Executive summary

At about 1345 on Monday 11 October 2004 freight train DS212, hauling 54 fully loaded coal wagons from the Bloomfield Colliery balloon loop, derailed approximately 300 metres north of Thornton station platform. The majority of the train passed through the crossing before seven wagons derailed over 103 and 104 points at the main rail corridor junction. Three of the derailed wagons tipped over fouling the Up¹ and Down² Coal Roads as well as the Down Main Line.

Thornton is located on the main northern railway corridor between Sydney and Brisbane, 182.2 kilometres from Central Railway Station, Sydney. Thornton is on the railway section³ between Broadmeadow and Maitland, prescribed as part of the Defined Interstate Rail Network (DIRN). The railway corridor contains four standard gauge lines, an Up and Down Coal Road, and an Up and Down Main Line.

The train crew tried contacting Broadmeadow Train Control Centre but were referred to Maitland signal box. The signaller at Maitland noticed the transit lights for the points flashing, indicating that the ‘points detection’ had been lost. The signaller immediately placed all signals in that location at stop and applied blocking facilities. An approaching passenger train 747 on the Down Main Line at Beresfield was stopped by signals before the section of fouled track. The track circuiting⁴ over the Down Main Line was not affected by the derailment.

To assist the investigation an independent railway engineering organisation, Interfleet Technology Pty Ltd, was engaged to inspect the track and train to determine what factors contributed to the derailment.

Other freight train traffic had traversed the line earlier in the day without incident. The first half of train DS212 passed safely over the site before wagon NHRH 50245C derailed. This indicated that although the track-based elements were significant causal factors, it was a combination of the vehicle-based and track-based factors that led to the derailment.

The investigation concluded that the derailment occurred as a result of gauge widening at about 182.527 km due to poor rail fastener condition, in between 104A points and 104B catch points which allowed the right wheel, in direction of travel, of the third axle on the trailing bogie of wagon NHRH 50245C to drop into the four foot⁵. The adjacent right wheel on axle number three rolled the high rail⁶ outwards.

Given the lack of awareness of the poor and deteriorating condition of the high rail stability, a derailment at this location was inevitable without remedial action.

A number of safety actions have been taken, or are underway, through the New South Wales Independent Transport Safety and Reliability Regulator, Pacific National, and the Australian Rail Track Corporation. These safety actions included: the inspection of all branch line turnouts/junctions used by Hunter Valley coal traffic; the implementation of a plan to remediate infrastructure identified as deficient and/or non-compliant; and the inspection of rollingstock for asymmetric wheel profiles. Of the eleven branch line turnout/junction sites inspected, one site was found to be deficient and/or non-compliant. Pacific National found one bogie with significant asymmetric wheel wear and four bogies in the early stages of asymmetric wheel wear, all were at the end of effective service life.

The report makes a number of recommendations on pages 33 and 34 relating to: track monitoring and maintenance; reviewing the effects of asymmetric wheel wear on lateral track forces and rail safety; and reinforcement of network rules and procedures regarding the protection of trains.

__________________________________

1. An Up line or train is a line or train heading towards Sydney.
2. A Down line or train is a line or train heading away from Sydney.
3. See appendix 6.1 for an area map.
4. An electric circuit that uses the rails of a railway track as conductors such that a train electrically connects them via its axles. The absence or presence of this rail-to-rail connection indicates the absence or presence of a train or item of rollingstock.
5. A colloquial term used to describe the gap between the rails (laid at standard gauge of 1435mm).
6. The high rail is the outer rail of railway track in a curve.

Executive summary

At 0738:06¹ on 20 September 2004 train 8868 passed signal FS66 when it was displaying a stop aspect. Train 8868 was a freight train from central Queensland bound for Fisherman Islands and was crewed by a driver who had signed on at Maryborough at 0050 the same day. Signal FS66 is about five kilometres from Fisherman Islands and about 100 metres from a busy road crossing at Pritchard Road.

Train 8868 reached Lytton Junction at about 0736 and was routed onto the Fisherman Islands branch line for the final section of the journey. The driver recollects passing through this junction and setting one of the train radios to the Fisherman Islands local control channel. He thinks he then fell asleep, as he remembers little until sensing that the train was travelling too slowly. The driver then applied full power until about 15 metres from signal FS66, by which time the train was travelling at 49 km/h. Being momentarily unaware of where he was, applying full power, noticing the cars on the level crossing before realising (when 15 metres away) that the signal was red, indicate only a partial state of arousal. A service rate reduction² of the brake pipe failed to stop the train from passing the signal and proceeding through the Pritchard Road crossing. Because of the rate of reduction and because the brake pipe pressure reduced to 229 kPa, well below the 350 kPa equalised pressure of a full service application, it is concluded that the brake handle was placed in the ‘handle out’ position and not in the emergency position. The ‘handle out’ position is the notch immediately before the emergency position.

As train 8868 passed signal FS66 the protection cycle for the level crossing was only partially complete and the boom barriers were not horizontal. Train 8868 stopped about 175 metres beyond signal FS66 and 74 metres beyond the level crossing.

The emergency response to passing FS66 on a red signal was initiated by the driver, who radioed the Mayne train control centre to tell them what had happened. The Mayne train controllers had no indication of what aspect signal FS66 was displaying, or whether the SPAD³ had occurred, as this signal was controlled by the area coordinator at Fisherman Islands. The area controller at Fisherman Islands did receive a SPAD alarm at his workstation but had not responded to it by the time the Mayne train controller called. Train 8868 remained across the level crossing for nearly 40 minutes until a relief driver arrived and moved the train.

The investigation found that the driver of train 8868 was probably experiencing microsleep episodes on the approach to signal FS66 and that this was the principal contributing factor in this incident. The investigation was unable to determine if this fatigue was ‘personally induced’ or ‘task induced’.

The investigation also found that the interface procedures between the Mayne train control centre and Fisherman Islands local control in combination with the structure of the Fisherman Islands area coordinator/station officer’s role have the potential to inhibit emergency response. Additionally, the lack of certain track and train secondary protection devices was considered to be an absent defence.

Safety actions recommended as a result of this investigation include the drafting of a fatigue management standard/policy, evaluation of secondary wayside safety devices, emergency procedure amendments, interface procedure amendments, training in these procedures and a review of attendance at the Fisherman Islands area coordinator/station officer workstation.

___________________________________________

1          0738:06 – Eastern standard time synchronised as described at section 2 of this report.

2          Service rate reduction – brake-pipe air vented to atmosphere at a controlled rate to apply consistent propagation throughout the length of the train.

3          SPAD – An acronym common to the rail industry that stands for ‘Signal Passed at Danger’.

Recommended Safety Actions

As a result of its investigation, the ATSB makes the following recommendations with the intention of improving railway operational safety. Rather than provide prescriptive solutions, these recommendations are designed to provide guidance to interested parties on the issues that need to be considered. Recommendations are directed to those agencies that should be best placed to action the safety enhancements intended by the recommendations, and are not necessarily reflective of deficiencies within those agencies.

RR20050036

The ATSB recommends that QR compile a fatigue management policy/standard to guide managers and workers in how to manage fatigue in planning and operational situations. This policy/standard should also provide guidance on how to deal with reported instances of fatigue.

RR20050037

The ATSB recommends that QR evaluate the installation of wayside secondary protective/prompt device/s in advance of signal FS66. Such device is to be compatible with all operators. This recommendation is made in light of the position of FS66 and the number of SPADs that have occurred at this signal.

RR20050038

The ATSB recommends that QR amend the emergency response procedures applicable to Fisherman Islands to ensure that initial notification of a SPAD is sent to the officer who has control of the signal in question.

RR20050039

The ATSB recommends that QR amend the interface procedures between Mayne train control centre and Fisherman Islands with the intention of ensuring that all trains are contactable by the Fisherman Islands area controller/station officer when in signalled territory controlled by the Fisherman Islands area coordinator/station officer. Boundaries of signalled territory should be clearly defined and current discrepancies in regard to radio channels and signal control should also be amended.

RR20050040

The ATSB recommends that QR undertake training of all concerned in regard to emergency response and interface procedures between the Mayne train control centre and Fisherman Islands.

RR20050041

The ATSB recommends that QR examine methods of ensuring continued attendance by appropriately qualified employees at the Fisherman Islands area coordinator/station officer workstation.

RR20050042

The ATSB recommends that the Queensland Railway Safety Regulator actively monitor the actions initiated by QR in response to this investigation.

Executive summary

Train Y245 passed signal MR5 at stop by 81.8 metres, about 500 metres short of a conflicting freight train movement. The driver of Y245 had previously passed a caution signal but, when only 439 metres from signal MR5, increased power as if proceeding on clear signals. When about 200 metres from signal MR5, an emergency brake application was made. The train controller supervising this section of track was temporarily absent from the workstation and this contributed to the 16 seconds it took to relay an emergency 'stop' radio transmission.

Train Y245 was crewed by a driver as the sole crew member and, apart from the locomotive vigilance devices, there were no secondary protection devices such as Automatic Warning System, Automatic Train Protection or Automatice Train Control at this location. The driver of train Y245 died on 26 October 2004 following a severe coronary episode. This and his previous involvement in SPAD incidents, where loss of concentration had been cited as a causal factor, led to the examination of this driver's state of health. This in turn led to an examination of the medical standards applicable to Queensland Rail drivers.

The investigation found that, while it was unlikely that partial incapacitation was a factor in the SPAD at signal MR5, the possibility could not be ruled out. It was also found that the investigations reports into the driver's previous SPAD incidents focused on the active factors in lieu of latent or systemic factors. Additionally, it was found that the process of returning this driver to full duties following previous SPAD's seemingly followed set patterns. Once returned to full driving duties, little evidence of additional monitoring or supervision was produced during this investigation.

Executive Summary

At 0335:08 on 15 March 2004, the trailing bogie of the second to last wagon of train 6SM9V, operated by Freight Australia, derailed as it departed the crossing loop at Alumatta. This train was a scheduled service from Sydney to Melbourne on the standard gauge network. There were no reported injuries as a result of this derailment either in the active or recovery stages, nor was there any adverse environmental impact.

6SM9V came to a stand at 0339:58, having travelled 5.579 kilometres from the point of departure at Alumatta. During this time four level crossings and one bridge were traversed and a maximum speed of 69 kph was attained. For 4.795 kilometres of this distance this train was travelling in a derailed state thereby inflicting damage to these installations, track, associated infrastructure and the two trailing wagons. At the final level crossing the derailed bogie became dislodged, causing the last wagon to lift, separate from the train and roll over to the eastern side of the track. The consequential loss of air and the actions of the driver brought the train to a stop about 895 metres beyond this level crossing. The train crew then used the local UHF radio to warn any broad gauge trains that may have been in the vicinity that train line air had been lost and the train may be foul.

Initially, the driver suspected a ruptured air hose and left the cab to find the fault and repair it. Advice of this occurrence and proposed actions were forwarded to the ARTC train controller at 0342:10. At 0353:10 the driver advised the ARTC train controller that he had reached the rear of the train and that the second last wagon was derailed and the last wagon was not in sight. At 0353:40 the ARTC train controller rang the Central train controller and instructed that no broad gauge trains be allowed into the vicinity.

This accident occurred on a section of the corridor where standard and broad gauge tracks parallel each other only metres apart. Both of these tracks are operated independently, having separate train control centres and differing safeworking systems with little readily identifiable transparency between them. The investigation has determined that the time taken to notify the Central train control centre was not in accordance with the existing safety management system requirements.

The investigation has determined that the probable cause of this derailment was the geometry of the track combined with the excessive speed of the rear of train 6SM9V as it exited the crossing loop. This led the left wheel of the third axle of the second last wagon to climb the eastern stock rail at the toe of the point blade of the cripple siding. It is probable though that neither of these two factors was, in itself, sufficient to cause the derailment.

Safety actions recommended as a result of this investigation are aimed at revising track standards and maintenance procedures, ensuring the rear of trains do not exceed speeds of curves or turnouts and improving communications between trains on what are essentially two separate rail corridors in one.

Executive summary

At 1303 on Wednesday 25 February 2004 an empty Endeavour passenger train D743 was placed on a collision course with a stationary loaded coal train LD166 at Sandgate. The crew of train D743 realised that number 157 points were in the wrong position and made an emergency brake application. Train D743 stopped about 75 metres short of train LD166.

Sandgate is located on the main northern railway corridor between Sydney and Brisbane about 170 kilometres from Central Railway Station, Sydney. The rail corridor contains four standard gauge tracks for non electric powered trains.

Train D743 was a relief train from Newcastle provided to render assistance to passenger train 604 which had broken down on the Up Main line at Beresfield station platform. The following two passenger trains 736 and 738 on the Up Main line were diverted to the Up Coal Road at Thornton to bypass train 604. Trains 736 and 738 were diverted back onto the Up Main line at Sandgate through number 157 points.

The wrong direction movement of train D743 via the Up Main line was controlled by a Special Proceed Authority (SPA) issued by the train controller at Broadmeadow and Yard Working authorised by the signaller. Train movements in the Sandgate area are controlled by a signaller located at Hanbury Junction signal box, 2.575 kilometres southeast from Sandgate station. Normal signal protection (unidirectional) could not be used for the wrong direction movement. The movement was manually controlled by the signaller. In addition to delayed trains and increased traffic movements, there were a number of factors competing for the signaller's attention.

The investigation concluded that the incident occurred because the signaller did not check the intended route for train D743 thoroughly or apply blocking facilities to the greatest effect to protect the route. Number 157 points were in the reverse position after the passage of train 738. Blocking facilities are physical devices used as memory aids for the signaller to prevent inappropriate issue of Proceed Authorities, or signalling or point equipment operation.

Summary

On 7 October 2004 two crew members were killed and three others seriously injured when Lowlands Grace's port lifeboat became detached from its falls during a lifeboat drill. The ship, a Hong Kong registered cape-sized bulk carrier, was at anchor off the port of Port Hedland, Western Australia, at the time of the accident.

Investigation revealed that the lifeboat's after hook had failed where it was attached to the keel while the boat was being lowered. The lifeboat's stern had then dropped, and the boat rotated around the remaining forward fall before the forward hook opened under the load of the swinging boat. The lifeboat then fell upside down into the sea approximately 16 metres below.

At about 1806 on 14 August 2004, three seamen on the forecastle head of the container ship Aotearoa Chief were struck by a wave that broke over the ship's port bow. One man was thrown against the windlass and mooring machinery and suffered injuries from which he later died. Another sustained severe lacerations to one leg while the third man was unhurt.

Abstract

On 25 September 2004 the general cargo ship Mellum arrived at Thevenard, South Australia, loaded cargo and on 28 September sailed for Melbourne.

At about 1217 the ship cleared Yatala Channel beacons one and two at a speed of about seven knots. Shortly thereafter the pilot disembarked as the ship approached the entrance beacon. Once the pilot had disembarked the master ordered the helmsman to steer a course of 222° by gyro compass.

At 1233, the ship grounded with the entrance beacon bearing 005½°, at a range of 0.52 miles.

The master de-ballasted the ship and with the assistance of the pilot, who had reboarded, refloated the ship. At 2309 the pilot reported that the ship was afloat.

At 1606 on 30 September, Mellum weighed anchor and sailed for Melbourne after it had been checked for damage and seaworthiness. The report found that a misunderstanding between the master and pilot and a lack of planning by the ship's crew were contributing factors. The master/pilot information exchange was found to be deficient. It was also found that the insets and scales of the navigational chart in use and the choice by the pilot to disembark before the ship arrived at the pilot boarding ground may have contributed to the grounding.

The report makes a recommendation relating to pilot training and practices.

Summary

At 2304 on 7 August 2004, the 34.56 m passenger vessel True North, with 38 people on board, ran aground in the entrance passage to St. George Basin in Western Australia's Kimberley region. At the time of the grounding, the master was alone on the bridge. The vessel was in automatic navigation mode with course adjustments initiated by means of a GPS navigation system linked to an Electronic Chart System (ECS), that transmitted course adjustment information to an autopilot. No pollution resulted from the grounding, although several passengers received minor injuries during the initial impact.

True North had draughts of 1.5 m and 2.2 m forward and aft respectively and remained aground for about one hour. It floated clear of the rock on which it had grounded on the incoming tide and moved under its own power to an anchorage south of one of the islands in St. George Basin. The vessel was beached by the master on the morning of 8 August and a full appraisal of the damage made.

On 9 August, a repair team and equipment was flown to the vessel by float plane from Broome, Western Australia. The passengers and non-essential crew were then flown to Broome and repair work commenced. Temporary repairs were carried out while the vessel was beached in between the periods of high water. These repairs were completed by 11 August and True North departed St. George Basin for Darwin, where permanent repairs were to be effected.

The vessel arrived at repair facilities in Darwin on 14 August. True North was able to depart Darwin for Wyndham on 26 August to resume its cruise schedule.

The report concludes that:

• True North grounded on or near Strong Tide Point in unsurveyed waters while being navigated by an auto helm unit and ECS receiving position information from GPS satellites.
• GPS derived positions plotted on the ECS differed from the vessel's true position, possibly caused by:
  • GPS system inaccuracy;
  • Geodetic datum ambiguity;
  • The ECS operating on a common user personal computer which allowed possible corruption of the ECS operating system;
  • A change in the GPS receiver parameters;
  • Loose wiring or connections in the on board systems; or
  • A combination of the above.
• Too much reliance was placed on the accuracy of GPS positions and the GPS positions plotted on the vessel's ECS were not adequately checked by other navigational means.
• A proper lookout was not maintained by visual, radar and other means to ensure the vessel remained in safe water.
• At the time of the grounding, it is probable that the master was suffering from some effects of fatigue as a result of his work routine.
• There were deficiencies in the procedures which dealt with the mustering of passengers in that:
  • The initial safety briefing was not sufficiently comprehensive;
  • company orders in respect of lifejackets and emergency signals were not followed; and
  • company practices in the provision of emergency information and procedures did not meet WA legislative requirements.

The report recommends that:

• Owners and operators of vessels running Electronic Chart Systems should do so on dedicated computers.
• Owners, operators and masters of vessels regularly operating in confined waters should revise procedures governing the use of GPS and auto pilot systems.
• Owners and operators of passenger vessels should ensure information concerning muster areas, emergency signals and instructions and diagrams of how to don lifejackets, are included in any passenger information kits in cabins, and on the inside of all cabin doors, in order to comply with state/territory regulations.
• True North Cruises should review the procedures associated with passenger briefings and the emergency information displayed in the passenger cabins and vessel common rooms.
• Owners, operators and masters of non-SOLAS vessels should consider the introduction of effective fatigue management policies and practices on board vessels.
• AMSA and state/territory marine authorities should carefully consider the type of operation and area of operation when determining the safety manning of vessels of any length.
• Western Australia's Department of Planning and Infrastructure should consider recommending the amendment of WA marine legislation to allow for the option to review and, if necessary, change the safety manning requirements for state registered vessel.

Executive summary

On 3 June 2004, the Australian sail training ship Windeward Bound was off the Victorian coast, heading northeast at about six knots. The wind was from the northwest and the vessel was heeled between 10 and 15 degrees to starboard. The upper and lower topsails and the main and fore staysails were set. A cold front was expected that afternoon. The watch officer was assisted by a watch leader and two general purpose hands, one of whom was at the helm.

At 1726, when the ship was about 30 miles south of Gabo Island, the wind speed increased. The helmsman was instructed to run the vessel downwind. Shortly thereafter, the vessel yawed to port and the helmsman put the rudder hard over to starboard. The watch leader then took the helm and, as the vessel was now swinging to starboard, applied a 'considerable amount' of port rudder. When the vessel had started to swing to port, the watch leader put the rudder hard over to starboard to arrest the swing. A gust of near-hurricane force wind then heeled the vessel about 68 degrees to starboard.

The starboard side of the main deck was submerged, and seawater entered a fuel tank through an open-air pipe on deck. A quantity of seawater also entered the deckhouse and accommodation before doors to the deckhouse and accommodation were shut. The vessel was righted after several minutes using the main engine and rudder and by letting the sheets go.

There was a minor injury to a crew member during the incident and the main engine had been damaged after being run with little or no lubricating oil pressure when the vessel was heeled.

During the next twenty-four hours contact was lost with authorities ashore. An air search was initiated in the afternoon on 4 June and Windeward Bound was found safe and heading for Jervis Bay, where it arrived late in the evening on 5 June.