In Issue 195 (Mar-April 2022) and Issue 196 (May-Jun 2022), Rail Engineer reported on a February 2022 IMechE event which featured several case studies on rolling stock cracks and fractures.
In his keynote address at a similar event in September 2025, Andrew Skinner, head of engineering at Great Western Railway (GWR) said that the day was about: (i) using data to avoid issues from cracks, materials, and fatigue including insights from aerospace; (ii) the importance of the wheel/rail interface; (iii) why standards matter; and (iv) using data to move from risk to condition-based maintenance. It would also include several case studies describing how cracks or wear issues were resolved.
This article will cover the standards, tools, and techniques aspects of the event, while a follow-up article to be published later will cover the case studies.
Aviation comparison
Dr Muhammad Khan, head of the Centre for Life-cycle Engineering and Management at Cranfield University, introduced fracture mechanics theory in the context of aerospace structures and machined components. There are many parallels between aerospace and rail, albeit aerospace has a much greater weight-saving imperative. This has led to very good understanding about material properties, and application of and the nature of loads and their consistency.
Aerospace structures are predominantly designed based on damage tolerance. This is a methodology that requires structures to tolerate damage safely for a predictable period, allowing time for detection during inspections before the damage becomes critical. Fracture mechanics formulations are used to predict remaining life (through crack growth and its rate) and quantify inspection intervals.
This is not easy. Modern airframes use fibreglass, aluminium, steel, titanium, and carbon composites, and the material used in various locations is mainly dictated by the material properties and applications/nature of loads. It is also influenced by the shape.
The situation is rather different for aerospace machined components which are designed based on weight optimisation and either:
- Safe-Life: Components such as wheels/axles, landing gear struts, and engine rotating parts are designed to be used only for a fixed number of cycles or flight hours and must be replaced before that life is reached. No cracks or damage are tolerated.
- Indefinite-Life: Components such as control levers, pins and hinges, small brackets, and fixings are designed so that the stress levels are well below the material’s endurance limit, meaning fatigue failure should not occur if limits are respected.
- That said, there is still a trade-off between the design safety factor and the ever-present aerospace challenge of weight reduction, which means that inspection is still required because hidden defects or corrosion and/or fretting fatigue can lead to failure.
Dr Khan concluded that the principles of fracture mechanics and formulation are same for both aerospace and rail, but while aerospace has a very good idea about the material properties, application of loads, and the nature of loads and their consistency, many sources of load in the rail sector are not fully codified and can be higher than allowed for in standards.
Standards
Neil Dinmore from RSSB outlined the UK standards regime (see panel) emphasising those applicable to the prevention and management of cracks and fractures on the railway. He said that most structural failures originate with fatigue, which can be caused by poor design, manufacture, and maintenance, as well as corrosion, damage, or ignorance.
Standards have been developed to help prevent most of these causes:
- Design: UIC 566 (carbodies), UIC 515 (bogies), British Rail Load Case Documents, CP/DDE/115, – these are now covered in EN 12663 (bodies), EN 13749 (bogies).
- Fatigue: BS 5400-10, BS 8118, BS 7608, EN 1993/1999 – these are now covered in the BS EN 17149 series.
- Manufacture: EN 15085 series (welding), component standards.
- Maintenance: NTSN, RIS-2004-RST.
- Corrosion: CR/PE0102 – BS EN ISO 9466, TN2302, TN2309.
Also, not necessarily covered in standards, all UK railways include training and communication as part of their safety management systems.
Transport for London crack risk management
Matt Brown, Senior Engineering Leader – Mechanical Principles Engineer and LU Asset Performance Mechanical team leader, TfL, outlined TfL’s fracture management process. This relies on identification and then safety assessment. For some of the older fleets there is an information source known as a fracture map which shows all high stress areas and/or areas which have experienced cracks/fractures in the past, providing an appropriate inspection periodicity for these locations.
Newer fleets are supplied having had much more attention to loads/stresses than before. All fleets, however, are subject to emerging issues leading to reliance on identification by maintenance staff during routine inspection. If a new crack is found it is subject to a process called ‘Assurance of Fractured Rolling Stock Components’ which involves collecting knowledge regarding fractures, understanding the initiation and growth of fractures (in general and railway specific), and developing a rationale for continued safe operation. This last factor involves proposing possible mitigations to allow a safe service to operate while determining the root cause and developing a permanent solution.
The output of this process feeds into a Case for Continued Safe Operation (CCSO), which is a document providing a structured argument for safe operation with supporting objective evidence containing:
- A risk assessment based on failure likelihood and consequence leading to a top-level event (e.g., derailment or collision).
- The likelihood is based on data from percentage fleet check and analysis of the impact of any mitigations.
- The consequence is determined through failure mode analysis of the event.
- Actions to be taken to manage the risk, broadly grouped into three stages: (i) Immediate: quantification of issue and immediate safety mitigation; (ii) Medium term: Safety mitigation to further lower risk; and (iii) Long term: Provide permanent solution and reduce risk as low as reasonably practicable.
Once a permanent solution has been implemented that reduces the safety risk to as low as reasonably practicable (ALARP), the CCSO can be closed.
Wheel Rail Interface
Freight wagon defects can and do lead to cracks and fractures both in wagon components and in rails. Mike Briggs, director of data science and AI at RSSB and Adam Bevan, professor of railway systems engineering at University of Huddersfield explored the use of existing lineside Wheel Impact Load Detector (WILD) data to assess network risk and prioritise interventions.
WILD installations were originally provided to detect extremely high impact loads which can lead to rail breaks. However, they collect the wheel load at each individual wheel and can therefore identify issues such as wheel damage, load imbalance, frame twist, damaged or failed primary springs, and other suspension defects. The wheel load data, together with wagon identification from RFID tags, is recorded and monitored over time to identify trends.
It was explained how this data is being used in a freight condition-based maintenance tool – an AI enabled tool to improve freight fleet maintenance. The tool can identify defects from measured wheel loads and had identified defects that would not otherwise have been detected during routine maintenance.
Brian Whitney, engineering expert (track and S&C) at Network Rail explained how the organisation can increasingly use information from measurement trains to provide track defect information accurate to 30mm, allowing work to be planned without first visiting the location, reducing ‘boots on the track’, and achieving efficiencies.
One very promising innovation is the Fault Navigator application. The whole railway has now been photographed and accurately located. It has also been found that a form of machine recognition of ballast layout acts as a unique identifier. The Fault Navigator app on a phone or tablet can be used to accurately locate faults. When looking for a fault, the phone is scanned over the area concerned. These images are compared with the central cloud database, and a virtual pin is dropped into the image when the phone is in the correct location.
Brian reported that the latest news is Network Rail’s vision to procure a comprehensive service that will replace the existing methods of collection, manage operations, and deliver the necessary output data to Network Rail. The intent is to replace the current monitoring fleet that has an average age of over 50 years.
Overview of Aluminium metallurgy
Dr David Howse, Technology Fellow – Arc Welding Engineering at TWI, was unable to attend the event but provided a presentation which explained some of the issues and challenges involved in welding aluminium alloys.
There are some important factors to be born in mind before getting into detail:
- Welds are always potential points of weakness in a structure and will always contain imperfections to some extent.
- The cost of implementing proper controls around welding necessarily involves time and money which can add to project costs.
- The cost of failure in meeting relatively basic requirements for welding can be in the order of hundreds of millions in reclamation and or repair costs.
- Failures often occur around welds but are not necessarily related to poor welding practice.
There are three main root causes of weld failures related to: (i) fabrication, e.g., lack of fusion and porosity; (ii) design, such as incorrectly specified joint types, sizes or materials; and (iii) environment, use or design including unexpected corrosion or higher than expected loading, particularly cyclic loading.
This begs the question: Why design in Aluminium?
David suggested five main reasons, especially in transport: (i) Aluminium is relatively lightweight at one third the density of steel; (ii) it generally has good corrosion resistance; (iii) it is easily formed and has ductile behaviour at relatively low temperatures; (iv) it has good electrical and thermal properties; and (v) its yield strength can be relatively high – up to 450 MPa for some heat treated alloys.
David outlined many of the ways in which welds can fail.
Cracking: Welding relies on heating to a molten state, ‘free’ mixing, and solidification. For any metals, highly alloyed types simply do not solidify in a homogenous manner and form very weak and brittle phases that crack on solidification. These alloy types can be more difficult to weld, with some simply considered unweldable.
Loss of strength: For the weldable alloy types, higher strength may be derived from either heat treatment and/or work hardening. The cycle of welding – i.e., heating to local high temperature and cooling – will effectively remove this effect leaving a lower strength state present at the fusion zone (>60% of original).
Lack of fusion: Aluminium has a high affinity for oxygen and forms a tenacious and more inert oxide film with a high melting point. The melting point for aluminium oxide is approximately 2,060oC, whereas for aluminium it is approximately 660oC.
The oxide layer is a benefit for corrosion resistance but more problematic for welding as this needs to be removed or broken down to allow the molten material to fuse to the underlying ‘clean’ metal and create a sound joint.
It also means that the molten metal needs to be protected from the atmosphere during welding. This is usually achieved by Inert gas welding, i.e. Tungsten Inert Gas (TIG) or Metal Inert Gas (MIG).
However, aluminium’s high thermal conductivity means that the heat needed to melt and weld the parts will be conducted away faster, requiring concentrated heat sources.
Porosity: Aluminium has a very high solubility for hydrogen in the liquid phase but very low solubility in the solid. Hydrogen dissolved in the molten weld metal is therefore expelled from the solidifying weld pool, forming bubbles of gas in the solid weld. Note: Some porosity is always expected, and allowed for, in most application standards, e.g. allowable 1% affected cross sectional area.
Mechanical Weld Failures: There are several causes including:
- Failure by static overload: This is determined by the strength of the weld metal and its size, so the size of the weld is important. Fillet weld strength is determined by its throat thickness. Lack of fusion will affect strength and can be very difficult to detect by surface inspection techniques.
- Cyclic load failure (fatigue): this is a progressive failure mode and earlier than expected failure may be caused by poor fit up during manufacturing, such as incorrect root gaps for fillet welds. The resultant reduced throat size and reduced stiffness can lead to reduced fatigue life. This may reduce stiffness to the point where fatigue propagation occurs unexpectedly. It can also move expected fatigue initiation from the toe of fillet (growing inwards) to root of fillet (growing outwards) and is not easily detectable until the failure is through to front surface. Excessive gaps are not easily inspected for if the weld is made on both sides, so it is best to inspect and record before welding. If the welds are well designed and made, the joint should only fail by fatigue from the toes of the external welds.
Mention of design led to a description of simple elements for welding control:
- Design: Clearly specify weld sizes and design in proper access to make the welds, having determined expected loading and environment.
- Fabrication: Standards apply to the development of the welding method and welder skill/competence. The method requirements are in BS EN ISO 15614 part 1 for Steel, BS EN ISO 15614 part 2 for Aluminium, and for welder skill in BS EN ISO 9606 parts 1 and 2. Completed welds are inspected to verify quality using a number of techniques including visual inspection; surface inspection e.g. Dye Penetrant Inspection; and/or volumetric inspection e.g., X-radiography or ultrasonic testing (UT). The method is determined by the criticality and possibly the complexity of the weld.
Railway welding is covered by the BS EN ISO 15085 series of standards: Railway applications – Welding of railway vehicles and components. This suite is in five parts, covering manufacture, design, production, inspection, testing, and documentation.
Summarising, David said that the design should clearly state the performance class and specify the types and sizes of welds required. The fabricator must assure the quality of the welded product in the specific environment where the component or structure is manufactured. Inspection must be defined by performance class, and welding co-ordination has to be carried out by competent and experienced personnel.
You should always question whether these elements are in place and remember there is no such thing as a perfect weld!
Vibration monitoring
David Vincent, a technology director (digital transformation) at Hitachi Rail, presented on ‘managing fractures by moving from risk to condition based maintenance. His premise was that although industry standards mandate that defects must be repaired when they reach a level of concern, these defects all start small and if we can detect and diagnose them soon enough then we can schedule in earlier repairs, maintaining better condition assets with less overall effort.
Vibration monitoring, as one component of monitoring and predicting changes in condition can deliver the asset knowledge that is necessary to maintain assets to a known good condition. Condition based maintenance is about having improved knowledge of the asset and maintaining it in good condition, using smaller planned interventions without stressing maintenance systems. Standards are frequently directed at limiting a defect size before it develops into a failure. This standard limited approach leads to larger maintenance actions with little option for planning, but many monitoring systems such as visual inspection are only capable of operating at this level.
Using vibration sensors in this way has led to improvements in rolling stock maintenance, as covered most recently in Rail Engineer 212 (Jan-Feb 2025). Can this approach also be used for infrastructure leading to better condition assets at less cost?
Rail Engineer concluded that while standards are key components of train design, it is equally important to understand the environment in which trains will operate, hence the work by Messrs Bevan and Briggs and the monitoring equipment described by David Vincent.
Indeed, during discussion, Neil Dinmore observed that some requirements in standards have been introduced and/or reinforced to require designers to take more account of the environment, especially track quality.
This was a lesson learned from the cracks in anti-roll bar/yaw damper bracket on Class 8XX trains, which we will examine in the second part of this article early next year.
What is a standard?
According to BS EN 45020 (itself a standard), a standard is:
- A document established and approved by a recognised body, that provides, for common and repeated use, rules, guidelines and characteristics for activities or their results, aimed at achieving the optimum degree of order in a given context.
- In other words, an agreed way of doing things which may include requirements and/or recommendations in relation to products, systems, processes or services.
- Standards can also be used to describe a measurement or test method or establish a common terminology within a specific sector.
- Standards can help facilitate trade between countries, create new markets n, enable innovation and cut compliance costs.
One of the earliest GB standards was created as a result of the expansion of railways and introduced the requirement to standardise time throughout the land. One of the first engineering standards was about metal fatigue, based on research by August Wöhler into rail axle fatigue in 1858.
For GB rail, the use of standards is mandated by legislation:
- The Railways (Interoperability) Regulations 2011 (as amended) – RIR
- The Railways and Other Guided Transport Systems (Safety) Regulations 2006 – ROGS
This legislation effectively mandates the standards regime in use on UK main line rail as shown in the diagram below from the RSSB website:
The former mandates use of National Technical Specification Notices (GB versions of Technical Specifications for Interoperability) which in turn call up European and International standards where appropriate. Despite Brexit, the UK continues to be involved in European (CEN/CENELEC) and International standards (ISO/IEC) with volunteers either supporting standard drafting committees or UK committees that feed into the drafting committees – known as mirror groups. Mirror groups report on CEN/CENELEC and ISO/IEC activities, prepare a common GB view for input to ENs and ISOs and review/update British Standards. RSSB provides the secretariat to railway related mirror groups.
Railways not subject to the Interoperability Regulations (Metros, Heritage, Light Rail) are not bound by any particular standards, but the overall legislative requirement of ROGS means that these railways have to develop their own standards, or adopt National, European and/or International standards in their safety management system and as part of managing risks ALARP.
Image credit: iStockphoto.com/MEDITERRANEAN