Tarefa Tecnologia dos Materiais

Mutton 2004 Mutton 2004 Contents Executive Summary6 1 Introduction7 2 Background on Rail Materials Rail joints8 21 Joint Welds8 22 Common Rail Failures9 221 Fatigue Cracking9 222 Thermal Buckling9 3 Documented Rail Failures10 31 Narwonah Derailment11 32 Broken rail at Dubbo Coonamble line12 33 Broken Rail at Sydney Metro Northwest line14 4 Literature Review16 41 Fatigue Behaviour and Failure Mechanisms17 42 Defect Formation and Thermal Effects18 43 Strategies for Improving Weld Reliability19 5 Detailed Failure Analysis20 51 Identification of Failure Mode20 52 Origin and Mechanisms of Failure21 521 Origin of Defects21 522 Mechanisms of Crack Initiation and Propagation21 523 Thermal Buckling Contribution22 53 Testing and Examination Techniques23 531 Scanning Electron Microscopy SEM23 532 Ultrasonic Testing UT24 533 Rail Geometry Analysis25 534 Dynamic Load Simulation26 6 Economic and Safety Implications27 7 Future Research Directions29 8 Conclusion31 References32 Table of Figures Figure 1 Rail Aluminothermic Weld at ManildraNSW 20259 Figure 2 Track Buckle10 Figure 3 Derailed grain train in Northwestern NSW11 Figure 4 Broken rail at Dubbo Coonamble line13 Figure 5 Rail crack detected at Sydney Metro Northwest line15 Figure 6 a Straight break failure in an AT weld initiated from the weld base and b horizontal split adapted from Salehi et al 201117 Figure 7 ANSYS simulation validated against Dick et al 201623 Figure 8 Scanning Electron Miscroscopy SEM24 Figure 9 Rail Flaw Detector Ultrasonic testing system from NDT image source NDT Global25 Figure 10 Track Geometry Measurement Systems from Mermec Source Mermec public material25 Summary of Tables Table 1 Hypothesis Air rail Temperature Correlation22 Executive Summary This report comprehensively examines rail material failures causing derailments specifically focusing on fatigue cracking and thermal buckling It explores detailed scientific and engineering aspects of rail materials failure modes causative factors testing methods and preventive strategies Detailed case studies including the high profile Narwonah derailment are discussed with indepth analysis providing robust recommendations for material selection processing and maintenance practices 1Introduction Rail infrastructure forms the backbone of global transportation systems necessitating stringent safety and reliability standards Rail failures particularly those resulting from material deficiencies lead to severe consequences including derailments economic loss and potential fatalities For practical handling and facilitation of construction rails are not supplied in continuous lengths but rather in lengths that vary from 137m to 110m To provide the continuity of running surface the rails are joined by either fishplates or field welding of the rail ends In NSW regional most of the railway network is joined by field welding 2Background on Rail Materials Rail joints Rails predominantly utilise highstrength carbon steels meeting international standards such as AS 10851 They endure continuous cyclic loading dynamic stress and environmental exposure The mechanical properties required include high tensile strength toughness fatigue resistance and thermal stability 21 Joint Welds Rail ends welded in the field are done so with either the flash butt or the aluminothermic welding process which relies on a premixed alumina welding portion fusing onto the rail steel and filling the gap inbetween the rail ends providing the continuity required Aluminothermic welds are subject to similar inclusions and intrusions that occur in the rail manufacturing process As such similar defects occur which can result in broken rails and derailments Ultrasonic testing is the best method of detection and is carried out by the hand held unit shortly after the weld is complete Figure 1 Rail Aluminothermic Weld at ManildraNSW 2025 22 Common Rail Failures 221 Fatigue Cracking Rail fatigue cracks typically initiate at points of high stress concentration notably weld joints or defects Fatigue crack propagation is accelerated by cyclic loading conditions inherent in railway operations 222 Thermal Buckling Rail buckling results from excessive compressive stress due to thermal expansion This typically occurs during extreme temperature variations and is exacerbated by inadequate rail fastening and ballast support systems Improper rail welds are also known contributing factor to rail buckling especially under high temperature conditions The weld acts as a structural discontinuity that when combined with compressive thermal stress can cause the track to lose lateral stability and buckle Figure 2 Track Buckle 3Documented Rail Failures In order to discuss documented cases of rail failures I will use as an example the Narwonah Derailment 2017 Australia which is a high profile derailment that has a good source of literature online I will also use an internal case where I currently work which is a Broken rail at Dubbo Coonamble line at km 5012 2024 Australiaand I will also use another case from another project I worked in Sydney NSW in 2022 which was a partial transverse fracture at a rail weld at the Sydney Metro Chatswood to Tallawong line Detailed analyses of these incidents highlight the interplay of material properties maintenance practices and environmental factors contributing to catastrophic failures 31 Narwonah Derailment On 1 October 2017 a Pacific National loaded grain train 8838N was travelling on the Australian Rail Track Corporation ARTC rail network from Nevertire to Manildra in northwestern New South Wales The train consisted of two locomotives and 23 wagons The train was travelling south at Narwonah when 11 loaded grain wagons located at the rear of the consist derailedFigure 3 An emergency brake application occurred due to the uncoupling which brought the front portion of the train to a stand There were no injuries but there was substantial damage to nine wagons and track infrastructure Bureau 2020 Figure 3 Derailed grain train in Northwestern NSW The Australian Transport Safety Bureau ATSB found that a train derailment occurred due to poor track conditions at a rail joint and a short twist defect at the point of mount These factors contributed to the vertical unloading of wheels on the twelfth wagon which led to the derailment of that wagon and ten trailing wagons It was noted that the train was traveling at approximately 80 kmh which was above the 60 kmh speed limit specified by ARTC for that section of track Bureau 2020 Previous track defects in the area had been identified but not effectively addressed Additionally postincident investigations revealed that two of the three wagons examined were overloaded beyond the stated 81 tonnes and the grain within them was unevenly distributed likely contributing to the instability As a result of the derailment approximately 300 metres of rail fasteners and sleepers were replaced along with 150 metres of new track formation Thirteen prefabricated 12metre track panels were installed onsite using cranes with all rail joints welded and the track realigned In response to broader concerns ARTC made systemic changes to its maintenance practices and initiated an Asset Management Improvement Program aimed at enhancing the effectiveness of its Enterprise Asset Management System DCN 2020 In coordination with the rolling stock operator the rail infrastructure manager also reinforced the importance of adhering to speed limits published in official notices and ensuring that train loading is conducted accurately and within allowed limits The incident ultimately highlighted the critical need to maintain track infrastructure in sound condition and to ensure trains operate at or below designated speed thresholds to prevent future derailments 32 Broken rail at Dubbo Coonamble line During a routine patrol inspection on 29 July 2024 a broken rail was discovered on the Dubbo to Coonamble Line at kilometre 501237 The defect was located on the Up Main track and was found at an aluminothermic weld approximately 24 weeks old part of rail asset TRW61461A UGLRL 2024 The failure occurred within a 60HH rail section under moderate temperature conditions 16C The weld break had resulted in a clean transverse fracture with an 18 mm gap forming between the rail ends Figure 4 Fortunately the break did not cause any immediate signal failure or derailment as the last recorded movement was a light freight train limiting dynamic stress on the defect A close inspection of the weld revealed a brittle fracture along the weld plane The failure mode was consistent with poor fusion and potentially inadequate weld execution which led to crack propagation from the base of the weld upward Field observations indicated a foul aluminothermic weld as a primary cause with possible contribution from impact damage or cyclic fatigue The rail head dimensions showed slight variation between sides possibly indicating uneven cooling or misalignment during the original weld Figure 4 Broken rail at Dubbo Coonamble line Following identification the rail break was reported to management and the Network Control Officer at 1416 shortly after its discovery Temporary speed restrictions were implemented limiting traffic over the section to 40 kmh The weld was plated on the same day and the defective section was removed and replaced the following day 30 July 2024 using a premeasured closure rail This rapid intervention mitigated the risk of service disruption or derailment UGLRL 2024 The broken sections of the rail were quarantined for further investigation However no investigation has been undertaken till date It would have been important to conduct metallurgical tests to determine the root cause of the failure These tests would typically include macro and microstructural analysis to assess grain formation and fusion integrity as well as hardness testing across the heataffected zone Scanning Electron Microscopy SEM fractography would also be a valid method to study the fracture surface and confirm whether the crack originated at a weld defect or from external impact Chemical composition testing would also be another important test to verify compliance with required material standards while ultrasonic testing could confirm if internal flaws extended beyond the visible fracture This case highlights the critical importance of weld quality and regular patrol inspections in preventing catastrophic rail failures If undetected the fracture could have progressed under the next loaded train potentially leading to a derailment signal circuit failure or more severe infrastructure damage The proactive identification of this defect exemplifies the value of human inspections alongside automated systems and reinforces the need for continuous weld integrity monitoring particularly in rural freight corridors It also emphasizes the importance of maintaining welder certification standards and following strict aluminothermic welding procedures 33 Broken Rail at Sydney Metro Northwest line On a routine inspection at the Sydney Metro Northwest line at km 467 an incipient weld failure was discovered by an automated vision system fitted on train set 22 pf the Sydney Metro Fleet operating on the network The system detected a defect at an aluminothermic weld where a partial fracture Figure 5 had developed along the lower portion of the rail head and web The defect was flagged in real time and escalated for immediate action Figure 5 Rail crack detected at Sydney Metro Northwest line The weld in question appears to have failed due to a combination of residual stress and thermal effects compounded by a potentially flawed initial weld process Aluminothermic welds while commonly used for intrack rail joining are highly sensitive to proper execution Poor mold alignment inadequate preheating slag entrapment or rapid cooling can all contribute to internal flaws In this case the fracture propagated across part of the weld but had not yet split the rail entirely classifying it as a partial or incipient brittle fracture The location and orientation of the crack suggest it may have originated from the weld toe a common site for stress concentration and fatigue under repeated traffic loading Metro systems although typically subjected to lower axle loads than heavy rail networks operate at high frequency with minimal downtime placing constant stress on rail components In confined tunnel environments thermal cycles due to braking ventilation and environmental temperature changes can create localized heating contributing to stress accumulation If left undetected this type of weld defect could have rapidly progressed into a full rail break This would pose a severe derailment risk especially in highdensity metro environments where trains operate under tight headways and rely heavily on automated systems and track integrity Following the detection the defective rail section was removed immediately and material was quarantine for further forensic investigation Typical followup tests in such cases would include macrostructure and microstructure analysis of the weld zone scanning electron microscopy SEM to identify the crack initiation site hardness mapping across the heataffected zone and possibly residual stress testing These tests would determine whether the defect was caused by poor workmanship material inclusion or fatiguerelated damage and help inform future quality control measures This incident emphasizes the importance of combining regular patrols historical defect monitoring and realtime automated inspection technologies in maintaining metro rail safety The fact that the failure was detected on the same day it progressed shows the effectiveness of advanced condition monitoring systems in mitigating risk It also highlights the need for stringent control of field weld quality and proactive maintenance strategies to prevent weldrelated failures in highfrequency rail networks 4Literature Review Aluminothermic AT welding remains a widely adopted rail joining technique particularly in heavy haul railway systems due to its flexibility and relatively low equipment cost It is commonly employed for intrack welding associated with defect removal rerailing and field repairs Despite its operational convenience the process is highly operatordependent and inherently prone to metallurgical and geometrical inconsistencies which significantly affect weld reliability under high cyclic loads ISalehi 2011 41 Fatigue Behaviour and Failure Mechanisms Fatiguerelated failures in AT welds are typically categorized into two primary modes straight breaks and horizontal split webs HSWs Straight breaks often originate from stress concentrations at the weld collar edgeparticularly in the underhead foot or base regionsand propagate vertically as Mode I fatigue cracks These are exacerbated by the combined effects of trafficinduced cyclic stresses residual tensile stresses from the welding process and seasonal thermal contractions especially during colder months ISalehi 2011 HSWs by contrast initiate as horizontal cracks within the web region and may eventually change direction toward the rail head or base increasing the risk of material separation and derailment Factors such as track curvature vehicle hunting behaviour and vertical residual stresses strongly influence HSW formation and propagation Figure 6 a Straight break failure in an AT weld initiated from the weld base and b horizontal split adapted from Salehi et al 2011 Finite element analysis FEA and fatigue modelling using the Dang Van multi axial criterion have demonstrated that conventional Type A welds are more susceptible to fatigue initiation at the underhead region especially under lateral loading and curving conditions Conversely Type B welds which feature a smoother transition at the collar and selective alloying in the rail head exhibit superior fatigue resistance due to lower flank angles larger toe radii and reduced residual stresses ISalehi 2011 42 Defect Formation and Thermal Effects Complementing the fatiguecentric analysis Chen 2006 investigated weld quality from a thermophysical and metallurgical perspective identifying four critical defect types that compromise weld integrity and fatigue life 1 Coldlap defects result from insufficient fusion between molten metal and the rail base particularly at the weld toe These defects emerge when the meltback depth is less than the railend stickout and are prime sites for fatigue crack initiation 2 Shrinkage cavities are internal voids that form due to volumetric contraction of weld metal during solidification especially when isolated liquid pools are trapped by prematurely solidified material 3 Centreline defect clusters consist of smaller shrinkage pores aligned along the weld axis Their formation is evaluated using Niyamas temperature gradient criterion where a threshold of indicates a high likelihood of defect formation due to poor feeding in the mushy zone 4 Microporosity typically found in the rail head results from gas evolution during solidification and is correlated to dendrite arm spacing which increases under slower solidification and lower temperature gradients Chen et al showed that thermal process parametersincluding preheating time liquid steel temperature and weld gaphave a strong influence on the occurrence and severity of these defects For instance increasing liquid steel temperature 2100C and preheating duration 6 minutes significantly improves meltback and fusion quality thereby reducing coldlaps and shrinkage cavities Similarly widening the weld gap to 3850 mm facilitates vertical heat flow and reduces the risk of centreline defects However the same conditions that suppress major defects also slightly increase microporosity which while less critical may still contribute to longterm fatigue under unfavourable conditions Therefore weld design involves a tradeoff between suppressing severe defects and managing tolerable levels of microporosity Chen et al 2006 43 Strategies for Improving Weld Reliability To optimize the structural performance of AT welds both studies recommend practical process modifications Transition from Type A to Type B weld geometries to minimize geometric stress concentrations Adopt optimized welding parameters such as o Weld gap 38 mm o Preheating time 6 minutes o Liquid steel temperature 2100C These adjustments reduce the risk of initiating fatigue cracks eliminate shrinkage and fusionrelated defects and improve the overall fatigue life and structural reliability of the weld particularly in heavy haul applications Furthermore maintaining adequate ballast support avoiding dynamic impact events and performing rigorous weld quality inspections are essential environmental and operational considerations to sustain weld performance Building on the insights from the literature and realworld examples the following chapter presents a detailed failure analysis of aluminothermic welds in rail applications integrating both documented case studies and simulation results to illustrate the mechanisms involved 5Detailed Failure Analysis 51 Identification of Failure Mode Examination of the affected rail section confirms that fatigue cracking was the dominant failure mechanism with crack initiation occurring at the aluminothermic AT weld joints The initiation sites were linked to material processing defects specifically Coldlap defects and centreline defect clusters as described by Chen et al 2006 Stress concentrations at the weld underhead and foot due to poor geometrical transitions as demonstrated by ISalehi 2011 These findings are consistent with Mutton 2004 who showed that AT welds are a primary source of fatigueinduced rail failures in heavy haul operations particularly under high axleload conditions These intrinsic flaws acted as microcrack nucleation points Under repeated highcycle loading of rail traffic cracks propagated progressively through the rail crosssection The deterioration was further exacerbated by thermal stresses seasonal temperature fluctuations induced tensile longitudinal stresses leading to lateral buckling and eventual catastrophic failure in the affected rail section Jeong 1997 This sequence of failure processing defect fatigue crack thermal stress buckling is consistent with mechanisms observed in both controlled laboratory studies Chen 2006 and field observations in heavy haul networks ISalehi 2011 52 Origin and Mechanisms of Failure 521 Origin of Defects In aluminothermic welds several common defect types are known to impair fatigue resistance Coldlap defects occur when insufficient fusion takes place between molten metal and the rail base or foot This typically results from inadequate preheating low liquid steel temperature or incorrect weld gap Chen 2006 Shrinkage cavities and centreline defect clusters originate during weld metal solidification due to volumetric contraction and interdendritic feeding limitations These defects align along the weld axis and significantly reduce fatigue life Chen 2006 Surface geometric discontinuities such as small toe radii and sharp flank angles in Type A welds introduce stress concentrations that promote crack initiation ISalehi 2011 Ross 2004 further demonstrated that such geometric factors directly correlate with reduced fatigue resistance in field welds 522 Mechanisms of Crack Initiation and Propagation Fatigue crack initiation typically occurs at subsurface or nearsurface defect sites under cyclic loading High cyclic longitudinal stresses exacerbated by residual stresses and vehicle curving forces create conditions ideal for fatigue crack initiation at the underhead radius of the weld Centreline pores and microporosity even if initially small can also grow into significant fatigue cracks through repeated stress cycling Chen 2006 It has been shown that microporosity while not a primary driver of initial cracking can accelerate fatigue crack growth in the absence of larger defects Barsom Imhof 1978 Crack propagation is accelerated by Tensile residual stresses induced during the welding process Thermal contraction during cold seasons which pulls the rail longitudinally and amplifies crackdriving forces Dynamic loads from rail traffic including bending and lateral loads from wheel hunting and track misalignments 523 Thermal Buckling Contribution Rails subjected to increasing thermal stresses during hot weather may experience lateral instability thermal buckling especially when local stiffness is reduced due to fatigue cracks or microstructural weakening at the weld As fatigue damage accumulates the rails critical buckling load decreases This makes it susceptible to track alignment failure during temperature extremes as shown in fracture tolerance studies by Jeong 1997 5231 Thermal rail simulation A rail to air temperature increase of 15C is typically observed in rail that is exposed to full sunlightTable 1 This observation is mirrored in work undertaken by the FRA Mathew Dick 2016 Table 1 Hypothesis Air rail Temperature Correlation Air Temperature Expected Rail Temperature Expected Variance in Rail Temperature compared to Air Temperature 0⁰C 0⁰C 5⁰C 5⁰C 5⁰C 5⁰C 10⁰C 15⁰C 5⁰C 15⁰C 25⁰C 5⁰C 20⁰C 30⁰C 7⁰C 25⁰C 40⁰C 7⁰C 30⁰C 45⁰C 10⁰C 35⁰C 50⁰C 10⁰C 38⁰C 53⁰C 10⁰C 40⁰C 55⁰C 10⁰C 50⁰C 65⁰C 10⁰C In order to validate this theory a simple steadystate thermal simulation was performed to evaluate temperature gradients across a rail crosssection under peak NSW summer conditions A heat flux of 1200 Wm² was applied to the railhead to represent typical peak solar irradiance levels 10001200 Wm² based on BOM and PVGIS data Surfacetoambient radiation was considered with an ambient temperature of 35 C an emissivity of 085 representative of oxidized steel surfaces and a film coefficient of 10 Wm²C consistent with natural convection in still air under lowwind conditions The rail was meshed using refined tetrahedral elements particularly around the railhead to accurately resolve thermal gradients Simulation results Figure 7 indicated a maximum temperature of 533 C at the railhead and a minimum temperature of 494 C at the lower rail base demonstrating a clear vertical thermal gradient from the rail base convectiondominated to the railhead solar exposuredominated These outcomes are consistent with observed field behaviour under comparable climatic conditions and closely match the expected result of 50⁰C from table 1 Figure 7 ANSYS simulation validated against Dick et al 2016 53 Testing and Examination Techniques 531 Scanning Electron Microscopy SEM SEM is employed to examine fracture surfaces and microstructural details of weld defects It allows visualisation of fatigue striations confirming that a crack was propagated via cyclic loading SEM has been used to verify that coldlap defects and centreline clusters in thermite welds are initiation sites for fatigue cracking SEM also provides highresolution images of interdendritic microporosity and confirms pore morphology Figure 8 Scanning Electron Miscroscopy SEM 532 Ultrasonic Testing UT UT is a nondestructive technique widely used for inservice rail inspection It is highly effective at detecting internal voids such as shrinkage cavities and centreline defects UT can also reveal fatigue cracks before they propagate to visible or surfacebreaking stages Modern phased array UT systems provide realtime 3D imaging of rail weld zones Ross 2004 Figure 9 Rail Flaw Detector Ultrasonic testing system from NDT image source NDT Global 533 Rail Geometry Analysis Measurement of rail vertical profile lateral alignment and crosssectional geometry is performed to detect subtle changes due to thermal buckling or fatigueinduced distortion Track geometry vehicles use laserbased systems or inertial profilers to record rail geometry with submillimetre accuracy Deviation patterns are correlated with known zones of weldinduced weakness as observed in ISalehi 2011 Figure 10 Track Geometry Measurement Systems from Mermec Source Mermec public material 534 Dynamic Load Simulation Advanced finite element models FEM and multiaxial fatigue criteria eg Dang Van criterion are used to simulate how dynamic train loads interact with preexisting weld defects Salehi et al 2011 demonstrated that cyclic dynamic loads cause stress amplification at the weld underhead correlating well with observed crack locations Dynamic simulations also account for track support degradation eg ballast pumping which was shown by Salehi to accelerate fatigue damage and trigger local buckling 6Economic and Safety Implications Rail material failures such as those documented in this report carry significant economic and safety consequences The direct costs of failure include rail and track component replacement service disruption rolling stock damage and emergency response operations The indirect costs including network delays contractual penalties reputational harm and reduced freight competitiveness can be even more substantial The Narwonah derailment Bureau 2020 illustrates this very well approximately 300 meters of rail and track formation were replaced requiring extensive crane operations and prefabricated panel installation Beyond material replacement there were large operational disruptions on the ARTC network including lost train paths and followon delays across the grain supply chain Similar repair interventions following the DubboCoonamble rail break and Sydney Metro incipient weld failure further highlight the costs of unscheduled maintenance and emergency welding in operational corridors From a safety perspective weldrelated rail failures pose critical risks to both heavy haul and passenger networks The Dubbo case showed how a weld break with an 18 mm transverse gap if undetected could have caused a derailment on the next loaded train UGLRL 2024 The Sydney Metro incident highlights the heightened derailment risks in highdensity metro environments where automatic train operations ATO and tight headways leave little margin for rail failure Aluminothermic weld failures also contribute to thermal buckling vulnerability As shown in this report fatigue cracking and weldplane defects reduce a rails critical buckling load Jeong 1997 When welds act as weakened points under high compressive thermal stresses lateral buckling events can occur even under otherwise compliant track geometry with serious derailment potential The cumulative safety risk is magnified by the fact that AT weld reliability remains operatordependent ISalehi 2011 and field practices still exhibit significant variability Welds with poor fusion centreline defects or coldlaps become latent failure points that degrade progressively under cyclic loading and seasonal stress cycles Mutton 2004 As modern networks increase axle loads and traffic density the economic and safety stakes of weld quality failures rise sharply However this report also demonstrates that welldocumented strategies to improve weld reliability including transition to Type B welds optimized thermal processing parameters and enhanced inspection regimes Ross 2004 Lawrence et al 2004 offer costeffective paths to reduce both economic losses and safety risks Moreover the integration of automated vision systems as in Sydney Metro and modern UT techniques Ross 2004 can further improve defect detection and prevent derailments In summary weldrelated failures not only incur high direct economic costs but can also trigger catastrophic safety events Proactive process control continuous monitoring and rigorous maintenance strategies are essential investments to mitigate these risks and safeguard both rail network efficiency and public safety 7Future Research Directions Building on the analysis of failure mechanisms and the economic and safety impacts outlined in this report this chapter identifies key areas where future research can further enhance the reliability of rail materials and weld performance While significant progress has been made in understanding the failure mechanisms of aluminothermic AT welds fatigue behaviour and the role of thermal stresses evolving operational demands and emerging technologies open new avenues for targeted research The analysis presented in this report highlights several critical knowledge gaps and opportunities for future research to further enhance rail safety and material reliability While significant progress has been made in understanding the failure mechanisms of aluminothermic AT welds fatigue behaviour and the role of thermal stresses evolving operational demands and new technologies warrant continued investigation Firstly future research should focus on developing advanced welding process monitoring and control systems Despite known best practices Salehi et al 2011 Chen et al 2006 field weld variability remains high The application of realtime thermal imaging combined with digital process logging could enable early detection of suboptimal weld parameters reducing the occurrence of coldlaps and shrinkage cavities Secondly there is a strong need for enhanced understanding of residual stress evolution during rail service life While initial residual stresses from welding are well documented Mutton 2004 their interaction with cyclic loads track settlement and temperature fluctuations over time is not fully understood Longterm field studies using embedded strain gauges and periodic ultrasonic stress mapping would improve predictive maintenance models and help prevent latestage failures Another promising avenue is the refinement of automated defect detection systems The success of automated visionbased crack detection in the Sydney Metro case Chapter 33 demonstrates the potential of machine learning for early intervention However further research is needed to integrate multi sensor platforms vision UT thermal imaging and to develop robust defect progression models that can better predict when interventions are required Additionally more work is required to evaluate the interaction between weld defects and modern operational profiles As axle loads increase and networks move toward higher speeds and denser traffic Ross 2004traditional fatigue life models may underestimate risk Expanded multiaxial fatigue testing and dynamic simulation studiesincorporating realworld loading scenarios and climate dataare needed to validate and update existing design criteria At the materials level further exploration of alternative weld compositions and postweld treatments offers additional opportunities for enhancing fatigue life Finally future research should explore alternative materials and weld filler compositions for AT welding Selective alloying of the weld metal combined with optimized postweld heat treatments may reduce residual stresses and improve fatigue life Chen 2006 Pilot trials of new alloy formulations in controlled test track environments should be prioritized In conclusion advancing the state of knowledge in process monitoring residual stress modelling automated inspection fatigue modelling under modern loads and weld metallurgy will be key to enhancing rail network safety and reliability Such research will not only reduce economic losses but also provide vital safeguards against future derailments in both heavy haul and passenger rail network 8Conclusion This assessment has demonstrated that rail material failures particularly weldrelated fatigue cracking exacerbated by thermal stresses remain a critical risk to network safety and reliability Through detailed failure analysis simulation and realworld case studies this report has outlined practical strategies to mitigate these risks and highlighted future research directions that will be essential to enabling safer more resilient rail infrastructure in the face of growing operational demands References Bureau 2020 Bureau 2020 Bureau 2020 Bureau 2020 DCN 2020 DCN 2020 UGLRL 2024 UGLRL 2024 UGLRL 2024 UGLRL 2024 ISalehi 2011 ISalehi 2011 Chen 2006 Chen 2006 Mutton 2004 Mutton 2004 Jeong 1997 Jeong 1997 ISalehi 2011 ISalehi 2011 Ross 2004 Ross 2004 Mathew Dick 2016 Mathew Dick 2016 Mutton 2004 Mutton 2004 TASKS In this assessment you will research and report on a realworld material failure of your choice The material failure must be due to either 1 Use of an improper material 2 Improperly processed materials 3 Unexpected mechanical failure 4 Service conditions You can choose any example and are encouraged to choose an example or application that is of interest to you eg bridge collapse rocket explosion It is recommended that you select an example with sufficient documentation for you to report on The elements of your report will include Literature review including Background on the component which failed what its purpose is what the design requirements are what the environmental factors are At least 3 citations which document the material failure most likely news sources and similar but can be scientific publications if you can find them At least 8 citations from scientific publications to support the remainder of your analysis and report Failure analysis including Identification of the failure mode Origin of the failure Discussion on how the failure occurred Any examination or testing which was done to determine the cause of the failure if applicable Suggestions for how the failure could have been avoided including as applicable What material choice should have been made Change to material processing Changes to component design Changes to loading assumptions Any other topics which may be relevant to your application NOTE Use Harvard AGPS referencing style Mutton 2004 Mutton 2004 Marking Rubric Criteria Level 1 Level 2 Level 3 Level 4 Level 5 Marks 1 Literature Review Demonstrate the ability to evaluate and select appropriate literature with critique and review 04 Marks No evaluation and selection of literature with critique and review Cited no references 58 Marks Poor evaluation and selection of literature with critique and review Cited less than 5 references 912 Marks Adequate evaluation and selection of literature with critique and review Cited less than 8 references 1316 Marks Good evaluation and selection of literature with critique and review Cited correct number of references 1720 Marks Excellent evaluation and selection of literature with critique and review Cited correct number of references 80 2 Background Demonstrate the ability to clearly and concisely describe the situation which led to the failure mode 02 Marks No evaluation of background information 34 Marks Poor evaluation of background information 56 Marks Adequate evaluation of background information 78 Marks Good evaluation of background information 910 Marks Excellent evaluation of background information 40 3 Failure analysis Demonstrate the ability to analyse and critique the potential failure modes and propose the likely cause of failure and a strong justification argument to back the assertions 08 Marks No attempt or the analysis and critique was limited andor without support or justification 916 Marks The analysis and critique was poorly articulated supported andor justified 1724 Marks The analysis and critique was somewhat articulated supported andor justified 2532 Marks The analysis and critique was clear and well articulated supported andor justified 3340 Marks The analysis and critique was clear comprehensive and extremely well articulated supported andor justified 160 4 Rectification Demonstrate the ability to propose rectification or prevention strategies 04 Marks No attempt or proposed strategies were limited andor without support or justification 58 Marks Proposed strategies were poorly articulated supported andor justified 912 Marks Proposed strategies were somewhat articulated supported andor justified 1316 Marks Proposed strategies were clear and wellarticulated supported andor justified 1720 Marks Proposed strategies were clear comprehensive and extremely wellarticulated supported andor justified 80 5 Writing Style Demonstrate the ability to professionally present the report including intext citation and referencing 02 Marks No attempt at intext citations and referencing andor professional presentation and language were limited 34 Marks Poor demonstration of in text citations and referencing andor professional presentation and language 56 Marks Adequate demonstration of intext citations and referencing andor professional presentation and language 78 Marks Good demonstration of in text citations and referencing andor professional presentation and language 910 Marks Excellent demonstration of intext citations and referencing andor professional presentation and language 40 No visible text to extract Contents Executive Summary8 1 Introduction9 2 Background on Rail Materials Rail joints10 21 Joint Welds10 22 Common Rail Failures11 221 Fatigue Cracking11 222 Thermal Buckling11 3 Documented Rail Failures12 31 Narwonah Derailment13 32 Broken rail at Dubbo Coonamble line14 33 Broken Rail at Sydney Metro Northwest line16 34 Broken Rail at Murrobo to Spring Hill line17 4 Literature Review21 41 Fatigue Behaviour and Failure Mechanisms22 42 Defect Formation and Thermal Effects23 43 Strategies for Improving Weld Reliability24 5 Detailed Failure Analysis25 51 Identification of Failure Mode25 52 Origin and Mechanisms of Failure26 521 Origin of Defects26 522 Mechanisms of Crack Initiation and Propagation26 523 Thermal Buckling Contribution27 53 Testing and Examination Techniques29 531 Scanning Electron Microscopy SEM29 532 Ultrasonic Testing UT30 533 Rail Geometry Analysis31 534 Dynamic Load Simulation32 6 Economic and Safety Implications33 7 Future Research Directions36 8 Conclusion38 References39 Appendix A Broken Rail at Murrobo to Spring Hill line Additional Photos extracted from the Bureau Veritas Report NO252662710441 Table of Figures Executive Summary8 1 Introduction9 2 Background on Rail Materials Rail joints10 21 Joint Welds10 22 Common Rail Failures11 221 Fatigue Cracking11 222 Thermal Buckling11 3 Documented Rail Failures12 31 Narwonah Derailment13 32 Broken rail at Dubbo Coonamble line14 33 Broken Rail at Sydney Metro Northwest line16 34 Broken Rail at Murrobo to Spring Hill line17 4 Literature Review21 41 Fatigue Behaviour and Failure Mechanisms22 42 Defect Formation and Thermal Effects23 43 Strategies for Improving Weld Reliability24 5 Detailed Failure Analysis25 51 Identification of Failure Mode25 52 Origin and Mechanisms of Failure26 521 Origin of Defects26 522 Mechanisms of Crack Initiation and Propagation26 523 Thermal Buckling Contribution27 5231 Thermal rail simulation28 53 Testing and Examination Techniques29 531 Scanning Electron Microscopy SEM29 532 Ultrasonic Testing UT30 533 Rail Geometry Analysis31 534 Dynamic Load Simulation32 6 Economic and Safety Implications33 7 Future Research Directions36 8 Conclusion38 References39 Appendix A Broken Rail at Murrobo to Spring Hill line Additional Photos extracted from the Bureau Veritas Report NO252662710441 Summary of Tables Executive Summary8 1 Introduction9 2 Background on Rail Materials Rail joints10 21 Joint Welds10 22 Common Rail Failures11 221 Fatigue Cracking11 222 Thermal Buckling11 3 Documented Rail Failures12 31 Narwonah Derailment13 32 Broken rail at Dubbo Coonamble line14 33 Broken Rail at Sydney Metro Northwest line16 34 Broken Rail at Murrobo to Spring Hill line17 4 Literature Review21 41 Fatigue Behaviour and Failure Mechanisms22 42 Defect Formation and Thermal Effects23 43 Strategies for Improving Weld Reliability24 5 Detailed Failure Analysis25 51 Identification of Failure Mode25 52 Origin and Mechanisms of Failure26 521 Origin of Defects26 522 Mechanisms of Crack Initiation and Propagation26 523 Thermal Buckling Contribution27 5231 Thermal rail simulation28 53 Testing and Examination Techniques29 531 Scanning Electron Microscopy SEM29 532 Ultrasonic Testing UT30 533 Rail Geometry Analysis31 534 Dynamic Load Simulation32 6 Economic and Safety Implications33 7 Future Research Directions36 8 Conclusion38 References39 Appendix A Broken Rail at Murrobo to Spring Hill line Additional Photos extracted from the Bureau Veritas Report NO252662710441 Executive Summary This report provides a comprehensive examination of rail material failures that can lead to derailments with a particular focus on fatigue cracking and thermal buckling It investigates the scientific and engineering aspects of rail materials key failure mechanisms contributing factors advanced testing methods and preventive strategies The report draws on detailed case studies including the highprofile Narwonah derailment and three additional recent cases from operational rail networks to illustrate how material properties maintenance practices and environmental conditions interact to cause rail failures Through indepth analysis the report offers robust recommendations for improving material selection welding practices inspection regimes and maintenance strategies with the aim of enhancing rail safety and reliability across both freight and passenger networks 1Introduction Currently rail transport stands out as one of the most commonly used means of transportation due to its high safety standards large carrying capacity and cost efficiency As railway networks continue to expand including urban rail systems there is an increasing focus on ensuring operational safety and effective risk management throughout the entire lifecycle This goal relies on the use of scientific tools and efficient railway operation management practices Luong T A 2022 Rail infrastructure forms the backbone of global transportation systems and requires stringent safety and reliability standards Rail failures particularly those arising from material deficiencies or welding defects can result in severe consequences including derailments economic loss and even fatalities In practice rails are not manufactured or supplied in continuous lengths They are delivered in sections ranging from 137 m to 110 m to facilitate handling and installation To provide a continuous running surface these rail sections are joined using either fishplates or field welding at the rail ends On the NSW regional network the majority of rail joints are formed by field welding This report explores how welding practices material behaviour and environmental factors contribute to the occurrence of rail failures It draws on both established literature and recent real world case studies to highlight critical risks and opportunities for improved engineering practice 2Background on Rail Materials Rail joints Rail tracks are predominantly made of carbon steels due to their high mechanical strength and compliance with international standards such as AS 10851 Standards Australia 2019 These rails are subjected to constant cyclic loading dynamic stresses and exposure to harsh environmental conditions The required mechanical properties include high tensile strength toughness fatigue resistance and thermal stability 21 Joint Welds For rail connections it is necessary for the rail ends to be welded in the field using two main processes flash butt welding or aluminothermic welding In the aluminothermic process a preprepared mixture of aluminum oxide reacts exothermically melting and bonding with the steel of the rails while filling the gap between the rail ends thus ensuring track continuity Aluminothermic welds are susceptible to defects similar to those found in the rail manufacturing process such as inclusions in the metallic microstructure and internal discontinuities These defects act as stress concentrators and can lead to fractures in the rail structure or even derailments The most effective technique for detecting this type of defect is ultrasonic testing which is typically performed manually right after the welding process is completed ensuring both the safety and quality of the weld Figure 1 Rail Aluminothermic Weld at ManildraNSW 2025 22 Common Rail Failures 221 Fatigue Cracking Fatigue cracks in rails typically initiate at stress concentration points especially at welded joints or locations with microstructural defects and develop due to the cyclic loads applied to the rails The propagation of these cracks is accelerated by the cyclic loading conditions inherent to railway operations 222 Thermal Buckling Rail buckling results from excessive compressive stress due to thermal expansion This typically occurs during extreme temperature variations and is exacerbated by inadequate rail fastening and ballast support systems Improper rail welds are also known contributing factor to rail buckling especially under high temperature conditions The weld acts as a structural discontinuity that when combined with compressive thermal stress can cause the track to lose lateral stability and buckle Figure 2 Track Buckle 3Documented Rail Failures In order to discuss documented cases of rail failures I will use as an example the Narwonah Derailment 2017 Australia which is a high profile derailment that has a good source of literature online I will also use an internal case where I currently work which is a Broken rail at Dubbo Coonamble line at km 5012 2024 Australiaand I will also use another case from another project I worked in Sydney NSW in 2022 which was a partial transverse fracture at a rail weld at the Sydney Metro Chatswood to Tallawong line and a very recent case that occurred in this May2025 at Murrobo to Spring Hill at km 305027 in NSW regional Detailed analyses of these incidents highlight the interplay of material properties maintenance practices and environmental factors contributing to catastrophic failures 31 Narwonah Derailment On 1 October 2017 a Pacific National loaded grain train 8838N was travelling on the Australian Rail Track Corporation ARTC rail network from Nevertire to Manildra in northwestern New South Wales The train consisted of two locomotives and 23 wagons The train was travelling south at Narwonah when 11 loaded grain wagons located at the rear of the consist derailedFigure 3 An emergency brake application occurred due to the uncoupling which brought the front portion of the train to a stand There were no injuries but there was substantial damage to nine wagons and track infrastructure Bureau 2020 Figure 3 Derailed grain train in Northwestern NSW The Australian Transport Safety Bureau ATSB found that a train derailment occurred due to poor track conditions at a rail joint and a short twist defect at the point of mount These factors contributed to the vertical unloading of wheels on the twelfth wagon which led to the derailment of that wagon and ten trailing wagons It was noted that the train was traveling at approximately 80 kmh which was above the 60 kmh speed limit specified by ARTC for that section of track Bureau 2020 Previous track defects in the area had been identified but not effectively addressed Additionally postincident investigations revealed that two of the three wagons examined were overloaded beyond the stated 81 tonnes and the grain within them was unevenly distributed likely contributing to the instability As a result of the derailment approximately 300 metres of rail fasteners and sleepers were replaced along with 150 metres of new track formation Thirteen prefabricated 12metre track panels were installed onsite using cranes with all rail joints welded and the track realigned In response to broader concerns ARTC made systemic changes to its maintenance practices and initiated an Asset Management Improvement Program aimed at enhancing the effectiveness of its Enterprise Asset Management System DCN 2020 In coordination with the rolling stock operator the rail infrastructure manager also reinforced the importance of adhering to speed limits published in official notices and ensuring that train loading is conducted accurately and within allowed limits The incident ultimately highlighted the critical need to maintain track infrastructure in sound condition and to ensure trains operate at or below designated speed thresholds to prevent future derailments 32 Broken rail at Dubbo Coonamble line During a routine patrol inspection on 29 July 2024 a broken rail was discovered on the Dubbo to Coonamble Line at kilometre 501237 The defect was located on the Up Main track and was found at an aluminothermic weld approximately 24 weeks old part of rail asset TRW61461A UGLRL 2024 The failure occurred within a 60HH rail section under moderate temperature conditions 16C The weld break had resulted in a clean transverse fracture with an 18 mm gap forming between the rail ends Figure 4 Fortunately the break did not cause any immediate signal failure or derailment as the last recorded movement was a light freight train limiting dynamic stress on the defect A close inspection of the weld revealed a brittle fracture along the weld plane The failure mode was consistent with poor fusion and potentially inadequate weld execution which led to crack propagation from the base of the weld upward Field observations indicated a foul aluminothermic weld as a primary cause with possible contribution from impact damage or cyclic fatigue The rail head dimensions showed slight variation between sides possibly indicating uneven cooling or misalignment during the original weld Figure 4 Broken rail at Dubbo Coonamble line Following identification the rail break was reported to management and the Network Control Officer at 1416 shortly after its discovery Temporary speed restrictions were implemented limiting traffic over the section to 40 kmh The weld was plated on the same day and the defective section was removed and replaced the following day 30 July 2024 using a premeasured closure rail This rapid intervention mitigated the risk of service disruption or derailment UGLRL 2024 The broken sections of the rail were quarantined for further investigation However no investigation has been undertaken till date those are allocated due to the level of defect and are determined by the internal SME It would have been important to conduct metallurgical tests to determine the root cause of the failure These tests would typically include macro and microstructural analysis to assess grain formation and fusion integrity as well as hardness testing across the heataffected zone Scanning Electron Microscopy SEM fractography would also be a valid method to study the fracture surface and confirm whether the crack originated at a weld defect or from external impact Chemical composition testing would also be another important test to verify compliance with required material standards while ultrasonic testing could confirm if internal flaws extended beyond the visible fracture This case highlights the critical importance of weld quality and regular patrol inspections in preventing catastrophic rail failures If undetected the fracture could have progressed under the next loaded train potentially leading to a derailment signal circuit failure or more severe infrastructure damage The proactive identification of this defect exemplifies the value of human inspections alongside automated systems and reinforces the need for continuous weld integrity monitoring particularly in rural freight corridors It also emphasizes the importance of maintaining welder certification standards and following strict aluminothermic welding procedures 33 Broken Rail at Sydney Metro Northwest line On a routine inspection at the Sydney Metro Northwest line at km 467 an incipient weld failure was discovered by an automated vision system fitted on train set 22 of the Sydney Metro Fleet operating on the network The system detected a defect at an aluminothermic weld where a partial fracture Figure 5 had developed along the lower portion of the rail head and web The defect was flagged in real time and escalated for immediate action Figure 5 Rail crack detected at Sydney Metro Northwest line The weld in question appears to have failed due to a combination of residual stress and thermal effects compounded by a potentially flawed initial weld process Aluminothermic welds while commonly used for intrack rail joining are highly sensitive to proper execution Poor mold alignment inadequate preheating slag entrapment or rapid cooling can all contribute to internal flaws In this case the fracture propagated across part of the weld but had not yet split the rail entirely classifying it as a partial or incipient brittle fracture The location and orientation of the crack suggest it may have originated from the weld toe a common site for stress concentration and fatigue under repeated traffic loading Metro systems although typically subjected to lower axle loads than heavy rail networks operate at high frequency with minimal downtime placing constant stress on rail components In confined tunnel environments thermal cycles due to braking ventilation and environmental temperature changes can create localised heating contributing to stress accumulation If left undetected this type of weld defect could have rapidly progressed into a full rail break This would pose a severe derailment risk especially in highdensity metro environments where trains operate under tight headways and rely heavily on automated systems and track integrity Following the detection the defective rail section was removed immediately and material was quarantine for further forensic investigation Typical followup tests in such cases would include macrostructure and microstructure analysis of the weld zone scanning electron microscopy SEM to identify the crack initiation site hardness mapping across the heataffected zone and possibly residual stress testing These tests would determine whether the defect was caused by poor workmanship material inclusion or fatiguerelated damage and help inform future quality control measures This incident emphasises the importance of combining regular patrols historical defect monitoring and realtime automated inspection technologies in maintaining metro rail safety The fact that the failure was detected on the same day it progressed shows the effectiveness of advanced condition monitoring systems in mitigating risk It also highlights the need for stringent control of field weld quality and proactive maintenance strategies to prevent weldrelated failures in highfrequency rail networks 34 Broken Rail at Murrobo to Spring Hill line During a routine inspection on 3 May 2025 a broken rail was detected on the Murrobo to Spring Hill line at kilometre 305300 Figure 6 within the UP rail The defect occurred at a flash butt weld located in a 507 m radius curve and involved an aged 53 kg rail rolled in 1966 The break triggered a track circuit failure on the PSB circuit leading to the issue being detected by signal electricians at 1033 hours during train movements WT27 had to be terminated early at Blayney and SPA working was implemented to manage operations Upon field inspection a horizontal break was observed across the weld area with a 10 mm gap at 24C rail temperature Initial track records indicated that the rail section had passed ultrasonic testing on 13 February 2025 with no defect detected and that the curve had been adjusted by MPM on 24 February 2025 An indepth metallurgical investigation was commissioned and carried out by Bureau Veritas to identify the root cause of the failure VERITAS 2025 The investigation included visual inspection fractography macro and microstructural analysis mechanical testing chemical composition testing and hardness testing Figure 6 Broken rail at Murrobo to Spring Hill line The results showed that the fracture initiated within the weld bond line where significant gas porosities were detected These porosities caused by entrapment of gases during the flashing phase of the flash butt welding process acted as stress concentrators initiating a brittle crack Furthermore it was found that the weld collar exceeded the maximum allowed profile deviation AS 1085202020 allows max 2 mm and that no postweld grinding had been performedboth clear indicators of poor weld execution and quality control issues Fracture analysis revealed a multistage failure mechanism Stage 1 Brittle crack initiation from the weld porosities Stage 2 Fatigue crack propagation along the rail web under cyclic loading Stage 3 Brittle transverse propagation through the web toward the rail head and foot Stage 4 Final sudden brittle fracture through the remaining cross section resulting in complete rail separation Metallography confirmed cracks both adjacent to and beneath the fracture surface consistent with stress risers from weld porosity Mechanical and chemical testing verified that the parent rail material was fit for purpose although the silicon content was slightly below the specified minimumbut this was not a primary factor in the failure Optical micrographs shown in Figure 7 taken close to the crack initiation site on the fracture surface reveal several cracks indicated by red arrows The crack initiated adjacent to the weld line where porosities were present in the excess weld metal These porosities may have acted as stress concentrators promoting crack initiation Figure 7 Optical micrographs a and b revealing cracks indicated by red arrows adjacent to the crack initiation site along the fracture surface Etchant 2 Nital Magnification x50 Remedial actions included the immediate plating of the break and removalreplacement of the defective section on the same day 3 May 2025 Bureau Veritas recommended that to avoid similar failures Postweld grinding must always be performed to remove upset metal and eliminate porosities Strict adherence to AS 2020AS 1085202020 profile tolerances must be enforced Welding process parameters should be optimised to minimise gas entrapment This case highlights the critical role of weld quality and inspection in ensuring rail integrity Despite passing ultrasonic testing only weeks prior the pre existing crack in this weld had already formed and remained dormant for a period before leading to final fracture If undetected this failure could have resulted in derailment or more severe operational impacts The combination of field vigilance through the detection of the signal fault and a thorough post failure investigation provided key insights to prevent recurrence and improve weld management practices across the network More details can be found at Appendix A 4Literature Review Aluminothermic AT welding remains a widely adopted rail joining technique particularly in heavy haul railway systems due to its flexibility and relatively low equipment cost It is commonly employed for intrack welding associated with defect removal rerailing and field repairs Despite its operational convenience the process is highly operatordependent and inherently prone to metallurgical and geometrical inconsistencies which significantly affect weld reliability under high cyclic loads ISalehi 2011 The process involves pouring molten steel into a mold positioned around the gap between the rail ends to be joined forming the bond through the solidification of the molten metal The generation of this molten steel occurs through a highly exothermic chemical reaction in which aluminum powder reacts with iron oxide releasing sufficient heat to melt the metal To execute the welding the rail ends must first be cut to create a precise gap and properly aligned Then a mold made of refractory material is fixed in place using metal clamps and sealed with a special cementitious material to prevent leakage After this setup the rail ends are preheated and the exothermic reaction is initiated inside a crucible which can be either reusable or singleuse The molten steel flows into the mold completely filling the joint while the resulting slag is separated and directed into specific collection containers Once the joint is fully filled the metal is left to cool Subsequently the excess material is removed followed by surface grinding to ensure proper leveling and a high quality finish of the rail Given that this technique is largely applied directly to rails in operation factors such as execution speed and operator safety become crucial to ensuring the efficiency quality and feasibility of the railway welding process RAIL WELD 2012 41 Fatigue Behaviour and Failure Mechanisms Fatiguerelated failures in AT welds are typically categorized into two primary modes straight breaks and horizontal split webs HSWs Straight breaks often originate from stress concentrations at the weld collar edge particularly in the underhead foot or base regionsand propagate vertically as Mode I fatigue cracks These are exacerbated by the combined effects of trafficinduced cyclic stresses residual tensile stresses from the welding process and seasonal thermal contractions especially during colder months ISalehi 2011 HSWs by contrast initiate as horizontal cracks within the web region and may eventually change direction toward the rail head or base increasing the risk of material separation and derailment Factors such as track curvature vehicle hunting behaviour and vertical residual stresses strongly influence HSW formation and propagation Figure 8 a Straight break failure in an AT weld initiated from the weld base and b horizontal split adapted from Salehi et al 2011 Finite element analysis FEA and fatigue modelling using the Dang Van multi axial criterion have demonstrated that conventional Type A welds are more susceptible to fatigue initiation at the underhead region especially under lateral loading and curving conditions Conversely Type B welds which feature a smoother transition at the collar and selective alloying in the rail head exhibit superior fatigue resistance due to lower flank angles larger toe radii and reduced residual stresses ISalehi 2011 42 Defect Formation and Thermal Effects Complementing the fatiguecentric analysis Chen 2006 investigated weld quality from a thermophysical and metallurgical perspective identifying four critical defect types that compromise weld integrity and fatigue life 1 Coldlap defects result from insufficient fusion between molten metal and the rail base particularly at the weld toe These defects emerge when the meltback depth is less than the railend stickout and are prime sites for fatigue crack initiation 2 Shrinkage cavities are internal voids that form due to volumetric contraction of weld metal during solidification especially when isolated liquid pools are trapped by prematurely solidified material 3 Centreline defect clusters consist of small shrinkage pores aligned along the weld axis Their formation is evaluated using the Niyama criterion which considers both the temperature gradient 𝐺how fast the temperature changes in space and the solidification rate 𝑅how fast the solidification front moves expressed as A value below this threshold indicates a high likelihood of defect formation as it suggests poor feeding of liquid metal in the mushy zone 4 Microporosity typically found in the rail head results from gas evolution during solidification and is correlated to dendrite arm spacing which increases under slower solidification and lower temperature gradients Chen et al showed that thermal process parametersincluding preheating time liquid steel temperature and weld gaphave a strong influence on the occurrence and severity of these defects For instance increasing liquid steel temperature 2100C and preheating duration 6 minutes significantly improves meltback and fusion quality thereby reducing coldlaps and shrinkage cavities Similarly widening the weld gap to 3850 mm facilitates vertical heat flow and reduces the risk of centreline defects However the same conditions that suppress major defects also slightly increase microporosity which while less critical may still contribute to longterm fatigue under unfavourable conditions Therefore weld design involves a tradeoff between suppressing severe defects and managing tolerable levels of microporosity Chen et al 2006 43 Strategies for Improving Weld Reliability To optimize the structural performance of AT welds both studies recommend practical process modifications Transition from Type A to Type B weld geometries to minimize geometric stress concentrations Adopt optimized welding parameters such as o Weld gap 38 mm o Preheating time 6 minutes o Liquid steel temperature 2100C These adjustments reduce the risk of initiating fatigue cracks eliminate shrinkage and fusionrelated defects and improve the overall fatigue life and structural reliability of the weld particularly in heavy haul applications Furthermore maintaining adequate ballast support avoiding dynamic impact events and performing rigorous weld quality inspections are essential environmental and operational considerations to sustain weld performance Similar fatigue mechanisms and defect formation risks have also been observed in flash butt welds in recent field cases such as the Murrobo to Spring Hill failure discussed in this report underscoring the need to apply these strategies across all weld types Building on the insights from the literature and realworld examples the following chapter presents a detailed failure analysis of aluminothermic welds in rail applications integrating both documented case studies and simulation results to illustrate the mechanisms involved 5Detailed Failure Analysis 51 Identification of Failure Mode Examination of the affected rail section confirms that fatigue cracking was the dominant failure mechanism with crack initiation occurring at the aluminothermic AT weld joints The initiation sites were linked to material processing defects specifically Coldlap defects and centreline defect clusters as described by Chen et al 2006 Stress concentrations at the weld underhead and foot due to poor geometrical transitions as demonstrated by ISalehi 2011 These findings are consistent with Mutton 2004 who showed that AT welds are a primary source of fatigueinduced rail failures in heavy haul operations particularly under high axleload conditions These intrinsic flaws acted as microcrack nucleation points Under repeated highcycle loading of rail traffic cracks propagated progressively through the rail crosssection The deterioration was further exacerbated by thermal stresses seasonal temperature fluctuations induced tensile longitudinal stresses leading to lateral buckling and eventual catastrophic failure in the affected rail section Jeong 1997 This sequence of failure processing defect fatigue crack thermal stress buckling is consistent with mechanisms observed in both controlled laboratory studies Chen 2006 and field observations in heavy haul networks ISalehi 2011 Recent field experience with the Murrobo to Spring Hill rail failure further validated this sequence confirming that fatigue cracking and brittle fracture mechanisms apply equally to flash butt welds in modern operational conditions 52 Origin and Mechanisms of Failure 521 Origin of Defects In aluminothermic or Flash butt welds several common defect types are known to impair fatigue resistance Coldlap defects occur when insufficient fusion takes place between molten metal and the rail base or foot This typically results from inadequate preheating low liquid steel temperature or incorrect weld gap Chen 2006 Gas porosity entrapped during the welding process VERITAS 2025 Shrinkage cavities and centreline defect clusters originate during weld metal solidification due to volumetric contraction and interdendritic feeding limitations These defects align along the weld axis and significantly reduce fatigue life Chen 2006 Surface geometric discontinuities such as small toe radii and sharp flank angles in Type A welds introduce stress concentrations that promote crack initiation ISalehi 2011 Ross 2004 further demonstrated that such geometric factors directly correlate with reduced fatigue resistance in field welds 522 Mechanisms of Crack Initiation and Propagation Fatigue crack initiation typically occurs at subsurface or nearsurface defect sites in or near the weld zone The combination of residual stresses cyclic operational loads and seasonal thermal stresses provides sufficient driving force for crack growth Initial fatigue cracks propagate longitudinally along the rail web or head Subsequent crack branching into transverse planes leads to rapid section weakening Final fracture often occurs suddenly when remaining crosssection integrity is lost Dynamic load interactions ISalehi 2011 further amplify these effects especially in curved track where lateral forces and rail bending are pronounced Realworld validation of this failure sequence was observed in the Murrobo to Spring Hill rail fracture at km 305300 VERITAS 2025 The Bureau Veritas investigation confirmed that The fracture initiated at gas porosities within the weld bond line of a flash butt weld Progressive fatigue crack growth occurred longitudinally through the web The fracture then propagated transversely to the rail head and foot resulting in final brittle fracture and rail separation SEM analysis macro and microexaminations confirmed this multistage fatigue mechanism fully consistent with the model described above This case provides strong practical evidence that the generic fatiguecrack propagation mechanisms described here are highly applicable in modern rail weld failures regardless of weld process type 523 Thermal Buckling Contribution Rails weakened by fatigue cracks are also more prone to thermal buckling as the rails lateral stiffness and critical buckling load are reduced In this degraded state seasonal temperature variations can trigger lateral instability particularly in curved sections Jeong 1997 As part of this study a thermal rail simulation was conducted to quantify temperature gradients and the potential impact on rail integrity see Section 5231 below 5231 Thermal rail simulation A rail to air temperature increase of 15C is typically observed in rail that is exposed to full sunlightTable 1 This observation is mirrored in work undertaken by the FRA Mathew Dick 2016 Table 1 Hypothesis Air rail Temperature Correlation Air Temperature Expected Rail Temperature Expected Variance in Rail Temperature compared to Air Temperature 0⁰C 0⁰C 5⁰C 5⁰C 5⁰C 5⁰C 10⁰C 15⁰C 5⁰C 15⁰C 25⁰C 5⁰C 20⁰C 30⁰C 7⁰C 25⁰C 40⁰C 7⁰C 30⁰C 45⁰C 10⁰C 35⁰C 50⁰C 10⁰C 38⁰C 53⁰C 10⁰C 40⁰C 55⁰C 10⁰C 50⁰C 65⁰C 10⁰C In order to validate this theory a simple steadystate thermal simulation was performed to evaluate temperature gradients across a rail crosssection under peak NSW summer conditions A heat flux of 1200 Wm² was applied to the railhead to represent typical peak solar irradiance levels 10001200 Wm² based on BOM and PVGIS data Surfacetoambient radiation was considered with an ambient temperature of 35 C an emissivity of 085 representative of oxidized steel surfaces and a film coefficient of 10 Wm²C consistent with natural convection in still air under lowwind conditions The rail was meshed using refined tetrahedral elements particularly around the railhead to accurately resolve thermal gradients Simulation results Figure 9 indicated a maximum temperature of 533 C at the railhead and a minimum temperature of 494 C at the lower rail base demonstrating a clear vertical thermal gradient from the rail base convectiondominated to the railhead solar exposuredominated These outcomes are consistent with observed field behaviour under comparable climatic conditions and closely match the expected result of 50⁰C from table 1 Figure 9 ANSYS simulation validated against Dick et al 2016 53 Testing and Examination Techniques A range of nondestructive and destructive testing techniques are essential for diagnosing rail failure modes and verifying fatigue mechanisms 531 Scanning Electron Microscopy SEM A highly effective method for analyzing material failures is Scanning Electron Microscopy SEM which generates highresolution images by detecting electrons either reflected from the surface or emitted from the nearsurface regions of the sample Due to the much shorter wavelength of electrons compared to visible light SEM offers substantially greater resolution than conventional optical microscopy The technique operates by directing an electron beam that scans the sample in a raster pattern Electrons are generated at the top of the column by an electron source when their thermal energy exceeds the work function of the source material These electrons are then accelerated towards a positively charged anode The entire electron column operates under high vacuum conditions which are essential to prevent interactions between electrons and gas molecules that could lead to beam scattering image distortion and loss of resolution Additionally the vacuum ensures the protection of the electron source from contamination vibrations and noise thereby preserving the quality and accuracy of the imaging process Thermo Fisher Scientific 2021 SEM is extensively applied in failure analysis particularly for examining fracture surfaces and microstructural defects in welds It enables precise visualization of fatigue striations confirming crack propagation under cyclic loading Furthermore SEM has been instrumental in identifying coldlap defects and centerline clusters in thermite welds as critical initiation points for fatigue cracking It also provides detailed imaging of interdendritic microporosity enabling confirmation of pore morphology Figure 10 Scanning Electron Miscroscopy SEM In the Murrobo case SEM analysis was crucial in detecting chevron marks and fatigue striations on the fracture surfaces which confirmed a multistage failure mechanism characterized by an initial fatigue crack growth followed by brittle fracture 532 Ultrasonic Testing UT UT is a nondestructive technique widely used for inservice rail inspection It is highly effective at detecting internal voids such as shrinkage cavities and centreline defects UT can also reveal fatigue cracks before they propagate to visible or surfacebreaking stages Modern phased array UT systems provide realtime 3D imaging of rail weld zones Ross 2004 Figure 11 Rail Flaw Detector Ultrasonic testing system from NDT image source NDT Global The Murrobo rail section had passed ultrasonic inspection shortly prior to failure February 2025 illustrating the limitation of UT in detecting certain earlystage subsurface fatigue cracks This reinforces the importance of complementary inspection techniques and rigorous weld process control 533 Rail Geometry Analysis Measurement of rail vertical profile lateral alignment and crosssectional geometry is performed to detect subtle changes due to thermal buckling or fatigueinduced distortion Track geometry vehicles use laserbased systems or inertial profilers to record rail geometry with submillimetre accuracy Deviation patterns are correlated with known zones of weldinduced weakness as observed in ISalehi 2011 Figure 12 Track Geometry Measurement Systems from Mermec Source Mermec public material 534 Dynamic Load Simulation Advanced finite element models FEM and multiaxial fatigue criteria eg Dang Van criterion are used to simulate how dynamic train loads interact with preexisting weld defects Salehi et al 2011 demonstrated that cyclic dynamic loads cause stress amplification at the weld underhead correlating well with observed crack locations Dynamic simulations also account for track support degradation eg ballast pumping which was shown by Salehi to accelerate fatigue damage and trigger local buckling Dynamic load effects were particularly relevant in the Murrobo case as the affected rail was located on a 507 m radius curve where lateral and cyclic forces would have accelerated fatigue crack propagation along the weld web 6Economic and Safety Implications Rail material failures such as those documented in this report carry significant economic and safety consequences The direct costs of failure include rail and track component replacement service disruption rolling stock damage and emergency response operations The indirect costs including network delays contractual penalties reputational harm and reduced freight competitiveness can be even more substantial The Narwonah derailment Bureau 2020 illustrates this very well approximately 300 meters of rail and track formation were replaced requiring extensive crane operations and prefabricated panel installation Beyond material replacement there were large operational disruptions on the ARTC network including lost train paths and followon delays across the grain supply chain Similar repair interventions following the DubboCoonamble rail break and Sydney Metro incipient weld failure further highlight the costs of unscheduled maintenance and emergency welding in operational corridors From a safety perspective weldrelated rail failures pose critical risks to both heavy haul and passenger networks The Dubbo case showed how a weld break with an 18 mm transverse gap if undetected could have caused a derailment on the next loaded train UGLRL 2024 The Sydney Metro incident highlights the heightened derailment risks in highdensity metro environments where automatic train operations ATO and tight headways leave little margin for rail failure Aluminothermic weld failures also contribute to thermal buckling vulnerability As shown in this report fatigue cracking and weldplane defects reduce a rails critical buckling load Jeong 1997 This phenomenon was directly observed in the Murrobo to Spring Hill failure where fatigue cracking originating from weld porosities significantly compromised the rails integrity prior to final fracture Despite having passed ultrasonic inspection weeks earlier the failure progressed rapidly under normal service loads illustrating the serious operational risk posed by latent weld defects In this case additional costs were incurred not only for emergency track repair and component replacement but also for the Bureau Veritas metallurgical investigation and temporary operational restrictions further illustrating the economic impact of latent weld failures When welds act as weakened points under high compressive thermal stresses lateral buckling events can occur even under otherwise compliant track geometry with serious derailment potential The cumulative safety risk is magnified by the fact that AT weld reliability remains operatordependent ISalehi 2011 and field practices still exhibit significant variability Welds with poor fusion centreline defects or coldlaps become latent failure points that degrade progressively under cyclic loading and seasonal stress cycles Mutton 2004 As modern networks increase axle loads and traffic density the economic and safety stakes of weld quality failures rise sharply However this report also demonstrates that welldocumented strategies to improve weld reliability including transition to Type B welds optimized thermal processing parameters and enhanced inspection regimes Ross 2004 Lawrence et al 2004 offer costeffective paths to reduce both economic losses and safety risks Moreover the integration of automated vision systems as in Sydney Metro and modern UT techniques Ross 2004 can further improve defect detection and prevent derailments In summary weldrelated failures not only incur high direct economic costs but can also trigger catastrophic safety events Proactive process control continuous monitoring and rigorous maintenance strategies are essential investments to mitigate these risks and safeguard both rail network efficiency and public safety 7Future Research Directions Building on the analysis of failure mechanisms and the economic and safety impacts outlined in this report this chapter identifies key areas where future research can further enhance the reliability of rail materials and weld performance While significant progress has been made in understanding the failure mechanisms of aluminothermic AT welds fatigue behaviour and the role of thermal stresses evolving operational demands and emerging technologies open new avenues for targeted research The analysis presented in this report highlights several critical knowledge gaps and opportunities for future research to further enhance rail safety and material reliability The findings from the Murrobo to Spring Hill case further underscore these gaps particularly in improving early detection of weld porosityinduced fatigue cracks developing better process monitoring for flash butt welding to prevent gas entrapment and refining UTbased inspection methods to more reliably detect subsurface fatigue initiation in curved track sections Firstly future research should focus on developing advanced welding process monitoring and control systems Despite known best practices Salehi et al 2011 Chen et al 2006 field weld variability remains high The application of realtime thermal imaging combined with digital process logging could enable early detection of suboptimal weld parameters reducing the occurrence of coldlaps and shrinkage cavities Secondly there is a strong need for enhanced understanding of residual stress evolution during rail service life While initial residual stresses from welding are well documented Mutton 2004 their interaction with cyclic loads track settlement and temperature fluctuations over time is not fully understood Longterm field studies using embedded strain gauges and periodic ultrasonic stress mapping would improve predictive maintenance models and help prevent latestage failures Another promising avenue is the refinement of automated defect detection systems The success of automated visionbased crack detection in the Sydney Metro case Chapter 33 demonstrates the potential of machine learning for early intervention However further research is needed to integrate multi sensor platforms vision UT thermal imaging and to develop robust defect progression models that can better predict when interventions are required Additionally more work is required to evaluate the interaction between weld defects and modern operational profiles As axle loads increase and networks move toward higher speeds and denser traffic Ross 2004traditional fatigue life models may underestimate risk Expanded multiaxial fatigue testing and dynamic simulation studiesincorporating realworld loading scenarios and climate dataare needed to validate and update existing design criteria At the materials level further exploration of alternative weld compositions and postweld treatments offers additional opportunities for enhancing fatigue life Finally future research should explore alternative materials and weld filler compositions for AT welding Selective alloying of the weld metal combined with optimized postweld heat treatments may reduce residual stresses and improve fatigue life Chen 2006 Pilot trials of new alloy formulations in controlled test track environments should be prioritized In conclusion advancing the state of knowledge in process monitoring residual stress modelling automated inspection fatigue modelling under modern loads and weld metallurgy will be key to enhancing rail network safety and reliability Such research will not only reduce economic losses but also provide vital safeguards against future derailments in both heavy haul and passenger rail network 8Conclusion This analysis has unequivocally demonstrated that rail material failures particularly those associated with fatigue cracking in welds exacerbated by seasonal thermal stresses and severe operational dynamics remain a critical and recurring risk to the safety reliability and operational sustainability of modern railway networks Through an approach that combines forensic failure analysis computational modeling thermal simulations dynamic load analysis and realworld case studies this report presents not only a highly detailed technical diagnosis but also practical and objective recommendations for mitigating the risks associated with structural rail failures Furthermore it identifies significant knowledge gaps and outlines strategic directions for future research which are essential for advancing welding practices inspection methods and railway infrastructure maintenance strategies The failure case of the flash butt weld at Murrobo to Spring Hill provides compelling realworld validation of the theoretical failure models discussed confirming that the classical mechanisms of crack nucleation from internal porosity fatigue propagation along the rail web and final brittle transverse fracture remain highly relevant under current operational conditions This event clearly reinforces the urgent need to revisit welding execution standards strictly enforce geometric tolerances mandate postweld grinding and enhance process control and traceability of field welds Additionally the documented economic impacts operational integrity risks and technical challenges associated with weld failures highlight that the adoption of predictive maintenance strategies supported by continuous monitoring automated inspection and advanced fatigue and residual stress modeling is no longer merely a best practice but has become an operational necessity for ensuring railway safety It is therefore evident that effective weld quality management optimization of fabrication and inspection processes coupled with the development of real time monitoring technologies and predictive analysis tools are key elements for building a railway network that is safer more resilient operationally efficient and fully aligned with the growing demands of freight and passenger transportation in the coming decades References AS 2020 AS 1085202020 Retrieved from Standards Australia httpswwwstandardsorgaustandardscataloguestandarddetails designationas1085202020 Bureau A T 2020 Derailment of grain train 8838N Canberra ACT OTSI Chen Y L 2006 Weld defect formation in rail thermite welds Proceedings of the Institution of Mechanical Engineers Part F Journal of Rail and Rapid Transit DCN 2020 May 27 Poor tracks and speed highlighted in grain train derailment Retrieved from The DCN httpswwwthedcncomaunewslaw regulationtradepoortrackandspeedhighlightedingraintrain derailment ISalehi P M 2011 Fatigue and Fracture Behavious of Aluminothermic Rail Welds under High Axle Load conitions Jeong D Y 1997 Damage tolerance analysis of detail fractures in rail Theoretical and Applied Fracture Mechanics 282 109115 Luong T A 2022 Research on the system safety management in urban railway Dissertation Technische Universität Dresden Fakultät Verkehrswissenschaften Friedrich List Professur für Verkehrssicherungstechnik Retrieved from httpscoreacukdownload553289906pdf Mathew Dick P R 2016 Comparison of Predicted and ENSCO Rail Mutton P J 2004 Failure modes in aluminothermic rail welds under high axle load conditions Engineering Failure Analysis 111 151166 RAIL WELD 2022 Characteristics of aluminothermic welding of rails Retrieved from Railweld httpswwwrailweldcomcnnewscharacteristicsof aluminothermicweldingofrailshtml Accessed June 19 2025 Ross E T 2004 A statistical study of improved thermite rail welds MSc thesis University of Illinois at UrbanaChampaign Standards Australia 2019 AS 108512019 Steel rails Retrieved from RISSB httpswwwrissbcomauwpcontentuploads201910AS10851 2019Stee RailsPreviewpdf Thermo Fisher Scientific 2021 What is SEM Scanning Electron Microscopy explained Retrieved from Thermo Fisher httpswwwthermofishercomblogmaterialswhatissemscanning electronmicroscopyexplained UGLRL 2024 Rail Fail From CM224 Dubbo NSW UGLRL VERITAS B 2025 Rail Fracture at 305300km UGL Regional Linx Bureau Veritas Appendix A Broken Rail at Murrobo to Spring Hill line Additional Photos extracted from the Bureau Veritas Report NO2526627104 Figure 1 Photograph of the rail section as received at Bureau Veritas The rail had fractured into three pieces arbitrarily numbered as a result of preexisting and final fracture zones The preexisting crack was centred on a weld bond line indicated by the white lines in the rail Figure 2 Photograph showing the new rail profile of 53 kg rail in agreement with AS 10851 Supplement 12017 black dashed line current rail profile solid black line and rail head wear measurements indicated by red dimension lines Photographs showing various views of piece 1 a front view showing weld bond line and preexisting crack profile b top view showing crack propagation in two directions along the web c and d side view showing crack propagation down into the rail foot Photographs showing the primary crack origin on piece 1 after cleaning with Safedescale solution The fracture surface exhibits chevron marks pointing toward the origin of the crack indicated by a white arrow located on the web The defect was positioned within the weld bond line Photographs a and b showing crack arrest marks at both ends of initially longitudinally propagated preexisting crack on piece 1 prior to the change in crack propagation direction into the transverse plane a Final fracture zone b Final fracture zone c Preexisting crack Figure 6 Photographs showing various views of piece 2 a front view b side view and c bottom view revealed the extent of preexisting fracture and final fracture fast final fracture zone Final fracture zone Preexisting cracking Final fracture zone Figure 7 Photographs of piece 3 showing the preexisting and final fracture surfaces The closeup view exhibits evidence of crack arrest marks from which the preexisting crack changed direction and final stage of cracking initiated leading to fast fracture White arrows on the web indicate the crack propagation direction The image also demonstrates that the preexisting fracture propagated up to the head followed by final fast fracture through the head Weld Figure 8 Schematic diagram illustrating the crack propagation sequence in the failed rail segment Stage 1 represents a brittle fracture initiated from a weldrelated defect origin marked as O followed by Stage 2 fatigue crack propagation under cyclic loading and Stage 3 involving brittle transverse propagation These three stages are considered indicative of preexisting cracking Stage 4 denotes a sudden brittle fracture through the remaining crosssection forming the final fracture zone FFZ which completed the rail separation Figure 9 Macro photograph of the polished and etched rail web crosssection adjacent to the weld interface and crack origin The closeup view highlights how weld porosity arrowed white acted as a notch ie stress raiser that initiated cracking Figure 10 Optical micrograph of the weld zone microstructure showing enlarged pearlite colonies consistent with a large prior austenite grain size Etchant 2 Nital Magnification x50 Figure 11 Optical micrograph of the bond line zone microstructure showing enlarged pearlite colonies surrounded by a network of grain boundary ferrite example indicated by orange arrows Etchant 2 Nital Magnification x100