Rail Industry Safety Report 2009 10

Rail Industry Safety Report 9 1

Contents Executive summary 1 1. Introduction 4 1.1 Industry overview 4 1. Rail safety reporting 7 Abbreviations ARTC Australian Rail Track Corporation BOCSAR Bureau of Crime Statistics and Research (NSW) Country Regional Network CSB Controlled Signal Blocking. Rail-related injury in NSW 9.1 Fatality 9. Non-fatal injury 11 3. Accidents and other key rail safety occurrences 15 3.1 Train derailment 15 3. Level crossing collision 18 3.3 Collision other than at level crossing 1 3.4 Fire 4 DIRN LPA NAR OTSI RSRP SPAD TOA TWA Defined Interstate Rail Network Local Possession Authority Metropolitan Rail Area No Authority Required Office of Transport Safety Investigations Rail Safety Regulators Panel Signal Passed at Danger (without authority) Track Occupancy Authority Track Work Authority 4. Precursor rail safety occurrences 8 4.1 Proceed authority irregularity 8 4. Safeworking irregularity 31 4.3 Signal and track irregularity 36 4.4 Rolling stock irregularity 4 4.5 Load irregularity 43 5. Drug and alcohol testing 45 6. Glossary of rail-related terms 47 7. List of notifiable occurrences 49 Independent Transport Safety and Reliability Regulator 11. This material may be reproduced in whole or in part, provided the meaning is unchanged and the source is acknowledged. ISSN 1836-318 January 11

Executive summary This is the seventh NSW annual rail industry safety report prepared by the Independent Transport Safety and Reliability Regulator (ITSRR). It summarises safety performance on the NSW rail network in 9 1 and considers trends over time. The information presented in the report is based primarily on occurrence notification records the initial written advice of a safety occurrence that a rail transport operator must submit to ITSRR in accordance with the Rail Safety Act 8. The report considers three aspects of rail safety performance: injuries: particularly fatal and serious injuries as direct measures of harm to people associated with railway operations accidents: those safety occurrences that result in injury or can lead directly to injury, with a focus on accidents with potential for multiple casualties such as passenger train collisions accident precursors: safety occurrences which do not directly cause harm but serve as warning signs of failures in safety risk controls, for example, a broken rail. The key findings in relation to each of these areas are summarised below. Injuries Passengers No rail-related passenger fatalities occurred on the NSW network in 9 1. The number of passenger fatalities has also decreased over the past two decades. The rate of serious passenger injury (per passenger journey) in 9 1 was lower than that for the previous two years. About 85% of injuries were associated with falls on railway property. Railway employees Two track maintenance workers were killed on separate parts of the network in 9 1. Several train crew were injured in train accidents in 9 1 including the driver of a freight train that derailed in the Hunter Valley; the guard of a passenger train that collided with a buffer stop at Newcastle and drivers of both monorail vehicles that collided with one another in Sydney. Members of the public No rail-related public fatalities (other than trespass and suspected suicide-related incidents) occurred in 9 1. The number of public fatalities in NSW has also decreased over the past two decades. One pedestrian was seriously injured when struck by a light rail vehicle in Sydney. There have been 15 similar occurrences over the past five years. Accidents Train derailment The rate of running line passenger and freight train derailment (per train km travelled) in 9 1 was consistent with that for the previous four years. Three running line passenger train derailments occurred in 9 1. One was a potentially higher risk derailment resulting from an interurban passenger service striking a landslide. Twenty-one running line freight train derailments occurred in 9 1 from a range of causes. Two derailments are being investigated one due to incorrectly set points and the other due to a failed wagon bearing. Several passengers received minor injuries in a collision between two monorail vehicles. This accident is the subject of investigation by ITSRR and the Office of Transport Safety Investigations (OTSI). Two passengers were seriously injured in separate occurrences when struck by a train at the train/platform interface. Five of the past six passenger deaths in NSW were of this nature. Executive summary Rail Industry Safety Report 9 1 1

Train collision Eight level crossing collisions between trains and road vehicles occurred in 9 1. One was a potentially higher risk occurrence a CountryLink passenger train struck a car at a crossing protected by signs. The number of level crossing collisions has decreased steadily over the past two decades. A contributing factor was the large number of crossing closures and level crossing upgrades. Six running line collisions between trains occurred in 9 1 which was consistent with that for the previous four years. Most were lower risk occurrences but two involved loaded passenger trains striking another train on the same line. Fire About 9% of passenger train fires in 9 1 were associated with arson on the Metropolitan Rail Area network (). The incident rate (per passenger journey) was comparable to the previous year. The rate of rolling stock-related passenger train fires (per train km travelled) in 9 1 was consistent with that observed over the previous four years. Twenty-one freight train fires were notified in 9 1, associated primarily with locomotive faults. The rate of rolling stock-related freight train fires (per train km travelled) has increased over the past five years. Accident precursors Rolling stock The rate of braking system irregularities and door faults on passenger trains (per train km travelled) has decreased over time. Signal and track The number of serious signal and level crossing irregularities in 9 1 was consistent with historical data. One signal irregularity causing a passenger train near miss is under investigation. The number of broken rails has decreased over time on the and remained steady for the other networks. The number of track misalignments has increased on the Defined Interstate Rail Network (DIRN) in 9 1 due, in part, to a major program of resleepering which has disturbed the track base. ITSRR is currently reviewing the adequacy of the Australian Rail Track Corporation s (ARTC) approach to managing this risk. Safeworking system A high number of serious failures in the systems to manage the safety of track workers occurred in 9 1, including two fatal incidents and several near miss occurrences. ITSRR undertook a major compliance campaign on track worker safety in 9 1 and will continue this in 1 11 as one of ITSRR s corporate priorities. The number of signals passed at danger (SPAD) on the has decreased over the past five years. Drug and alcohol The amount of drug and alcohol testing performed by industry in 9 1 was the highest of the past four years. A large increase in the rate of testing was observed for freight and infrastructure maintainers. The rate of both drug and alcohol detection in 9 1 (.95% and.9% respectively) was lower than that for the previous year (1.3% and.13% respectively). This improvement was associated primarily with rail infrastructure managers and maintainers. The rate of braking system irregularities on freight trains (per train km travelled) has increased over time. Also, bearing-related faults on freight wagons were the cause of two derailments in 9 1. Load The rate of freight train load irregularities in 9 1 (per train km travelled) was markedly lower than that of the previous year. A previously noted increase in open freight container doors over time appears to have steadied. Independent Transport Safety and Reliability Regulator

Many aspects of the 9 1 safety performance analysis point towards improvement in rail safety. Of the higher risk occurrence types, there has been an ongoing decrease in both passenger fatalities and level crossing risk over time. Importantly, improvements were also suggested in several areas which previously exhibited adverse trends. The rate of passenger-related train fire and collision risk associated with open freight container doors both appear to have steadied. Favourable trends in key areas of railway risk do not mean that catastrophic accidents will not occur. Despite the reduction in risk to the public at level crossings, one of the eight collisions in 9 1 involved an intercity passenger train with potentially severe consequences. Many other accidents progressed to a point just short of serious harm such as the collision between two monorail trains in July 9 and a collision of a passenger train with a buffer stop in May 1. Several failures in signals and procedural systems for managing train movements resulted in near misses with passenger trains. Under the co-regulatory model of regulation used across Australia, rail transport operators are responsible for assessing and managing safety risks associated with their operations. A significant part of ITSRR s regulatory activity is therefore devoted to verifying and enhancing the competency of rail transport operators in this regard. For example, previous reports have noted potentially serious train incidents associated with the heritage sector. In 9 1 ITSRR completed a three-year strategy to assist this sector in understanding and managing its safety risks. risks highlighted in 9 1 operate across sectors and networks. The continued high number of serious irregularities in the management of track worker safety highlights the need for improvement in worksite protection. In this case, findings from incident data were combined with intelligence from other sources (accident investigation, inspections, etc) to direct and target ITSRR s regulatory focus. In 9 1 ITSRR undertook a campaign of briefings, hundreds of inspections and incident assessments as well as formal compliance investigations of several worksite protection incidents, with this campaign continuing into 1 11. Prosecution is also an option for serious breaches of legislation. In 9 1 ITSRR initiated a successful prosecution for breaches of the Rail Safety Act in relation to an occurrence at Singleton in July 7 in which two track workers were fatally injured when struck by a freight train. A total of 1 rail safety workers were also successfully prosecuted in 9 1 for offences under the drug and alcohol provisions of the NSW rail safety legislation. ITSRR s 9 1 regulatory program and operational priorities are described fully in ITSRR s Annual Report 9 1 1. Education and promotion is not an appropriate regulatory option in all cases. Adverse trends in some freight rolling stock irregularities such as fires and brakes have, and will continue to be, the subject of targeted compliance inspections to address deficiencies in the systems for managing the rolling stock asset. Some particularly serious incidents in 9 1 are the subject of formal investigation to determine whether breaches of legislation have occurred. These include a collision between two monorails in Sydney and a freight train derailment at Whittingham which was the result of human error. 1 See ITSRR s website <http://www.transportregulator.nsw.gov.au> Executive summary Rail Industry Safety Report 9 1 3

1. Introduction Section 6 of the Rail Safety Act 8 requires the Independent Transport Safety and Reliability Regulator (ITSRR) to submit an annual safety report to the Minister for Transport. The Rail Industry Safety Report 9 1 summarises key rail safety incidents on the NSW rail network for the year and also considers historical trends. Most of the statistical summaries in this report are based on the national incident categorisation of the Guideline for the Top Event Classification of Notifiable Occurrences: Occurrence Classification Guideline One (OC G1). 1.1 Industry overview For ITSRR s purposes, a railway is a guided system designed to transport passengers or freight on a railway track, together with related infrastructure. This includes the heavy rail systems of RailCorp, Pacific National and the Australian Rail Track Corporation (ARTC) as well as light rail, monorail and various tourist and heritage operations. The heavy rail sector in NSW comprises four segments: below-rail or track and infrastructure networks above-rail or fleet and train operations on the networks ancillary facilities including terminals, stations and sidings support services such as construction and maintenance of infrastructure. Below-rail There are four primary below-rail networks in NSW (Figure 1): The Metropolitan Rail Area network () is centred in Sydney and comprises about, km of track. It is under the management and control of RailCorp and mainly used by CityRail passenger trains. It also services national and intrastate freight transport. The Defined Interstate Rail Network (DIRN) covers almost 3,3 km of track. The DIRN extends beyond NSW and connects the with the other mainland state capitals. It is under the management and control of ARTC and is used primarily for freight transport and long distance passenger trains including CountryLink. The Hunter Valley network (Hunter) is about 7 km in length and runs to the port of Newcastle. It is under the management and control of ARTC and mainly used for coal freight. It is also used for other freight and passenger services. The Country Regional Network () covers some,7 km of operational and 3, km of non-operational lines. It is under the management and control of ARTC on behalf of the Rail Infrastructure Corporation. The main traffic on this network is bulk commodities including grain. It is also used for other freight and passenger services. Above-rail The above-rail segment comprises rail fleets for passenger and freight tasks. A summary of passenger and freight movements over the past five years is shown in Figures to 4. The passenger segment is dominated by RailCorp s CityRail (Table 1). This is a commuter railway with about 1,7 carriages, 37 stations and providing about,6 services per weekday. CityRail provided 33 million passenger journeys in 9 1 (Table 1), representing almost 98% of all passenger journeys in NSW. CountryLink, also owned by RailCorp, is the principal provider of long distance passenger rail services. It provides about daily train services. In 9 1 the CountryLink fleet undertook some 1.8 million passenger journeys to regional NSW and interstate destinations. Heritage railways also carry passengers in NSW (Table ). Operators include Zig Zag Railway Co-op Ltd and Sydney Tramway Museum. Some operators use their own track, while others operate on the and the DIRN. The freight task comprises two main elements bulk freight such as coal and grain and intermodal freight such as containerised goods. In NSW the largest freight task is coal haulage in the Hunter Valley. Pacific National is the biggest rail freight hauler in NSW (Table 3), representing almost 8% of total freight train km travelled in NSW. While there has been a decrease in total freight movements in NSW over the past five years (Figure 4), coal haulage in the Hunter Valley has increased 3. In this report, the allocation of incidents to networks (for example, DIRN, ) is generally based on the nearest reported location as supplied in the occurrence notification. The statistics for each network type may include incidents in yards that adjoin the running line of each network. 3 Australian Rail Track Corporation, 9 Annual Report, ARTC, Mile End, 9 4 Independent Transport Safety and Reliability Regulator

Ancillary facilities Ancillary facilities include stations, terminals and private line connections to the networks. Nearly all operational stations in NSW are controlled by RailCorp. The light rail and monorail tracks and facilities in Sydney, including respective stations, are controlled by Veolia Transport. Some terminals are owned by above-rail freight operators, for example the Sydney Freight Terminal at Chullora is owned by Pacific National. terminals are controlled by third parties, for example Port Waratah (Newcastle) by Port Waratah Coal Services. Privately owned lines and sidings connected to the NSW network include balloon loops from coal mines, grain silo sidings and lines in industrial areas such as Port Kembla. Regulation Rail safety regulation in Australia is based on a co-regulatory model. Under this model, rail transport operators are responsible for assessing the safety risks associated with their railway operations and establishing a safety management system to identify and control these risks. Rail safety regulators are responsible for accrediting rail transport operators. The purpose of accreditation is to attest that a rail transport operator has demonstrated it has the systems, capacity and competency to manage risk and assure safety. Regulators also monitor the safety performance of rail transport operators and undertake compliance and enforcement activities such as education and awareness programs, audits, inspections and investigations. Support services Organisations that construct or maintain infrastructure or rolling stock also need to comply with the Rail Safety Act 8. Some rail transport operators conduct a substantial amount of maintenance in-house. However, for some tasks such as infrastructure maintenance, a comparatively high proportion of work is undertaken by contracting organisations. Narrabri Armidale Coffs Harbour Broken Hill Cobar Werris Creek Dubbo Orange Lithgow Broadmeadow SYDNEY Griffith Goulburn Junee N Kilometres Albury CANBERRA Legend 5 1 3 4 Rail Centreline Copyright RailCorp 7 Figure 1: Major rail networks in NSW Introduction Rail Industry Safety Report 9 1 5

3 Passenger journeys (million) 8 6 4 Jul 5 Oct 5 Jan 6 Apr 6 Jul 6 Oct 6 Jan 7 Apr 7 Jul 7 Oct 7 Jan 8 Apr 8 Jul 8 Oct 8 Jan 9 Apr 9 Jul 9 Oct 9 Jan 1 Apr 1 Figure : Passenger journeys on the NSW rail network, 5 1 Shows monthly total passenger journeys and smoothed trend. Based on data provided by NSW rail transport operators. 4. Passenger train km (million) 3.8 3.6 3.4 3. 3. Data not comparable (different method of calculation).8 Jul 5 Oct 5 Jan 6 Apr 6 Jul 6 Oct 6 Jan 7 Apr 7 Jul 7 Oct 7 Jan 8 Apr 8 Jul 8 Oct 8 Jan 9 Apr 9 Jul 9 Oct 9 Jan 1 Apr 1 Figure 3: Passenger train km on the NSW rail network, 5 1 Shows monthly total passenger train km and smoothed trend. Based on data provided by NSW rail transport operators. Data prior to 8 not shown due to a change in method of calculation by RailCorp which limits comparability over time. 1.8 Freight train km (million) 1.7 1.6 1.5 1.4 1.3 1. Jul 5 Oct 5 Jan 6 Apr 6 Jul 6 Oct 6 Jan 7 Apr 7 Jul 7 Oct 7 Jan 8 Apr 8 Jul 8 Oct 8 Jan 9 Apr 9 Jul 9 Oct 9 Jan 1 Apr 1 Figure 4: Freight train km on the NSW rail network, 5 1 Shows monthly total freight train km and smoothed trend. Based on data provided by NSW rail transport operators. 6 Independent Transport Safety and Reliability Regulator

Table 1: Top five commercial passenger rolling stock operators in NSW (by passenger journeys) 9 1 Operator Passenger journeys Passenger train km RailCorp CityRail 33,19,555 36,48,379 Veolia Transport Light Rail,768,65 454,73 Veolia Transport Monorail,473,186 17,74 RailCorp CountryLink 1,87,156 5,4,439 Perisher Blue Pty Ltd 194,515 73,85 Table : Top five heritage rolling stock operators in NSW (by passenger journeys) 9 1 (operating as at 3 June 1) Operator Passenger journeys Passenger train km Zig Zag Railway Co-op Ltd 76,48 3,39 Sydney Tramway Museum 51,588 5,633 NSW Rail Transport Museum 17,495 3,93 381 Limited 9,17 14,5 Australian Railway Historical Society (ACT Div) 8,99 7,81 Table 3: Top five freight rolling stock operators in NSW (by freight train km travelled) 9 1 Operator Freight train km Pacific National Pty Ltd 13,17,171 Interail Australia Pty Ltd 634,8 Independent Railways of Australia Pty Ltd 596,464 Queensland Rail Pty Ltd 555,378 South Spur Rail Services Pty Ltd 478,46 1. Rail safety reporting Background Section 63 of the Rail Safety Act 8 requires rail transport operators to notify ITSRR of safety-related incidents on the NSW rail network. The specific types of incidents to be notified, termed notifiable occurrences, are defined in clause 7 of the NSW Rail Safety (General) Regulation 8 and summarised in section 7. Notifiable occurrences must be reported to ITSRR within 7 hours of the rail transport operator becoming aware of the occurrence taking place. Occurrences are notified using a standard form. The specific information to be provided to ITSRR is based on the national Guideline for the Reporting of Notifiable Occurrences: Occurrence Notification Standard One (ON S1 8) 4. Notifiable occurrence statistics For the purpose of analysis and reporting, each notifiable occurrence report received by ITSRR is classified into one of the standard national occurrence categories defined in OC G1 5. Classification is generally done at the time of notification based on the information available at that time. However, in some circumstances the classification of an occurrence record may change at a later date as further information comes to hand. The actual unit of classification under the national classification scheme is the top event of an occurrence. The top event is essentially the event with the greatest adverse outcome. A notifiable occurrence may often comprise several events, for example, a signal passed at danger (SPAD) followed by a derailment at catch points. Under the national classification scheme this example would be classified (and counted) as a derailment. 4 Rail Safety Regulators Panel, Guideline for the Reporting of Notifiable Occurrences: Occurrence Notification Standard One (ON-S1), RSRP, June 8 5 Rail Safety Regulators Panel, Guideline for the Top Event Classification of Notifiable Occurrences: Occurrence Classification Guideline One (OC G1), RSRP, June 8 Introduction Rail Industry Safety Report 9 1 7

Changes in occurrence classification The national occurrence classification scheme was revised in June 8 as OC G1 and implemented by NSW rail transport operators in January 9. All notifiable occurrence records held by ITSRR prior to January 9 were classified to a former version of the scheme 6. ITSRR continues to reclassify historical records for important types of occurrences to OC G1. Most statistical summaries in this report are based on the occurrence category definitions of OC G1. This includes summaries of associated factors such as person type (for example, passenger, public, employee) and train type (for example, freight train, passenger train). Where categories or related data items differ to OC G1 these are noted in the text. Safety trends through time ITSRR s notifiable occurrence records extend back over years. In this time there have been major changes in the legislative requirements for the type of occurrences to be notified, their classification and data custodianship. Consequently, some of the observed changes in occurrence rates over time and between sectors are artefacts of changes in the processes governing incident data capture, coding and reporting. For this report, most historical data series are restricted to the most recent five years, that is, from 1 July 5. This period of data excludes most of the major changes to reporting processes in NSW. For example, on 1 January 4, the administration and enforcement of the Rail Safety Act was transferred to ITSRR. From this time a larger number of rail transport operators undertaking a greater range of operations began notifying occurrences directly to ITSRR. Also, the most recent structural changes to the rail industry in NSW occurred during 4 and the national occurrence classification scheme was introduced in NSW at the end of that year. However, the period of data in this report is still subject to changes in reporting process and these issues are noted in the text of relevant chapters. Informing safety management Occurrence-based rail safety statistics (on their own) represent safety issues that have progressed to an unacceptable state such as a broken rail, or an adverse outcome such as injury or damage. Also, because of the need to notify in a timely manner, notifications often lack information on causal factors, which usually only come to light after investigation. In order to prevent serious incidents such as train collisions and derailments, safety performance monitoring must also incorporate measures of the safety-related systems and activities that are designed to prevent harm. Knowledge of the state of organisational, procedural and behavioural systems designed to prevent accidents ensures that any safety deficiencies are identified at an early stage, before they progress to the point of a failure or adverse outcome. Such measures are referred to as lead indicators of safety because they help to flag potential safety issues. Examples of lead indicators include the proportion of required training delivered and the proportion of scheduled safety audits performed. ITSRR therefore uses a variety of safety information, in addition to notifications, to measure safety performance and guide regulatory activity. These include findings from investigations of rail accidents both within and outside NSW to identify hazardous events and contributing factors relevant to NSW railways. Safety risk modelling also plays a critical role, particularly in relation to lower frequency, high consequence risks that may not be apparent over the decade or two that occurrence data is available. A large part of ITSRR s regulatory activity comprises safety audits and compliance inspections, which focus on rail transport operators capacity and competency to identify safety risks relevant to their operations and ensure they are appropriately controlled 7. 6 Rail Safety Regulators Panel, Occurrence Categories and Definitions (ON-S1), RSRP, August 4 7 ITSRR s regulatory activities are described in ITSRR s Annual Report 9 1. See ITSRR s website <http://www.transportregulator.nsw.gov.au> 8 Independent Transport Safety and Reliability Regulator

. Rail-related injury in NSW.1 Fatality Summary There were no injury-related passenger fatalities notified in 9 1 and the annual number of fatalities has decreased over the past two decades. Two track workers were fatality injured in separate occurrences in 9 1. There were no fatalities to members of the public (other than trespass and suspected suicide) in 9 1, and the annual number of fatalities has decreased over the past two decades. Passengers There were no injury-related passenger fatalities on the NSW rail network in NSW in 9 1. The number of passenger fatalities has decreased over the 1-year period of record (Figure 5) from a median of 5.5 fatalities per year in the first decade to 1. fatalities in the decade to 9 1. The nature of fatal occurrences has changed over time. The main cause of passenger fatality in the first decade was falls from trains, until a program of central locking in urban and interurban passenger services in the late 199s virtually eliminated this type of occurrence 8. There have been three multi passenger fatality accidents in NSW in the 1 year period of available data. The last of these was the Waterfall accident in January 3. Since this time, the number of passenger fatalities has remained stable at about one per year despite successive increases in annual patronage over this time. Five of the past six passenger fatalities were the result of falls from platforms into the path of trains. The remaining fatality was the result of an assault. Employees Two railway employees were killed on the NSW rail network in 9 1 in separate occurrences. Both were track workers fatally injured as a result of being struck by trains that entered the section of track where work was taking place. Both accidents are under investigation (Table 4). About one-third of all employee fatalities were associated with accidents of this nature, based on the 1-year fatality history (Figure 5). The death of six track workers over a two-year period in the late 199s led to a number of revisions to the rules and procedures to improve track worker safety. However, since this time there have been another four deaths including the two in 9 1. The safety of track workers and the effectiveness of the available methods for their protection was a regulatory priority for ITSRR in 9 1 and this issue is discussed further in section 4.. Table 4: Fatalities on the NSW rail network, 9 1 Excludes fatalities associated with ill-health, trespass and suicide. Date Category Location Description 13 April 1 Train collision with person Kogarah 1, () CityRail interurban passenger train travelling to Central Station struck a worker who was part of a station cleaning team. 5 May 1 Train collision with train (track machine) Newbridge,3 () CountryLink passenger train travelling from Sydney to Orange collided with a track mounted excavator. The operator of the excavator was fatally injured. 1 Subject of investigation by OTSI Subject of compliance investigation by ITSRR 3 Subject of investigation by Australian Transport Safety Bureau (ATSB) 8 Independent Transport Safety and Reliability Regulator, Train door emergency egress and access and emergency evacuation procedures, ITSRR, November 4 Rail-related injury in NSW Rail Industry Safety Report 9 1 9

Members of the public (other than trespass/suicide) There were no fatalities to members of the public in 9 1. The 1-year history of public fatalities on the NSW rail network is shown in Figure 5. There has been a gradual decrease in public fatalities over this period from a median of 5.5 fatalities per year in the first decade to 1.5 fatalities per year in the decade to 9 1. This improvement was associated with a gradual reduction in collisions at level crossings which is the primary means by which members of the public (other than passengers) are exposed to rail-related risk. This issue is discussed further in section 3.. Trespass and suicide A note on trespass and suicide statistics Suicide occurrences are distinct from those of trespass because they are associated specifically with intentional self harm. Trespass occurrences comprise a broader range of motives including suicide. Fatality statistics for the two occurrence types are combined because, in most cases, the specific intent of the deceased cannot be reliably determined at the time of the occurrence. A coronial inquiry is required to confirm a death as suicide and even then a statement of intent may not be made by the coroner. With this caution in mind, based on limited description of circumstances of each occurrence at the time of notification, about 8% of occurrences in 9 1 were suspected to be acts of suicide. 3 Passenger 1 Employee Number of fatalities 5 15 1 5 Hawkesbury River Glenbrook Waterfall Train accident Fall Strike Number of fatalities 8 6 4 Train accident Fall Strike 89-9 9-91 91-9 9-93 93-94 94-95 95-96 96-97 97-98 98-99 99- -1 1- -3 3-4 4-5 5-6 6-7 7-8 8-9 9-1 89-9 9-91 91-9 9-93 93-94 94-95 95-96 96-97 97-98 98-99 99- -1 1- -3 3-4 4-5 5-6 6-7 7-8 8-9 9-1 14 Public (other than trespass/suspected suicide) 8 Trespass/suspected suicide Number of fatalities 1 1 8 6 4 Level crossing Fall Train accident Number of fatalities 6 4 Strike Fall 89-9 9-91 91-9 9-93 93-94 94-95 95-96 96-97 97-98 98-99 99- -1 1- -3 3-4 4-5 5-6 6-7 7-8 8-9 9-1 89-9 9-91 91-9 9-93 93-94 94-95 95-96 96-97 97-98 98-99 99- -1 1- -3 3-4 4-5 5-6 6-7 7-8 8-9 9-1 Figure 5: Rail-related fatality on the NSW rail network, 1989 9 to 9 1 Vertical bar is annual fatality count. Employee includes volunteer and contractor. Excludes fatalities associated with ill-health for example, stroke. Strike is collision of train with person. Labelled bars for passenger fatalities indicate multi-fatality train accidents. 1 Independent Transport Safety and Reliability Regulator

There were 6 fatalities associated with acts of trespass or suspected suicide in 9 1. About 9% of these fatalities were the result of being struck by a train. Twenty-two occurrences were on the. The number of fatalities in 9 1 was higher than that in each of the previous two years but remained consistent with a longer term decrease in this type of occurrence since the late 199s (Figure 5). While these types of occurrences are beyond the direct control of railways, rail transport operators seek to influence behaviour and reduce trespass by various means including surveillance and enforcement, education and maintenance of fencing along higher risk areas of the corridor. The decrease in fatalities may also reflect broader societal change the number of suicides on a national basis has also gradually decreased in the decade to the end of 8 9. Assault was the second most common cause of passenger injury in 9 1 (about 15% of passengers injured; Figure 6). Longer term data on assault-related incidents 11 is compiled by the NSW Bureau of Crime Statistics and Research (BOCSAR) based on crimes reported to, or detected by, police 1. The monthly rate of recorded assault on NSW railway premises (per million passenger journeys) has not changed significantly over the period of the BOCSAR data (Figure 8). Two passengers were injured as a result of being struck by trains in 9 1 (Table 5). Both occurrences occurred at the train platform interface. Five of the past six passenger fatalities in NSW were associated with this type of occurrence, highlighting it is a significant risk. However, excessive consumption of alcohol was also a contributing factor in more than one of the occurrences. A note on injury severity. Non-fatal injury Summary A total of 556 people were transported to hospital because of injury on railway property in 9 1. About 9% of people injured in 9 1 were passengers. Most injuries were associated with falls and the rate of fall-related injury (per million passenger journeys) has decreased over time. Two train crew members were injured in separate train accidents in 9 1 a freight train derailment and a passenger train collision with buffer stops. Passengers A total of 485 passengers were transported to hospital as a result of injury on railway property in 9 1. This corresponded to 1.6 episodes per million passenger journeys which was lower than that for the previous two years in NSW and lower than the comparable figure for Great Britain (. for 9) 1. Passenger injuries are summarised by occurrence type in Figure 6. The major cause of injury was slips, trips and falls which is consistent with the general risk profile of passenger railways in other jurisdictions. Most passenger falls occurred at the busier stations of the during the morning and afternoon commuting peak. Two-thirds of passenger falls occurred on stairs, escalators or platforms. There was a gradual decrease in the rate of fall-related injuries (per million passenger journeys) over the previous three years (Figure 7). Under OC G1 a person injured in a notifiable occurrence is counted as a serious injury if they are admitted to hospital as a result of their injury. A person is counted as a minor injury if they require medical attention but do not require admittance to hospital. The grading of injury severity is therefore based on the nature of medical care received. Unfortunately, rail transport operators nationally have great difficulty in determining the specific nature of health care received by people after they have left railway premises, for example, determining whether or not a person transported to hospital was formally admitted. Alternative and various criteria are used by rail transport operators to grade injury severity, for example, workplace injury scales 13, and length of time of incapacitation. These criteria vary over time and between operators introducing inconsistency into data and limiting comparability 14. In this report, instead of using the hospital admission definition of serious injury, injury statistics are based on a criterion of ambulance transported. Information to support this criterion is generally available in the occurrence narrative and provides an opportunity to delineate more significant injuries in a consistent manner. This criterion is far broader (more inclusive) than the hospital admittance criterion as less than half of all persons presenting at hospital will be admitted. It is, however, similar to definitions used elsewhere 15. 9 Australian Bureau of Statistics 1, 137. Measures of Australia s Progress, 1, viewed September 1 <http://www.abs.gov.au> 1 Calendar years, Office of Rail Regulation, National rail trends 9 1 yearbook, ORR, London, 1 11 As defined in: Australian Bureau of Statistics, Australian standard offence classification, cat. no. 134., ABS, 1997 1 NSW Bureau of Crime Statistics and Research, unpublished quarterly criminal incident statistics (quarterly update to end June 1), 8 September 1 13 Standards Australia, Workplace Injury and Disease Recording Standard, AS 1885.1 199, SA, Homebush, 199 14 Issues with injury reporting and classification are within the scope of the National Data Strategy, led by the Rail Safety Regulators Panel 15 For example, Office of Rail Regulation, National rail trends 9 1 yearbook, ORR, London, 1 Rail-related injury in NSW Rail Industry Safety Report 9 1 11

Employees There were 5 employees injured on railway premises in 9 1 (Figure 6). Most employee injuries were distributed equally amongst three occurrence categories slip, trip, fall; alleged assault and other. The latter category comprised workplace-type incidents including crush injuries, persons struck by plant and electrocution. Two employees were injured in separate train accidents, a derailment and collision with stop blocks at a terminal station (Table 5). Passenger Employee Slip, trip, fall Alleged assault 7 79 1 4 49 445 Slip, trip, fall Alleged assault 6 6 9 9 9 15 Strike Collision/ derail Fire 11 7 18 5 1 1 9 1 8 9 7 8 Strike Collision/ derail Fire 8 9 13 9 1 8 9 7 8 1 3 4 5 Number of injured persons 4 6 8 1 1 14 16 Number of injured persons Public (other than trespass/attempted suicide) Trespass/attempted suicide Slip, trip, fall Alleged assault 1 9 Slip, trip, fall Alleged assault 1 1 3 4 Strike Collision/ derail Fire 1 1 1 4 7 9 1 8 9 7 8 Strike Collision/ derail Fire 5 6 7 1 11 15 9 1 8 9 7 8 4 6 8 1 Number of injured persons 4 6 8 1 1 14 16 Number of injured persons Figure 6: Number of people injured and transported to hospital on the NSW rail network, 9 1 Totals for earlier years may differ from previously reported figures due to revisions. Employee includes volunteer and contractor. Strike is collision of train with person. 1 Independent Transport Safety and Reliability Regulator

On stairs/ escalators On platform Platform/ train interface On train situation Falls by situation 4 8 1 16 Number of injured passengers 9 1 8 9 Number of passengers injured per million passenger journeys.1 1.9 1.7 1.5 1.3 1.1.9 Falls over time Jul 7 Sep 7 Nov 7 Jan 8 Mar 8 May 8 Jul 8 Sep 8 Nov 8 Jan 9 Mar 9 May 9 Jul 9 Sep 9 Nov 9 Jan 1 Mar 1 May 1 Figure 7: Fall-related passenger injury on the NSW rail network Injured passengers transported to hospital. Falls over time shows monthly rate and smoothed trend. Members of the public (other than trespass/ attempted suicide) There were 19 members of the public injured in 9 1. This was higher than the previous two years, however there was marked year to year variation in the counts for some categories (Figure 6). This volatility largely reflects uncertainty in the nature of some public-related occurrences and associated occurrence coding, for example, the relationship of the person to railway at the time of the occurrence and the relevance of the occurrence to railway operations (such as falls in areas at the interface of railway and non-railway property). Trespass and attempted suicide Twenty-seven people required transport to hospital for injuries on railway premises associated with trespass or attempted suicide in 9 1 (Figure 6). At least 18 occurrences were suspected to be acts of self harm with most involving train strikes. Notwithstanding the above, longer term data shows the most common rail-related risk for the general public is level crossing collisions. There were no injuries to members of the public at level crossings in 9 1. Another relevant risk to the public in NSW is associated with pedestrians moving about the Sydney light rail system. One member of the public was struck by a light rail vehicle in 9 1. There have been 15 such occurrences notified to ITSRR over the five years to June 1. Rail-related injury in NSW Rail Industry Safety Report 9 1 13

Table 5: Non-fatal injury (excluding falls) on the NSW rail network, 9 1 Persons injured and requiring transport to hospital as a result of railway-related accidents. Excludes slip, trip or fall; trespass, attempted suicide, alleged assault and occurrences not classified under OC G1, for example, some workplace-type injuries. Date Category Location Description 14 July 9 Train collision running line with person 3 July 9 Train collision running line with person 13 November 9 Train collision running line with person 13 March 1 Derailment running line 3 May 1 Train collision running line with infrastructure 1 Subject of investigation by ITSRR Subject of investigation by OTSI Sydney (Light Rail) Flemington () Granville () Whittingham 1, (Hunter) Newcastle () Light rail vehicle travelling at low speed struck pedestrian who then rolled underneath the train. Pedestrian taken to hospital. Person walked into passenger train as it was departing the station. Train stopped. Passenger taken to hospital. Intoxicated person on platform fell between platform and passing freight train. Person with serious injuries taken to hospital. Empty grain train derailed two locomotives and eight wagons at incorrectly set points. Train driver with shoulder injury taken to hospital. Passenger service struck stop blocks. No passengers injured. Train guard with arm injuries taken to hospital. Number or recorded assaults per million passenger journeys 6. 5.5 5. 4.5 4. 3.5 3. Jan 4 Apr 4 Jul 4 Oct 4 Jan 5 Apr 5 Jul 5 Oct 5 Jan 6 Apr 6 Jul 6 Oct 6 Jan 7 Apr 7 Jul 7 Oct 7 Jan 8 Apr 8 Jul 8 Oct 8 Jan 9 Apr 9 Jul 9 Oct 9 Jan 1 Apr 1 Figure 8: Assault on rail premises in NSW, 4 to 9 1 Shows monthly rate and smoothed trend. Based on recorded criminal incidents from BOCSAR and passenger journey data reported to ITSRR by rail transport operators. 14 Independent Transport Safety and Reliability Regulator

3. Accidents and other key rail safety occurrences This section presents statistics for four OC G1 occurrence categories that are most closely associated with significant injury: running line derailment, level crossing collision, running line train collision and fire. A subset of incidents within each of these national categories represent high risk occurrences, that is, with potential to cause multiple injuries and fatalities. It is the higher risk occurrences that are the focus of this chapter. 3.1 Train derailment Summary A total of 45 derailments on or affecting the safety of running lines were notified in the 9 1. Three running line passenger train derailments were notified in 9 1 with no injures reported for any of these occurrences. A total of 1 freight train derailments were notified in 9 1. The driver of a train that derailed at Whittingham suffered an injury and was taken to hospital. A total of 45 running line and 15 yard derailments were notified to ITSRR in 9 1. Running line derailments are typically the higher risk of the two subcategories and are summarised in Figure 9. Three running line derailments in 9 1 involved passenger trains. This count was consistent with that observed over the previous four years (Figure 1). Each of these occurrences is summarised in Table 6. The most significant occurrence was the derailment of an intercity passenger service in the Blue Mountains, west of Sydney. The leading wheels of the train derailed after hitting a landslide during wet weather. Twenty-one running line derailments in 9 1 involved freight trains (Figure 9). Eighteen of the 1 derailments were associated with the main rail networks. About half of these occurrences were associated with low speed movements of trains to/from yards. In some cases the point of derailment was within a yard but the safety of a running line was affected. ITSRR is investigating one occurrence at Blayney in July 9 in which wagons were shunted past a signal at stop and blocked a level crossing. Freight train 1 (,1) Track machine 15 (,) rolling stock Passenger train 3 (,) 6 (,) 5 1 15 5 Figure 9: Derailment (running line) on the NSW rail network, 9 1 In brackets are number of fatalities and number of injured people transported to hospital respectively. rolling stock comprises light engines and non-powered rolling stock such as freight wagons. Accidents and other key rail safety occurrences Rail Industry Safety Report 9 1 15

Table 6: Passenger train derailments (running line) on the NSW rail network, 9 1 Date Train type Location Description 13 September 9 Tourist Sydney Olympic Park (Millenium Parklands Railway) Passenger train derailed on points. Train was rerailed by train crew while passenger remained on carriage. No injuries reported. 4 February 1 Passenger interurban 4 April 1 Passenger urban Woodford () Penrith () Interurban passenger train en route to Sydney derailed lead wheels after striking landslide during torrential rain. No injuries reported. Empty passenger train passed signal at stop without authority and derailed at catchpoints in yard. Derailed train fouled adjacent running line. No injuries reported. The remaining freight train derailments in 9 1 are summarised in Table 7. While the cause of derailment was not stated in all cases, a range of factors are highlighted, including track and rolling stock irregularities as well as human error. Two of the derailments were the result of a seized wheel bearing leading to a sheared axle. This type of occurrence is discussed further in section 4.4. Two other occurrences were associated with the uneven loading of coal wagons. ITSRR also issued a rail industry safety notice in relation to one of these occurrences (Coggan Creek, March 1), the cause of which was suspected to be an empty coal wagon placed between loaded wagons creating forces that lifted the unloaded wagon off the rails. Previous rail industry safety reports have noted a decrease in the number of freight train derailments over time, based on data extending back over a decade. This trend was not apparent in the most recent five years of data (Figure 11), however the number of running line freight train derailments remains at a historical low. Table 7: Freight train derailments (running line) on the NSW rail network, 9 1 Excludes derailments associated with entry to/from yards, shunting and track possession. Date Location Description 14 July 9 Mt Victoria () August 9 Wyee () 1 November 9 Combarra () 4 January 1 Exeter 1 (DIRN) 13 March 1 Whittingham,3 (Hunter) 19 March 1 Coggan Creek (Hunter) March 1 Antienne (Hunter) 8 June 1 Coniston () Freight train parted between two wagons. It was subsequently discovered that one of the affected wagons had derailed and rerailed itself over points. Suspected cause not stated. Freight train derailed one wagon. Suspected cause was a sheared axle due to wheel bearing fault. Freight train derailed six wagons. Suspected cause was a broken rail. Freight train derailed one wagon. Suspected cause was a sheared axle due to wheel bearing fault. Two locomotives and eight wagons of freight train derailed. Suspected cause was the incorrect setting of points. Freight train derailed one wagon. Suspected cause was associated with a single unloaded coal wagon within a consist of loaded wagons. Freight train arrived in yard with missing front bogie. Bogie from same wagon derailed the previous day at Coggan Creek. Suspected cause not stated. Freight train derailed two wagons. Suspected cause was associated with a single partially loaded wagon within a consist of empty wagons. 1 Subject of investigation by ATSB Subject of compliance investigation by ITSRR 3 Subject of investigation by OTSI 16 Independent Transport Safety and Reliability Regulator

8 6 4 Derailment by line type Running line Yard 5 15 1 5 Running line derailment by network 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 1: Derailment on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Solid line is quarterly occurrence rate (count per million train km travelled). Derailment passenger train Derailment freight train 5 4 3 1 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q.4.. per million train km 15 1 5 3 1 per million train km 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 11: Passenger and freight train derailment (running line) on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Solid line is quarterly occurrence rate (count per million train km travelled). Accidents and other key rail safety occurrences Rail Industry Safety Report 9 1 17

3. Level crossing collision Summary Eight level crossing collisions between a train and road vehicle were notified in 9 1. One collision resulted in minor injuries to a train driver and truck driver. The annual number of collisions at level crossings has decreased gradually over the past two decades. A total of 1 level crossings have been closed since and another 55 level crossings were upgraded from passive to active control in 9 1. Level crossings provide the main point of interaction between rail and road traffic and therefore present a high risk for serious collisions between trains and road vehicles. There are about 3,8 level crossings in NSW 16 and most are located in regional areas. About two-thirds of all crossings in NSW are associated with private roads and/or rail yards and the majority of these are passively controlled. The remaining crossings intersect public roads and running lines of the NSW rail network. About one-quarter of these are actively controlled and the remainder are passively controlled. There were eight collisions between train and road vehicle at separate level crossings in 9 1, summarised in Table 8 and Figure 1. Six of the eight collisions in 9 1 occurred at passively controlled crossings which was consistent with the historical record and reflected, in part, the far larger number of passive crossings in NSW compared to active crossings. However, results of risk modelling and accident research also show the chance of collision is reduced with higher levels of protection in place, for example, boom gates rather than stop signs. The most important type of level crossing collision in terms of potential for multiple casualties is that between a passenger train and a road vehicle. As well as the threat to road users posed by such incidents (56 killed in 5 separate occurrences over the 1 years to 9 1), an individual collision exposes multiple passengers to potential harm. There was one such occurrence in 9 1 a CountryLink train struck the front of a car at a passive crossing. In this case there were no injuries and damage to the train was described as minor. The seven remaining collisions in 9 1 involved freight trains or light locomotives. Three of these were relatively lower risk incidents associated with shunting in or adjacent to yards. Interestingly, in each of these cases an employee road vehicle was involved. One incident within a steelworks at Port Kembla yard involved a locomotive striking a truck and then derailing. A contributing factor to this incident was the active controls being disabled at the time of the occurrence. Road Vehicle (passive) 6 (,) Road Vehicle (active) (,) Person (passive) Person (active) 1 3 4 5 6 7 8 Figure 1: Level crossing collision on the NSW rail network, 9 1 In brackets are number of fatalities and number of injured people transported to hospital respectively. 16 NSW Level Crossing Strategy Council, Yearly Report 9/1, Level Crossing Safety Improvement Programs, LCSC, 1 18 Independent Transport Safety and Reliability Regulator

Table 8: Level crossing collisions between trains and road vehicles on the NSW rail network, 9 1 Date Crossing control Location Description 15 August 9 Passive Nemingha () 3 September 9 Passive Sandown 1 () 14 September 9 Passive Yennora Industrial Yard () 7 November 9 Passive Botany Yard () 6 February 1 Passive Wee Waa () 17 March 1 Active Port Kembla Steelworks () 1 April 1 Active Narromine (DIRN) 7 June 1 Passive Stockingbingal (DIRN) CountryLink passenger service travelling at 7km/h struck the bullbar of a car that had failed to stop at the level crossing. There were no injuries reported and only minor damage to both the train and the road vehicle. During shunting in the Sandown branch line, a freight train struck and damaged an employee s car parked on a level crossing. Freight train shunting at 7 km/h struck an employee s four wheel drive which moved onto the crossing. No injuries reported. The road vehicle sustained minor damage. Freight train collided with rail employee s road vehicle on level crossing. No injuries reported. Freight train was struck by a car that failed to stop in time. No injuries reported. Train unable to continue. Road vehicle extensively damaged. Locomotive struck a B-Double. Locomotive derailed and extensively damaged. Active control not operational at time of incident due to track work. Locomotive driver and truck driver treated for minor injuries. Freight service struck a car while shunting at Narromine. Leading locomotive collided with the rear of the car which continued through the crossing and failed to stop. Car struck wagons of stationary freight train. Driver of road vehicle returned positive reading to a road side breath test. 1. Subject of compliance investigation by ITSRR The full history of level crossing collisions over time and across networks is shown in Figures 13 and 14 respectively. Historically, most collisions at passive crossings occurred on the and most collisions at active crossings on the and DIRN. There was a significant decrease in the number of collisions at both active and passive crossings over the 1-year record, across all networks. The number of fatalities associated with level crossing collisions has also dropped (see section.1). A contributing factor to the improvement in level crossing safety in NSW over the longer term is a number of safety initiatives undertaken by NSW rail and road authorities under the coordination of the NSW Level Crossing Strategy Council 17. These included 18 : Level crossing closures: A total of 1 level crossings have been closed since. During 9 1 the Minister for Transport approved the closure of eight level crossings. Level crossing upgrades: A total of 55 passive crossings were earmarked for upgrade from passive to active control in 9 1. This program finished on schedule with 53 crossing upgrades completed by ARTC and the remaining two completed by RailCorp. Awareness and enforcement actions: The vast majority of level crossing collisions are the result of road user behaviour. Targeted awareness and enforcement actions are a critical element of managing the risk at level crossings. In 9 1 the NSW Police Force was involved in ongoing and targeted patrols of level crossings to improve road user behaviour. 17 See Transport NSW s level crossing website <http://www.transport.nsw.gov.au/levelcrossings> 18 NSW Level Crossing Strategy Council, Yearly Report 9/1, Level Crossing Safety Improvement Programs, LCSC, 1 Accidents and other key rail safety occurrences Rail Industry Safety Report 9 1 19

Level crossing near miss Near miss incidents were not routinely reported as notifiable occurrences until their introduction in the national occurrence classification scheme in 5 and subsequent definition within legislation in 6. Near miss incidents represent a sequence of events that has progressed to a point just short of actual harm being realised. Patterns in near miss data, such as multiple incidents at a specific crossing, serve as potential warning signs of a situation requiring attention. There was a slight decrease in the overall number of reported near misses over time which is consistent with the decrease in collisions. However, the relationship between near miss incidents and collisions is not straightforward. For example, crossings with a history of near misses are not necessarily ones where collisions occur and vice versa. Further, two-thirds of collisions occur at passive crossings whereas active crossing dominate the near miss statistics. The complex relationship between near miss and collision incidents highlight the fact that many factors contribute to collisions at level crossings. The monitoring and assessment of risk therefore requires the use of various types of information in addition to near miss data, for example, risk modelling and site assessment. 6 5 4 3 1 Active control Passive control Near miss with road vehicle, 5 6 to 9 1 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Near miss with road vehicle by control type, 9 1 1 3 4 5 6 7 36 57 Collision with road vehicle active control Collision with road vehicle passive control 3 5 15 1 5 3 5 15 1 5 89-9 9-91 91-9 9-93 93-94 94-95 95-96 96-97 97-98 98-99 99- -1 1- -3 3-4 4-5 5-6 6-7 7-8 8-9 9-1 89-9 9-91 91-9 9-93 93-94 94-95 95-96 96-97 97-98 98-99 99- -1 1- -3 3-4 4-5 5-6 6-7 7-8 8-9 9-1 Figure 13: Level crossing train and road vehicle collision on the NSW rail network, 1989 9 to 9 1 Vertical bar is annual occurrence count. Independent Transport Safety and Reliability Regulator

Namoona Camurra Narrabri Armidale Coffs Harbour Cobar Werris Creek Broken Hill Dubbo Griffith Parkes Broadmeadow Orange Lithgow SYDNEY Goulburn Legend Number of collisions at active crossings 5 4 3 1 Junee Number of collisions at passive crossings 9 N CANBERRA 7 6 5 4 3 1 Albury NSW rail network Kilometres 8 16 3 Rail Centreline Copyright RailCorp 7 Figure 14: Level crossing train and road vehicle collision across the NSW rail network, 1989 9 to 9 1 Shows level of control at the time of the accident. Some crossings have now been closed or upgraded. 3.3 Collision other than at level crossing Summary Six running line collisions between trains were notified in 9 1. A collision between an intercity passenger train and a track machine on the resulted in the death of the track machine operator. A collision between two monorail vehicles on the Sydney monorail system resulted in minor injuries to several passengers and the drivers of the two trains. Under OC G1, train collisions on running lines are classified separately to those in yards. Collisions on or affecting the safety of running lines are higher risk occurrences and these are summarised in Figure 15. The categories of running line train collision vary markedly in terms of the typical safety risk they pose. For example, about 9% of occurrences in 9 1 comprised trains hitting animals and smaller objects on or about the track. While these types of incidents are relatively frequent, they usually result in no injuries and little or no damage. There were exceptions however in February 1 a passenger train derailed after striking a landslide in the Blue Mountains, west of Sydney (Table 6). The most important classes of running line collision within the OC G1 scheme in terms of safety risk are the less frequent occurrences such as train to train collision and train colliding with person (strike). These occurrences, particularly strikes, often result in injury or fatality. Accidents and other key rail safety occurrences Rail Industry Safety Report 9 1 1

Obstruction 35 (,) Animal 174 (,) Infrastructure Person Between trains Rolling stock Road vehicle (not at level crossing) 3 (,1) 14 (4,7) 6 (1,) 1 (,) (,) 1 3 4 Figure 15: Running line train collision (other than level crossing) on the NSW rail network, 9 1 In brackets are the number of fatalities and number of injured people transported to hospital respectively. Excludes occurrences classified as suspected/attempted suicide and collision running line other. There were six running line collisions between trains in 9 1, summarised in Table 9. There were three train collisions involving passenger trains in 9 1. Accidents involving in service/loaded passenger trains are of particular concern due to their potential for multiple casualties in a single occurrence. There was one such occurrence on the at Newbridge where a CountryLink passenger service struck a track machine 19. While no passengers were injured in this occurrence, the operator of the track machine was fatally injured. This accident resulted from a failure in the application of a specific procedure used to protect track workers from trains. This risk is discussed further in section 4.. A second collision on the Sydney monorail system involving two monorail vehicles resulted in minor injury to five people and is the subject of investigation. The number of running line train collisions in 9 1 was consistent with that observed over the previous four years. Two of the six collisions in 9 1 involved open container doors on freight trains protruding into the path of oncoming trains. This was the single most common form of collision over the past five years, representing one-third of all collisions. Most of these occurred on the DIRN between Sydney and Junee (Figure 17). Table 9: Collision between trains (running line) on the NSW rail network, 9 1 Date Location Train type Description 6 October 9 Bullocks Flat (Perisher) 6 November 9 Campsie () 4 February 1 Epping () Passenger other Freight freight Passenger freight Low speed collision between shunting locomotive and stationary passenger train. Passenger train was empty at time of occurrence. Open container door on freight train struck mirror of freight train locomotive passing on adjacent line. Minimal damage reported. Open container door on freight train struck front of passenger train passing on adjacent line, smashing its windscreen. Monorail monorail Monorail vehicle struck rear of another monorail vehicle. 7 February 1 Darling Park (Sydney Monorail) 1, Passengers from both monorail vehicles evacuated. Three passengers and two employees received minor injuries. 3 April 1 Walgett () 5 May 1 Newbridge,3 () 1 Subject of compliance investigation by ITSRR Subject of investigation by OTSI 3 Subject of investigation by ATSB Track machine track machine Passenger track machine Collision between two track machines. One employee received a minor injury. CountryLink passenger train struck an on-track excavator. Operator of excavator was fatally injured. 19 On-track excavator is train type track machine under ON-S1 8. See footnote 4. Independent Transport Safety and Reliability Regulator

There was a total of 14 collisions with person notified in 9 1. Caution must be attached to these data because they include trespass but exclude suspected suicide and the distinction between the two is not reliable (see note section.1). However, within this incident category there are several types of occurrences under the direct control of railways, namely, trains striking passengers and employees. One of the four fatalities was of this nature a track worker fatality at Kogarah in April 1 (Table 4). Two of the seven serious injuries involved passengers struck by trains (Table 5). There were 3 collisions between train and infrastructure on running lines in 9 1 which was consistent with that observed over the previous four years (Figure 16). A range of circumstances and causes were involved including trains striking high ballast and out of gauge rolling stock striking infrastructure such as platforms. One collision in May 1 involved a passenger train striking buffer stops at the terminal station at Newcastle. While no passengers were injured in this occurrence, the train guard was transported to hospital. 1 Train to train collision Running line Yard 1 Rolling stock collision Running line Yard 1 1 8 6 4 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 8 6 4 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 1 Train to person collision Running line Yard 35 Train to infrastructure collision Running line Yard 1 8 6 4 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 3 5 15 1 5 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 16: Train collision (running line) on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Train to person collision excludes occurrences classified as suspected/ attempted suicide and occurrences at level crossings. Accidents and other key rail safety occurrences Rail Industry Safety Report 9 1 3

Armidale Boggabri Werris Creek Dubbo Broadmeadow Legend Lithgow Orange 1 Picton SYDNEY Goulburn Between freight trains 1 Between passenger and freight trains 1 Between track machine and passenger train 1 Between passenger and other trains 1 Junee Between passenger trains Between track machine and freight train 1 CANBERRA NSW rail network Albury Bullocks Flat N Kilometres 4 8 16 Rail Centreline Copyright RailCorp 7 Figure 17: Collision between train (running line) across the NSW rail network, 5 6 to 9 1 Excludes collisions between track machines. Mapped location is based on nearest reported location as notified by rail transport operators. Collision at Bullocks Flat occurred on an isolated tourist line. 3.4 Fire Summary A total of 541 fires were notified in 9 1. None of the occurrences resulted in fatality or injury requiring transport to hospital. A total of 165 passenger train fires were notified in The greatest fire-related risk on the NSW rail network is passenger train fires, which have the potential to cause multiple casualties in a single occurrence. A total of 165 passenger train fires were notified in 9 1 (Figure 18). About 9% of passenger train fires in 9 1 were associated with arson on the. Most incidents involved 9 1. About 9% of occurrences were associated the lighting of newspapers or attempted lighting of seats with arson. A previously reported increase in these types with about three quarters of incidents occurring between the of occurrences appears to have steadied. hours of 6pm and 1am. These types of fires occur across the A total of 1 freight train fires were notified in 9 1. There was a significant increase in the rate of freight train fires (per million train km travelled) over time. network, although a lower number of incidents were observed for the lines running to the north of Sydney (Figure ). None of the fires in 9 1 resulted in injury requiring transport to hospital. Several occurrences led to the evacuation of passengers from the train. The train guard in one occurrence required medical attention after using a fire extinguisher. 4 Independent Transport Safety and Reliability Regulator

Situation Train fire Lineside 89 (,) Passenger Rolling stock irregularity 17 (,) 148 (,) Train Station 65 (,) 187 (,) Freight Track machine Rolling stock irregularity Rolling stock irregularity 1 (,) 1 (,) 4 8 1 16 4 8 3 36 5 1 15 Figure 18: Fire on the NSW rail network, 9 1 In brackets are number of fatalities and number of injured people transported to hospital respectively. The five-year history of arson-related passenger train fires is shown in Figure 19. Previous reports have noted a significant increase in arson-related train fires over time, based on data extending back to 4. However, analysis of data for the most recent five-year period indicates the rate of incidents appears to have steadied. This suggests that previously reported measures implemented by RailCorp to manage this risk may be having a positive effect..8.4 6 4 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 19: Arson-related passenger train fire on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Solid line is quarterly occurrence rate (count per million passenger journeys).. per million passenger journeys Seventeen passenger train fires in 9 1 were associated with rolling stock irregularities. Most fires were associated with traction motors and air-conditioning units. The count for 9 1 was less than that for the previous year (Figure 1). There was also the suggestion of a longer-term decrease in the rate of rolling stock-related passenger train fires (per million train km travelled) although the length of the record has been truncated by reporting changes. Despite the suggested improvement over time, several occurrences in 9 1 were serious and required the evacuation of passengers from trains. There were 1 freight train fires notified in 9 1. While the 9 1 count was comparable to the previous year (3), it does not take account of the overall decrease in freight train movements over time. The rate of fire had in fact increased over the previous 5 years from.8 per million train km travelled in 4 5 to 1.3 in 9 1 (Figure 1). All but two fires in 9 1 were associated with locomotives. The remaining fires were associated with wagons (sticking brakes). Rail transport operators compliance with rolling stock standards was a corporate priority for ITSRR in 9 1 (see section 4.4). Accidents and other key rail safety occurrences Rail Industry Safety Report 9 1 5

Inset B To Broadmeadow (See Inset B) Wyong Richmond To Lithgow (See Inset A) Hornsby Penrith St Marys Blacktown Parramatta Inset A Lidcombe Katoomba Cabramatta Bankstown Liverpool Yagoona Strathfield Town Hall CENTRAL Redfern Glenfield Hurstville Minto Sutherland Campbelltown Macarthur Legend Inset C N Number of fires 9+ 6 8 4 5 3 1 NSW rail network Dunmore To Bomaderry (See Inset C) Kilometres 4 8 16 Rail Centreline Copyright RailCorp 7 Figure : Passenger-related train fire across the, 8 9 to 9 1 Mapped location is nearest reported location as notified and for some occurrences will represent the location where evidence of fire was first detected. 6 Independent Transport Safety and Reliability Regulator

Passenger train Freight train 15 1 5 1 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q per million train km 1 8 4 3 1 per million train km 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 1: Rolling stock-related train fire on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Solid line is quarterly occurrence rate (count per million train km travelled). About two-thirds of all fire occurrences notified to ITSRR in 9 1 were lineside and station fires with most occurring on the (Figure ). About one-third of lineside fires were classified as affecting safety-related infrastructure, the majority of which involved sleepers. Most station fires were associated with vandalism or careless acts, for example, discarded cigarettes. The longer term pattern in the number of lineside fires (Figure ) was characterised by higher counts towards the end of the period although counts for the last two quarters returned to historical lows. The count of station fires had not changed appreciably over the five-year period. Lineside Station 14 1 1 8 6 4 4 35 3 5 15 1 5 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure : Lineside and station fire on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Accidents and other key rail safety occurrences Rail Industry Safety Report 9 1 7

4. Precursor rail safety occurrences Most of the rail safety occurrences notified to ITSRR each year do not result in an adverse outcome such as injury or damage. Many of these occurrences are precursor events incidents that could, in combination with other events progress to accidents and actual harm. Accident precursors serve as warning signs of failures in safety risk controls employed by rail transport operators. They are particularly important in providing insight to the underlying risks of infrequent, but serious, accidents that are still relevant to a railway under its current level of risk control but have not occurred in the period of available accident data. This section summarises five key groups of precursor occurrence within the national occurrence classification scheme. Assessing longer-term trends in precursor occurrences is complicated by changes in the legislative requirements to report them over time. These effects have been minimised in this report by restricting the analysis to more recent years, after the period of major changes. However, more recent data will still be subject to some reporting changes associated with improvement to rail transport operators safety management systems and associated reporting processes. 4.1 Proceed authority irregularity Summary The number of SPADs on the has decreased over time. This occurred despite an increase in train movements on this network over the same period. The rate of SPADs (per million train km travelled) for freight trains was about 5% higher than that for passenger trains, despite a greater proportion of freight train movements occurring on less densely signalled or unsignalled track. In NSW there are four principal systems used to manage the movement of rail traffic in a way that ensures adequate separation of trains and prevents conflicting movement. An integral part of each of these systems is a means to authorise the movement of a train from one section of track to another. For the and the majority of the DIRN and Hunter the authority for a train to proceed is given by a signal indication. For much of the and the western section of the DIRN, an authority to proceed is given via the train driver s possession of some form of token, for example, a metal rod (staff) and/or the issue of a written or verbal authority. Figure 3 summarises occurrences in 9 1 that involved a train exceeding the limit of its authorised movement, either by passing a signal or other form of limit. Signal passed at danger signal restored as train approached 33 Signal passed at danger non technical Limit of authority exceeded 13 177 1 3 4 Figure 3: Limit of authority exceedance on the NSW rail network, 9 1 Signal passed at danger non technical comprises driver misjudged, completely missed and starting against signal. Excludes occurrences classified as signal passed at danger other. There is another series of systems to authorise train movements at times when the normal systems of safeworking are not available (see section 4.) 8 Independent Transport Safety and Reliability Regulator

Signals passed at danger without authority SPADs are important precursor events to train collisions and derailments. For example, the passenger train derailment at Penrith in April 1 (Table 6) was preceded by the train passing a signal at stop without authority. A note on the signals passed at danger without authority The number of SPADs alone does not provide a complete picture of the risk posed by occurrences of this type. This is because the actual risk of collision following a SPAD depends on many factors, including whether the signal is equipped with engineering defences which automatically stop a train once it has passed a red signal, how far the train travelled into another section and whether that section was occupied. The two most important classes of SPAD in the national occurrence classification scheme in terms of accident risk are driver misjudged and completely missed. However, the information provided in an occurrence notification is usually insufficient to distinguish between the two. For this reason, data for these two categories (together with starting against signal) are combined and reported as SPAD non technical in Figures 3 to 5 1. There were 177 non-technical SPADs in 9 1 (Figure 3). Almost 7% of these occurred on the, reflecting its greater volume of rail traffic and higher signal density compared to other networks. The count on the (1) in 9 1 was markedly lower than the previous year (165) and consistent with a decreasing trend over the previous four years (Figure 4). Most of this change was associated with a decrease in CityRail train SPADs although the number of freight train SPADs on the also fell in 9 1. Forty-two of the 177 non-technical SPADs occurred on the DIRN and Hunter networks and most (75%) involved freight trains. About % of occurrences were associated with passenger trains. One of these incidents is the subject of investigation by the ATSB in September 9 a CountryLink train passed a signal at danger when departing Junee by about three carriage lengths. Figure 5 shows non-technical SPADs for passenger and freight trains. The SPAD rate for freight trains over the period of comparable data (4.4 per million train km travelled) was about 5% greater than that for passenger trains (.8), despite freight trains undertaking a greater proportion of their movements on unsignalled track. In 9 1 ITSRR continued targeted compliance inspections of freight rolling stock operators to ensure each SPAD was investigated and SPADrelated risks were adequately managed. SPAD (non-technical) SPAD (restored as train approached) 1 8 6 4 1 8 6 4 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 4: Signals passed at danger on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. 1 SPADs can rarely be attributed to a single cause such as driver error and are better understood by looking at the full range of contributing factors which may make human error more likely Information on better practice in SPAD management and associated tools for investigating SPADs are available for download from ITSRR s website <http://www.transportregulator.nsw.gov.au> Precursor rail safety occurrences Rail Industry Safety Report 9 1 9

Passenger train Freight train 6 4 6 4 per million train km 6 4 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 6 4 per million train km 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 5: Signals passed at danger (non technical) on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Solid line is quarterly occurrence rate (count per million train km travelled). Technical SPADs The other principal category of SPAD notifications to ITSRR is signal restored as train approached. These occurrences involve a signal indication changing from proceed to stop as a train approaches, with insufficient time given to the driver to stop the train. They are associated with a variety of causes including power failures, signaller error and track circuit failures. These SPADs do not pose a collision risk as the route ahead of the signal will be cleared for the train. However, they still pose a safety hazard because rapid deceleration associated with emergency braking may cause load shifts on freight trains and falls on passenger trains. There were 33 such occurrences in 9 1 (Figure 3). The count of occurrences in 9 1 was higher than the previous year and consistent with an overall increase since mid-5 (Figure 4). This rise in the number of occurrences was associated with the DIRN, specifically the Hunter and the main southern corridor between Sydney and the Victorian border. In the case of the Hunter, the observed increase may have reflected, in part, an increase in coal-related freight movements on this network. Proceed authority exceeded There were 13 notifications of trains exceeding the limit of their authorised movement under non-signalled systems of authorisation (Figure 3). While this was approximately double that of the previous year it remains within the bounds of the year to year fluctuation observed over the past four years (Figure 6). 1 8 6 4 Authority exceeded 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 6: Authority exceeded on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. 3 Independent Transport Safety and Reliability Regulator

4. Safeworking irregularity Summary Two track workers were killed in separate occurrences in 9 1. There was a high number of serious irregularities in the application of rules and procedures to manage the safety of worksites on track and protect track workers from trains. ITSRR initiated a successful prosecution for breaches of the Rail Safety Act in relation to an occurrence at Singleton in July 7, in which two signal maintenance workers were fatally injured when struck by a freight train. Safeworking systems comprise the procedures and technology governing the safe operation of trains and protection of people on or about the track. Section 4.1 of this report described a specific subset of irregularities associated with safeworking systems those where a train exceeded the limit of its authorised movement. However there are many other important elements of safeworking systems in which irregularities and threats to safety occur. In 6 7 ITSRR began a broad and detailed review of irregularities in various safeworking systems. The following chapter is based on the data collected as part of this review. The summaries in this section are not based on the OC G1 scheme which only introduced a coarse level categorisation of these types of safety occurrences in 8 9. What is procedural safeworking? On rail vehicle detection track in NSW, the authorisation for a train to proceed is provided by fixed signals. The signal displays are determined largely by the status of interlockings and by track occupancy information provided by track circuits and axle counters. As such, the integrity of the authority communication system does not rely on verbal communications or manual procedures. On non-rail vehicle detection track, or when signalled track develops a fault, manual safeworking methods are used (albeit supported by hardware and/or software in some cases) and involve communications by telephone, train radio or other verbal forms. These manual methods of safeworking are referred to as procedural safeworking. Procedural safeworking There were 34 irregularities in procedural safeworking in 9 1 (Figure 7). This was consistent with the historical pattern of behaviour for these types of occurrences since mid 6 (Figure 8). Most irregularities in 9 1 were associated with degraded working rail vehicle detection territory. These irregularities occur at the time when the normal signalled system of safeworking has been suspended, for example, due to signal fault, and an error has occurred in the alternative procedural-based system for managing train movements such as hand signalling. About 7% of degraded working irregularities in 9 1 occurred on the. The relatively high failure rate on this network reflected, in part, the large proportion of signalled track compared to other networks. When the normal signalled system becomes unavailable, only procedural methods of working trains are available, but these generally lack the inbuilt protections of the hardware-supported systems. Moreover, railway staff are required to work to these procedural methods only occasionally, and will have less experience and familiarity with the system compared to the normal signalled system of safeworking. This increases the likelihood that something will be missed or overlooked. Only four degraded working irregularities occurred on the DIRN and Hunter networks. However, three of these are the subject of compliance investigation by ITSRR. In one of these irregularities (Whittingham, March 1) the error resulted in the derailment of a freight train (Table 7). The two other occurrences both involved passenger trains and in both cases derailment or collision was only narrowly avoided. There were comparatively few irregularities for the other forms of procedural safeworking. However, the absence of trend or high counts does not necessarily translate to an acceptable level of risk. There was a serious train order occurrence at Manildra in February 1 in which an empty passenger train had a train order to travel in a section already occupied by a freight train. This occurrence is being investigated by both ITSRR and the ATSB. In such cases, a breakdown in procedure or a miscommunication of information can lead to situations where the blocking required to keep trains separated and on-track has, in effect, been compromised, increasing the potential for accidents such as derailments or collisions 3. 3 Procedural safeworking incidents that escalate to a train exceeding the limit of its authorised movement are also summarised in section 4.1 Precursor rail safety occurrences Rail Industry Safety Report 9 1 31

Degraded working Train order 5 Electric staff 5 Staff and ticket 4 5 1 15 5 Figure 7: Procedural safeworking irregularity on the NSW rail network, 9 1 Irregularities are those with potential for collision or derailment. 3 5 15 1 5 Degraded working Staff and ticket Electric staff Train order 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 8: Procedural safeworking irregularity on the NSW rail network, 6 7 to 9 1 Vertical bar is quarterly occurrence count. Irregularities are those with potential for collision or derailment. Worksite protection What is worksite protection? Worksite protection refers to a specific set of safeworking rules and procedures to manage the safety of worksites on track. In NSW there are five different methods of worksite protection prescribed in the network rules. These range from low levels of protection such as no authority required (NAR), where lookouts provide notification of approaching trains, through to local possession authority (LPA) where the protection arrangements are advertised in advance and no trains (other than work trains) are permitted to enter the worksite. The number of irregularities associated with each method of worksite protection in 9 1 and the longer term pattern of irregularities are summarised in Figure 9. These occurrences represent a situation where either the block between trains and worksite has failed or an insufficient level of protection was used 4. There was a total of 13 occurrences across the five protection methods in 9 1 which was consistent with the available historical record. The occurrences are associated with a variety of contributing factors and consequences. Two cases resulted in the death of a track worker. In several other cases, harm was narrowly avoided. A summary of four near miss occurrences subject to investigation is provided in Table 1. The nature of irregularity varied between networks. The most problematic method of protection on the DIRN, Hunter and was track occupancy authority (TOA), consistent with it being the most common method of protection used on these networks. The NAR method of protection recorded the second highest number of irregularities, in line with it being the second most common type of protection used on these networks. For the, just under half of all irregularities were associated with NAR. However, the four other forms of protection had a significant proportion of irregularities recorded. These results are consistent with the use of a wider variety of protection methods on the compared to the other main networks. 4 The concept of a block applies to the forms of protection such as LPA, TOA, TWA and CSB. It does not however apply to NAR, where track workers are required to move out of the path of an oncoming train. For NAR the definition of a serious irregularity is the report of insufficient level of protection being used or the warning method was compromised. 3 Independent Transport Safety and Reliability Regulator

Irregularity by protection type, 9 1 Irregularity by network, 6 7 to 9 1 Track occupancy authority 44 35 3 No authority required Local possession authority Controlled signal blocking Track work authority 8 15 17 39 5 15 1 5 1 3 4 5 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 9: Worksite protection irregularity on the NSW rail network NAR is excluded from time series because this form of protection does not block a train from traversing the worksite and so differs from the other types of protection. Irregularities associated with each of the five methods of worksite protection have been further categorised by ITSRR into specific failure types. These are summarised in Figures 3 to 34. Almost two-thirds of failures for NAR (Figure 3) 5 involved failing to provide a lookout, or positioning the lookout in an unsuitable location. Lookouts are a critical element of NAR protection because they are responsible for watching for approaching trains and warning workers to move away from the track in sufficient time. Another 4% of failures involved situations where a higher form of protection was required, for example, when earthmoving vehicles were being used on or near the track. Controlled signal blocking (CSB) involves the use of signals to exclude traffic from a worksite. It was the method of protection applied when a track worker was struck by a train at Kogarah in April 1. There were no dominant failure types for CSB (Figure 31) the apparent differences between failure types are unreliable due to the small number of failures across a large number of categories. Track work authority (TWA) was associated with the lowest number of irregularities of all methods of protection which reflects, in part, its limited use. It is used primarily on the and authorises occupation of track by workers between rail traffic movements. There were two dominant failure types for the TWA method of protection (Figure 3) incorrect location of detonators and hand signallers (33%) and hand signallers failing to maintain their post (8%). Like NAR, hand signallers have a crucial role in the protection of workers under TWA in managing a train s approach towards, and passage through, a worksite in a safe manner. Table 1: Worksite protection near miss occurrences on the NSW rail network, 9 1 Selected occurrences which are the subject of investigation by ITSRR or OTSI. Date Location Protection type Description 4 August 9 Gunnedah () NAR Crew of freight train on approach to Gunnedah Station observed track work involving use of an excavator. Freight crew sounded whistle and slowed before workers moved off track. 9 September 9 Kilbride (Hunter) NAR Crew of freight train on approach to Kilbride observed a front end loader loaded with ballast on track. Freight crew made emergency brake application and sounded whistle coming to stand just short of front end loader. 6 March 1 Westmead () 1 April 1 Strathfield () NAR NAR Crew of passenger train on approach to Westmead Station observed workers on track. Train crew sounded horn and workers jumped out of the path of the train. CityRail interurban service nearly struck track maintenance staff working on a set of points. 5 These data may not represent the complete risk associated with the different methods of worksite protection. For example, misjudgements by lookouts or other irregularities in NAR may never be formally reported unless an accident or near miss occurs. Precursor rail safety occurrences Rail Industry Safety Report 9 1 33

Figures 33 and 34 summarise the various types of failures for the two possession types of track work, namely, TOA and LPA respectively. TOA and LPA represent the highest forms of worksite protection and are distinct from others in that they are intended to completely exclude trains from the protection area. However, these higher levels of protection are not without risk the fatality at Newbridge in May 1 (Table 4) was associated with the TOA method of protection. While there was no dominant form of failure for LPA irregularities, the primary failure type for TOA was track vehicles exceeding the limit of their authorised movement (7%). No lookout used or lookout at unsuitable location NAR used when a higher level of protection required Lookout fails to detect or warn of approaching train Warning by lookout fails to be heard or acted on by workers Work conducted outside protected area Track not effectively cleared after warning acted on PO does not provide safe places Lookout does not remain at post Lookout not in safe place 1 3 4 5 6 7 Proportion of total failures Figure 3: No authority required failure on the NSW rail network, 6 7 to 9 1 NAR = No authority required. PO = Protection officer. Work conducted without CSB Fail to establish block Track work conducted outside protected area Work performed using CSB when higher form of protection required PO does not protect all means of access to worksite PO sets up on wrong track (unprotected) Failure to maintain block NCO does not identify train between blocking signal and worksite, PO does not check Signaller sets wrong signal to stop or selected signal does not protect worksite Train SPADs signal 1 3 4 5 6 7 Proportion of total failures Figure 31: Controlled signal blocking failure on the NSW rail network, 6 7 to 9 1 CSB = controlled signal blocking. NCO = network control officer. PO = protection officer. 34 Independent Transport Safety and Reliability Regulator

Detonator and handsignallers not located correctly Handsignaller fails to maintain post Inner handsignaller authorises train to proceed when unsafe to do so Authority to work to TWA not obtained from NCO NCO fails to set and hold signal(s) at stop Train too fast through worksite Track work conducted outside protected area Train travelling too fast to stop at inner handsignaller PO does not identify and protect all points of entry or exit Violation - work performed using TWA when higher form of protection required 1 3 4 5 6 7 Proportion of total failures Figure 3: Track work authority failure on the NSW rail network, 6 7 to 9 1 TWA = track work authority. NCO = network control officer. PO = protection officer. Track vehicle exceeds limits of authority Track work conducted outside protected area NCO fails to establish block Work performed with no authority NCO fails to maintain block Protection applied to wrong area of track PO authorises end of possession with track still obstructed Protection not put in place/removed Train enters protected area Equipment left foul of track (other than detonators/red flags) PO absent from duty 1 3 4 5 6 7 Proportion of total failures Figure 33: Track occupancy authority failure on the NSW rail network, 6 7 to 9 1 NCO = network control officer. PO = protection officer. Protection applied to wrong area of track Track vehicle exceeds limits of authority Track work conducted outside protected area Train enters protected area Protection not put in place/removed NCO fails to establish block NCO fails to maintain block PO authorises end of possession with track not fully restored Equipment left foul of track (other than detonators/red flags) 1 3 4 5 6 7 Proportion of total failures Figure 34: Local possession authority failure on the NSW rail network, 6 7 to 9 1 NCO = network control officer. PO = protection officer. Precursor rail safety occurrences Rail Industry Safety Report 9 1 35

Effective management of worksite safety risks by rail transport operators was a corporate priority for ITSRR in 7 8 and 8 9. In previous years, ITSRR undertook a range of compliance activities including industry briefings, random inspections of track work and targeted inspections of specific issues, for example, rail safety worker training and competency. However, analysis of 9 1 occurrence statistics and findings from ITSRR s compliance activities show that track worker safety continues to be an area requiring a high level of regulatory activity. The death of two rail safety workers in 9 1 together with a continuing high number of near miss incidents highlights the need for improvement in rail transport operators management of safety risks associated with track work. In late 9, ITSRR further expanded its compliance activities in relation to worksite protection safety. This strategy included a major compliance campaign on track worker safety, resulting in 149 rail worksite inspections, 111 incident assessments, several investigations and notices to produce information. The safety of track workers will continue to be a corporate priority in 1 11. ITSRR is planning further compliance activity and engagement with NSW rail infrastructure managers and registered training organisations. 4.3 Signal and track irregularity Summary The number of broken rails on the has decreased over time. The number of track misalignments on sections of the DIRN has increased over time due, in part, to the development of mud holes on sections of the network. Signal irregularity The OC G1 category signal and other proceed authority system irregularity comprises irregularities in the systems, components and equipment used to authorise train movements in a safe manner. There are five principle subcategories reflecting each of the main systems used nationally for issuing an authority to proceed. The categories are defined in a way that limits them to serious irregularities, such as wrong side failures. Irregularities in the systems used to authorise movements of trains are shown in Figure 35. This figure also includes serious defects in level crossing protection equipment which are associated with signalling systems. There were six signal irregularities notified on the NSW rail network in 9 1, which was consistent with that observed over the previous four years (Figure 36). One of the occurrences on the DIRN resulted in a near miss a signal indicated proceed for a CountryLink train to change tracks, however, the CountryLink driver observed the last wagon of a freight train to be obstructing the path of his train. The passenger service s emergency brakes were applied and it stopped just short of the freight train. This occurrence is being investigated by the ATSB. There was one irregularity in the token-based system of authority in 9 1. This occurrence at Turrawan on the in April 1 resulted in two electric staffs being removed for the same track section a situation representing the potential for two trains to enter the one section of track. Various types of level crossing irregularities are defined within OC G1. The most significant of these in terms of risk under the direct control of railways are occurrences involving wrong side failures, for example, when the level crossing lights and bells do not operate as intended to warn road users of an approaching train. There were six occurrences in 9 1 that comprised actual or potential serious failures in the level crossing equipment (Figure 35). Four of the six occurrences were on the DIRN. The overall total was consistent with that observed over the previous four years (Figure 36). Due to the serious nature of the failures summarised in Figure 35, every occurrence was reviewed by ITSRR s specialist support team to ensure it was investigated by the rail infrastructure manager and associated risks were appropriately managed. As a result of these reviews, ITSRR commenced a compliance investigation for the occurrence at Turrawan. Level crossing failure (wrong side failure) Authority system irregularity (signal-based) 6 6 Authority system irregularity (token-based) Authority system irregularity (communications-based) 1 1 3 4 5 6 7 8 Figure 35: Signal and other proceed authority system irregularity on the NSW rail network, 9 1 36 Independent Transport Safety and Reliability Regulator

Signal failure Level crossing failure 8 6 4 8 6 4 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 36: Signal and level crossing irregularity on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Track irregularity Track condition is a key indicator of the safety of the rail network. Notified track defects provide a valuable measure of the way the asset is being maintained in addition to their role as precursors for more serious incidents. At least one freight train derailment in 9 1 was the result of a track irregularity (see section 3.1). A note on the definition of track irregularities The definition of broken rail and misaligned track in the OC G1 scheme is different to the engineering standards used by industry for track inspection, maintenance and reporting. Broken rail (OC G1 category 11.1) includes, but is not restricted to, complete breaks through the entire cross section of the rail. It may also include fractures and broken joints. Misaligned track (OC G1 category 11.) encompasses misalignment in both the horizontal and vertical planes as well as failures in the formation the ground upon which the track is laid. Misaligned track therefore comprises a diverse range of incident types that vary markedly in terms of their causes, consequences and safety risks posed. There were 119 broken rails notified in 9 1 (Figure 37). About 8% of these occurred during the cooler months of the year, specifically April to September. This seasonal pattern was also a feature of the longer term record (Figure 38) and reflects a seasonal increase in stresses associated with rail contraction during very cold weather. At such times, rail is more likely to break under load from rolling stock. The distribution of broken rails across the NSW rail network over the past five years is shown in Figure 39. About % of all broken rails notified in 9 1 occurred on the. Previous reports have noted a gradual decrease in the number of broken rails over time for this network. A decreasing trend was still evident over the five-year period to the end of June 1 although the rate of decrease has levelled off in recent years (Figure 38). Almost two-thirds of broken rails in 9 1 occurred on the DIRN and Hunter, which was consistent with that observed over the previous four years (Figure 38). The greatest concentration of breaks was on the Hunter between Maitland and Muswellbrook, and the main southern corridor of the DIRN between Macarthur and Junee (Figure 39). The bulk of the remaining broken rails occurred on the. The rate of broken rail (per track length) was comparable between networks (about 1 per 1, track km) with the exception of the Hunter region which was far higher (about 4 per 1, track km). ITSRR s regulatory activity in relation to track issues such as broken rail included monitoring rail infrastructure managers adherence to applicable engineering standards and also verifying that the standards themselves provided the necessary margin of safety. In specific cases, ITSRR initiated regulatory action to obtain information and ensure systems for identifying and managing track-related risks were adequate. Precursor rail safety occurrences Rail Industry Safety Report 9 1 37

Track obstruction 76 Track misalignment Broken rail 119 315 4 6 8 1 Figure 37: Track-related irregularity (running line) on the NSW rail network, 9 1 Spread track not shown (usually associated with a higher event such as derailment). Excludes OC G1 track-related categories points irregularity and other. The occurrence category misaligned track encompasses a particularly diverse range of defects (see note on the definition of track irregularities). It includes defects in horizontal geometry often described as buckles and kicks that occur more often in the warmer months when higher temperatures expand rail and increase its chance of buckling. The category also includes vertical irregularities, some of which are associated with defects in the ballast and underlying formation. While the seasonality of vertical irregularities is less pronounced, some such as bog holes and mud holes are more frequent in the cooler months when rainfall and drainage issues become relevant. About 4% of all misalignments in 9 1 were associated with the. The record of misalignments on this network has been truncated due to a significant change in reporting of occurrences in late 6. However, the number of occurrences in 9 1 was consistent with historical record since the change in reporting. Approximately 5% of notified misalignments in 9 1 were associated with the DIRN. Consistent with historical data (Figure 4) the greatest concentration of occurrences was on the main southern corridor between Macarthur and the Victorian border and the Hunter. There was an increase in misalignments on the DIRN over the five-year period (Figure 38) associated primarily with the Hunter region and main southern corridor between Sydney and the Victorian border. In the Hunter, the increase was gradual and may have reflected, in part, increased coal freight movements on this network. In the case of the main southern corridor, the increase occurred towards the end of a major program of concrete resleepering, completed in 8 9 by ARTC. The specific method of resleepering employed on this section of the DIRN led to disturbance of the track base, drainage problems and the formation of mud holes. An increase in train partings also occurred at the same time on this section of the DIRN (see section 4.4). The number of misalignments on the (3) was lower than the previous year but consistent with that observed over the past four years. Historical data shows the occurrences were concentrated on two sections of the network, from Wallerwang to Orange and from Werris Creek to Narrabri (Figure 4). In response to the increased number of misalignments on the DIRN and associated risks to trains, ITSRR undertook a review of ARTC s risk-based approach for managing mud holes and progress towards developing effective procedures, standards and rectification programs. Follow up actions are ongoing. Broken rail Misaligned track 1 8 4 1 8 4 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q Figure 38: Track irregularity (running line) on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Data excluded for misaligned track on the prior to 7 due to changes in RailCorp s reporting process. 38 Independent Transport Safety and Reliability Regulator

The national occurrence category track obstruction contains a broad range of incidents that vary markedly in terms of the typical safety risk they pose. Under the national definition, occurrences are not restricted to obstructions of track but include objects that interfere with the broader rail infrastructure. There was a total of 76 occurrences classified as track obstruction in 9 1. The vast majority of occurrences did not pose a collision risk. Almost half of the occurrences involved animals in the rail corridor. The remaining occurrences included trees obstructing signals and vandalism-related occurrences, for example, shopping trolleys placed on tracks. A small proportion of occurrences were associated with rail operations, for example, material that had fallen from fixed infrastructure. Narrabri Armidale Coffs Harbour Broken Hill Cobar Werris Creek Muswellbrook Dubbo Broadmeadow Orange Lithgow SYDNEY Griffith Goulburn Legend Junee Yass Junction CANBERRA Number of broken rails + 11 19 6 1 3 5 1 N Kilometres 65 13 6 Albury NSW rail network Rail Centreline Copyright RailCorp 7 Figure 39: Broken rail (running line) across the NSW rail network, 5 6 to 9 1 Mapped location is based on nearest reported location as notified by rail transport operators. Precursor rail safety occurrences Rail Industry Safety Report 9 1 39

Narrabri Armidale Coffs Harbour Cobar Werris Creek Broken Hill Muswellbrook Dubbo Orange Lithgow Broadmeadow SYDNEY Griffith Goulburn Legend Junee Yass Junction CANBERRA Number of misalignments + 11 19 6 1 3 5 1 N Kilometres 65 13 6 Albury NSW rail network Rail Centreline Copyright RailCorp 7 Figure 4: Misaligned track (running line) across the NSW rail network, 5 6 to 9 1 Misalignment comprises vertical and horizontal misalignments as well as formation failures. Mapped location is based on nearest reported location as notified by rail transport operators. 4.4 Rolling stock irregularity Summary The rate of braking system irregularity on passenger trains (per train km travelled) has decreased over time. The rate of braking system irregularity on freight trains (per train km travelled) has increased over time. The number of freight train partings has increased over time due, in part, to track-related irregularities on a segment of the DIRN. Rolling stock irregularities are important precursors to serious safety incidents. The national occurrence categories braking system, wheel/axle and defective bearing are the most significant in this regard. For example, defective brakes may lead to runaway trains and defective bearings may escalate to sheared axles and derailment. Rolling stock irregularities for 9 1 are summarised in Figure 41 by train type. Passenger door Braking system 115 115 Train parting Defective bearing Wheel/axle 9 44 14 Passenger Freight 3 6 9 1 15 Figure 41: Rolling stock irregularity on the NSW rail network, 9 1 Braking system includes occurrences with a top event of scaled or skidded wheels (which are classified as rolling stock other under OC G1). wise excludes OC G1 rolling stock other. 4 Independent Transport Safety and Reliability Regulator

Passenger train A single wheel irregularity was notified in 9 1 which was consistent with that observed over the previous four years. Six bearing-related irregularities were notified in 9 1. These occurrences were spread across three sectors urban, interurban and heritage operators. The overall count was also consistent with historical data (Figure 4). About 95% of brake-related irregularities on passenger trains were associated with CityRail passenger trains, which provide the vast majority of the state s passenger train km. While the length of historical data was truncated due to changes in reporting, there was a significant decrease in the rate of notified braking system irregularities from 17.8 occurrences per million train km travelled in 8 to 1.4 in 9 1. Faulty passenger train doors are a lower significance precursor in terms of safety risk. However, they are a major contributor to fleet-related delays on the. Past reports have described a range of measures undertaken by RailCorp to address the rate of door failures. The rate of door faults (per million train km travelled) has decreased significantly over the period for which comparable data are available. 4 3 1 Braking system Passenger Freight 4 per million train km 4 3 1 Defective bearing Passenger Freight 8 6 4 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 4 3 1 Train parting 1 8 6 4 per million train km 5 4 3 1 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q per million train km Passenger Freight Faulty passenger door 6 4 per million train km Figure 4: Rolling stock irregularity on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Solid line is quarterly occurrence rate (count per million train km travelled). Braking system includes occurrences with a top event of wheel scale or wheel skid (classified as rolling stock other under OC G1). Data for passenger brakes and passenger doors prior to the third quarter of 6 was excluded due to changes in RailCorp s reporting process. Passenger train rates for defective bearing and train parting not shown due to low counts. Precursor rail safety occurrences Rail Industry Safety Report 9 1 41

Freight train Bearing-related occurrences are important precursors to axle-related failures and subsequent derailments. Two running freight train derailments in 9 1 were the result of bearing failures. There were 35 bearing-related notifications in 9 1. Most of these incidents are detected by trackside monitoring equipment. This equipment plays a critical role in the early detection of bearing faults and prevention of derailments although false alarms are sometimes notified. The rate of bearing-related occurrences in 9 1 (.1 per million train km travelled) was comparable to the historical record excluding a period in 7 8 affected by a suspected rise in false alarms. A total of 49 braking system irregularities were notified in 9 1 which was higher than the previous year and consistent with an increase in these types of occurrences over the past four years. The rise was evident in both raw occurrence counts and rate (per million train km travelled). While the majority of irregularities were associated with rolling stock defects, for example, sticking brakes, they did include crew and vandalism-related incidents such as interference with handbrakes. Train parting occurrences comprise the unintentional separation of rolling stock. These are associated with a range of causes including coupler faults, crew error and vandalism. There were 136 train parting occurrences notified in 9 1. They are typically lower risk occurrences with regard to the train involved because brakes automatically apply in such situations, preventing runaways. However, both the number (136) and rate (8. per train km travelled) of partings in 9 1 was higher than the previous year (15 and 5.9 respectively). The rate of increase was greatest in 9. 1 1 Train parting 8 6 4 Jul 5 Oct 5 Jan 6 Apr 6 Jul 6 Oct 6 Jan 7 Apr 7 Jul 7 Oct 7 Jan 8 Apr 8 Jul 8 Oct 8 Jan 9 Apr 9 Jul 9 Oct 9 Jan 1 Apr 1 Track misalignment 16 1 8 4 Jul 5 Oct 5 Jan 6 Apr 6 Jul 6 Oct 6 Jan 7 Apr 7 Jul 7 Oct 7 Jan 8 Apr 8 Jul 8 Oct 8 Jan 9 Apr 9 Jul 9 Oct 9 Jan 1 Apr 1 Figure 43: Train parting and track misalignment on the main southern corridor of the DIRN, 5 6 to 9 1 Shows monthly total occurrences and smoothed trend. 4 Independent Transport Safety and Reliability Regulator

The increase in partings was associated specifically with the Hunter region and the main southern corridor of the DIRN. In the case of the Hunter, the increase was gradual and occured in conjunction with increased coal freight movements. In the case of the main south, partings rose markedly in late 9 and at a time of decreasing freight traffic, suggesting causes other than changes in traffic volume. A separate analysis of these occurrences shows a strong correlation between the location of partings and the location of a specific type of track irregularity referred to as a mud hole (see section 4.3). The pattern of parting occurrences through time also corresponded with that for misalignments (Figure 43). In 9 1 ITSRR undertook targeted compliance inspections of specific rolling stock operators to ensure safety risks associated with rolling stock irregularities were being appropriately managed and rolling stock complied with relevant engineering standards. A corporate priority in 9 1 was also a program of inspections to ensure the engineering standards themselves provided the necessary safety assurance for the design, commissioning and operation of rolling stock. In the case of train partings, where track condition was a contributing factor, the management of track-related safety risks was reviewed by ITSRR (see section 4.3). 4.5 Load irregularity Summary A total of 17 load irregularities were notified in 9 1. The rate of load irregularity in 9 1 (per million train km travelled) was markedly lower than previous years. A previously reported increase in door open occurrences over time appears to have steadied. Load irregularities are specific to freight trains and are precursors to a range of rail accidents. A load that shifts beyond the train envelope poses a collision hazard to other trains and railway infrastructure. The loss of materials such as coal from trains can obstruct tracks and derail trains. A total of 17 load irregularities were notified for the period 9 1 (Figure 44). About 6% of load irregularities occurred on the DIRN, which carries most freight traffic. While there was no evidence of trend in load irregularities over the five-year period of record, the rate of load irregularity in 9 1 (1. occurrences per million train km travelled) was significantly lower than that observed over the past four years (1.3). Loose load fastening 6 Door open 51 Load shift Out of gauge Uneven distribution of load 3 13 43 4 6 8 1 Figure 44: Load irregularity on the NSW rail network, 9 1 Excludes OC G1 category load irregularity other. Precursor rail safety occurrences Rail Industry Safety Report 9 1 43

An increase in door open occurrences over time had been noted in previous reports. This issue was the subject of past regulatory activity by ITSRR due to the collision risk posed by these occurrences. About one-third of all running line collisions between trains involved open doors swinging into the path of trains on adjacent lines. There was no evidence of a trend in the most recent five years of data for this type of irregularity (Figure 45). This suggests the steps taken by freight operators to reduce such occurrences, for example, wagon and door modifications, have had a positive effect. ITSRR continues to monitor this issue to ensure the problem does not escalate again. Loads that shift in transit may protrude from wagons and pose a collision hazard for passing trains, infrastructure and passengers on platforms. Load shifts may be precursor events to out of gauge loads, however, the information provided in occurrence notifications is often insufficient for determining whether or not a load has shifted to a point where it is out of gauge. There were 56 occurrences classified as load shift or out of gauge in 9 1. The combined total of these two occurrence types in 9 1 was consistent with that observed over the previous four years (Figure 45). One occurrence at Woy Woy on the north of Sydney is currently under investigation by OTSI in April 1 a metal gate fell from a container wagon and struck a glass partition on the platform, narrowly missing nearby passengers. The category uneven distribution of load covers two types of occurrence load distributed unevenly within an individual freight wagon which can cause a wagon to lean, and an uneven distribution of load across wagons within a train consist. There were just three of these occurrences in 9 1. While infrequent, they do pose a safety risk. The presence of empty coal wagons amongst a consist of loaded wagons at Coggan Creek in March 1 was suspected to have been the cause of a derailment. This incident was the subject of a rail industry safety notice issued by ITSRR in 9 1. The most frequent, but lowest risk type of load irregularity in 9 1 was loose load fastening. This category comprises incidents such as dragging chains, loose ropes and flapping tarpaulins. There were 6 occurrences reported in 9 1, which was consistent with that observed over the past four years (Figure 45). Door open Load shift/out of gauge 4 3 1 8 6 4 per million train km 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 4 3 1 8 6 4 per million train km Loose load fastening 4 3 1 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 5Q3 5Q4 6Q1 6Q 6Q3 6Q4 7Q1 7Q 7Q3 7Q4 8Q1 8Q 8Q3 8Q4 9Q1 9Q 9Q3 9Q4 1Q1 1Q 8 6 4 per million train km Figure 45: Load irregularity on the NSW rail network, 5 6 to 9 1 Vertical bar is quarterly occurrence count. Solid line is quarterly occurrence rate (count per million train km travelled). 44 Independent Transport Safety and Reliability Regulator

5. Drug and alcohol testing Summary Almost 14,3 drug and 7, alcohol tests were conducted by rail transport operators in 9 1. The amount of testing in 9 1 was the highest on record. The rate of drug and alcohol detection for rail infrastructure managers/maintainers in 9 1 was markedly lower than previous years while the rate of detection for other sectors remained consistent with previous years. A total of 1 rail safety workers were successfully prosecuted for offences under the drug and alcohol provisions of the NSW rail safety legislation in 9 1. The requirement for rail transport operators to conduct drug and alcohol testing of rail safety workers was introduced in the Rail Safety Act and further developed under the Rail Safety (Drug and Alcohol Testing) Regulation 3 and the Rail Safety (Drug & Alcohol Testing) Regulation 8. The Rail Safety Act 8 requires all rail transport operators to have formal drug and alcohol programs in place that comply with the Regulations. Since 6 it has been mandatory for each operator s random testing program to test 5% or more of its rail safety workers every year. In addition, testing may also include targeted and post-incident testing. All operators are required to notify ITSRR of positive test results as well as any instance where an employee refused to undergo testing. Commercial rail transport operators are also required to submit quarterly summaries of testing activity to ITSRR. These extra reporting requirements do not apply to heritage operators. Program activity At the time of writing, over 99% of expected quarterly summaries for 9 1 had been submitted to ITSRR. Quarterly summaries show that almost 14,3 drug and 7, alcohol tests of rail safety workers were conducted during 9 1. The number of drug tests in 9 1 was about 3% higher than that for 8 9 (1,915) while the number of alcohol tests in 9 1 was slightly higher than the previous year (7,878). A breakdown of random testing activity by industry sector (Figure 46) shows the average rate of testing in 9 1 for each sector, as a whole, exceeded the minimum required under the Regulation (a minimum of 5% of an operator s rail safety workers tested per year). This also shows that freight operators and infrastructure managers/maintainers were largely responsible for the increased rate of drug testing observed in 9 1. ITSRR also undertook drug and alcohol testing of rail safety workers on two separate occasions in 9 1. This resulted in 45 random drug and 95 random alcohol tests. While the testing focused on rail transport operators not previously subjected to testing by ITSRR, it also included a range of small and large operators. Program results Table 11 presents summary statistics for 9 1 based on all forms of testing. As in previous years, the overall detection rate for drugs (.95%) was higher than that for alcohol (.9%). The overall detection rate for both in 9 1 was significantly lower than the corresponding figures for the previous year (1.3% and.13% respectively). Drug testing covers a range of drug classes including amphetamines (for example, speed), opiates (for example, heroin) and cannabiniods (derived from cannabis). As in previous years, cannabis was the most common drug associated with positive drug tests. The overall detection rate is sensitive to the influence of larger operators (who conduct the majority of tests) and any operators reporting anomalous results. For this reason, Table 11 also includes a summary of detection rates on an operator basis. It shows about 58% of rail transport operators testing for drugs and 66% of rail transport operators testing for alcohol did not return a positive result in 9 1. Another limitation of the overall detection rate is that it is not necessarily representative of the general railway safety worker population because it includes results of non-random testing. Non-random testing may yield higher rates of detection than random testing because it includes for-cause testing testing conducted on the basis that there is reason to believe an individual could be affected by drugs or alcohol. A breakdown of random test results by industry sector over four years is shown in Figure 46. Significant differences are apparent between sectors. As in previous years, the detection rates for the freight sector in 9 1 were low compared to other sectors. Historically, both the infrastructure and rolling stock maintainers sectors recorded comparatively high detection rates. However, there was a significant decrease in the rate of both drug and alcohol detection for infrastructure maintainers in 9 1 (.66% and.7%) compared to previous years (1.99% and.5% respectively for 8 9). Drug and alcohol testing Rail Industry Safety Report 9 1 45

During 9 1, ITSRR completed 1 successful prosecutions 6 of rail safety workers for offences under the Rail Safety (Drug and Alcohol Testing) Regulation 8. Five matters involved rail safety workers providing samples of synthetic urine, four involved instances of exceeding the prescribed concentration of alcohol, two matters related to the failure to supply a sample and one involved a rail safety worker being under the influence of a drug. These offences occurred on separate occasions at various locations across NSW. Table 11: NSW rail industry drug and alcohol testing results, 9 1 Statistics based on rail transport operators that were accredited and required to submit quarterly returns for at least part of 9 1. Excludes refusal to be tested (n=15). Includes all testing reasons, for example, for-cause, random, post-incident. Excludes the heritage sector which has different testing and reporting requirements. Description Drug Alcohol Number of rail transport operators testing 38 41 Approximate random component (% of all tests) 94.1 98.5 Overall detection rate (% of all tests) 1.95.9 Median rail transport operator detection rate (% of all tests).. Number of rail transport operators reporting no positive results 7 Number of rail transport operators reporting exactly one positive result 3 7 Number of rail transport operators reporting more than one positive result 13 7 1 Total positive tests (all rail transport operators) divided by total tests (all rail transport operators) multiplied by 1 (excluding refusals) Rail transport operator detection rate is rail transport operator s total positive tests divided by rail transport operator s total tests multiplied by 1. Median is the middle ranked value of all rail transport operator detection rates. Drug testing rate Alcohol testing rate Rate (no. tests/no. workers) 1.4 1. 1..8.6.4. 9 1 8 9 7 8 6 7 Rate (no. tests/no. workers) 8 7 6 5 4 3 1 9 1 8 9 7 8 6 7. Freight operator Infrastructure manager/maintainer Passenger operator Rolling stock maintainer Freight operator Infrastructure manager/maintainer Passenger operator Rolling stock maintainer Drug detection rate Alcohol detection rate Rate (%) (no. detections/no. tests) 3..5. 1.5 1..5 9 1 8 9 7 8 6 7.5..15.1.5 Rate (%) (no. detections/no. tests).3 9 1 8 9 7 8 6 7. Freight operator Infrastructure manager/maintainer Passenger operator Rolling stock maintainer. Freight operator Infrastructure manager/maintainer Passenger operator Rolling stock maintainer Figure 46: NSW rail industry random drug and alcohol testing activity by sector Dotted line is industry average over four years 6 7 to 9 1. Excludes refusals to be tested. Excludes the heritage sector which has different testing and reporting requirements. All rates are averages for the sector as a whole, whereby an individual rail transport operator s contribution to a sector rate is proportional to the operator s size. 6 ITSRR initiates prosecution of offences in accordance with its Prosecution Policy which is available on ITSRR s website <http://www.transportregulator.nsw.gov.au> 46 Independent Transport Safety and Reliability Regulator

6. Glossary of rail-related terms Source of definitions Most definitions have been sourced from OC G1 and the Glossary for the National Codes of Practice (Australasian Railway Association <http://www.ara.net.au>). Level crossing definitions are from NSW Staysafe Committee 7. Some descriptions may differ from definitions contained in the legislation for compliance purposes refer to section 4 (Interpretation) of the Rail Safety Act 8. Accreditation requirements are outlined in the Rail Safety Act 8. NSW rail transport operators must be accredited by ITSRR or exempt from the requirement to be accredited under the Act. The granting of accreditation indicates that a rail transport operator has demonstrated it has the competence and capacity to manage the risks to safety associated with the railway operations for which it is accredited. Ballast refers to material, usually stone, that surrounds the sleepers to hold them properly in place. Balloon loop is track forming a loop usually at the end of a railway line where empty wagons are loaded from overhead bins, or full wagons are emptied through hopper doors in to under-track grates, all while the train is moving at low speed. The loop enables the train to effectively do a U turn. Commercial operations are railway services operated for reward, but generally excluding the tourist and heritage sector. Freight trains are designed and used for carrying goods such as coal and minerals, grain, fuel, livestock and containers. Infrastructure generally includes the track and its components, for example, rails, sleepers, bridges, ballast, and signalling equipment. Generally the term does not include stations or terminals. Interlocking is an arrangement of signal equipment that prevents conflicting movements of trains through junctions or crossings. It is designed so that it is impossible to give clear signals to trains unless the route to be used is proved to be safe. Level crossing is any crossing of a railway at grade, providing for both vehicular traffic and other road users including pedestrians. The control of railway crossings is classified as either active or passive according to the following criteria: active control control for the movement of vehicular or pedestrian traffic across a railway crossing by devices such as flashing signals, gates or barriers, or a combination of these, where the device is activated prior to and during the passage of a train through the crossing passive control control for the movement of vehicular or pedestrian traffic across a railway crossing by signs and devices, none of which are activated during the approach or passage of a train and which rely on the road user, including pedestrians, detecting the approach or presence of a train by direct observation. In addition to active and passive controlled crossings there are also occupational or accommodation crossings between private property and public roads; maintenance crossings; and illegal crossings. Light locomotive(s) means one or more locomotives coupled together without any vehicles attached. Near miss is any occurrence where the driver of a moving train takes emergency action, or would have if there was sufficient time, to avoid impact with a person, vehicle or other obstruction and no collision occurred. Emergency action includes continuous audible warning and/or brake application. Network refers to the connected track and infrastructure, generally owned, managed and maintained by a single organisation, for example RailCorp. Network rules are rules issued to mandate the requirements for safe operation on a rail network. Passenger journeys in urban areas measures the number of point to point journeys for each passenger, irrespective of the number of vehicles or mode used for the trip. For non-urban areas, it measures the number of point to point journeys for each passenger, but each change of vehicle along the route is a separate journey. Passenger trains are trains designed and used for carrying passengers. 7 NSW Staysafe Committee, Report on updating progress on railway level crossing safety, rep. no. 54/9, Staysafe Committee, Parliament of NSW, June 9 Glossary of rail-related terms Rail Industry Safety Report 9 1 47

Rail infrastructure manager is the person who has effective management and control of rail infrastructure of a railway, whether or not the person owns the rail infrastructure or has a statutory or contractual right to use the rail infrastructure or to control, or provide, access to it. Rail safety worker is a person who has carried out, is carrying out or is about to carry out rail safety work. Classes of rail safety work are defined under section 7 of the Rail Safety Act 8. Rail transport operator is a rail infrastructure manager, a rolling stock operator, or both. Rail vehicle detection track is a portion of track formed into an electric circuit where current is carried through the rails and used to detect the presence of trains. Track circuits are used in the operation and control of points and signalling equipment. Risk is exposure to the chance of harm. In a safety context, risk is expressed in terms of likelihood of an adverse event and the consequences of that event. Rolling stock means any vehicle that operates on or uses railway track. Rolling stock operator is a person who has effective management and control of the operation or movement of rolling stock on rail infrastructure for a particular railway, but does not include a person merely because the person drives the rolling stock or controls the network or the network signals. Staff is a metal rod which represents the authority for rail traffic to proceed into a section. Staff and ticket is a system of safeworking involving issue of a written authority to proceed into a section after driver has seen the staff for the section. Terminals are places where freight is loaded onto or unloaded from trains. A passenger terminal is a place where passenger trains commence or terminate for passengers to board or alight. Track machines are specialised pieces of rolling stock used on the rails to maintain infrastructure. Train km refers to the total km travelled in NSW by a rolling stock operator s trains. Train order is an instruction, on a prescribed form, issued by a train controller to direct the movement of traffic. Wheel flat is the loss of roundness of the tread of a wheel caused by wheelslip or wheel slide. Wheel scale is the build up of metallic material on a wheel tread s surface. Wrong side failure refers to a failure in the signalling system which results in the signal displaying a less restrictive aspect than required, for example, showing a proceed indication when the correct indication should be stop. Yard is track other than running line. Running line is railway track used primarily for the through movement of trains. Safeworking system is an integrated system of operating procedures and technology for the safe operation of trains and the protection of people and property on or in the vicinity of the railway. Shunt is to move trains or vehicles on lines for purposes other than through movement. Siding is a portion of railway track, connected by points to a running line or another siding, on which rolling stock can be placed clear of the running line. 48 Independent Transport Safety and Reliability Regulator

7. List of notifiable occurrences Under the Rail Safety Act 8 (section 63), rail transport operators are required to report to ITSRR, or another authority specified by ITSRR, all notifiable occurrences that happen on, or in relation to, the operator s railway premises or railway operations. Under the Rail Safety (General) Regulation 8 (clause 7) notifiable occurrences are defined as follows: Category A 8 (a) an accident or incident that has caused death, serious injury or significant property damage, (b) a running line derailment, (c) a running line collision between rolling stock, (d) a collision at a road or pedestrian level crossing between rolling stock and either a motor vehicle or a person, (e) a fire or explosion on or in rail infrastructure or rolling stock that affects the safety of railway operations or that endangers one or more people, (f) a suspected terrorist attack, (g) any accident or incident involving a significant failure of a safety management system that could have caused death, serious injury or significant property damage, (h) the theft of or from rolling stock or railway premises of a rail transport operator of security sensitive dangerous goods (within the meaning of the Australian Dangerous Goods Code prepared by the National Transport Commission as in force on the commencement of this regulation) or the tampering with any such goods on rolling stock or railway premises of a rail transport operator, (i) any other accident or incident that is likely to generate intense public interest or concern. Category B (a) a derailment, other than a running line derailment, (b) a collision involving rolling stock, (c) any accident or incident at a road or pedestrian level crossing arising from a failure of rail infrastructure or that caused a risk to safety or damage to a person or property, (d) the passing of a stop signal, or a signal with no indication, by rolling stock without authority, (e) any accident or incident where rolling stock exceeds the limits of authorised movement given in a proceed authority, (f) any failure of a signalling or communications system that endangers, or that has the potential to endanger, the safe operation of trains or the safety of people, or that causes or could cause damage to adjoining property, (g) any slip, trip or fall by a person on, to or from a train, railway track, railway bridge, station, platform, escalator, lift or stairs, or any person being caught in the door of any rolling stock, (h) any situation where a load affects, or could affect, the safe passage of trains or the safety of people, or causes or could cause damage to adjoining property, (i) any accident or incident involving dangerous goods that affects, or could affect, the safety of railway operations or the safety of people, or that causes or could cause damage to adjoining property, (j) any breach of a safe working system or procedure, or the detection of any irregularity or deficiency in such a system or procedure, (k) any irregularity in any rail infrastructure (including electrical infrastructure and any obstruction on a running line) that could affect the safety of railway operations or the safety of people, (l) any irregularity in any rolling stock that could affect the safe operation of the train or the safety of people, or cause damage to the rolling stock, (m) any fire or explosion that causes damage to rail infrastructure or rolling stock, or both, or that causes the disruption or closure of a railway (even if the closure is only a precautionary measure), (n) any accident or incident on railway premises where a person inflicts, or is alleged to have inflicted, an injury on another person, (o) a suspected attempt to suicide, (p) if a rail safety worker employed by a rail transport operator has returned a result to a test designed to determine the concentration of alcohol or other drugs in a sample of blood or urine that suggests that the worker was in breach of a relevant safety requirement concerning the use of alcohol or other drugs at a relevant time, (q) the infliction of any wilful or unlawful damage to, or the defacement of, any rail infrastructure or rolling stock that could affect the safety of railway operations or the safety of people, (r) any accident or incident in a rail corridor that indicates that the security of the corridor is compromised and that affects, or may affect, the safety of railway operations. 8 Category A occurrences are the more serious than Category B occurrences and therefore must also be notified to OTSI as soon as practicable after the operator becomes aware of the occurrence List of notifiable occurrences Rail Industry Safety Report 9 1 49

ITSRR contact details Independent Transport Safety and Reliability Regulator Level, 1 Elizabeth Street Sydney NSW Phone: 863 71 Fax: 863 7 Email: contact@transportregulator.nsw.gov.au Web: www.transportregulator.nsw.gov.au Business and service hours ITSRR s hours of business are from 8.3am to 5pm, Monday to Friday (except public holidays). Reception is open from 9am to 5pm. After hours contact (for emergencies only) 863 711