At about 1055 on 21 August 2014, train 3DA2, carrying containerised freight (including dangerous goods), arrived at Snowtown, South Australia. While at Snowtown, the crew noticed smoke coming from one of the containers conveying dangerous goods. The crew contacted ARTC Network Control to arrange for the Country Fire Service to attend the site. An exclusion zone was set up around the site and the fire was brought under control with minimal damage sustained.
What the ATSB found
The ATSB found that freight within the affected container, including undeclared dangerous goods, had been packed in a manner that was not in accordance with the code of practice for Transport of Dangerous Goods by Road or Rail, or the Genesee & Wyoming Australia dangerous goods policy.
Genesee & Wyoming Australia had a documented policy on the transportation of dangerous goods, including a Standard Condition of Carriage, which documented the obligations of GWA’s customers when providing freight for transportation. However, GWA had no active verification processes in place to check and confirm compliance with those requirements (either in total or in part through random selection). Such a process may have provided the opportunity to detect any undeclared dangerous goods or inappropriately packed freight before an incident or accident resulted.
What's been done as a result
GWA has undertaken an independent audit of their policies and procedures for consigning freight, including the adequacy of training in receiving handling and storage of dangerous goods. GWA has also undertaken to improve communications with customers at their Alice Springs and Darwin terminals to identify where deficiencies may exist and how they might be best addressed.
Safety message
This incident illustrates the importance of freight forwarders and rail operators ensuring that dangerous goods freight accepted for carriage meets the relevant requirements of the Transport of Dangerous Goods by Road or Rail code of practice. All rail operators should ensure that their policies and procedures for the acceptance of dangerous goods are effective in ensuring that the goods accepted have been appropriately packed to minimise the risk of incidents during transportation.
Fire damage to goods
The occurrence
At about 09201 on 19 August 2014, train 3DA2, operated by Genesee & Wyoming Australia (GWA), departed from the Berrimah freight terminal (near Darwin, Northern Territory) bound for Islington Freight Terminal, South Australia. The train was scheduled to pick up and drop off freight at various locations during the journey.
At about 1015 on 20 August, the train arrived at the Alice Springs freight terminal where two freight wagons were dropped off and two picked up; one of which (FQAY0017H) was carrying dangerous goods. The two added wagons were marshalled into the second and third position behind the trailing crew van. Train 3DA2 departed the Alice Springs freight terminal at about 1105 and continued on towards Islington Freight Terminal with a train crew change scheduled at Snowtown, SA.
Figure 1: Location of Snowtown SA
Source: Geoscience Australia annotated by the ATSB
At about 1055 on 21 August, train 3DA2 arrived at Snowtown, South Australia (Figure 1). The relief crew boarded 3DA2 and was preparing the train for departure when the driver noticed smoke coming from a container being conveyed on the centre platform of wagon FQAY 0017H. He checked the hazardous goods manifest and ascertained that the wagon contained hydrochloric acid solution. He then advised the Australian Rail Track Corporation (ARTC) Network Control Officer (NCO) that there was a fire in a wagon which had dangerous goods and to call the Country Fire Service (CFS) immediately. The crew then set up a 250 m exclusion zone and secured the train as part of their emergency procedure.
The NCO contacted the CFS at 1100 and also advised GWA Transport Control. The CFS arrived at 1130 and took control of the site. At about 1150 the train crew were removed from the train by the CFS. By 1310, the fire had been extinguished. There was some damage to goods in the direct vicinity of the fire but there was no damage to the wagon or rail infrastructure.
Through consultation with the CFS coordinator, the NCO authorised other train movements to pass the incident site via the loop track at about 1320 and 1508. The site was declared safe with no hazard for the public or the environment at about 1605.
The wagon was inspected by a GWA safety representative and deemed to be safe to continue its travel to Islington Freight Terminal. Train 3DA2 departed Snowtown at about 1624, arriving at Islington at about 1818 later that day.
Context
Train and crew information
Train 3DA2 was a GWA super freighter, consisting of two locomotives (GWU007 leading and CLP14 trailing) hauling 28 wagons of containerised freight, some of which included dangerous goods. It had a length of 1166 m and a trailing mass of 1821 t.
Train crew
The train crew possessed relevant competencies to operate the train and to transport dangerous goods. They were also trained in emergency management procedures.
Freight loading unloading at Alice Springs
The wagons attached to train 3DA2 at Alice Springs were loaded with containers packed by Toll Australia and delivered to the GWA freight terminal for transport to Islington Freight Terminal.
Container number 4CB215, loaded on wagon FQAY0017H, had been packed by Toll Australia with consignments from multiple customers, each with a consignment note signed by the customer. The consignment listed on the train’s manifest as dangerous goods consisted of hydrochloric acid solution, noted as 4 x 1000 l Intermediate Bulk Containers (IBC). The remaining consignments were general freight, though one also included an unsecured lead-acid battery2 which was not accompanied by a dangerous goods declaration.
The containers were received by GWA at their Alice Springs yard but were not inspected.
Fire on 3DA2 at Snowtown
GWA policy for train crew dealing with a fire event
GWA’s response plan for “on-rail’ emergencies involving a fire or dangerous goods event, required the train driver to take control of the site until they were relieved by a GWA Incident Response Coordinator or Incident Controller. To ensure train safety and integrity in accordance with the safe working regulations, the driver was required to remain at the controls in the lead locomotive. The response plan also required people to be evacuated if there were fire, smoke, vapour clouds or fumes and move to a safe distance uphill/up wind if there were dangerous goods involved. Transport (train) control was to be notified and required to constantly monitor the communications (radio) onsite.
In this case, once the relief crew noticed the smoke from the container, they obtained the train’s manifest to determine the nature of the freight inside. They identified that the smoke was coming from wagon FQAY0017H (container 4CB215), checked the hazardous goods manifest and determined the container contained hydrochloric acid solution. The driver immediately called the Network Control Officer (NCO), advised of the smoke coming from a container of dangerous goods and requested the fire brigade attend. The driver then set up a 250 m exclusion zone and remained with the train. The NCO contacted the Country Fire Service (CFS) advised them of the fire and the quantity of hydrochloric acid solution that the manifest indicated was within the container.
The CFS contacted the customer of the hydrochloric acid solution and established that the quantity listed on the manifest differed to the actual amount being transported on board train 3DA2 – the containers being empty, with only residual acid product within.
The CFS further assessed the situation and determined that there was nothing leaking from the container. The decision by the CFS to physically access the container was deferred until the arrival of specialist equipment. The CFS permitted the NCO to operate other services past 3DA2 via the crossing loop track while waiting for the equipment. Once the equipment had arrived, the CFS requested that the NCO close the track again as they prepared to enter the container.
After an internal inspection of the container, the CFS determined that there were no further threats to personal safety or to the environment and allowed the track to be re-opened.
An inspection of the rolling stock was conducted by GWA to ensure the integrity of the train. Once it was determined that the risk of fire and contamination had been mitigated, and that the wagon was safe to continue, train 3DA2 departed Snowtown at 1624.
Emergency response to the dangerous goods fire
Network Control made the call to the emergency services and gave details about the type of dangerous goods and the amount involved. The initial assessment by the CFS was to approach the container fire with the expectation that there were substantial quantities of hydrochloric acid solution present.
The initial limited information available to the CFS on site hindered them in determining the possible ignition sources or the exact contents of the container. This compromised their most efficient response to the situation, including determining the extent of exclusion area or whether it was appropriate to commence a full evacuation. As such, the initial (default) response was to plan for the worst case and then to scale back the effort once more was known about the dangerous goods.
Throughout the incident, the NCO continued to coordinate with the emergency services, GWA and the train crew at the scene. The train crew and local staff managed the site according to their policy and procedures for on-rail emergencies. The implementation of the GWA “on-rail emergency” plan and the coordination with ARTC effectively managed the incident and provided assistance and information to the emergency services as required.
Fire origin
Inspection of the container found a number of plastic tubs, one of which contained a lead-acid battery. At some point during the journey, an unsecured metal-framed portable chair had fallen across the exposed terminals of the battery (Figure 2).
The metal of the chair frame had shorted out the battery causing high current to flow and heat the metal frame. The heat transmitted from the chair frame resulted in the melting and combustion of the plastic tubs (Figure 3). The smoke produced by the burning plastic was subsequently observed coming from the container by the GWA train crew.
Figure 2: Tub with battery and metal frame of the chair.
Source: Genesee & Wyoming Australia annotated by the ATSB.
Figure 3: Damaged freight in situ in the container
Source: Genesee & Wyoming Australia annotated by the ATSB
Dangerous goods
Transportation of dangerous goods at the time of this incident was governed by the Australian Dangerous Goods (ADG) Code 7.3 edition. The ADG Code is subject to regular reviews that keep pace with international developments and local industry best practice. The current edition is based on the recommendations of the Dangerous Goods Model Regulations 17th edition, United Nations, 2011.
The Code must be read in conjunction with the Model Act on the Transport of Dangerous Goods by Road or Rail3.The model act sets out, in general terms, the legal requirements for transporting dangerous goods by road and rail in Australia. It establishes the broad regulatory framework for the transport of dangerous goods and is complemented by the Model Subordinate Law.
Dangerous goods on train 3DA2
The train consist included wagon FQAY 00017H; a 5-pack4 48-foot artic skeletal wagon designed to carry containers, with an overall length of 79.1m. (Figure 4)
Source: Genesee & Wyoming Australia
The train manifest showed the wagon was located 6th in the sequence and was loaded with seven containers, including container number 4CB215, which contained dangerous goods and was the subject of the fire.
The container manifest for 4CB215 showed that it was loaded with 19 different consignments under individual consignment notes. There were three consignments listed in the category of dangerous goods, although these were not listed in the first part of the documentation as required by the ADG Code. The goods were listed as hazardous class 2.1 (gas sample) and hazardous class 8 (empty intermediate bulk containers of hydrochloric acid).
Genesee & Wyoming requirements for transporting dangerous goods
Genesee & Wyoming requirements for the transportation of dangerous goods are set out in their Domestic Intermodal Infopak. Within the Domestic Intermodal Infopak, Pacific National’s Freight Loading Manual (FLM) indicates how the freight should be handled during its transport.
Pacific National requirements for transporting dangerous goods
Pacific National sets out in its Freight Loading Manual, requirements for freight forwarders when consigning dangerous goods. Section FLM-01-_05 Transit of Dangerous Goods by Rail details the stowage of dangerous goods and also goods that are transported along with dangerous goods.
Safety analysis
Receiving of dangerous goods for transport by rail
The Dangerous Goods Code required that customers declare dangerous goods, and that those goods were adequately described, packaged and secured. The code further provided that batteries were classified as dangerous goods within the meaning of the code (and thus needed to be declared), if they were not secured within equipment. If they were carried with other goods, then all the goods were to be adequately secured to prevent movement or damage, or in the case of batteries, to prevent the terminals bridging and shorting.
Transporting dangerous goods by Genesee & Wyoming Australia
Genesee & Wyoming Australia policy on the transportation of dangerous goods is set out in their Intermodal Domestic Infopak document, which referred to Pacific Nationals Freight Loading Manual (FLM) for transporting freight.
Pacific National’s requirement for dangerous goods states that:
Stowage of non-dangerous goods within the container must be done in a manner to prevent damage to any dangerous goods packing. All other loading must be secured within the container; there must be no loose items.
The FLM provides for the inspection of goods and freight before the freight is accepted – in particular if there is any ambiguity to the type of freight to be conveyed. There is also a function within the FLM to undertake random auditing of the freight to ensure compliance with the relevant stowage and declaration requirements.
While not referred to in GWA’s documentation, Pacific National’s Load Inspection Authority also states that random audits can be conducted while freight is still attached to the prime mover at the terminal gate or yard. Their standard condition of carriage also states that the customer authorises Pacific National to:
…open any container or inspect the goods to determine their nature, conditions, ownership or destination.
GWA’s equivalent document, Standard Condition of Carriage, makes no provision for inspection of freight that may contain undeclared dangerous goods. However, there is a requirement that:
The customer authorises GWA to open any container or inspect the goods to determine their nature, condition, ownership or destination.
and that:
the customer has not asked GWA to handle, transport or store the goods in any way that could be unlawful.
While the intent was that any goods provided to GWA by customers shall be in a condition that is lawful, and while the provisions for inspection were present in the various associated documents, there was no procedure in place for the actual verification of compliance (either in total or in part through random selection). If such a procedure had been in place, it may have provided an opportunity to discover any undeclared dangerous goods (such as the exposed battery) and any loosely and inappropriately packed freight.
Findings
From the evidence available, the following findings are made with respect to the fire on the fourth wagon of train 3DA2 that occurred near Snowtown, SA, on 21 August 2014. These findings should not be read as apportioning blame or liability to any particular organisation or individual.
Safety issues, or system problems, are highlighted in bold to emphasise their importance. A safety issue is an event or condition that increases safety risk and (a) can reasonably be regarded as having the potential to adversely affect the safety of future operations, and (b) is a characteristic of an organisation or a system, rather than a characteristic of a specific individual, or characteristic of an operating environment at a specific point in time.
Contributing factors
When consigned for transport, freight container 4CB215 contained an undeclared lead-acid battery packed alongside other freight that had not been secured in accordance with the operator’s requirements or those of the Australian National Code of Transport of Dangerous Goods by Road and Rail.
During transport, it was likely that an unsecured metal-framed chair shorted the battery terminals, with the resulting heat melting and igniting the plastic tubs surrounding the battery.
Other factors that increased risk
Genesee & Wyoming Australia had no procedure in place to verify (either in total or by random selection) that the nature or condition of freight provided by their customers, complied with their Standard Condition of Carriage. [Safety issue]
Other findings
The procedures followed by Genesee & Wyoming Australia and ARTC Network Control, in their management of the fire emergency, were effective and followed their documented response protocols.
Safety issues and actions
The safety issues identified during this investigation are listed in the Findings and Safety issues and actions sections of this report. The Australian Transport Safety Bureau (ATSB) expects that all safety issues identified by the investigation should be addressed by the relevant organisation(s). In addressing those issues, the ATSB prefers to encourage relevant organisation(s) to proactively initiate safety action, rather than to issue formal safety recommendations or safety advisory notices.
All of the directly involved parties were provided with a draft report and invited to provide submissions. As part of that process, each organisation was asked to communicate what safety actions, if any, they had carried out or were planning to carry out in relation to each safety issue relevant to their organisation.
Genesee & Wyoming Australia procedures for acceptance of dangerous freight
Genesee & Wyoming Australia had no procedure in place to verify (either in total or by random selection) that the nature or condition of freight provided by their customers, complied with their Standard Condition of Carriage.
Genesee & Wyoming Australia had no procedure in place to verify (either in total or by random selection) that the nature or condition of freight provided by their customers, complied with their Standard Condition of Carriage.
Proactive safety action taken by Genesee & Wyoming Australia
Action number: RO-2014-015-NSA-006
As a result of this occurrence, Genesee & Wyoming Australia advised that they have undertaken an independent audit of their policies and procedures for the receiving, storage, handling, transporting and delivery practices at the Alice Springs and Darwin yards – relative to dangerous goods on GWA-accredited intermodal trains. The audit encompassed the following aspects;
Communications with customers re DGs at the time bookings are made – chain of responsibility.
Receipt of DGs at terminal and adequacy of the receipting process. Documentation relative to GWA’s responsibilities and gaining assurance as to nature of the goods – chain of responsibility.
Segregation and placarding of loaded and compliance of same with code of practice.
Adequacy of training delivered to train crews, transport controllers and supervisors.
Existence and adequacy of emergency/evacuation plans relative to DGs at the terminal.
ATSB comment in response
While the likelihood of such events is relatively low, Genesee Wyoming Australia’s review of policy and procedures will help identify opportunities to further mitigate the risks of transporting dangerous goods. The ATSB is satisfied that, when fully implemented, the actions proposed by Genesee & Wyoming Australia should address this safety issue.
Current status of the safety issue:
Issue status: Safety action pending Justification: At the time of this report release, the safety action advised by Genesee & Wyoming Australia had not yet been fully implemented.
Sources and submissions
Sources of information
The sources of information during the investigation included:
Genesee & Wyoming Australia
Australian Rail Track Corporation
Pacific National
Toll Ipec
References
Australian code for the Transport of Dangerous Goods by Road & Rail Edition 7.3, August 2014
Pacific National Freight Loading Manual FLM 00-97 Transit of dangerous goods by rail FLM 01-06_05.
Pacific National Freight Loading Manual FLM 00-97 Provisional load process FLM 01-12
Pacific National Freight Loading Manual FLM 00-97 Acceptance of loading FLM 01-02
Pacific National Freight Loading Manual FLM 00-97
Genesee & Wyoming Australia Domestic Intermodal Infopak, October 2013
Genesee & Wyoming Standard Condition of Carriage Rev A
Genesee & Wyoming Australia Provisional Load Procedure GWA-FLP-03-006_01
Genesee & Wyoming Australia Response plan for “on rail” emergencies RS-PRC-006 version 008
Submissions
Under Part 4, Division 2 (Investigation Reports), Section 26 of the Transport Safety Investigation Act 2003 (the Act), the Australian Transport Safety Bureau (ATSB) may provide a draft report, on a confidential basis, to any person whom the ATSB considers appropriate. Section 26 (1) (a) of the Act allows a person receiving a draft report to make submissions to the ATSB about the draft report.
A draft of this report was provided to Genesee & Wyoming Australia, The Office of National Rail Safety Regulator (ONRSR), Toll Ipec, the Australian Rail Track Corporation, Pacific National and the train crew of 3DA2.
Responses and/or submissions were received from Genesee & Wyoming Australia, The Office of National Rail Safety Regulator (ONRSR), Toll Ipec, the Australian Rail Track Corporation, Pacific National and the train crew of 3DA2. The submissions were reviewed and where considered appropriate, the text of the report was amended accordingly.
Purpose of safety investigations
The objective of a safety investigation is to enhance transport safety. This is done through:
identifying safety issues and facilitating safety action to address those issues
providing information about occurrences and their associated safety factors to facilitate learning within the transport industry.
It is not a function of the ATSB to apportion blame or provide a means for determining liability. At the same time, an investigation report must include factual material of sufficient weight to support the analysis and findings. At all times the ATSB endeavours to balance the use of material that could imply adverse comment with the need to properly explain what happened, and why, in a fair and unbiased manner. The ATSB does not investigate for the purpose of taking administrative, regulatory or criminal action.
Terminology
An explanation of terminology used in ATSB investigation reports is available here. This includes terms such as occurrence, contributing factor, other factor that increased risk, and safety issue.
Publishing information
Released in accordance with section 25 of the Transport Safety Investigation Act 2003
Ownership of intellectual property rights in this publication
Unless otherwise noted, copyright (and any other intellectual property rights, if any) in this report publication is owned by the Commonwealth of Australia.
Creative Commons licence
With the exception of the Coat of Arms, ATSB logo, and photos and graphics in which a third party holds copyright, this publication is licensed under a Creative Commons Attribution 3.0 Australia licence.
Creative Commons Attribution 3.0 Australia Licence is a standard form licence agreement that allows you to copy, distribute, transmit and adapt this publication provided that you attribute the work.
The ATSB’s preference is that you attribute this publication (and any material sourced from it) using the following wording: Source: Australian Transport Safety Bureau
Copyright in material obtained from other agencies, private individuals or organisations, belongs to those agencies, individuals or organisations. Where you wish to use their material, you will need to contact them directly.
_____________________
1. The 24 hour clock is used in this report to describe the local time of day, Central Standard Time (CST)
2. The Dangerous Goods Code provides that cells and batteries, except when installed in equipment, must be packed in inner packaging that completely enclose the cell or battery or must be protected so as to prevent short circuits.
3. Source: Australian Code for the Transport of Dangerous Goods by Road and Rail available at www.ntc.gov.au
4. An articulated wagon comprising five platforms, the adjacent ends of individual units being supported on a common bogie and permanently connected by a device which permits free rotation in all planes. RISSB, Glossary of Railway Terminology - Guideline
On 20 August 2014, the pilot of a Robinson R66 helicopter, registered VH-JRX, conducted a site inspection from a vehicle prior to commencing aerial spraying near Giru, Queensland. The pilot identified powerlines running along the eastern, southern and northern boundaries of the paddock to be sprayed. He then conducted a flight over the paddock and assessed the hazards in the area and confirmed he was able to see all of the powerlines he had identified from the ground.
At about 1500 Eastern Standard Time (EST), the pilot commenced aerial spraying in an east-west direction. At the end of each run, the pilot climbed the helicopter up and over the powerlines, turned then descended once clear of the powerlines and sprayed the paddock in the opposite direction.
At about 1645, the pilot commenced the final clean-up run. The helicopter was operating along the southern boundary of the paddock, parallel to the powerlines running along the southern and northern borders. The pilot was aware of those powerlines, however when about 5 m from the eastern boundary, he sighted the powerlines running perpendicular to the direction of flight. The pilot assessed that it was too late to climb over the powerlines and elected to fly underneath them and to ensure the helicopter remained above the crop.
The main rotor blade hub struck the powerlines and the helicopter collided with the ground. The helicopter was substantially damaged, and the pilot sustained minor injuries.
This incident provides a reminder to pilots to conduct an aerial reconnaissance to confirm wire locations and other hazards.
On 24 August 2014, the pilot of a Gippsland Aeronautics Airvan aircraft, registered VH-XHV had just completed a parachute drop and was on descent to land at the Pinjarra Skydiving Airstrip, Western Australia.
At about 1,000 ft above the ground, the aircraft was configured for landing with full flap selected. The pilot reported that the airspeed was about 80 kt on base and that day, and during the landing roll noticed the wing of a Classic Aircraft Corporation WACO aircraft was over the runway.
The pilot applied some braking but could not stop the aircraft in the 836 metres available. In an effort to avoid a collision, the pilot manoeuvred the aircraft to the left of the runway. Shortly after, the wing of the Airvan struck the WACO causing the Airvan to spin rapidly. The propeller struck a parked trailer and then an unoccupied 4WD.
The pilot was not injured, however two people on the ground sustained minor injuries and both the Airvan and Waco sustained substantial damage.
This accident serves as an important reminder of the need to initiate a go-around as soon as there is doubt about the suitability of the approach and landing. In this instance, the pilot had assumed that the full length of the runway was available and had left the option of a safe go-around too late.
The information contained in this Preliminary report is released in accordance with section 25 of the Transport Safety Investigation Act 2003 and is derived from the ongoing investigation of the occurrence. Readers are cautioned that new evidence will become available as the investigation progresses that will enhance the ATSB's understanding of the accident as outlined in this Preliminary report. As such, no analysis or findings are included in this report.
What happened
On the afternoon of 17 August 2014, the Grand Pioneer and AAL Fremantle were berthed, stern to stern at North Quay berths 11 and 12 respectively, in the port of Fremantle.
At about 2200, while both ships were discharging cargo, a storm front passed over Fremantle bringing with it a rapid increase in wind speed and gusting to about 55 knots. Both ships' sterns broke away from the berth and they swung bow to the wind. AAL Fremantle’s starboard side subsequently made contact with the Parmelia I and the Fremantle Rail Bridge.
The Parmelia I, Grand Pioneer and AAL Fremantle suffered minor damage.
A stanchion supporting the catenary wire for the 25 kV AC electrified train network on the rail bridge was destroyed along with large sections of temporary scaffolding along the side of and under the bridge. The track alignment was also disturbed. The bridge was closed to rail traffic until repairs could be undertaken and the bridge deemed safe.
By about 0300 both ships were safely berthed.
What the ATSB found
The ATSB’s initial analysis of the incident indicates that a bollard, with all of AAL Fremantle’s and two of Grand Pioneer’s five stern lines, failed. This allowed the stern of AAL Fremantle to swing to starboard. The loss of the stern lines increased the tension on Grand Pioneer’s remaining stern lines, which were not able to hold. Grand Pioneer’s stern then swung to port.
Investigation direction
The investigation is ongoing and will focus on the:
bollard failure mechanism
preparedness of the ships leading up to the incident
port authority procedures to avoid weather related incidents
effect of weather on water levels in the port and the movement of ships
risk analysis and management in relation to relevant marine and rail infrastructure
risk management and emergency response between relevant agencies.
Final report
Safety summary
What happened
On 17 August 2014, Grand Pioneer and AAL Fremantle broke away from their berths when a thunderstorm passed across the Port of Fremantle. A bollard on the wharf holding both ships’ stern lines failed, most likely after Grand Pioneer’s vehicle ramp contacted it. AAL Fremantle contacted a ship at an adjacent berth, and parts of the Fremantle Rail Bridge nearby.
The ships were berthed again with tug assistance. The ships had suffered minor damage. The rail bridge, however, was closed for 3 weeks for inspection and repairs to track alignment and other non-structural damage.
What the ATSB found
The ATSB investigation concluded that Grand Pioneer moved slightly as the tension in its mooring lines increased, in response to the high winds associated with the passing thunderstorm. It is likely that its vehicle ramp then made contact with the bollard that held the stern lines of both ships. As a result of the contact, the bollard broke away from the wharf.
The investigation found that Fremantle Port Authority’s (Fremantle Ports) examination of the risks associated with a ship contacting the rail bridge contained limited analysis on keeping ships alongside in adverse weather, particularly at berths 11 and 12 where the wind is predominantly on ships’ beams. There was also no analysis of the means to assist a ship that got close to Wongara Shoal and the rail bridge.
The ATSB also found that Fremantle Ports’ adverse weather procedures were triggered only by specific Bureau of Meteorology (BoM) forecast categories and terms. There was no guidance for vessel traffic service (VTS) officers to take action based on actual weather conditions, or certain other severe weather terms used in BoM forecasts.
Another investigation finding was that BoM’s marine forecast at the time of the incident did not describe expected wind speeds using recognised marine terms, such as ‘gale force’. Further, the forecast title understated the wind speeds expected.
What's been done as a result
Fremantle Ports has put into service 12 ‘shore tension’ devices in the inner harbour. These devices maintain a constant tension in a mooring line to assist keeping a ship alongside its berth.
The port has subscribed to a customized weather prediction service for its area, and upgraded its weather station and VTS equipment to enhance monitoring. The VTS officers have been trained to use the new and upgraded equipment. The port has updated its ship weather warnings to include a broader range of meteorological terms and descriptions.
Fremantle Ports has also revised its weather warning distribution list to include the manager of the rail bridge, the Public Transport Authority (PTA). The port and the PTA have established direct, high level, points of contact with communication processes in place to manage any emergencies that could involve the rail bridge.
The BoM safety action includes undertaking to use standard terminology in marine forecasts and implementing a formal process to consult stakeholders to better identify and meet their needs.
Safety message
When analysing the risks to a port’s operation, its operator needs to consider the risk controls to avoid a serious incident as well as the recovery controls in case an incident does occur.
Findings
From the evidence available, the following findings are made with respect to the breakaway of Grand Pioneer and AAL Fremantle from their berths when a forecast, severe thunderstorm passed across the Port of Fremantle on 17 August 2014. These findings should not be read as apportioning blame or liability to any particular organisation or individual.
Safety issues, or system problems, are highlighted in bold to emphasise their importance. A safety issue is an event or condition that increases safety risk and (a) can reasonably be regarded as having the potential to adversely affect the safety of future operations, and (b) is a characteristic of an organisation or a system, rather than a characteristic of a specific individual, or characteristic of an operating environment at a specific point in time.
Contributing factors
The wind direction on 17 August 2014, during the thunderstorm, was nearly abeam of both Grand Pioneer and AAL Fremantle with wind gusts reaching 55 knots.
The rapid increase in wind speed as the thunderstorm passed did not allow the crew of either ship to respond to the significantly worsening weather conditions.
The parting of a single line at the stern of Grand Pioneer, when the storm front passed over, increased the load on the remaining stern lines.
As the wind loading on Grand Pioneer increased, it is likely that the stern moved sufficiently off the wharf so that the ship’s vehicle ramp made contact with the adjacent bollard.
It is very likely that the ramp’s impact resulted in the bollard’s failure, leaving Grand Pioneer with only two stern lines that payed out under the increasing load. The ship’s stern then broke away from the wharf causing the two aft spring lines to part.
The bollard’s failure resulted in AAL Fremantle losing its three stern lines. Consequently, the ship’s stern broke away from the wharf and then the two aft spring lines parted.
After breaking away, AAL Fremantle contacted Parmelia I, a ship at an adjacent berth, and then the Fremantle Rail Bridge.
Fremantle Ports’ assessment of risks associated with a ship contacting the Fremantle Rail Bridge as a result of a breakaway (particularly from berths 11 and 12) was limited. Preventing a breakaway from berths where the wind was likely to be on a ship’s beam had not been considered. Similarly, the impediments to assisting a ship near Wongara Shoal after a breakaway had not been assessed. [Safety issue]
Other factors that increased risk
The Bureau of Meteorology (BoM) marine forecast title of ‘strong wind warning’ understated the ‘damaging winds’ expected during the ‘severe thunderstorm’. The forecast did not use recognised marine weather terms for wind speed, such as ‘gale force’. [Safety issue]
Fremantle Ports’ staff did not understand the significance of some wind and weather terminology used in the BoM forecast. Consequently, port procedures triggered by a BoM ‘gale’ or ‘severe weather’ warning such as preparing the tugs and calling the harbour master were not followed. [Safety Issue]
Fremantle Ports’ procedures for adverse weather were not adequate for weather that could be reasonably be expected to occur. Some procedures could not be reasonably implemented and others were not monitored for compliance. [Safety issue]
Other findings
A ‘meteo-tsunami’ did not contribute to the breakaway.
While the failed bollard was compliant with its documented specification, examination showed that it had been made from a low-quality cast iron.
While the loads imposed solely by the mooring lines could be calculated, the overall load imposed on the bollard could not. As such, the contribution of the low quality cast iron to the failure could not be determined.
Analysis
Breakaway
On 17 August 2014, Grand Pioneer and AAL Fremantle broke away from their berths when a thunderstorm passed across the Port of Fremantle. A bollard on the wharf holding both ships’ stern lines failed; most likely after Grand Pioneer’s vehicle ramp contacted it. AAL Fremantle contacted a ship at an adjacent berth, and then parts of the Fremantle Rail Bridge.
The ships suffered minor damage; however, the rail bridge was closed for 3 weeks for inspection and repairs to track alignment and non-structural damage.
Bollard failure
Scrape marks, running parallel to the wharf’s edge, were found on the wharf apron that indicated minor movement of Grand Pioneer’s ramp footplate due to the passage of vehicular traffic and changes to the ship’s draught and the height of tide.
A distinct rubber mark (Figure 7 and Figure 8), at right angles to the wharf’s edge near the failed bollard had a circular pattern at the inboard end consistent with being produced by the outboard tyre under the ramp’s footplate. There were no other marks on the wharf apron to indicate that the ramp had moved in any direction other than off the wharf. The tyre mark shows that the ramp only had to move a short distance (Figure 8) for the sheave to be able to make contact with the bollard.
Figure 7: Tyre marks on the wharf near the failed bollard
Source: ATSB
Figure 8: Indicative set up of the ramp at the time of the incident
Source: ATSB
When the ramp dropped off the wharf, it fell about 4 m to the water’s surface. The size and shape of the groove and burr (Figure 10, left) was consistent with being caused by the wire rope jumping off the sheave and rubbing on the cheek plate under heavy load as the ramp dropped. Further, the damage was recent. When examined a few days after the event, it only had a covering of light surface rust (Figure 9 and Figure 10) which is consistent with having happened in Fremantle.
Figure 9: Ramp set up
Source: ATSB (photo taken in Melbourne)
The ATSB considered the possibility that the contact damage to the sheave could have been made by a means other than the bollard, such as a vehicle. This possibility was discounted for two reasons: there was insufficient space for equipment to approach the sheave (Figure 8) and a large force would have been required to distort its 25 mm steel plate (Figure 10).
Figure 10: Damage to the ramp’s outboard flap sheave
Source: ATSB
Closed circuit television footage
Fremantle Ports has closed circuit television (CCTV) cameras at strategic points at wharves and other locations in the inner harbour to assist with monitoring shipping and cargo operations. The following images (frames) are from footage from a CCTV camera on the North Quay wharf apron.
Figure 11: CCTV footage
The 2200 frame, just before the storm’s arrival, shows Grand Pioneer’s two stern lines (red arrow) on the shared bollard that failed. The bollard itself is just outside the frame.
By 2202, the wind speed has increased, and rain has started to fall, with the approach of the thunderstorm.
At about this time, the driver towing the hay baler heard a mooring line part. The two stern lines (red arrow) on that bollard are still visible.
This frame also shows that AAL Fremantle has moved slightly off the wharf, as its mooring lines take the weight and then stops moving.
The CCTV footage over the next minute shows heavier rain. Grand Pioneer has moved slightly away from the berth. Both stern lines on the shared bollard lines are no longer visible. This is consistent with the failure of the bollard at that time.
The footage that follows shows both ships then moving further away from the wharf.
Collectively, the evidence is consistent with the ship’s ramp contacting the bollard at about 2203. It is very likely that the contact caused the bollard to fracture and be pushed from its mounts.
Bollard examination
Fremantle Ports engaged Exceed Consulting to conduct an independent examination of the bollard and determine the reasons for its failure. The following relevant information comes from Exceed Consulting’s examination report.
The specification for the casting metal[12] in the original bollard design drawings specified a tensile strength but not a chemical composition. The bollard material met the tensile strength requirements and was therefore compliant with the specification.
However, the examination report stated that the bollard had been made from low quality cast iron with a high phosphorus and sulphur content. Due to these high levels, particles of manganese sulphide and steadite with iron phosphide formed. The brittle steadite phase would have contributed to the fracture of the cast iron bollard, particularly in the areas of the structure where it was present as a network. The network was predominantly at the austenitic grain boundaries and was found to be associated with the fracture surface. The report did not offer any information on how the low quality casting material affected the tensile strength of the bollard.
The bollard was held down with grade 4.8 bolts,[13] which had a higher tensile strength than the bollard’s cast iron parent material.
Examination of the bollard’s fracture surfaces identified the corner opposite the grout hole as the point of fracture initiation (Figure 12). This area of the fracture had a finer texture and showed characteristic ‘river marks’ that propagated away from it.
Figure 12: The bollard
Source: Exceed Consulting, annotated by ATSB
Exceed Consulting noted that the crack originating from the grout hole was not present before the bollard failed and, thus, had not contributed to the failure. The rust marks around the crack were the result of immersion in seawater after the failure.[14]
Bollard loading
Fremantle Ports engaged Det Norske Veritas – Germanischer Lloyd (DNV-GL) to calculate the load imposed by the mooring lines on the failed bollard. The DNV-GL analysis[15] and calculation of the overall force on the bollard took into account the:
windage areas of each ship
recorded wind speed
angle of the wind relative the ships’ beams, and
angles of the mooring lines relative to the bollard.
This analysis concluded that the mooring line-induced loads on the bollard were less than its expected failure load, thus, it did not fail solely due to the load from the five mooring lines.
Bureau of Meteorology forecasting
The Bureau of Meteorology (BoM) issues land and marine forecasts twice a day[16] with predicted wind speeds and directions. The forecasts use descriptive terms, such as ‘afternoon’ or ‘evening’, to indicate different periods during the day (Appendix C). Every forecast also clearly states at its beginning that the gust strength can be up to 40 per cent higher than the predicted wind speed.
On 16 August, the Perth local waters forecast changed during the day. The predicted wind speed was increased, and the updated forecasts’ terms were upgraded to include strong winds and later thunderstorms.
On 17 August, the BoM issued a ‘severe thunderstorm’ warning for the Perth area, (including Fremantle Port), at 0401, 1001, 1214, 1453, 1751 1601 and 2053. These warnings advised of damaging winds in the late evening.
The 0401 forecast highlighted the potential for gusty winds with thunderstorms. At 1601, the forecasts became more specific, including reference to the severity of the winds associated with the expected thunderstorms:
Strong wind warning. Winds NW 15 to 25 knots, reaching 30 knots during the evening. Scattered thunderstorms in the late evening, possibly severe with damaging winds.[17]
Local waters forecasts for use by professional and recreational mariners. Some terms used in these forecasts, such as ‘strong wind’, indicate a specific wind speed range to mariners.[18] However, the wind force scale used by mariners does not similarly define other terms, such as ‘severe’ or ‘damaging’ winds.
At the time of the incident, Fremantle Ports’ operational staff considered ‘severe’ and ‘damaging winds’ as forecast terminology for shore-based users. Other than having a general understanding of these terms, staff were not aware that there was a specific BoM-definition. Therefore, the significance of the terms was not fully understood, and they did not take any action suited to the predicted ‘storm’ conditions.
The port’s operational staff most likely saw the title of the forecast, ‘strong wind warning’, as the dominant piece of information. However, the forecast included other defined terms for much stronger weather categories. It is likely that a lack of understanding of these terms, and the use of other, weather specific terms in the port’s weather-related procedures,[19] resulted in them placing importance only on the category of the warning issued.
Weather event management
Fremantle Ports’ Port Information Guide includes the following weather-related advice for the Fremantle area.
Gales - The barometer is a good indicator of the weather, as a general rule rising with southerly and falling with northerly winds. It invariably gives several hours notice of the approach of bad weather.
During the winter months, gales generally commence from the north and rapidly shift W’ly with a falling barometer.
When the wind shifts to the NW with the barometer still falling the wind will be approaching its maximum and may reach gale force. When the wind shifts to the West, or WSW, it generally increases. As the wind shifts S’ly, with a rising barometer and the weather moderating, as it generally does, it continues to back to the NE quarter.
Should the wind after backing SW, veer to the W or NW, the gale is not over, but will probably blow harder than before, the barometer keeping below 1016 hPa.
Fremantle Ports
Fremantle Ports’ Vessel Traffic Service – Operational Procedures sets out operating parameters for ships, including size, draught and others, using the inner and outer harbours. The procedures also detail the roles and responsibilities of the duty vessel traffic services officer (VTSO) and the harbour master.
The weather-related procedure included the following:
Strong wind, gale and severe warnings that are formally issued by the BoM …are received automatically by all VTSOs via e-mail.
These weather warnings are further distributed automatically to relevant port users including…
Upon receipt of a severe weather or gale warning[20] the VTSOs are to ensure that a standby tug has been fully crewed both in the Inner Harbour and Outer Harbour and that Port Service Officers have completed or are in the act of distributing such warnings to vessels in the Inner Harbour.
Additionally the bunker vessel located at berth 12A, Inner Harbour is also to be fully crewed at all times.
VTSOs are to ensure that vessels are compliant in securing for bad weather especially those vessels along North Quay, Inner Harbour where the outboard anchor must be lowered to the sea floor, all mooring lines doubled up and self-tensioning winches secured.
Another procedure required the VTSO to notify the duty harbour master when the tugs and bunker barge were advised of a gale warning. This procedure aimed to ensure that gale force wind forecasts came to the attention of the duty harbour master.
The weather forecast for the night of 17 August included a ‘strong wind warning’ (Appendix C). As a result, the port issued a strong wind warning advice to the masters of berthed ships.
The port’s weather-related procedures were triggered by the type of BOM-issued warning; that is, a strong wind, gale or severe weather warning. There was no requirement, or guidance, to implement any procedure based on sustained local wind speeds or gusts recorded by the port’s own equipment.
While VTSOs are required to ensure compliance with the port’s procedures, it appears there was no specific process for this. For example, Grand Pioneer’s anchor was not lowered as required by the procedures, and this non-compliance was not followed up or addressed.
Similarly, some other procedural requirements are not, or could not be, effectively implemented. The requirement to ‘double up’ on all mooring lines is difficult, if not impossible. Most ships cannot double up on all of their mooring lines, as they have neither the number, nor the length of lines, to do so. In the case of Grand Pioneer and AAL Fremantle, it was not possible for the ships, and the port, to be able run and make fast twice as many lines.
It is not unusual for a ship’s bridge to remain unattended in port. Further, visibility in the direction of an approaching storm, from the bridge, can be obscured by land or infrastructure, as was the case at berths 11 and 12. However, the VTS tower is manned on a 24-hour basis and the duty VTSO has an uninterrupted view of an approaching storm from the tower. Radar, anemometer and barometer provide VTSOs with useful tools for effectively monitoring the weather, giving them the ability to provide advance warning to ships in port waters.
At the time of the incident, Fremantle Ports’ weather-related procedural triggers were written with the aim of having precautionary measures in place in advance of adverse weather. However, they did not adequately cover the weather that could not be predicted, or allow for a response to actual wind speeds experienced. Further, those precautionary measures that were put in place were not adequately monitored to ensure compliance.
Grand Pioneer
Grand Pioneer’s master had monitored the weather forecasts before berthing and was expecting strong winds while alongside. As a precaution, he had an additional mooring line run at either end when berthing. Based on his experience, the master decided that a tug would also be required to keep the ship alongside if the winds from abeam exceeded 35 knots.
Both berths 11 and 12 have bollards along the wharf edge, but no bollards further back from the wharf face. Therefore, there was no provision for running effective breast lines from the winches, which were about 10 m above the wharf.
As the forecast only gave a predicted time of ‘late evening’ for the arrival of the adverse weather, the master did not know when, or even if, a tug would be required. Hence, the master waited until the wind reached 35 knots before calling for a tug. Had he been advised, by VTS, that the storm was approaching he could have called for a tug earlier so that it was ready to assist when the storm arrived.
When berthing, the master had instructed the mates to maintain equal tension in the mooring lines (as is usual) while alongside. When he received Fremantle Ports’ strong wind warning advice, he re-issued his earlier instruction. The master stated that he did not lower the outboard anchor as required because he was concerned about the possibility of the anchor fouling on the seabed.
The second and third mates stated that both additional lines were left on the drum ends to assist with maintaining an equal tension as, when transferring the lines on to the bits, the lines would lose tension. The additional line forward was ‘locked off’[21] so it could not slip. However, the additional line aft was not, and once the bollard failed, this line payed out with little resistance. The remaining stern line also payed out, as the brake did not, or could not, hold the line fast.
AAL Fremantle
When AAL Fremantle’s master received the port’s strong wind warning advice, he had the outboard anchor lowered. Based on his experience, he was confident the ship would remain safely alongside with the mooring lines used. Further, with a ship berthed astern, there were no available options to use a different mooring arrangement, or to move the ship to use other bollards.
Rail bridge risk assessment
Fremantle Ports
In May 2011, Parmelia I contacted the Fremantle Rail Bridge.[22] In response to a Western Australian Department of Transport investigation into that incident, Fremantle Ports completed the ‘Fremantle rail bridge impact assessment’ risk assessment. The assessment identified the following categories of operational risk:
the operation of Parmelia I at berth 12A
ships with a draught of less than 7 m
ships with a draught of greater than 7 m.
Those three categories were further divided into:
berthing
unberthing
a severe weather event while alongside.
The risk assessment determined that ships with draughts greater than 7 m could not contact the rail bridge because Wongara Shoal provided a barrier. The assessment determined that contact with the rail bridge was possible for the operations of Parmelia I and ships with a draught less than 7 m.[23] However, the analysis deemed the risks to be as low as reasonably practicable (ALARP) due to the existing operational risk controls.
In October 2013, the Public Transport Authority (PTA, the bridge operators) engaged DNV-GL to conduct a qualitative assessment (workshop) of Fremantle Ports’ risk mitigating measures for protecting the rail bridge.[24] That workshop considered Fremantle Ports’ and PTA’s procedures, hydrographic survey data, and previous risk assessments. The investigation report into the 2011 bridge contactby Parmelia I, and the resulting procedural changes, were also taken into account. The qualitative assessment was based on the ‘best judgement’ of workshop participants, who were considered subject matter experts.
When investigating risks associated with weather events, both Fremantle Ports’ and DNV-GL’s risk assessments dealt with the consequences of a ship breaking away from the berth. Neither assessment considered mitigating measures associated with keeping a ship alongside, or the ability of tugs and the port to assist a ship that was at risk of contacting the rail bridge – particularly when in the shallow water close to Wongara Shoal.
During winter in Fremantle, there is an increase in frequency and severity of weather events. A rise in the number of storms will directly affect the number of weather related occurrences in the port, and therefore, the outcome of the risk assessment. However, the frequency of severe weather was not considered when assessing the level of risk.
Fremantle Ports’ had undertaken computer-based ship modelling for berths 11 and 12 as part of its risk assessment. The process modelled the simultaneous failure of all mooring lines at both berths, with and without the outboard anchor lowered, which was considered the worst-case scenario. Bow-in berthing, port side alongside, was not modelled as it was considered unusual.
In summary, the risk assessments concluded that, while it was possible for a ship to make contact with the rail bridge. The existing procedures and practices, in combination with protection given by Wongara Shoal, indicated a risk rating of ‘tolerable’.[25]
Berth 11 and 12 operations
AAL Fremantle’s cranes were on the port side, so it berthed port side alongside, bow in. When the ship broke away, some of its forward mooring lines remained intact, and were under tension. As a result, the ship’s stern swung outward and towards the rail bridge.
As such, the ship’s breakaway was significantly different to the modelled scenario.
Similarly, Grand Pioneer’s breakaway event was not as modelled, as there was no simultaneous failure of all mooring lines. Further, the car carrier had a much more uniform wind loading than a ship with an aft accommodation superstructure. This characteristic would have resulted in the ship moving quite differently to the modelling. Grand Pioneer’s movement was consistent with the modelling only because its forward mooring lines remained in place.
Fremantle Ports’ risk assessment of operations, particularly at berths 11 and 12, did not assess all of the various ship/berthing operations carried out there. Consideration was only given to what was deemed to be the worst-case scenario. No consideration was given to methods of preventing ships breaking away from their berths. Further, there was no analysis of possible impediments to assisting a ship on, or close to, Wongara Shoal and the rail bridge.
Tidal variation
Shortly after the incident, Fremantle Ports put forward that a ‘meteo-tsunami’[26] had resulted in a sudden, 0.5 m tidal rise, causing Grand Pioneer to surge 12 m aft. The port claimed that this surging led to the ship’s ramp contacting the bollard, which then failed.
Low water (0.6 m) was predicted at 1612 on 17 August and high water (1.0 m) at 0323 on the following day. The height of tide in the hours before the storm had maintained a steady positive residual of between 0.35 and 0.40 m. It was not until after AAL Fremantle had contacted the bridge that the positive residual increased, peaking shortly before 2230 (Appendix B).
A buoy located near the bow of Parmelia I fitted with measuring instruments provided local current direction and flow. The recorded data indicated a consistent direction and flow of about 1 knot, until 2105, after which the data became erratic.[27]
The evidence shows that Grand Pioneer did not surge and only moved directly away the wharf, and there were no reports of any other ships surging at that time.
As such, the ATSB investigation concluded that a ‘meteo-tsunami’ or other related tidal variation did not contribute to the breakaway.
AAL Fremantle is a 140 m long general cargo ship with two 350 t cranes mounted on its port side. At the time of the incident, the ship was managed by Columbia Ship Management, Singapore.
The ship had a crew of 20 Russian, Ukrainian, Chinese and Philippine nationals. The master first went to sea in 1982 and had sailed as master since 2004; holding a Russian master’s certificate of competency. He had completed three, 4-month assignments on board AAL Fremantle before the incident.
The ship was moored with ten, 80 mm diameter, eight-strand polypropylene lines, each with a 95 t rated breaking load. All lines were run onto self-stowing winch drums and held by a manual brake. The ship’s mooring arrangement in Fremantle (Figure 4) did not include any breast lines due to the unsuitable position of the available bollards.
Grand Pioneer
Grand Pioneer is a 190 m long pure car and truck carrier. The ship’s main vehicle ramp is on its starboard quarter. Two smaller ramps are located amidships on the port and starboard sides. At the time of the incident, the ship was managed by Cido Shipping, South Korea.
The ship had a crew of 20 Philippine nationals. The master first went to sea in 1991 and since then had sailed mainly on car carriers. He held a Philippine master’s certificate of competency and had sailed as master since 2010.
The ship was moored with ten, 72 mm diameter, eight-strand polypropylene lines run onto split drum winches and held by a manual brake. Each mooring line had a 97 t rated breaking load. Two additional mooring line, one forward and one aft, were of the same type and left on the winches’ drum ends. The ship’s mooring arrangement in Fremantle (Figure 4) did not include breast lines due to the position of the available bollards, and the height (about 10 m) of the winches above the wharf.
Themain vehicle ramp was about 9 m wide and 30 m long. The ramp comprised three sections, and weighed approximately 90 t. When landing it on the wharf, the standard practice of the ship’s crew was to place two automotive tyres under the ramp’s footplate (Figure 8 and Figure 9). The ramp had no float[7] function, which meant its entire weight was on the wharf apron. The wires to fold/unfold and hoist/lower the ramp were left slack. The slack wires allowed for the vertical movement of the ship relative to the wharf without the need for continuous adjustment.
Parmelia I
Parmelia I is a 65 m long tanker servicing the bunker (fuel) requirements of ships in the Port of Fremantle. At the time of the incident, it was managed by Teekay Shipping, Australia, and manned by United Maritime, Australia.
On the night of 17 August, a skeleton crew manned the barge overnight, as there were no bunker deliveries scheduled.
Port of Fremantle
Fremantle is the principal commercial port for Western Australia and is situated at the mouth of the Swan River. The port comprises the inner harbour, within the estuary of the Swan River, and an outer harbour with three open roadsteads where ships can anchor.
Fremantle Ports maintains a 24-hour vessel traffic service (VTS)[8] which, in accordance with international guidelines aims to provide vessels using the port with an information service (INS)[9] and a traffic organization service (TOS).[10] As part of its function, the service organises and manages traffic within its VTS area, and provides essential information related to shipping movements.
The duty VTS officer (VTSO) is required to maintain a continuous watch: monitoring shipping traffic and providing ships, pilots, tugs and others with information. The VTSO also provides service users information of shipping movements, allocation of berths, weather and other matters related to the safety of navigation within port waters.
Amongst the inner harbour berths, numbers 11 and 12 are common-user berths providing 429 m of berthing space. The reinforced wharf aprons have substantial stacking areas, making the wharf suitable for handling general, break-bulk and heavy lift cargoes. The wharf also provides suitable parking space for large numbers of vehicles.
The proximity of Wongara Shoal to berth 12 meant that only smaller ships could use that berth, while larger ships use berth 11 and are generally berthed stern-in. Berth 12 A, adjacent to berth 12 and near the Fremantle Rail Bridge, is reserved for Parmelia I (Figure 3).
Fremantle Rail Bridge
The Fremantle Rail Bridge is a 260 m long, double line, dual gauge (down direction only), rail-only bridge that spans the Swan River at the eastern end of the Port of Fremantle. Two steel girders (transoms), which sit across eight concrete piers, support the track 10 m above the water. The ends of the wooden railway sleepers overhang the sides of the transoms.
The bridge services the electrified[11] suburban train system, and the freight requirements of Fremantle’s inner harbour. Typically, it caters for 158 suburban and 30 freight movements per day.
The bollard
The bollard that failed was a Bean Type II 10-A-21 (Figure 6). It had been manufactured from flake graphite grey cast iron by EJ Bean, England. It weighed approximately 585 kg, and had a safe working load of about 95 t.
The bollard was mounted in a recess in the wharf apron and held down by bolts. Once in place, grout was pumped into the space under the bollard through a grouting hole (Figure 12) that was cast into the bollard’s top face. Steel plates covered the gap between the wharf apron and the base of the bollard.
No records were available to indicate when the bollard had been manufactured or installed. Production of this bollard model started in 1959 and stopped some time before 2001.
Fremantle Ports’ ongoing maintenance program includes a periodic visual inspection of bollard sections above the wharf apron and their fixings underneath the wharf. A 65 t bollard pull harbour tug is used for load testing the bollards.
Figure 6: Side, front and plan elevations of the bollard
At 1712[1] on 17 August 2014, the 190 m long car carrier, Grand Pioneer (Figure 1), berthed in the port of Fremantle at berth 11 (Figure 3). Berthed astern of the car carrier was the 140 m long general cargo ship, AAL Fremantle (Figure 2). It had been in port for the past 10 days, unloading and then loading cargo using its two, port-side mounted cranes.
Figure 1: Grand Pioneer
Source: Marcus, Shipspotting
Figure 2: AAL Fremantle
Source: Australian Transport Safety Bureau
At the time, a Bureau of Meteorology (BoM) strong wind (26 to 33 knots)[2] warning was in place for Perth local waters, which include Fremantle. The warning, issued that morning, included the caution ‘scattered thunderstorms in the late evening, possibly severe with damaging winds’.[3]Grand Pioneer’smaster had been monitoring the weather forecasts, and in response to the strong wind warning, the ship’s mooring arrangement had been enhanced to include an additional head line and stern line.
Fremantle Port Authority (Fremantle Ports) had taken its standard precautions in response to the BoM forecasts – advising the masters of berthed ships of the strong wind warning.[4] The port’s advice included asking that masters lower their ship’s outboard anchors to the seabed. AAL Fremantle’s master had its anchor lowered but Grand Pioneer’s master decided not to do so.
Figure 3: Section of navigational chart Aus 113 showing involved ships at their berths, and the movement of ships that broke away
Source: Australian Hydrographic Service, annotated by ATSB.
At 1900, Grand Pioneer’s cargo unloading started using the vehicle ramp on its starboard quarter. As forecast, the wind was from the north at 28 to 33 knots, on the ship’s starboard beam.
At 2200, Grand Pioneer’s master was on the ship’s navigation bridge attending to paperwork as the forecast thunderstorm approached. He saw that the anemometer (above the bridge front windows) indicated that the wind had increased to 35 knots. He radioed the duty mate to stop vehicle traffic on the ramp in case the ship moved off the wharf. He then telephoned the ship’s agent, and as a precaution, requested a tug to assist in keeping the ship alongside if required.
At 2202¾, AAL Fremantle’s stern moved slightly off the wharf and the tension in the mooring lines increased.
At that time, Grand Pioneer’s third mate was at the top of the ramp. As he finished checking the aft spring lines, a stern line parted. The third mate then noticed that two other stern lines, on a different bollard to the parted line, were slack and in the water. He then hurried up the ramp, attempting to stop the cars on it in case the ramp fell off the wharf.
At about 2203, Grand Pioneer’s master noted that the wind was now gusting to 55 knots. Looking aft, he saw the ship’s stern moving rapidly away from the wharf and the ramp sliding over the edge. He immediately called Fremantle vessel traffic service (VTS) via VHF radio to request immediate assistance. When he received no response, he contacted the agent and asked for assistance. He then instructed his crew to standby at their forward and aft mooring stations and requested the ship’s main engine and bow thruster to be prepared for immediate use.
AAL Fremantle’s master was in his cabin when he heard the sudden rise in wind speed. He went to the aft facing cabin window and saw his ship’s stern moving away from the wharf. As he could not see the ship’s three stern lines, the master immediately went to the bridge and started preparations to manoeuver the ship.
Figure 4: Indicative diagram showing the ships’ mooring arrangements (not to scale)
Source: ATSB
By about 2203, Grand Pioneer’s vehicle unloading had ceased, with one car stopped on the ramp after its driver ran back up to the vehicle deck. At this time, a northbound commuter train was passing over the nearby Fremantle Rail Bridge.
At 2203½, Grand Pioneer’s stern had moved sufficiently away from the wharf to allow the ramp to fall off the wharf. A stern line, after running off its winch drum, fell onto the ramp. The line then caught under a rear tyre of the stranded car, pulling the car down the ramp and turning it about 90 degrees before the line came free and fell into the water.
Shortly after, AAL Fremantle had swung to lie at about right angles to its berth. As the ship moved astern, a fender near the lower lead light (Figure 4) fouled the ship’s starboard anchor cable and the anchor was dragged onto the wharf apron. At about 2207, as the anchor was dragged across the apron, it brought down the lead light.
At about 2210, AAL Fremantle’s starboard quarter struck a stanchion[5] on the rail bridge, bringing it down and cutting power to that section of track. Temporary scaffolding on the side and under the bridge was also damaged, and the rail track’s alignment was disturbed (Figure 5). Moments later, AAL Fremantle, about midships, contacted the starboard shoulder of Parmelia I (Figure 3) at the adjacent berth. AAL Fremantle’s two remaining head lines were slack and the remaining forward spring line was still under tension (Figure 4). At that point, the ship’s astern movement stopped in the proximity of Wongara Shoal (Figure 3).
Response
Shortly after, AAL Fremantle’s master ran the main engine at dead slow ahead to maintain the ship’s position, and avoid putting any further weight on the rail bridge.
By 2210, Grand Pioneer had swung until its port shoulder was about 15 m from Singapore Bridge, a container ship at berth 10 (Figure 3). Grand Pioneer’s head lines and forward spring lines had held. The master let go the port anchor and used the bow thruster and main engine to maintain the ship’s position.
At 2215, the Public Transport Authority (PTA) train controller received a phone call from Brookfield Rail. [6] Brookfield Rail advised that the police had notified it of a ship striking the rail bridge.
The PTA then enacted their safety plan, which included terminating all southbound trains at North Fremantle station, and calling out overhead-staff to inspect the section of track. At this time, the next train due to cross the bridge was a southbound train was 13 minutes away from the rail bridge. The train was stopped at North Fremantle station.
At 2235, a tug arrived and began assisting Grand Pioneer. At 2245, a pilot boarded AAL Fremantle and other tugs arrived to assist both ships.
At about 2300, a pilot boarded Grand Pioneer and a second tug made fast to the ship in preparation for returning the ship to its berth. The weather had moderated and the wind had decreased to about 25 knots (Appendix B).
Figure 5: AAL Fremantle near the Fremantle rail bridge
Source: Fremantle Ports, annotated by ATSB
At about 0100 on 18 August, Grand Pioneer was all fast at its berth. One tug was retained to stand by the high-sided car carrier in the windy conditions. At about this time, AAL Fremantle’s fouled anchor was cleared from the lower lead light’s tower.
At 0115, two tugs had been made fast to AAL Fremantle. The aft tug’s towline was made fast to a ship’s mooring line, as there was insufficient water over Wongara Shoal for the tug to approach the ship.
By 0300, AAL Fremantle was all fast at berth H on South Quay (Figure 3).
Following the incident, the PTA closed the rail bridge to all traffic until necessary repairs and inspections were completed. Passenger rail services resumed on 1 September and freight services on 5 September 2014.
Charitha Pattiaratchi, Sarath Wijeratne, Meteo-tsunamis along the West Australian coastline, University of Western Australia.
Exceed Consulting, 9 September 2014, Failure investigation of bean type bollard from Fremantle wharf.
Submissions
Under Part 4, Division 2 (Investigation Reports), Section 26 of the Transport Safety Investigation Act 2003 (the Act), the Australian Transport Safety Bureau (ATSB) may provide a draft report, on a confidential basis, to any person whom the ATSB considers appropriate. Section 26 (1) (a) of the Act allows a person receiving a draft report to make submissions to the ATSB about the draft report.
A draft of this report was provided to the Australian Maritime Safety Authority (AMSA), Fremantle Ports, the master, managers and crew of AAL Fremantle, the master, managers and crew of Grand Pioneer, the Public Transport Authority (PTA), the Bureau of Meteorology (BoM) and Dr Charitha Pattiaratchi.
Submissions were received from AMSA, the managers of Grand Pioneer and AAL Fremantle, the master of AAL Fremantle, BoM, the PTA and Dr Charitha Pattiaratchi. The submissions were reviewed and where considered appropriate, the text of the report was amended accordingly
The safety issues identified during this investigation are listed in the Findings and Safety issues and actions sections of this report. The Australian Transport Safety Bureau (ATSB) expects that all safety issues identified by the investigation should be addressed by the relevant organisation(s). In addressing those issues, the ATSB prefers to encourage relevant organisation(s) to proactively initiate safety action, rather than to issue formal safety recommendations or safety advisory notices.
Depending on the level of risk of the safety issue, the extent of corrective action taken by the relevant organisation, or the desirability of directing a broad safety message to the marine industry, the ATSB may issue safety recommendations or safety advisory notices as part of the final report.
Where relevant, safety issues and actions will be updated on the ATSB website as information comes to hand. The initial public version of these safety issues and actions are in PDF on the ATSB website.
Fremantle Ports’ risk assessment
Fremantle Ports’ assessment of risks associated with a ship contacting the Fremantle Rail Bridge as a result of a breakaway, particularly from berths 11 and 12, was limited. Preventing a breakaway from berths where the wind was likely to be on a ship’s beam had not been considered. Similarly, the impediments to assisting a ship near Wongara Shoal after a breakaway had not been assessed.
The Bureau of Meteorology (BoM) marine forecast title of ‘strong wind warning’ understated the ‘damaging winds’ expected during the ‘severe thunderstorm’. The forecast did not use recognised marine weather terms for wind speed, such as ‘gale force’.
Fremantle Ports’ procedures for adverse weather were not adequate for weather that could reasonably be expected to occur. Some procedures could not be reasonably implemented and other were not monitored for compliance.
Fremantle Ports’ staff did not understand the significance of some wind and weather terminology used in the BoM forecast. Consequently, port procedures triggered by a BoM ‘gale’ or ‘severe weather’ warning such as preparing the tugs and calling the harbour master were not followed.
Appendix A – Qualitative analysis report risk matrix
Appendix B – Wind, tide and current information[31]
Appendix C – Bureau of Meteorology weather service and terms
The Australian Bureau of Meteorology (BoM) provides the maritime community with weather forecasts, warnings and observations for coastal waters areas and high seas around Australia.
Marine and local waters forecasts
Forecasts for wind speed and direction, and sea and swell heights are issued routinely every 12 hours. When a marine wind warning is in force, these forecasts are updated routinely every 6 hours. These forecasts provide information on mean wind, sea height, swell height and direction and weather conditions:
coastal waters – areas within 60 miles of the coast
high seas – areas beyond the coastal waters.
The following terminology is used to describe time periods in these forecasts:
Early in the morning - expected to occur before 6am.
In the morning - expected to occur between 7am and 10am.
Middle of the day - expected to occur between 11am and 1pm.
During early afternoon - expected to occur between 2pm and 3pm.
In the afternoon - expected to occur between 4pm and 6pm.
During the evening - expected to occur between 7pm and 8pm.
Later in the evening - expected to occur after 9pm
All BoM marine forecasts and warnings predict wind speed as the average (or mean) speed expected over any given 10 minute period at a height of 10 m above the ground. Hence, marine forecasts provide a summary of the average wind speed expected within the forecast area, not the maximum wind gust expected.
Wind warnings are only issued when mean winds (winds speed averaged over 10 minutes) are expected to exceed 25 knots for an extended period of time across more than 10 per cent of the marine area. A wind warning is not issued for individual wind gusts, which typically last seconds.
In the absence of severe weather, such as thunderstorms or squalls, it is normal for wind speed to vary by up to 40 per cent over a 10 minute period. Hence, a statement is added to BoM marine forecasts to remind users that wind gusts at any time can be 40 per cent higher than the mean wind predicted in BoM marine forecasts and warnings. Wind gusts can be much higher than 40 per cent during severe weather, such as thunderstorms or squalls.
Mariners should refer to the weather section of the coastal or local waters forecast to determine the likelihood of thunderstorms occurring. Mariners should be aware that there is a risk of damaging wind gusts occurring with any thunderstorm and should actively monitor weather conditions when they are forecast. Weather conditions can be monitored by checking current weather observations, weather watch radar and satellite imagery
For marine operations near the shore, BoM’s severe thunderstorm warning service should also be used. These warnings provide short term advanced notice of thunderstorms that are likely to be accompanied with wind gusts in excess of 48 knots, as well as other hazardous conditions that can accompany thunderstorms.
More information about BoM marine wind forecasts is available on its website
Marine weather warnings
Marine weather warnings are issued whenever strong winds, gales, storm force or hurricane force winds are expected. The following warnings are provided:
coastal waters wind warnings
ocean wind warnings – issued to ships at sea whenever gale, storm or hurricane force winds are expected
severe weather warnings – provided for potentially hazardous or dangerous weather that is not directly related to severe thunderstorms, tropical cyclones or bushfires.
Gust speed can be 40 per cent greater than the predicted wind speed
Marine wind warnings aim to provide a 24-hour lead-time and are normally reviewed every 6 hours and issued every 12 hours.
Wind speed warnings
These warnings are issued whenever strong winds, gale, storm or hurricane-force winds are expected. They provide around 24 hours’ notice and are updated every 6 hours. These warnings are based on expected 10 minute mean winds (not maximum wind gusts) and are only issued if mean winds are expected to exceed thresholds for extended periods.
The following 10 minute mean thresholds are used:
Strong wind warning: 26 to 33 knots, Force 6 to 7
Gale warning: 34 to 47 knots, Force 8 to 9
Storm force wind warning: 48 to 63 knots, Force 10 to 11
Hurricane force wind warning: 64 knots or more, Force 12.
Wind directionis given using the eight compass points for forecasts and 16 points for observations.
Recipients are further cautioned that maximum wave heights may be up to twice the height of those forecast (average). More information is available via BoM's website.
Severe thunderstorm warnings
They provide short term advance warning of the likelihood of severe thunderstorms impacting on the region and are updated every three hours and can produce any of the following:
Winds having a sustained wind speed between 63 km/h (34 knots) and 88 km/h (47.5 knots) with gusts between 90 km/h (48.5 knots) and 125 km/h (67.5 knots).
Severe weather
Sustained winds of 34 knots or more with gust of 48 knots or more and very heavy rain.
Purpose of safety investigations & publishing information
Purpose of safety investigations
The objective of a safety investigation is to enhance transport safety. This is done through:
identifying safety issues and facilitating safety action to address those issues
providing information about occurrences and their associated safety factors to facilitate learning within the transport industry.
It is not a function of the ATSB to apportion blame or provide a means for determining liability. At the same time, an investigation report must include factual material of sufficient weight to support the analysis and findings. At all times the ATSB endeavours to balance the use of material that could imply adverse comment with the need to properly explain what happened, and why, in a fair and unbiased manner. The ATSB does not investigate for the purpose of taking administrative, regulatory or criminal action.
Terminology
An explanation of terminology used in ATSB investigation reports is available here. This includes terms such as occurrence, contributing factor, other factor that increased risk, and safety issue.
Publishing information
Released in accordance with section 25 of the Transport Safety Investigation Act 2003
Ownership of intellectual property rights in this publication
Unless otherwise noted, copyright (and any other intellectual property rights, if any) in this report publication is owned by the Commonwealth of Australia.
Creative Commons licence
With the exception of the Coat of Arms, ATSB logo, and photos and graphics in which a third party holds copyright, this publication is licensed under a Creative Commons Attribution 3.0 Australia licence.
Creative Commons Attribution 3.0 Australia Licence is a standard form licence agreement that allows you to copy, distribute, transmit and adapt this publication provided that you attribute the work.
The ATSB’s preference is that you attribute this publication (and any material sourced from it) using the following wording: Source: Australian Transport Safety Bureau
Copyright in material obtained from other agencies, private individuals or organisations, belongs to those agencies, individuals or organisations. Where you wish to use their material, you will need to contact them directly.
On 6 August 2014, at about 0500 Eastern Standard Time (EST), the pilot of a Fairchild SA227 aircraft, registered VH-UZI, conducted a pre-flight inspection and weight and balance calculations for a freight charter flight from Rockhampton to Townsville, Queensland.
During the take-off run, the pilot reported all indications and performance were normal passing 70 knots. Passing V1, as the pilot increased the back pressure on the control yoke to rotate the aircraft for take-off, the control column felt heavy and the aircraft nose wheel did not lift off the ground. The pilot continued to increase the back trim and back pressure on the control yoke and the ‘out of trim’ warning sounded. The pilot rejected the take-off, applied maximum braking and reverse thrust.
After taxiing the aircraft back to the parking bay, the pilot requested the freight be re-weighed. The pilot then recalculated the aircraft weight and balance with the actual freight distribution and found the centre of gravity slightly more forward than the original load sheet position.
The pilot set the stabiliser trim gauge in the cockpit to read a nose up attitude, and then externally inspected the position of the stabiliser. He observed the stabiliser in a neutral position and therefore determined that the gauge did not accurately indicate the stabiliser position. The pilot assessed that the combination of the incorrectly loaded freight causing a more forward centre of gravity and the inaccurate stabiliser trim gauge led to the out of trim warning and overly heavy control pressure required for the attempted take-off.
The stabilizer trim potentiometer in the aircraft was replaced which resolved the fault.
The ATSB has investigated a number of accidents and incidents while maintenance work was being performed on or near railway tracks. The ATSB SafetyWatch, introduced in 2012 to emphasise broad transport safety concerns in Australia, also highlighted 'safe work on rail'.
The ATSB continues to receive notifications of safe working incidents involving worksite protection arrangements for work on track. These notifications suggest the existence of broader safety issues associated with work on track that continue to increase risk to worker safety. In 2017, safe work on track continues to be an ATSB SafetyWatch priority.
This safety issue investigation reviews available data from across Australia of incidents and accidents relating to work on track. It is designed to provide industry with insights into the protection arrangements that are failing, and the reasons why, across many occurrences so that safety action can be designed to reduce future safe work on track occurrences.
What the ATSB found
The ATSB analysis grouped the notifiable occurrence data into eleven categories. The analysis indicated the most common events exposing track workers to highest risk, were:
the incorrect removal of the worksite protection
the incorrect positioning of the worksite protection
the type of protection being insufficient or incorrect, and
the incorrect identification of the worksite location.
The results of this safety issue investigation were largely reflective of the safety factors identified from previous ATSB occurrence investigations. That is, incidents were predominately a result of errors during the implementation or dissolution stage of providing track protection. Protections were either removed incorrectly or prematurely, or key communication exchanges failed to establish the location of the worksite with respect to approaching rail traffic.
What's been done as a result
The outcome of this ATSB safety issue investigation suggests that the rail industry should consider the event types identified above in determining areas in which to target effort for maximising the effectiveness of safety arrangements for work on track.
Rail transport operators continue to enhance arrangements within their networks that facilitate safe work on track. Work to share learnings between operators and to deliver better safety outcomes across the industry is also ongoing through industry initiatives such as the National Track Worker Safety Forum.
This forum has identified priority areas and is exploring improvements in worker competencies, technologies for worksite protection systems, compliance with critical communications protocols and addressing interface arrangements where differing rules and procedures exist between adjoining networks – particularly in sidings and yards.
Safety message
To minimise risk, rail transport operators must ensure systems for safe work on track encourage workers accessing the rail corridor to communicate sufficient information to validate their worksite location, the adequacy of the protections in place, and their positioning in relation to any approaching train movements.
Context
Conducting work on or near a railway track can be dangerous, especially if safe working rules and procedures have not been correctly implemented to protect the worksite. Trains cannot stop quickly and any breakdown in the communication or management of a worksite can leave workers extremely vulnerable to dangerous situations.
The Australian Transport Safety Bureau (ATSB) has investigated a number of accidents and incidents while maintenance work was being performed on or near railway tracks (Appendix A). As a result, the ATSB SafetyWatch, introduced in 2012 to emphasise broad transport safety concerns in Australia, highlighted 'safe work on rail'.
The ATSB continues to receive notifications of safe working incidents involving worksite protection arrangements for work on track. These notifications suggest the existence of broader safety issues associated with work on track that continue to increase risk to worker safety. In 2017, safe work on track continues to be an ATSB SafetyWatch priority.
This safety issue investigation reviews available data from across Australia of incidents and accidents relating to safe work on track. It is designed to provide industry with insights into the protection arrangements that are failing, and the reasons why, across many occurrences so that safety action can be designed to reduce future safe work on track occurrences.
Safe work on track background
The provision of a safe workplace for workers to undertake construction, maintenance or inspection tasks is a fundamental objective of a rail transport operator (RTO)[1] in managing their railway network. It is also a legislated requirement under both the Rail Safety National Law as well as the harmonised Work Health and Safety (WHS) legislation.
It is impractical for a RTO to exclude rail traffic on every occasion that work on track is required. The railway manager must therefore implement risk controls to ensure the safe separation of workers and rail traffic. The risk controls are generally hierarchical and form an integral part of the operational rules, procedures and instructions within the RTO’s safety management system.
There are many RTO’s within Australia, each with management systems typically tailored to address the risks, and the operational/infrastructure requirements of their network. This environment of multiple RTO’s and multiple safety management systems can result in a variation in risk controls (rules and procedures). Although there may be variances in the methodologies, the overall objective and approach to ensuring safe separation of workers and rail traffic is generally consistent between RTO’s.
RISSB Australian Network Rules and Procedures (ANRP)
The hierarchical safety measures adopted by all RTO’s are generally consistent with the published suite of Australian Network Rules and Procedures (ANRP) developed by the Rail Industry Safety Standards Board (RISSB). Track workers planning work in a rail corridor must assess the implications to safety and the potential for their work to intrude into the danger zone (Figure 1). Usually, any work in the danger zone must not:
be carried out unless there is a safe place that can be easily reached
begin until the required safety measures are in place.
Work in the danger zone must be carried out using the safety measures detailed within the RTO’s selected arrangements for providing an authority or a means of protection. The safety measures selected are heavily dependent on the intended type of work activity and machinery used.
Figure 1: The danger zone
The RISSB have published five sets of work on track rules and procedures. These range from authorities, which provide for exclusive occupancy for workers and rail service traffic is excluded from the worksite, to methods of protection, where systems are implemented to ensure workers are provided sufficient warning before the arrival of a train.
Local possession authority (LPA)
A LPA authorises the closure of a defined portion of track for a specified period, giving exclusive occupancy to a possession protection officer (PPO).[2] Work or the movement of rail vehicles within the portion of track must only be done with the agreement of the possession protection officer.
Track occupancy authority (TOA)
A TOA authorises a protection officer (PO)[3] to occupy a track section for an agreed time period. A TOA gives exclusive occupancy, though joint occupancy is permitted under defined conditions. Only rail traffic authorised and associated with the TOA may enter the specified limits of the TOA.
Track work authority (TWA)
A TWA authorises a protection officer the occupation of a defined portion of track between rail traffic movements. A TWA does not give exclusive occupancy of the defined portion of track. Rail traffic crews must follow instructions given by hand signallers or the protection officer.
Absolute signal blocking (ASB)
An ASB excludes rail traffic from a portion of track by placing and holding a controlled signal[4] at stop. It is used for work in the danger zone where a risk assessment shows a work on track authority is not necessary.
Lookout working
In lookout working, the protection officer assigns workers whose sole task is to look for approaching rail traffic and provide an immediate warning to any workers in the danger zone. The warning must be provided in sufficient time to allow workers to clear the danger zone at least 10 seconds before the arrival of the oncoming rail traffic.
RISSB note that lookout working is the least preferred protection method and ASB should be used if practical.
Work on track investigations and identified safety factors
Between September 2006 and October 2015, the ATSB investigated 12 work on track occurrences. A summary of each ATSB investigation, and the associated safety factors, are detailed in Appendix A – ATSB work on track occurrence investigations.
Of the 12 investigations, about half involved a collision at the worksite location of rolling stock with a person. Two resulted in the death of a track worker.
The safety factors identified from the ATSB investigations related predominately to errors that occurred during the implementation or dissolution stage of providing track protection. However, there were occasions where a worker intentionally acted in violation of the prescribed rules or procedures.
A significant portion of the errors occurred during the key communication exchanges between the protection officer and the network controller when establishing the location of the worksite with respect to approaching rail traffic. Errors also occurred during communication between these parties resulting in incorrect or premature removal of protections.
A number of safety factors were associated with limitations within rules or procedures to prevent/detect mistakes or omissions. Although rail transport operators have implemented improvements to arrangements for safe work on track, the ATSB has continued to receive notifications of safe working incidents involving worksite protection arrangements.
Comparable research
In August 2015 the United Kingdom Rail Accident Investigation Branch (RAIB) published a report entitled Class investigation into irregularities with protection arrangements during infrastructure engineering work. The investigation analysed occurrences involving engineering work carried out, or planned to be carried out, in a two year period between April 2011 and April 2013. A total of 714 events involving operating irregularities during infrastructure engineering work were identified for this time period. Of these events, 86 per cent were able to be grouped according to the type of operating irregularity. Of these, 71 per cent were categorised into one of nine significant event categories. Although the event categories uses in the RIAB report differed from those used in this report (discussed below), there were similarities and some overlap of the categories chosen for both reports. For example the most common event category identified in the RIAB report (33%) was ‘protection equipment incorrectly placed’. This was followed by ‘work carried out without protection’ (13%), and ‘working outside a protected area’ (12%).
PPO: Possession protection officer. Qualified worker to whom train control transferred control of a section of track and who is responsible for coordinating protection of worksites under a LPA.
PTA: Public transport authority
RIM: Rail infrastructure manager
RISSB: Rail Industry Safety Standards Board
RTO: Rail transport operator. A rail infrastructure manager, or a rolling stock operator, or a person who is both a rail infrastructure manager and a rolling stock operator.
TOA: Track occupancy authority
TRI: Train running information
TWA: Track working authority
Findings
From the analysis of notifiable occurrence records for the 5 year period between July 2009 and June 2014 that were classified as a safe working (network) rule or procedure breach, the following findings are made with respect to the occurrence notifications related to safe work on track. These findings should not be read as apportioning blame or liability to any particular organisation or individual.
Analysis of occurrence data indicated the most common events exposing track workers to highest risk, were:
the incorrect removal of the worksite protection
the incorrect positioning of the worksite protection
the type of protection being insufficient or incorrect, and
the incorrect identification of the worksite location.
If an occurrence taxonomy with sufficient fidelity was employed consistently to occurrences reported by all states and territories, future analysis, such as that presented in this report, could be available to industry in a more timely manner. With such a system in place, this analysis could also include occurrence-type trend analysis to monitor high risk occurrences, and inform stakeholders of any new developing trends.
Sources and submissions
Sources of information
The sources of information during the investigation included the:
Department of Transport and Main Roads (Rail Unit), Queensland
Independent Transport Safety Regulator, New South Wales
Office of the National Rail Safety Regulator
Office of Rail Safety Department of Transport, Western Australia
Rail Industry Safety and Standards Board
Transport Safety Victoria, State Government of Victoria.
References
Classifying Notifiable Occurrences, Occurrence Classification Guideline (OC-G1), Office of the National Rail Safety Regulator, Version 1.1 Submissions.
Submissions
Under Part 4, Division 2 (Investigation Reports), Section 26 of the Transport Safety Investigation Act 2003 (the Act), the Australian Transport Safety Bureau (ATSB) may provide a draft report, on a confidential basis, to any person whom the ATSB considers appropriate. Section 26 (1) (a) of the Act allows a person receiving a draft report to make submissions to the ATSB about the draft report.
A draft of this report was provided on the ATSB website for public review before a final report is issued.
Submissions were received from the Office of the National Rail Safety Regulator. The submission was reviewed and where considered appropriate, the text of the report was amended accordingly.
Appendices
Appendix A – ATSB work on track occurrence investigations
RO-2015-002[8] - Collision between track worker and passenger train at Guildford, WA
On 10 February 2015, a Public Transport Authority (PTA) maintenance crew commenced work at Meadow Street, Guildford, Western Australia. The crew’s assigned tasks included maintaining the pedestrian gates adjacent to the level crossing. At about 1035, one of the track workers was struck by a Perth bound suburban passenger train. The track worker sustained fatal injuries.
RO-2015-019 - Track worker struck by a passenger train near Laverton station, Vic.
On 2 October 2015, track workers planned to undertake dog-spike removal works in preparation for sleeper replacement works. At around 0910, the supervisor commenced marking the track to identify those dog-spikes to be removed. A lookout had been stationed for his protection. At about 0916, a suburban commuter train approached the worksite. The lookout observed the train, warned workers of its approach and signalled to the driver that the track was clear. However, as the train took the crossover, the supervisor was foul of the track, and was struck by the train that was travelling at about 59 km/h. The supervisor suffered serious injuries.
RO-2013-025 - Transport Safety Safe working irregularity at Glenrowan, Vic.
On 29 October 2013, ballast redemption and drainage maintenance works was undertaken on the bidirectional west track between Benalla and Wangaratta in Victoria. Flagmen and audible track warning signals (ATW) were placed on the east track at either end of the worksite to protect workers from any rail traffic travelling on the adjacent east track.
At about 0730, a passenger train approached the worksite on the east track. The track force coordinator[9] instructed the flagmen on the east track to remove the ATWs and allow the train to pass the work site unrestricted. No notice was provided to the work site supervisor or any of the track workers that protections were removed or that rail traffic was approaching.
On this occasion, the work site supervisor saw the approaching rail traffic and alerted the track workers who took action to ensure that they and their machines were clear of the track. The driver of the train was not aware of the near-miss incident and passed the worksite without incident.
RO-2013-018 - Safe working Breaches at Blackheath, Newcastle and Wollstonecraft, NSW
During June and July 2013, three separate safe working breaches occurred on the Sydney Trains network in NSW involving the application of Network Rule NWT 308 absolute signal blocking and Network Procedure NPR 703 Using Absolute Signal Blocking. The incidents occurred at Blackheath on 13 June 2013, Newcastle on 13 July 2013 and Wollstonecraft on 17 July 2013. In each case, trains were to be excluded from worksites, as part of worksite protection arrangements, using the absolute signal blocking (ASB) rule and procedure.
The rule and procedure were not adhered to by the signaller or protection officer during the authorisation of the ASB resulting in trains entering or passing through the worksites from which they should have been excluded.
RO-2013-017 - Safe working breach involving a Local Possession Authority, Revesby, NSW
On 10 July 2013, a pre-planned and advertised local possession authority (LPA) was implemented on the Up Main line between Revesby and Turrella, NSW. Approximately 30 minutes after the LPA was implemented, a passenger train entered the limits of the possession area and immediately ran over railway track signals (detonators) and was brought to a stand.
RO-2011-018 - Collision between an empty coal train and a track mounted excavator near Maitland, NSW
On 20 December 2011, an empty coal train collided with an excavator that was being used for scheduled maintenance of rail lines near the High Street Station at Maitland, NSW. The excavator was extensively damaged. Neither the train drivers nor the track workers were injured in the collision. The lead locomotive incurred only minor damage and was able to continue on its journey after a crew change.
RO-2011-011 - Collision between freight train and road-rail vehicle near Menindee, NSW
On 13 July 2011, a freight train collided with a road-rail vehicle in the Kaleentha to Menindee section of track, western NSW. The road-rail vehicle, a station wagon, was extensively damaged. The lead locomotive of the train incurred only minor damage.
RO-2011-006 - Collision between freight train and a track mounted excavator near Jaurdi, WA
On 28 March 2011, a collision involving a freight train and a track mounted excavator occurred between Jaurdi and Darrine, WA. The train driver sustained a minor injury. There was significant damage to the lead locomotive and the excavator, and minor damage to the track as a result of the accident.
On 4 August 2010, a safe working incident occurred within the Junee station yard limits when a locomotive was moved from one road to another without authority, while a track occupancy authority (TOA) was in force. The drivers were aware that the southern end of the Junee yard was closed to rail traffic but were unaware that a TOA was also applicable at the northern end of the yard.
RO-2010-004 - Collision between passenger train and a track-mounted excavator near Newbridge, NSW on 5 May 2010
On 5 May 2010, a collision occurred between a passenger train and a track-mounted excavator near Newbridge, NSW. The operator of the track-mounted excavator was fatally injured. During the course of the investigation, a similar incident occurred near Wards River, NSW (17 March 2011), where two work groups had to hurriedly vacate their on-track worksite due to an approaching train. There were no injuries. Both incidents occurred despite the fact that the work groups had been authorised, under a TOA, to occupy and work on the track.
20060011 - Collision between freight train and an elevated platform vehicle at North Geelong, Vic
On 26 October 2006, an Adelaide to Melbourne bound freight train collided with an elevated platform vehicle at the Separation Street overpass at North Geelong, Vic. At the time of the collision, an employee was undertaking maintenance on the support beams of the overpass. He was working from within the basket of the elevated platform vehicle underneath the overpass and directly above the standard gauge track when struck by the train. The maintenance worker received serious injuries as a result of the collision.
2006008 - Collision between freight train and a track mounted excavator, Vic
On 25 September 2006, a Melbourne bound freight train travelling from Perth, collided with a track mounted excavator conducting track-work near Inverleigh, Vic. There were no injuries and only minor damage to both the train and the excavator as a result of the collision.
Common safety factors identified across ATSB investigations
The investigations above identified a variety of safety factors associated with each occurrence. The processes for planning to safely access the danger zone were routinely called upon by work crews. It was apparent that in implementing selected method of authority or means of protection applicable, common errors were repeated.
The identification of antecedents to these common errors is essential to improve the effectiveness of safe working procedures for work on track.
To highlight the likely origin of the safety factors identified with the breakdown, the relevant work on track method were grouped based on the factor being a characteristic of the organisation or system (organisational influence), a characteristic of an individual (human influence) or the operational environment at the time (environmental influence).
Local Possession Authority rules
Organisational influence:
Validation processes were not effective in detecting errors in safe work documents prior to the implementation by key stakeholders (RO-2013-017).
Conflicting information in safe work documents used during the planning of worksite possessions was not detected (RO-2013-017).
Control centre handover procedures were inadequate in ensuring the identification of correct possession authority limits (RO-2013-017).
Site pre-work briefings did not include dissemination of train running information or site protection arrangements (RO-2011-006).
Human influence:
Rules and procedures were not followed during the implementation by key stakeholders (RO-2013-017, RO-2011-006).
Non-compliances to the repeat back provisions when implementing a worksite possession (RO-2013-017). Similar factor was present in occurrence associated with CSB (RO-2011-018).
Environmental influence:
On site communications equipment was inadequate to effectively communicate with the key stakeholders (RO-2011-006).
Track Occupation Authority rules
Organisational influence:
Procedure did not require protection officer to inform all persons or workgroups within the possession boundary of its existence (RO-2010-007).
Procedure did not accurately identify boundary of possession or type of work being undertaken (RO-2010-004).
Track workers provided with insufficient training in relation to the hazards and required protections (RO-2010-004).
At times throughout the network controller roster, fatigue levels were conducive to performance degradation (RO-2010-004).
Human Influence:
Track workers were using unauthorised reproductions (uncontrolled copies) of forms (RO-2010-004).
Protection officer and network controller incorrectly concluded that the train had passed beyond the limits of the worksite (RO-2010-004). Similar factor was present in occurrence associated with CSB (RO-2011-008).
Train driver had not adequately prepared the train to stop prior to the worksite given the prior knowledge of the track work location, route experience and knowledge of the trains handling characteristics (2006008).
Environmental Influence:
Consideration of track local conditions (grade, line of sight) when determining the location of inner or outer lookouts (RO-2014-004, 2006008).
Controlled Signal locking/Absolute Signal Blocking
Organisational influence:
Rule and procedure did not provide any guidance on acceptable methods for determining the location of rail traffic in the section or confirming the clearance of rail traffic past a proposed work location (RO-2013-018).
There were no forms or checklists to provide practical guidance for completing the steps required to implement Absolute Signal Blocking or to provide an auditable record of the process (RO-2013-018).
Procedures did not require the coordination between network control officers when the CSB affected more than one controller’s area of responsibility (RO-2011-018).
Human Influence:
The protection officers and signallers did not effectively communicate all information that was critical to the implementation of Absolute Signal Blocking (RO-2013-018).
Differences exist in the way signallers and protection officers identify trains to each other (RO-2013-018).
Network controller misunderstood the location of the worksite and information provided by the Protection Officer related to the locomotive that has passed the worksite (RO-2011-018). Similar factor was present in occurrence associated with TOA (RO-2010-004).
Non-compliances to the repeat back provisions when implementing a worksite possession (RO-2011-018). Similar factor was present in occurrence associated with LPA (RO-2013-017).
Procedures were misconstrued by key stakeholders when implementing worksite protection (20060011).
Variations in safe working arrangements to the pre-determined plan were not formally communicated to key stakeholders (20060011).
Track Work Authority rules
Human Influence:
Track force coordinator allowed the removal of protections for the passage of a train without informing key stakeholders at the worksite (RO-2013-025).
Lookout working
Organisational Influence:
The organisation did not have any documented work instructions to ensure a consistent and safe approach to maintaining automatic pedestrian crossing equipment (RO-2015-002).
Human Influence:
Worksite protection had not been adequately implemented to ensure workers were protected against inadvertently stepping into the path of a train while undertaking maintenance work (RO-2015-002).
It was common practice for maintenance personnel to adopt a process that was inherently less safe than an alternative when maintaining automatic pedestrian crossing equipment
The track was accessed by the work crew without an assessment of the risks and without the establishment of appropriate risk controls (RO-2015-019).
The lookout gave an ‘All-Right’ hand signal to the train driver before the supervisor had moved to a position of safety clear of all tracks (RO-2015-019).
The supervisor did not move to a position of safety as the train approached. He probably expected the train to proceed directly along the adjacent line and not take the crossover to his location (RO-2015-019).
Appendix B – Safe work on track event/condition type taxonomy
Table 2: Event/condition categories and their corresponding subcategories
Category
Definition
Subcategories
Collision
Any reports of a train colliding with something
Infrastructure
Person
Rolling stock
Other
Danger zone not cleared
Workers or equipment left in danger zone after work zone cleared for rail movement.
Protection officer not clearing workers from danger zone for controlled train movement through work zone
Protection officer /flagman within danger zone during passage of controlled train movement
Worker re-entering danger zone after initially being cleared
Workers slow in responding to clearing work zone instruction
Other
Detonators exploded
Includes occurrences where detonators were encountered by rail movement and movement stopped.
Protection officer omitted to remove following completion of worksite protection
protection officer placed as worksite protection zone active
Other
Near miss rolling stock
Reports of rolling stock nearly colliding with other rolling stock
Entering into a worksite occupied by a rail vehicle
Operating within a worksite
Other
Near miss track worker
Reports of rolling stock nearly colliding with track workers
A report by a driver of a near miss (word near miss used in occurrence description)
Workers or plant reported in danger zone within 100 m or 10 seconds of approaching rail movement.
Other
Protected work zone exceeded
Workers or equipment/plant in danger zone outside of area protected by current track authority
A machine/rolling stock entering into a protected work zone from an adjacent worksite
A machine/rolling stock/worker entering into a protected work zone
A worker/machine/rolling stock moving outside the protected work zone
Other
Protection incorrectly removed
Includes occurrences where the protections were removed inadvertently by protection officer or network control officer while workers or plant were operating in the danger zone.
By network control officer managing adjacent worksites (protections on incorrect worksite removed)
By network control officer managing overlapping/nested worksites (protections compromised or removed on nested worksites)
By protection officer of an adjacent worksite (work on adjacent site completed)
By protection officer of overlapping/nested worksite (work on overlapping/nested worksite completed)
By the protection officer of the worksite prematurely (work on track not complete or equipment foul)
Hand signaller procedure breach
Train cleared into worksite
Unauthorised removal of protections
Other
Protection location incorrectly positioned
Includes occurrences where protection officer or network control officer erred in locating either 'in field' or 'control system' protections.
By the network control officer not placing electronic protections (blocking) on correct signalling infrastructure
By the worker (protection officer /flagman) positioning protections in incorrect location
Flagman/lookout/signalman in wrong position/looking wrong way/absent
TOA issued while train was in location
Other
Protection type insufficient/incorrect
Includes occurrences where no protection was implemented or the protection method implemented was incorrect for the type of work undertaken in the danger zone.
Implementing inadequate protections for the level of work undertaken
Not fully implementing the procedures associated with the worksite protection selected
Not implementing any protections
Other
Protections not cleared
Flags, detonators or signs left in situ after work on track authority cleared.
he protection officer omitting to cancel work site protection arrangements with network control officer following completion of work
The protection officer omitting to remove signs/ATW after completion of works
Other
Worksite location incorrectly identified
Includes occurrences where protection officer or network control officer incorrectly located the worksite
By the protection officer not correctly identifying the worksite limits
By the protection officer not identifying the correct worksite location
Other
Appendix C - SQL relational database structure
Figure 11: Safe work on track SQL relational database structure
Appendix D – ATSB analysis of incident data, 2009 – 2014
Appendices D1 to D11 illustrate the assessed data in both diagram and tabular form for each of the 11 occurrence categories and their corresponding subcategories. Also illustrated are the occurrences assessed as featuring in other occurrence categories. In this way, factors contributing to, or associated with, certain occurrence outcomes can be observed.
D1 - Protections incorrectly removed
Figure 12 illustrates the types of events (likely antecedents) which resulted in occurrences reports (219) where the protections were incorrectly removed. The diagram also illustrates the 20 recorded occurrences which featured in three other occurrence categories.
Detailed distribution is shown in Table 3.
Figure 12: Protections incorrectly removed
Table 3: Protections incorrectly removed
The 219 recorded occurrences where the danger zone had not been cleared and the subcategories describing the likely antecedents.
Occurrence category
Protection incorrectly removed
219
Subcategories
Other
78
By the PO of the worksite prematurely (work on track not complete or equipment foul)
58
Train cleared into worksite
41
Hand signaller procedure breach
22
By NCO managing adjacent worksites (protections on incorrect worksite removed)
8
By PO of an adjacent worksite (work on adjacent site completed)
7
Unauthorised removal of protections
3
By PO of overlapping/nested worksite (work on overlapping/nested worksite completed)
1
By NCO managing overlapping/nested worksites (protections compromised or removed on nested worksites)
1
The 20 recorded occurrences which featured in three other occurrence categories, and their respective subcategories
Occurrence categories associated with protection incorrectly removed
Near miss track worker
11
A report by a driver of a near miss (word near miss used in occurrence description)
7
Workers or plant reported in danger zone within 100m or 10 seconds of approaching rail movement
4
Detonators exploded
8
PO placed as worksite protection zone active
7
PO omitted to remove following completion of worksite protection
1
Protection type insufficient/incorrect
1
Implementing inadequate protections for the level of work undertaken
1
Examples - Protections incorrectly removed
Examples of de-identified occurrence reports are provided below grouped relative to the condition that was present at that time:
Removed protections
At 1856, the network controller identified that when contacted to fulfil a TOA that they had in fact removed the blocking facilities for this TOA at 1710.
A passenger train had authority from train control to depart the station. The train driver later observed rail safety workers on the track ahead and applied emergency brakes. Work on track protections were in place under a TWA. It was reported that the passenger train was brought to a ‘stop’ in the work site 20 m short of track welding activities on the West track. On observing the approaching train, the rail safety workers scattered to a place of safety. Train control confirmed the incident and reported that the TWA on the East track had finished earlier in the day, however, the ‘blocks’ were removed on the West track in error.
Track protection lifted without notification to the work group supervisor and a train proceeded through the down track without advice to workers.
Multiple (or nested) possessions
The controller advised the PO of requirement to suspend the TOA to which the PO acknowledged and requested time to clear the track of workers. The TOA was suspended to allow two track machines to depart another closure area. The PO later advised network controller that they had failed to liaise with another work group inside the TOA, which were still on track and proceeding with their work. The network controller advised track machines to stop immediately and do not proceed in either direction until advised. Network controller contacted the work group still on track and advised them to move off the track immediately due to the suspension of the TOA. The work group advised that they were half way through a weld and if the weld was not completed the track would be unserviceable.
The PO was advised that when his protections were removed from his worksite, a second nested worksite was left exposed.
A qualified employee working in rail corridor between A and B advised he was clear of area at 1413. A TOA, which was issued to another qualified employee working in the same section, was incorrectly fulfilled at the same time.
Trains through worksites
Protection lifted without authority of the project worksite coordinator when six track machines entered worksite without authority. An adjoining PO failed to contact the worksite coordinator before lifting the protection.
Train crew advised that they had come to an emergency stop due to maintenance personnel on the track. Driver advised that he had been given the all clear by lookout 100 m prior. Driver applied emergency brakes at 72 km/h, stopping train 250 to 300 m past work site.
D2 - Protection location incorrectly positioned
Figure 13 illustrates the types of events (likely antecedents) which resulted in occurrences reports (267) where the protections were incorrectly positioned. The diagram also illustrates the 41 recorded occurrences which featured in five other occurrence categories.
The 267 recorded occurrences where the danger zone had not been cleared and the subcategories describing the likely antecedents.
Occurrence category
Protection location incorrectly positioned
267
Subcategory
Other
98
By the worker (PO/flagman) positioning protections in incorrect location
70
Flagman/lookout/signalman in wrong position/looking wrong way/absent
61
By the NCO not placing electronic protections (blocking) on correct signalling infrastructure
30
TOA issued while train was in location
8
The 42 recorded occurrences which featured in six other occurrence categories, and their respective subcategories.
Occurrence categories associated with protection location incorrectly positioned
Detonators exploded
28
PO placed as worksite protection zone active
22
Other
6
Near miss track worker
9
A report by a driver of a near miss (word near miss used in occurrence description)
9
Worksite location incorrectly identified
2
By the PO not identifying the correct worksite location
2
Danger zone not cleared
1
Other
1
Collision
1
Person
1
Examples - Protections incorrectly positioned
Examples of de-identified occurrence reports are provided below grouped relative to the condition that was present at that time:
Train approaching worksite
Gang requiring to do work on the points in a yard requested a train running information (TRI) from train control. Train driver reported through the yard at 1330 and gang was the given a TRI to work up to 1400 for the yard. Gang working in the yard reported that the train had not cleared the yard and were surprised when the train then approached the points where their work location was a couple of minutes later.
Network controller had given the gang TRI permission to work on the track between the XX km and YY km. The network controller had endorsed the TRI in the C to D section on the train control graph where the actual kilo-meterage was in the A to B section. This resulted in the gang being given permission to work on track in front of the train. The gang sighted the train and cleared the track when the headlight was sighted.
Network controller issued the track inspector a TOA to work one track machine between 99.500 km and 100 km. A train crew was issued with a train authority (TA) to proceed from 0.000 km to 108.100 km. The track inspector observed approaching train and made contact with the train crew. Movement stopped and network controller contacted.
Position of trackside signage/flagmen trackside
Report of worksite approximately 82.618 km had lookout incorrectly positioned with his back to the approaching train. The protection officer failed to address a number of key areas with the protection for the worksite.
The driver of a train that travelled through a work site reported he was not notified of the work site and was travelling at a speed of 55 km/h were his speed should have been 20 km/h. There was no outer flagman, no audible warning devices in place and the inner flagman was located 50 m from the work site.
Train driver reported that the hand signaller was located at the incorrect location. Hand signaller located at 81.100 km and should have been located at 81.000 km as per worksite protection form.
Application of electronic/mechanical isolation at the control centre or signal box
Train control had not placed all required signals at stop for CSB.[10] No protection was applied to bi-directional signals in the down direction.
At 1515 it was identified that blocking facilities had not been placed in conjunction with TOA.
Incorrect placement of blocking protection by network controller caused a near miss with gang personnel working on points crossovers. The network controller did not place blocking protection in the correct location and or confirm or repeat the information or request a repeating of the information by the PO.
PO requested a TOA. PO called and reported a train approaching his work site. Control advised that a train was approaching his protecting signal and the train should be coming to a halt. The PO then stated that he had used the wrong signal that being the down distant automatic signal as his limit. The PO advised to fulfil TOA and off-track. The PO stated that train was roughly 1 km from the work site when he cleared the track.
D3 - Protection type insufficient/incorrect
Figure 14 illustrates the types of events (likely antecedents) which resulted in occurrences (581) reported as having insufficient or incorrect protection. The diagram also illustrates the 70 recorded occurrences which featured in seven other occurrence categories.
Detailed distribution is shown in Table 5.
Figure 14: Protection type insufficient / incorrect
Table 5: Protectiontypeinsufficient / incorrect
The 581 recorded occurrences where the danger zone had not been cleared and the subcategories describing the likely antecedents.
Occurrence category
Protection type insufficient/incorrect
581
Subcategory
Not implementing any protections
275
Implementing inadequate protections for the level of work undertaken
168
Not fully implementing the procedures associated with the worksite protection selected
112
Other
26
The 73 recorded occurrences which featured in eight other occurrence categories, and their respective subcategories.
Occurrence categories associated with protection type insufficient/incorrect
Near miss track worker
31
A report by a driver of a near miss (word near miss used in occurrence description)
26
Workers or plant reported in danger zone within 100m or 10 seconds of approaching rail movement
5
Protected work zone exceeded
30
A worker/machine/rolling stock moving outside the protected work zone
20
A machine/rolling stock/worker entering into a protected work zone
8
A machine/rolling stock entering into a protected work zone from an adjacent worksite
1
Other
1
Detonators exploded
5
PO placed as worksite protection zone active
4
PO omitted to remove following completion of worksite protection
1
Protections not cleared
1
The PO omitting to remove signs/ATW after completion of works
1
Protection incorrectly removed
1
Hand signaller procedure breach
1
Danger zone not cleared
1
Workers slow in responding to clearing work zone instruction
1
Collision
1
Rolling stock
1
Examples - Protection type insufficient / incorrect
These occurrences include a broad spectrum of occurrences from grass cutting within corridor with no authority to the wrong worksite protection being selected. Examples of de-identified occurrence reports are provided below.
The driver reported he came very close to striking two maintenance personnel working the points on the up road approaching a station. The driver advised the train was closer than 100 m from the workers and travelling at 90 km/h when they cleared the track. The driver advised he was shaken by the incident. The controller had no knowledge of workers in in that area and the matter referred to the network manager. The driver advised he could not see any lookouts in the area for the workers. The driver was too distressed to drive the train back.
The manager at the network control centre received a phone call from the track PO to advice of a possible near miss that morning. The contractors had started work before the PO had arrived on site.
D4 - Worksite location incorrectly identified
Figure 15 illustrates the types of events (likely antecedents) which resulted in occurrences (114) where the worksite location was incorrectly identified. The diagram also illustrates the 26 recorded occurrences which featured in five other occurrence categories.
Examples of de-identified occurrence reports are provided below.
At 1002, the PO fulfilled a CSB. At 1005, a train driver advised there were three detonators and red flag on the up main on the up side of the signal. Investigations revealed that protection was placed on the wrong track for TOA on the down main (protection placed on up main).
Welding gang was enroute to location ‘B’ to repair a rail defect in the main line. They were advised that there was spare rail at the site to facilitate the repair. The gang drove to location ‘C’ instead of location ‘B’ (which was 36 km away). Gang entered the corridor and contacted network control to obtain a TWA in preparation for work identifying their location as ‘B’ main line. Network control required that trains be able to use the loop road to travel through ‘B’ and therefore permanent way protection was placed within the main line allowing trains to safely traverse the points into the loop. However as network control thought the gang was at ‘B’, the route for the loop was set at that station and not at location ‘C’. The route at ‘C’ was actually still set (and clipped) for the main line where the gang was working. The supervisor identified a section of rail and a rail defect in the main line at ‘C’. As work progressed, the supervisor of the gang received a phone call from another employee enquiring as to the gang’s location so that he could undertake a spot safety audit. The supervisor advised they were at ‘B’ and the auditor proceeded to ‘B’. On arrival at ‘B’ the auditor could not find the gang and again rang the supervisor to enquire as their whereabouts. The supervisor again advised he was at ‘B’ at which time the auditor stated that he was standing at ‘B’ and nobody was in attendance. It was at this point that the supervisor realised that the gang was mistaken as to their location and with the weld already completed; the supervisor removed the personnel from the track and advised network control of their actual location. There was no defect found at ‘B’ and the defect reported to the gang was actually the one at ‘C’.
Driver reported track workers possibly working in wrong location. Driver advised gang working at ‘A’ which was not known to train control. After investigation, it was established the track supervisor had indicated via his work on track request he was in the ‘B’ area. The track supervisor had made the same error the day before which went undetected. The worksite was taken over by a senior qualified worker, and closed down.
Train struck three detonators and stopped short of a red flag in the four foot. The PPO advised that the protection placed at wrong line.
At 0100 mid-week night possession (LPA), a train struck detonators on the down local line (which was not part of the LPA). The driver reported this to network control. It was discovered that the flagman put out the LPA protection (detonators) at 11.672 km on the down local instead of the down main.
Gang not advised of train movement. The gang was working on ‘A’ Line. On the control graph the line was plotted incorrectly (showing gang to be on ‘B’ line) passage.
Workers entered the wrong tunnel unprotected (tunnel A) instead of the (tunnel ‘B’) which was under occupation.
Table 7: Protected work zone exceeded
The 201 recorded occurrences where the danger zone had not been cleared and the subcategories describing the likely antecedents.
Occurrence category
Protected work zone exceeded
201
Subcategory
A worker/machine/rolling stock moving outside the protected work zone
109
A machine/rolling stock/worker entering into a protected work zone
50
Other
22
A machine/rolling stock entering into a protected work zone from an adjacent worksite
20
The 54 recorded occurrences which featured in seven other occurrence categories, and their respective subcategories.
Occurrence categories associated with protected work zone exceeded
Protection type insufficient/incorrect
30
Not implementing any protections
23
Not fully implementing the procedures associated with the worksite protection selected
3
Implementing inadequate protections for the level of work undertaken
3
Other
1
Worksite location incorrectly identified
16
By the PO not correctly identifying the worksite limits
3
By the PO not identifying the correct worksite location
2
Other
11
Near miss rolling stock
2
Operating within a worksite
1
Other
1
Near miss track worker
2
A report by a driver of a near miss (word near miss used in occurrence description)
2
Detonators exploded
2
PO placed as worksite protection zone active
2
Protection location incorrectly positioned
1
By the worker (PO/flagman) positioning protections in incorrect location
1
Collision
1
Rolling stock
1
Examples - Protected work zone exceeded
Examples of de-identified occurrence reports are provided below.
A signal passed at danger (SPAD) alarm was activated for a signal. Upon investigation it was found that a track worker was working outside of the TOA limits and had placed a spirit level onto the tracks setting off the SPAD alarm. TOA limits were from the signal, begin train order sign, to the next yard limit board. The gang was cleared from the track
D6 - Near collision (‘near miss’) with a track worker
Figure 17 illustrates the types of events (likely antecedents) which resulted in occurrences (236) where rolling stock nearly collided with track workers. The diagram also illustrates the 82 recorded occurrences which featured in seven other occurrence categories.
Detailed distribution is shown in Table 8.
Figure 17: Near miss track worker
A site audit and compliance check noticed a work group was crossing tracks to access the worksite outside protection limits. This had been occurring from the start of the possession. The work party was stopped and placed in a position a safety. Existing protection was assessed and was extended to cover access area.
Table 8: Near miss track worker
The 236 recorded occurrences where the danger zone had not been cleared and the subcategories describing the likely antecedents.
Occurrence category
Near miss track worker
236
Subcategory
A report by a driver of a near miss (word near miss used in occurrence description)
194
Workers or plant reported in danger zone within 100m or 10seconds of approaching rail movement.
40
Other
2
The 84 recorded occurrences which featured in seven other occurrence categories, and their respective subcategories.
Occurrence categories associated with near miss track worker
Protection type insufficient/incorrect
31
Implementing inadequate protections for the level of work undertaken
11
Not fully implementing the procedures associated with the worksite protection selected
6
Not implementing any protections
13
Other
1
Danger zone not cleared
29
PO not clearing workers from danger zone for controlled train movement through work zone
18
Other
5
Workers slow in responding to clearing work zone instruction
4
PO/Flagman within danger zone during passage of controlled train movement
2
Protection incorrectly removed
11
Other
6
By the PO of the worksite prematurely (work on track not complete or equipment foul)
4
Train cleared into worksite
1
Protection location incorrectly positioned
9
Flagman/lookout/signalman in wrong position/looking wrong way/absent
5
Other
3
By the NCO not placing electronic protections (blocking) on correct signalling infrastructure
1
Protected work zone exceeded
2
A machine/rolling stock/worker entering into a protected work zone
1
A worker/machine/rolling stock moving outside the protected work zone
1
Examples - Near miss track worker
Examples of de-identified occurrence reports are provided below.
The driver of a train reported a near miss with track workers following receipt of a green hand signal by the flagman, who appeared to operating without detonators even though the gang was still working on the track.
The driver of a train loaded with 4500 tonnes of coal and 844 metres long travelling at about 25 km/h reported having a near miss with a track. The track worker had walked out from behind another train travelling in the opposite direction. The crew of the coal train estimated that they missed the person by approximately 1 metre.
D7 - Detonators exploded
Figure 18 illustrates the types of events (likely antecedents) which resulted in occurrences (163) where detonators exploded. The diagram also illustrates the 131 recorded occurrences which featured in seven other occurrence categories.
Detailed distribution is shown in Table 9.
Figure 18: Detonators exploded
Table 9: Detonators exploded
The 163 recorded occurrences where the danger zone had not been cleared and the subcategories describing the likely antecedents.
Occurrence category
Detonators exploded
163
Subcategory
PO omitted to remove following completion of worksite protection
94
Workers or plant reported in danger zone within 100m or 10seconds of approaching rail movement.
54
Other
15
The 131 recorded occurrences which featured in seven other occurrence categories, and their respective subcategories.
Occurrence categories associated with detonators exploded
Protections not cleared
83
The PO omitting to remove signs/ATW after completion of works
83
Protection location incorrectly positioned
28
By the worker (PO/flagman) positioning protections in incorrect location
15
Flagman/lookout/signalman in wrong position/looking wrong way/absent
6
Other
7
Protection incorrectly removed
8
Train cleared into worksite
3
Hand signaller procedure breach
2
Other
2
By the PO of the worksite prematurely (work on track not complete or equipment foul)
1
Protection type insufficient/incorrect
5
Implementing inadequate protections for the level of work undertaken
4
Not implementing any protections
1
Worksite location incorrectly identified
4
By the PO not correctly identifying the worksite limits
1
Other
3
Protected work zone exceeded
2
Other
2
Near miss track worker
1
A report by a driver of a near miss (word near miss used in occurrence description)
1
Examples - Detonators exploded
Examples of de-identified occurrence reports are provided below.
The driver reported they had come to a stand after striking 3 three detonators on the down main line as they approached a signal. The PO was contacted and he advised that he had staff out in that area that were preparing for a planned TWA to be granted and one of them must have put the detonators out early without permission.
The driver advised they ran over three detonators after being given a steady green flag to proceed. The driver also advised he had received an all clear signal from outer flagman after proceeding over two detonators. The driver brought the train to a stand after hitting three detonators and PO advised him that he was okay to proceed at normal speed as the worksite had been cleared.
D8 - Danger zone not cleared
Figure 19 illustrates the types of events (likely antecedents) which resulted in occurrences (103) where the danger zone had not been cleared when required. The diagram also illustrates the 36 recorded occurrences which featured in five other occurrence categories.
Detailed distribution is shown in Table 10.
Figure 19: Danger zone not cleared
Table 10: Danger zone not cleared
The 103 recorded occurrences where the danger zone had not been cleared and the subcategories describing the likely antecedents.
Occurrence category
Danger zone not cleared
103
Subcategory
PO not clearing workers from danger zone for controlled train movement through work zone
38
Other
32
Workers slow in responding to clearing work zone instruction
20
PO/Flagman within danger zone during passage of controlled train movement
7
Worker re-entering danger zone after initially being cleared
6
The 36 recorded occurrences which featured in five other occurrence categories, and their respective subcategories.
Associated Occurrence categories
Associated Subcategories
Near miss track worker
29
A report by a driver of a near miss (word near miss used in occurrence description)
18
Workers or plant reported in danger zone within 100m or 10seconds of approaching rail movement.
11
Near miss rolling stock
4
Operating within a worksite
2
Other
2
Protection location incorrectly positioned
1
Flagman/lookout/signalman in wrong position/looking wrong way/absent
1
Protection type insufficient/incorrect
1
Not fully implementing the procedures associated with the worksite protection selected
1
Collision
1
Other
1
Examples - Danger zone not cleared
Examples of de-identified occurrence reports are provided below.
The driver of a passenger train reported a near hit with three workers within the danger zone. The driver stated that the horn was sounded and acknowledged, however the workers failed to place themselves in a safe place, and a second warning whistle was sounded. The second warning was not acknowledged, as one of the workers had left a ballast spade beside the line. Believing that the ballast spade may have been foul of the approaching train, the PO directed the worker to remove it. The PO stated that at this time the driver sounded the second horn warning which was not acknowledged as the worker was fetching the spade and if an ‘alright’ hand signal was displayed it would have signified to the driver that his work crew were clear and in a safe place. The driver then sounded a third horn warning before the workers cleared themselves from the danger zone, at which point the train was estimated by the driver to be around 25m from the workers.
The driver of a freight train reported that they had been in a near miss with an unidentified track worker. The driver reported that he was travelling at 60 km/h when he saw a worker who was facing in the opposite direction and speaking on a mobile phone. The train was approximately 50 to 100 metres away when the driver noticed the worker and blasted the horn to warn the worker. The worker, who was visibly shaken, was later identified to be a PO who had recently fulfilled a TOA at the location.
D9 - Collisions
Figure 20 illustrates the types of events (likely antecedents) which resulted in reported collisions (6). The diagram also illustrates the four recorded occurrences which featured in four other occurrence categories.
Detailed distribution is shown in Table 11.
Figure 20: Collision occurrences
Table 11: Collision occurrences
The six recorded occurrences where the danger zone had not been cleared and the subcategories describing the likely antecedents.
Occurrence category
Collision
6
Subcategory
Rolling stock
3
Person
2
Other
1
The four recorded occurrences which featured in four other occurrence categories, and their respective subcategories.
Occurrence categories associated with collision
Protection type insufficient/incorrect
1
Implementing inadequate protections for the level of work undertaken
1
Danger zone not cleared
1
Other
1
Protected work zone exceeded
1
A machine/rolling stock/worker entering into a protected work zone
1
Protection location incorrectly positioned
1
Other
1
Examples - Collision occurrences
Six occurrences in the data set were recorded as collisions. Three of these involved collisions with rolling stock. One of these involved a collision between a freight train and an (unoccupied) excavator (ATSB investigation RO-2011-018), the second occurrence involved a collision between a ballast regulator and road rail vehicle, and the third involved a collision between a train and a road rail (Hi-Rail) vehicle. Of the two collisions with track workers, one occurrence involved a collision between a passenger train and an occupied hi-rail excavator. This occurrence is the only fatality identified in the occurrence notification data set and was investigated by the ATSB (investigation number RO-2010-004). The other occurrence involved a train striking a lookout. The collision occurrences coded as collision – other, involved a train striking a rail drill with no injuries or damage to track.
D10 - Near collision (‘near miss’) with rolling stock
Figure 21 illustrates the types of events (likely antecedents) which resulted in reported near misses with rolling stock (21). The diagram also illustrates the seven recorded occurrences which featured in three other occurrence categories.
Detailed distribution is shown in Table 12.
Figure 21: Near miss rolling stock
Table 12: Near miss rolling stock
The 21 recorded occurrences where the danger zone had not been cleared and the subcategories describing the likely antecedents.
Occurrence category
Near miss rolling stock
21
Subcategory
Other
11
Operating within a worksite
9
Entering into a worksite occupied by a rail vehicle
1
The seven recorded occurrences which featured in three other occurrence categories, and their respective subcategories.
Occurrence categories associated with near rolling stock
Danger zone not cleared
4
PO not clearing workers from danger zone for controlled train movement through work zone
2
Worker re-entering danger zone after initially being cleared
1
Other
1
Protected work zone exceeded
2
A machine/rolling stock/worker entering into a protected work zone
1
A worker/machine/rolling stock moving outside the protected work zone
1
Protections not cleared
1
The PO omitting to remove signs/ATW after completion of works
1
Examples - Near miss rolling stock
Examples of de-identified occurrence reports are provided below.
The driver of a train reported a near miss with a work crew near the track who were unloading sleepers off a truck with an excavator. The arm of excavator swung over track in front train and was cleared when train was 150 metres from work group.
A train crew reported a near miss with a tip truck with trailer that was backed up against shoulder ballast removing old sleepers. The crew reported missing the truck by approximately 2 feet.
D11 - Protections not cleared
Figure 22 illustrates the types of events (likely antecedents) which resulted in reported near misses with rolling stock (155). The diagram also illustrates the 85 recorded occurrences which featured in three other occurrence categories.
Detailed distribution is shown in Table 13.
Figure 22: Protections not cleared
Table 13: Protections not cleared
The 155 recorded occurrences where the danger zone had not been cleared and the subcategories describing the likely antecedents.
Occurrence category
Protections not cleared
155
Subcategory
The PO omitting to remove signs/ATW after completion of works
140
The PO omitting to cancel work site protection arrangements with NCO following completion of work
9
Other
7
The 85 recorded occurrences which featured in three other occurrence categories, and their respective subcategories.
Occurrence categories associated with protections not cleared
Detonators exploded
83
PO omitted to remove following completion of worksite protection
83
Protection type insufficient/incorrect
1
Not implementing any protections
1
Near miss rolling stock
1
Other
1
Examples - Protections not cleared
Examples of de-identified occurrence reports are provided below.
The driver reported that they had exploded three detonators and was sitting facing a red flag. A PO had taken out a verbal TOA for a worksite but the work had ceased with the possession fulfilled.
The driver reported running over three detonators. Detonators were left behind by a PO who had a TOA for a worksite which was fulfilled.
The driver reported that he had struck three detonators and a flag at the. A track worker reported that he had sent a worker out to retrieve the detonators and flag when he fulfilled his TOA, however the worker had travelled in the wrong direction.
A freight train came to a stand after running over three detonators and red light left on the line after track possession finished. The PO had delegated the task to another PO for completion, but he did not acknowledge the location of the protection.
Train ran over three detonators and red flag. The PO had TOA that was fulfilled. The network controller rang the PO to ask had he had lifted his detonators when he fulfilled his TOA; he said he had not. The driver was then informed that line ahead clear and to continue.
Under the occurrence classification guideline,[5] a safe working rule or procedure breach is defined as any breach of an operational safe working system or procedure that endangers or has the potential to endanger the safety of railway operations and/or persons. It includes:
human failures (intentional and unintentional acts) in the application of safe working procedures
failure to communicate or act on vital information to protect trains and personnel
failure to comply with a hand signal
irregularities in the management of train separation (for example, a near miss)
any breach of a network rule
any breach of the work scheduling practices and procedures set out in the rail transport operator’s fatigue risk management program.
Since occurrences coded under this category include such a broad range of event types, the ATSB provided regulators with a list of keywords relating to safe work on track to search for in their occurrence databases. In total, 12,146 occurrence records were obtained from all state and territory rail safety regulators for the 5 year period between July 2009 and June 2014. The ATSB further filtered these records, and about 15 per cent (1,779) were found to involve a safe working authority or a means of protection for a work on track related activity. The remaining records were not associated with work on track, but related to the operation of train movements.
In order to extract as much meaningful information as possible, the description field for each record was examined and the events/conditions (where described) grouped into eleven categories (Table 1). In order to best describe what happened during the occurrence, the events/conditions were further grouped into subcategories. The complete table of categories and subcategories is shown in Appendix B – Safe work on track event/condition type taxonomy.
Of the 1,779 occurrence records related to a work on track activity, 100 per cent were able to be categorised into one of the eleven event/condition categories. In addition, 80 per cent could also be categorised into one of the respective subcategories. The remaining 20 per cent were grouped into subcategories labelled ‘Other’ within each category. These included occurrences where insufficient information was provided to identify likely antecedents to the unsafe act, and those where the likely antecedents were not addressed within the taxonomy.
To preserve data integrity and assist with the analysis (and any potential future rail occurrence analysis), all data for this project was imported into a relational database.[6] Data was then extracted from this database for analysis. A database diagram showing the seven tables in the database, and how they relate to each other, is shown in Appendix C - SQL relational database structure.
Table 1: Event/condition category definitions
Category
Definition
Collision
Any reports of a train colliding with something.
Danger zone not cleared
Workers or equipment left in danger zone after work zone cleared for rail movement.
Detonators exploded
Includes occurrences where detonators were encountered by rail movement and movement stopped.
Near miss rolling stock
Reports of rolling stock nearly colliding with other rolling stock.
Near miss track worker
Reports of rolling stock nearly colliding with track workers.
Protected work zone exceeded
Workers or equipment/plant in danger zone outside of area protected by current track authority.
Protection incorrectly removed
Includes occurrences where the protections were removed inadvertently by protection officer or network control officer while workers or plant were operating in the danger zone.
Protection location incorrectly positioned
Includes occurrences where protection officer or network control officer erred in locating either 'in field' or 'control system' protections.
Protection type insufficient/incorrect
Includes occurrences where no protection was implemented or the protection method implemented was incorrect for the type of work undertaken in the danger zone.
Protections not cleared
Flags, detonators or signs left in situ after work on track authority cleared.
Worksite location incorrectly identified
Includes occurrences where protection officer or network control officer incorrectly located the worksite.
Overall analysis results
The distribution of the 1,779 safe work on track occurrences between the 11 categories is shown in Figure 2. It can be seen that occurrences where the protection applied was either insufficient or incorrect were the most commonly featured event category between July 2009 and July 2014. This occurrence category (580 occurrences) was nearly twice as common as the next most frequent, where protection systems were incorrectly positioned (267 occurrences).
Figure 2: Counts of work on track occurrences by event/condition categories category, June 2009–June 2014
Closer examination of occurrences associated with insufficient or incorrect protection found that the occurrence reports were largely a result of drivers observing workers in or adjacent to the rail corridor. Although reported as having none or insufficient protections in place, it was not clear whether the workers were infringing in the danger zone or what, if any, work on track protections were actually required by these workers.
The next most frequent category involved occurrences where protection systems were incorrectly positioned. A large proportion related to errors associated with administering the protection arrangements at the worksite location itself or errors between the protection officer trackside and the network controller in the control centre. Also of note was the number of instances where the protections provided by a flagman/lookout broke down.
In reviewing the notifications associated with each category it was evident that the magnitude of the category was not indicative of the potential increase in risk exposure to track workers. That is, in some categories there was a high proportion of events where the risk exposure to track workers was nil or minimal. In other categories, there were a high proportion of events with increased risk exposure, evident since the controls in place to protect a worksite were significantly compromised.
Consequently, further analysis was required to ensure the safety risk was considered within each category.
Higher risk event/condition types
To establish the likelihood of an increased risk to track workers, the description provided for each of the 1,779 reported work on track occurrences were reviewed further to evaluate the effectiveness of the risk controls in capturing the unsafe condition or event. The assessed risk was coded into four risk categories:
Risk neutral - where the train crew observed workers or equipment within the rail corridor unexpectedly and reported the occurrence, or where protections were not cleared correctly after works were completed.
Low risk - where the processes used in work on track started to break down, but still relatively low risk to workers or train crew.
Risk exposure increased - similar to low, but where a number of processes (defences) had broken down resulting in an increased likelihood that an adverse consequence could occur.
High risk exposure - significant failure in process (defences). The detection of the breakdown or lack of protection occurred late in the sequence of events, resulting in a heightened risk to workers or rail crew of adverse consequences.
Figure 3 shows the number of occurrences per risk level for each of the 11 event/condition categories. Figure 4 shows the risk ratings as a proportion of the total number of occurrences within that category.
Figure 3: Occurrence counts as a function of event/condition category and risk rating
Figure 4: Risk rating as a proportion of occurrences in each event/condition category
Although events where protection was considered insufficient or incorrect was by far the most commonly reported safe work on track occurrence category, nearly two thirds (62%) were classified as risk neutral occurrences. Similarly, many of the events where protections had not been cleared were also classified as neutral risk (about 86%).
While most of the event/condition categories had a high proportion of risk neutral classified occurrences, of particular concern were the occurrences where the defences that were in place failed to maintain track worker safety. These higher risk occurrences, (red in Figure 3 and Figure 4), give an indication of the event/condition categories more likely to result in injury to track workers.
Notable was where events resulted in a collision. Since the highest consequence had occurred, most were categorised as high risk events (four of six occurrences). However, this category only accounted for six events out of a total of 1,779 examined (about 3%).
Again, further analysis was required to ensure both safety risk and event frequency was considered when comparing event categories.
Frequency and risk analysis
The classification of a risk category for each occurrence allows the sum of risk for each of the occurrence categories to be considered. The sum of risk takes into account both the frequency of occurrences shown in Figure 2 as well as risk rating, shown in Figure 3.
Since 2012, the ATSB has utilised an Event Risk Classification (ERC) process to risk assess aviation occurrences. The ERC is based on the Aviation Risk Management Solutions Event Risk Classification ERC framework.[7]
Using a similar methodology, the four risk categories discussed above were assigned a risk value and multiplied by the number of occurrences to obtain the sum of risk. In this way the sum of risk integrates both the frequency and severity of potential outcomes for each of the occurrence categories. Figure 5 illustrates the occurrence categories in order of sum of risk.
Figure 5: Sum of risk for each event/condition category
The four categories assessed as having the highest sum of risk were:
protection incorrectly removed
protection location incorrectly positioned
protection type insufficient/incorrect
worksite location incorrectly identified.
Detailed analysis of categories
The following illustrates the data analysis in further detail. Complete tables of the data analysis are documented in Appendix D, along with some examples of de-identified occurrence reports for each category.
Protections incorrectly removed
This category included a range of events where the protection officer (PO) or network control officer (NCO) removed safeguards inadvertently while the plant or workers were operating in the danger zone. The consequence of the removal of safeguards, without an appropriate warning, understandably resulted in the highest assessed risk exposure to track workers.
There were 219 occurrences where the protections were incorrectly removed. About 22 per cent of these occurrences exposed workers to a high risk. Figure 6 illustrates the most common (subcategory) events that contributed to protections being incorrectly removed.
Figure 6: Protections incorrectly removed and subcategories
About a quarter (26%) of occurrences where protection was incorrectly removed were attributable to the protection officer prematurely removing site protections before the work on track was complete or before work equipment was removed from the danger zone. Another 41 (19%) were as a result of trains inadvertently being cleared into active worksites. Twenty-two occurrences (10%) were a result of incorrect procedures used by hand signallers.
Twenty occurrences involving protections incorrectly removed also featured in three other occurrence categories. Eleven of these occurrences led to a near collision, four of which were within 100 m or 10 seconds of a collision with track workers and/or equipment.
A high percentage occurrences in this category (36%) were assessed as ‘Other’. These were occurrences where insufficient information was provided to identify likely antecedents that led to the incorrect removal of the protections, or where the likely antecedents were not addressed within the taxonomy.
Protection location being incorrectly positioned
This category included events where the protection officer or network control officer incorrectly placed signage at the worksite or electronic blocks preventing signals from clearing. The consequence of incorrectly applying safeguards resulted in increasing risk exposure, particularly in situations where protections were positioned with a train still operating between the protection and the worksite.
There were 267 occurrences where the protections were incorrectly located. About 14 per cent of these exposed workers to a high risk. Figure 7 illustrates the most common (subcategory) events that contributed to the protections being incorrectly located.
Figure 7: Protections incorrectly positioned and subcategories
About a quarter (26%) of occurrences where the protections were incorrectly located were due to either the protection officer or flagman putting the protections in the wrong location. There were 61 instances (23%) where the flagman was facing the wrong way, absent, or in the wrong position, (these also include accounts of flagmen in their vehicles, and/or having lunch, and/or on telephones, and/or asleep while on duty). There were 30 accounts (11%) where network control officers did not implement electronic (blocking) protections in the correct location. The remaining eight were due to the issuing of TOA’s while trains were still in the section.
Forty-two recorded occurrences where protections were incorrectly positioned also featured in six other occurrence categories. Detonators exploded in 28 of these occurrences and of these, 22 were at active worksite.
A high percentage of incorrectly positioned protection occurrences (37%) were assessed as ‘Other’. These were occurrences where insufficient information was provided to identify likely antecedents that led to the incorrect positioning of protections, or where the likely antecedents were not addressed within the taxonomy.
Protection type insufficient/incorrect
Protection type insufficient or incorrect was the most commonly featured (Figure 2) of all the work on track occurrences. Although there were a large number of occurrences (581), the assessed risk exposure for majority were either neutral or low. However, the number of occurrences when multiplied with the assigned risk value resulted in the relatively high ranking (Figure 5) of this event category.
Figure 8 illustrates the most common (subcategory) events that contributed to reports of incorrect or insufficient protection.
Figure 8: Protection type insufficient / incorrect and subcategories
Occurrences relating to the protection type being insufficient or incorrect contained a broad spectrum of reported events and conditions. Almost half (47%) related to train drivers observing persons or vehicles within the rail corridor, without any advice to the train crew or discernible trackside protection arrangements in place. Many reports by the drivers arose from the unauthorised entry by rail maintenance crews, utility providers or contractors into the rail corridor, but not necessarily within the danger zone itself.
Seventy-three recorded occurrences relating to insufficient or incorrect protection also featured in eight other occurrence categories. The most common was a near collision with a track worker (31 events), but only five were reported as being within 100 meters or 10 seconds of a collision with track workers or machinery. Occurrence where workers or machinery moved beyond the work zone limits was the next most commonly co-occurring (30 events), most (20 events) involved unauthorised machinery entering into a protected work zone.
Worksite location incorrectly identified
Accurately determining and communicating the geographic location of the worksite is essential in interfacing with other worksite locations, qualified workers providing onsite protections, control centre staff at locations remote to the worksite, and train drivers.
There were 114 occurrences where the worksite location had been incorrectly identified. Figure 9 illustrates the most common (subcategory) events that led to the incorrect identification of worksite locations.
Almost two-thirds of records in this category (63%) were assessed as ‘Other’, due to insufficient information or the likely antecedents not being addressed within the taxonomy. However, there was sufficient information to determine that almost 60% of these occurrences exposed track workers to increased risk, with about 26% of these considered to have exposed workers to a high risk.
Figure 9: Worksite location incorrectly identified and subcategories
Thirty-one occurrences (27%) reported that the protection officer did not identify the correct worksite location. In another 11 occurrences (10%), the protection officer did not correctly identify the worksite limits.
Twenty-six recorded occurrences where the worksite location had been incorrectly identified also featured in five other occurrence categories. The most common was where workers or machinery moved outside the protected work zone, with 10 reported.
Protected work zone exceeded
Occurrences where the protected work zone was exceeded included where workers, machines or equipment were in the danger zone but outside of the protected area. The category also picked up occurrences where other workers or machines had entered a worksite without the knowledge of the protection officer.
Figure 10: Protected work zone exceeded and subcategories
Of the 201 occurrences, just over half (54%) were as a result of a workers or machines moving outside the protected work zone. Fifty (25%) were due to workers or machines entering into a protected work zone and another 20 (10%) due to workers or machines entering from an adjacent worksite.
Fifty-four occurrences also featured in seven other occurrence categories. The most common was related to insufficient or incorrect protection with 23 of those a result of protections not being implementing at all.
Near collision with track worker (‘near miss’)
This category included 236 reports where rolling stock nearly collided with track workers. The majority of occurrences (82%) were unspecified reports of a ‘near miss’. Only about 13 per cent of reports specifically mentioned that the track workers were within 100 metres or 10 seconds of a collision.
Eighty-four occurrences were also associated with other occurrence categories. The most frequent (31) involved insufficient or incorrect protection, although occurrences where workers had not cleared the danger zone also featured prominently (29).
Detonators exploded
This category included 163 occurrences where detonators were encountered by a rail movement and the movement stopped as a result.
In 94 (58%) of these, the detonators were left on site by the protection officer at the completion of work. Consequently, a large proportion of these occurrences also feature in the category of protections not cleared.
However in 54 occurrences (33%), the detonators were correctly placed and protecting an active worksite. A significant proportion of these were the result of protections having been incorrectly positioned.
Danger zone not cleared
This category includes 103 occurrences where workers or equipment were in the danger zone after the work zone was cleared for a rail movement.
These occurrences were most commonly attributed to the protection officer not clearing workers from the danger zone before a train was authorised to pass through the worksite (37%). This was followed by workers being slow to respond when instructed to clear a work zone (19%). A small proportion were due to a flagman remaining within the danger zone (7%) or workers re-entering the danger zone after having cleared (6%).
Notably, about 28 per cent of occurrences were also reported as a near collision with a track worker. Of these, 11 per cent were reported to be within 10 seconds or 100 meters of a collision.
About a third of the occurrence reports for this category were assessed as ‘other’, due to insufficient information or the likely antecedents not being addressed within the taxonomy.
Collision
Of the six recorded collisions, three involved vehicles, two involved people and one involved equipment.
Of the three involving vehicles, one was a collision between a freight train and an (unoccupied) excavator (ATSB investigation RO-2011-018). The second occurrence involved a low speed collision between a ballast regulator and road rail vehicle, and the third involved a collision between a train and a road rail vehicle.
Of the two collisions with track workers, one involved a collision between a passenger train and an occupied hi-rail excavator. This occurrence is the only fatality identified in the data set and was investigated by the ATSB (investigation RO-2010-004). The other occurrence involved a train striking a lookout.
The sixth collision involved a train striking a rail drill with no injuries or damage to track.
Near collision (‘near miss’) with rolling stock
There were 21 near collisions with rolling stock. About half of the reports were assessed as ‘other’. Of the remainder, nine occurred while the rolling stock (road rail vehicles, tampers, ballast regulators) were operating within a worksite, with only one occurrence involving a near collision due to rolling stock entering into an occupied worksite.
Protections not cleared
Protections not cleared refers to occurrences where flags, detonators or signs were left in situ after the work on track authority was cleared.
Just over 90 per cent of the 155 occurrences where protections were not cleared were as a result of the protection officer not removing them at the completion of work. Eight occurrences were as result of the protection officer omitting to cancel work site protections with the network control officer. In over half of the occurrences, it was also reported that detonators were exploded, all of which were left on site by the protection officer.
Summary of analysis
The ATSB analysis grouped the notifiable occurrence data into eleven categories. Each category was further grouped into subcategories in order to better describe the events/conditions that led to the occurrence. Further analysis took into consideration both safety risk and event frequency, allowing the data to be presented in order of risk exposure to track workers.
The four categories assessed as exposing track workers to highest risk were:
protection incorrectly removed
protection location incorrectly positioned
protection type insufficient or incorrect
worksite location incorrectly identified.
It was noted that the category of protection type insufficient/incorrect resulted in a relatively high ranking due to its high number of occurrences, even though the assessed risk exposure for majority of individual occurrences were either neutral or low. Almost half (47%) related to train drivers observing persons or vehicles within the rail corridor, but not necessarily within the danger zone itself.
The results of the study were largely reflective of the safety factors identified from previous ATSB investigations. That is, incidents were predominately a result of errors during the implementation or dissolution stage of providing track protection. Protections were either removed incorrectly or prematurely, or key communication exchanges failed to establish the location of the worksite with respect to approaching rail traffic.
The outcome of this ATSB safety issue investigation suggests that the rail industry should consider the event types identified above in determining areas in which to target effort for maximising the effectiveness of safety arrangements for work on track.
In addition, the rail industry should continue to monitor occurrence data to develop long-term trends, which would likely aid in the identification of other areas for improvement and in gauging the effectiveness of industry initiatives to reduce certain factors (events/conditions) from occurring.
While all of the occurrence records related to a work on track activity could be categorised into one of the eleven event/condition categories, about 20 per cent could not be subcategorised to better describe what happened during the occurrence. A proportion of these were a result of insufficient description of the conditions present at the time. The absence of this information limited the detailed analysis to identify likely antecedents to the unsafe act.
To facilitate future analysis, the current limitations in the descriptor information provided with each occurrence needs to be addressed. With the availability of more detailed narrative within the descriptor field, the ability to categorise an event will improve the accuracy of any future trend analysis.
Data limitations
The initial data collection for this report yielded 12,146 occurrence records from all states and territories. Due to variations in the state and territory occurrence coding practices, usage of the data as provided was significantly limited.
In order to undertake any of the analysis shown in this report, a new three-tiered occurrence event/condition type taxonomy was created (see Appendix B – Safe work on track event/condition type taxonomy). A significant proportion of the effort involved in this investigation went into the manual coding of data into these occurrence categories and subcategories. Additionally, a SQL database was created for this investigation (see Appendix C - SQL relational database structure) and the reclassified data was input into this. Once this was achieved, the actual analysis process was comparatively straight-forward.
Industry work on track initiatives
In November 2013, RISSB instigated the National Track Worker Safety Forum (NTWSF). The NTWSF is ongoing and includes a cross-section of representatives from rail infrastructure managers. Representation also may include rail regulators, contractors, the ATSB, unions or equipment suppliers as required.
The objective of the forum is to:
share learning
create a community of practice
develop a voice for this segment of industry
develop benchmarking opportunities
build relationships
work as an industry to deliver better safety outcomes.
The NTWSF members identified the following three priority issues that should be addressed to achieve improvement in the effectiveness of track worker safety arrangements:
competencies and consistency of training, local knowledge and currency of practice
engineering controls/sharing technology
possession management and complexity.
The NTWSF work is progressing with members exploring technologies for worksite protection systems and addressing interface arrangements where differing rules and procedures exist between adjoining networks – particularly in sidings and yards and developing nationally consistent track site safety inductions.
On 9 July 2014, at about 1340 Eastern Standard Time, a Cessna 206, registered VH-NCR, departed from Dubbo, New South Wales on a private flight to Gold Coast and Archerfield, Queensland, under the visual flight rules (VFR), with three passengers on board. When about 15 NM south of Inverell, the pilot observed the weather deteriorating and low cloud about the ranges, and elected to climb and operate VFR on top of the cloud. As the aircraft climbed above 5,000 ft, the pilot observed a widespread frontal mass of cloud with tops around 12,000 ft. He contacted Brisbane Centre air traffic control (ATC) and requested navigation assistance and ATC provided updated weather information.
The pilot initially considered a diversion to Moree, however he was able to descend through a break in the cloud and elected to divert to Inverell. When about 5 NM from Inverell, the pilot was unable to sight the airport and was concerned about the lowering cloud base in the area. He commenced a turn, but passing about 3,800 ft during the turn, the aircraft entered cloud. The pilot immediately applied full power and commenced climbing until the aircraft became clear of cloud at about 5,000 ft.
The pilot diverted to Gunnedah and was able to remain in visual meteorological conditions for the duration of the flight. The aircraft landed in Gunnedah at about 1650.
This incident provides a reminder to pilots to make conservative decisions when considering how forecast weather may affect their flight. If poor weather is encountered en route, timely and conservative decision making may be critical to a safe outcome.
On 2 August 2014, the pilot of a de Havilland DHC-2 (Beaver) aircraft, registered VH-AWI, calculated the fuel required for two return charter flights from Shute Harbour to Whitehaven Beach, Queensland as specified on the company schedule for that day.
The pilot planned to fly one group of passengers to Whitehaven Beach, return solo to Shute Harbour, and take a second group of passengers to Whitehaven Beach. The pilot would then wait at Whitehaven beach before returning to Shute Harbour with the first group of passengers. The pilot would then return solo to Whitehaven Beach, collect the second group of passengers and return to Shute Harbour. During the planning, the pilot had omitted to include the solo ferry flights in the fuel calculations.
During the final flight to Whitehaven Beach, the pilot realised that he had planned for two return flights and omitted to allow additional fuel for the two empty sectors. He expected to land back at Shute Harbour with about 7 minutes of fuel remaining and elected to collect the passengers from Whitehaven Beach and return to Shute Harbour as planned.
The pilot landed the aircraft close to the passenger pick-up point to reduce taxi time, collected the passengers for the flight and conducted a short taxi and take-off for the estimated 12 minute flight. The aircraft landed at Shute Harbour and the pilot and passengers disembarked. About 6 L of fuel was in the tank when the aircraft landed, significantly less than the required 50 L reserve.
This incident highlights the importance of establishing a known fuel status regularly and the impact distractions can have at critical times including during flight planning.
On 2 August 2014, the pilot prepared the C206, VH-TND for the first of three scenic flights. As the aircraft departed from runway 03, the wind was from the east at about 8 knots.
On return to Rawnsley Park, the pilot conducted a straight in approach onto runway 21. The pilot did not compensate for the crosswind during the flare and touchdown, and after landing retracted all stages of flap, and applied light pressure to the brakes.
The aircraft then began to veer to the left. The pilot applied right rudder, but was unable to correct the situation, so initiated a go-around. During the go-around the horizontal stabiliser struck a shrub, causing significant damage to the aircraft. The pilot lowered the aircraft nose. After clearing a small tree, the pilot raised the aircraft’s nose in an attempt to clear the windsock.
Moments later, the aircraft spun rapidly to the left and collided with the ground. The pilot shut down the aircraft and assisted the passengers to exit. One passenger sustained serious injuries and another sustained minor injuries. The pilot was not injured. The aircraft was substantially damaged.
It appears that an incorrect technique was applied during the crosswind landing and subsequent go-around. The suddenness of the unexpected events during an otherwise routine exercise caught the pilot unprepared, resulting in a loss of situational awareness.
The Civil Aviation Safety Authority (CASA) has available on their website a booklet and DVD looking at situational awareness. Chapter 6 specifically looks at Losing Situational Awareness.
This is available through CASA’s online store: www.casa.gov.au
The European General Aviation Safety Team (EASA) published a Safety Promotion Leaflet on Decision Making for General Aviation Pilots. This leaflet suggests that many accidents are the result of pilot action, including the decisions that they make.
On 1 August 2014, at about 0935 EST, a Reims Aviation F406 aircraft, registered VH-EYQ, departed Emerald, Queensland, on an aerial survey task with a pilot and navigator on board. The aircraft was fitted with an oxygen system to allow unpressurised operations above 10,000 ft. The pilot tested the oxygen system for normal operation prior to the flight.
During the climb, the pilot turned on the aircraft oxygen supply and connected and donned his oxygen mask. The pilot then monitored his blood oxygen saturation level on an oxygen pulse meter as the aircraft continued to climb, and monitored his flow of oxygen by reference to a flow indication in the supply tube. All appeared normal until about flight level 180, when the pilot noticed that his blood oxygen saturation level had fallen significantly.
The pilot attempted to increase the amount of oxygen he was receiving by making an adjustment to his oxygen system controller, but the accuracy with which he was controlling the aircraft deteriorated, and his speech became slurred. The navigator encouraged the pilot to maintain control and descend, and air traffic control prompted the pilot to ensure that he was receiving an adequate supply of oxygen. The pilot was ultimately able to reconnect a fitting in his oxygen supply system that had become disconnected. When he reconnected the fitting, the pilot sensed almost immediate relief and was able to make a controlled descent. The crew returned to Emerald for an uneventful landing.
This incident highlights the importance of careful attention to aircraft oxygen systems, particularly with respect to connecting and monitoring oxygen system performance. Pilots are reminded that hypoxia is an insidious condition and the time of useful consciousness is often very limited. A prompt and decisive response to the first indication of an oxygen supply problem is imperative.
