Witnesses at the airfield observed the aircraft between gaps in the trees to the north-east of the airfield banked to the right in a steep descent and then heard the sound of an impact.

Runway 35 was 1,000 m long and sloped down slightly to the north. The first half of the runway was bitumen and the second half was a mix of hard clay/gravel. Trees about 10 m high surrounded the runway, and sheltered it during crosswind conditions. At the time of the accident, the wind was gusting from the south-west.

The investigation found that the aircraft had been descending steeply in a right turn when it impacted the ground in a westerly direction. Data recovered from a global positioning system found in the wreckage supported other evidence, which indicated that the aircraft entered a spin during a right turn after take-off.

Examination of the wreckage revealed no evidence of pre-existing mechanical defects that may have contributed to the accident. The wing flaps were in the 10 degrees extended position at the time of the accident. The aircraft stall warning system was recovered from the wreckage, tested and found serviceable. The aircraft had sufficient fuel onboard for the planned flight.

An assessment of the aircraft weight indicated that it was approximately 30 kg above maximum allowable take-off weight, and the centre of gravity was calculated to have been towards the aft limit of its normal centre of gravity range. That extra weight would have increased the aircraft stall speed by 1.4% (less than 1 kt), and reduced its climb performance slightly.

Wedderburn was an uncontrolled airfield. It was normal procedure at uncontrolled airfields to maintain runway heading after take-off until the aircraft had reached a height of 500 ft above the airfield, and then to turn left.

As the aircraft climbed out of the shielding effect of the trees beside the runway, the crosswind from the left would have increased markedly. If the aircraft was continuing to climb as it turned to the right, it would have then been in an increasing tailwind. An increasing tailwind will lead to a momentary reduction in aircraft indicated airspeed. The wind was also strong and gusting, and if there had been a wind gust at the same time, it would have caused a greater momentary reduction in aircraft indicated airspeed.

When an aircraft turns away from the wind, at low level, the groundspeed increases. In such circumstances, the view of the ground accelerating below an aircraft may give an illusion of an increased airspeed. For any given nose attitude, an aircraft will fly slower if wing flaps are extended. With 10 degrees of wing flap extended, the aircraft was therefore flying at a lower airspeed in the climb than if the wing flaps had been retracted, if the same attitude was maintained. The aircraft's climb performance would also have been reduced by using 10 degrees of wing flap, compared with using no wing flap. The aircraft take-off performance data in the aircraft operating handbook indicated take-off performance with wing flap retracted, and provided information that 10 degrees of wing flap should be used for take off from a soft surface.

The amount of aerodynamic lift produced by a wing in flight can be changed by a pilot in a number of ways. If the speed of the air flowing over the wing is increased, aerodynamic lift normally increases. The shape of the wing can be changed to an extent by moving the control surfaces, which adjusts the amount of aerodynamic lift. The angle at which the airflow impinges on the wing can be adjusted, which will also change the aerodynamic lift. This angle is known as the angle of attack. In normal flight, if the angle of attack is increased, the aerodynamic lift is also increased, up to a certain angle of attack known as the stalling angle. In contrast, if the angle of attack is increased beyond that stalling angle, the amount of aerodynamic lift decreases. An aircraft flown at a greater angle of attack than the stall angle is commonly described as being aerodynamically 'stalled'.

In steady flight there is a relationship between speed and the angle of attack of the wing. The stall speed is the speed at which the angle of attack coincides with the stall angle for a given configuration.

One effect of flap extension is to increase the relative pitch angle (incidence) between the wing and the fuselage. As a result, in steady flight, an aircraft with flaps extended will fly at a lower speed than one at the same attitude with flaps retracted. If the aircraft was flown at the normal flapless climb attitude, but with 10 degrees of flap extended, it would fly at a lower airspeed. The additional drag produced by the flap extension would also have reduced the climb performance. The aircraft would also have stalled at a lower nose attitude when 10 degrees of wing flap was extended, compared with when the wing flap was retracted.

An aircraft's stall speed increases by a factor of the square root of the secant of the angle of bank, all other things being equal. The aircraft was in a gentle turn at the time the stall occurred. The wing is generally less effective at producing aerodynamic lift when an aircraft is flown out of balance. There is also an increased likelihood of one wing stalling before the other leading to a roll input at the onset of the stall. The accident aircraft had been flown out of balance shortly before the onset of the stall, but it was not known if it was out of balance at the onset of the stall.

The aircraft was observed flying slowly during its climb after takeoff. If the aircraft's airspeed became sufficiently slow in a steady climb, the aircraft would stall. The circumstances were consistent with the aircraft entering a stall and a spin at a height from which it was considered impossible to recover. Some or all of the following factors could have contributed to the aircraft entering a stall:

• The aircraft exceeded the maximum allowable take off weight, which would reduce its climb performance
• The aircraft was climbing into an increasing tailwind, which would create a momentary reduction in airspeed
• The wind was gusting, which could have created a further momentary reduction in airspeed
• The takeoff was downwind, which would have led to a higher groundspeed that would give an illusion of higher airspeed. The pilot may have compensated for this illusion by raising the aircraft's nose
• The aircraft was turned away from the wind at low level, which could have led to an illusion of increasing airspeed. The pilot may have compensated for this illusion by raising the aircraft's nose
• The aircraft was flown out of balance for parts of the flight, which would have reduced its performance
• The use of 10 degrees of wing flap would have reduced its climb performance, and meant that the aircraft would have been flying slower for any given nose attitude.

On 27 May 2003, at about 1650 Western Standard Time, the pilot of a Cessna 172P (C172) aircraft, registered VH-AUC, was conducting circuits on runway 06 right (06R) at Jandakot. An instructor and student pilot of a Piper PA-38-112 (Tomahawk) aircraft, registered VH-FIG, were also conducting circuits on runway 06R.

While on downwind for runway 06R, the pilot of the C172 requested a landing on runway 06 left (06L). The aerodrome controller responsible for runway 06R (ADC1) acknowledged that request and instructed the pilot of the C172 to follow the Tomahawk, which was also on downwind for runway 06R. After coordinating with the aerodrome controller responsible for runway 06L (ADC2), the ADC1 cleared the pilot of the C172 to make an approach to runway 06L and instructed the pilot to transfer to the ADC2 frequency. The C172 subsequently passed in close proximity to the Tomahawk while the Tomahawk was on final for runway 06R and the C172 was on right base leg for runway 06L.

Radar data indicated that the crew of the Tomahawk had extended downwind for sequencing with a preceding aircraft and did not turn base for runway 06R until close to the control zone boundary. Radar data also indicated that the pilot of the C172 had turned right base for runway 06L from a late downwind position and had flown an oblique base leg to join final for runway 06L. Sun glare may have contributed to the C172 pilot losing sight of the Tomahawk ahead after it had turned onto the base leg.

The Tomahawk was at about 500 ft above ground level and descending on long final approach to runway 06R when the instructor observed the C172 tracking towards them. The instructor in the Tomahawk attempted to contact the pilot of the C172, but used the callsign of another aircraft believed to be operating in the circuit at the time and received no response. Regardless, the pilot of the C172 would not have heard any transmissions from the instructor, as the pilot was operating on a different frequency, as instructed by ADC1.

The Tomahawk instructor increased the rate of descent of the aircraft and monitored the position of the C172 before it passed directly overhead, left to right, about 50 ft above their aircraft. The instructor reported that the C172 was sighted approximately 20 seconds before it passed overhead. The C172 pilot was unaware of the incident until after landing when it was brought to his attention by the surface movement controller. Due to the distance from the control tower and the angle of observation, the aerodrome controllers could not accurately judge the relative positions between the aircraft on base and final approach.

Following this incident, Airservices Australia issued instructions to Jandakot aerodrome controllers to delay, where practicable, the transfer of aircraft onto another frequency when facilitating a change in landing runway.

On 20 June 2003 at approximately 0840, a Robinson Helicopter Company Model R22 helicopter, registered VH-OHA (OHA), was being used to conduct flying training in the Bankstown training area with an experienced flight instructor and student pilot. The helicopter was observed and heard flying in a normal manner. Witnesses reported subsequently hearing a number of loud bangs and one witness observed what appeared to be a main rotor blade separating from the helicopter. The helicopter descended to the ground in an inverted attitude and both occupants were fatally injured.

Examination of the accident site and helicopter wreckage confirmed that one main rotor blade had failed in-flight. Examination of the helicopter and its systems did not reveal any other abnormality that would have contributed to the loss of the main rotor blade.

The helicopter had recently re-entered service following maintenance which included the fitting of an overhauled engine and the completion of a 100-hourly inspection. The helicopter also underwent maintenance action to rectify a main rotor blade vibration. This maintenance action involved a number of experienced R22 helicopter engineers being consulted about the possible reasons for the main rotor blade vibration. Rectification action was completed in accordance with normal maintenance practices and the manufacturer’s maintenance manual. Subsequent examination of the maintenance manual for the R22 helicopter revealed that it did not contain any information in the tracking and balancing section that indicated that a vibration may be the result of a crack in the main rotor blade. The manufacturer had produced other documentation containing this information, but these documents did not formally form part of the maintenance manual.

The helicopter had been manufactured in 1991 and had been imported into Australia in 1996. In the time prior to the accident it had been owned and operated by a number of organisations and individuals, and was operated both commercially and privately.

Following the accident, industry suggestions about the possible under-recording of time in service on the helicopter led the Australian Transport Safety Bureau to concentrate part of the investigation to the recording of time in service of the helicopter. Coincident with this investigation, a separate investigation of the recording of time in service on the helicopter was conducted by the Australian Civil Aviation Safety Authority (CASA).

Both investigations examined a wide range of documentation and records from numerous sources. The conclusion of both investigations was that the helicopter had not exceeded the mandatory time in service life of 2,200 hours, nor had it exceeded the mandatory calendar time in service life of 12 years. The final time in service of the helicopter was calculated to be 2,055.6 hours and the calendar time in service was 11 years and 8 months.

An examination of the main rotor blade in the ATSB laboratories revealed that it had failed as a result of fatigue crack growth in the blade root fitting at rotor station 10.35. The fatigue crack initiated as a result of localised pitting corrosion in the counterbore of the inboard bolthole. The examination also revealed that while the fatigue failure was in a similar position to two previous main rotor blade failure accidents in Australia, in OHA’s case, there was an area of adhesive disbonding between the main rotor blade skin and blade root fitting. This adhesive disbonding meant that the crack in the blade root fitting did not propagate into the blade skins and so was undetectable using visual means. The two previous failures were linked to under-recording of hours.

The material failure analysis found that the disbonding present on the failed main rotor blade was also present in a number of other main rotor blades that were examined. As a result, the ATSB issued a safety recommendation to the United States Federal Aviation Administration (FAA) and to the Robinson Helicopter Company, seeking that they conduct further testing on main rotor blade root fittings to evaluate the extent of adhesive disbonding in the blade root fitting. This examination was conducted on a total of 51 main rotor blades that had between zero and 2,200 hours time in service. Results of the examination revealed that adhesive disbonding between the spar and root fitting was present in all blades and that the extent of the disbonding was variable.

Subsequent to this accident there was another in-flight failure of a main rotor blade. In February 2004, an R22 helicopter being operated in Israel sustained an in-flight failure of a main rotor blade. This blade had failed as a result of fatigue in the same location as the failure in the Australian accident. The Israeli failure exhibited a similar loss of adhesion and corrosion. Both blades had failed before their mandatory time in service retirement lives and represented a failure of the fatigue fracture control plan. A third failure occurred in New Zealand in November 2004. Preliminary investigations have revealed that the failure may be the result of loadings on the blade that may have exceeded those intended by the manufacturer. The investigation of that accident is continuing.

The manufacturer has issued a safety letter and a service bulletin relating to revised retirement lives for main rotor blades, and has introduced a redesigned main rotor blade into service. The manufacturer indicated that it intends to publish safety alerts and notices on its Internet website as an additional means of bringing safety related information to the notice of owners, operators and maintenance organisations.

The R22 maintenance manual has also been amended by the manufacturer as a result of this investigation. The main rotor blade tracking and balancing section now contains information, which alerts maintenance personnel to the fact that a main rotor blade vibration may be the result of a developing crack.

Safety action taken by the CASA as a result of this accident was to amend an existing airworthiness directive to take into account the findings from the examination of the blade and to introduce additional amendments to the directive, when updated information became available from the manufacturer. They also introduced a discussion paper on the installation of mandatory time in service recorders for helicopters. As at October 2005, CASA was still evaluating the public comments on the discussion paper.

In addition, CASA has drafted a Notice of Proposed Rulemaking (NPRM 0503CS) in which it is proposed to require the retirement of similar main rotor blades by 1 March 2006 on Australian registered Robinson R22 helicopters.

The United States FAA issued a special airworthiness information bulletin and an emergency airworthiness directive.

As a result of several accidents involving main rotor blade failures, the European Aviation Safety Agency, issued an airworthiness directive on 5 July 2005 mandating compliance with the Robinson service bulletin.

The ATSB has contracted research to assess the validity of the usage spectrum assumptions that were used for certification of the Robinson R22 helicopter. A research investigation report on the project is planned to be released in 2006.