Impact case study database
- Submitting institution
- The University of Huddersfield
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Technological
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
Research in surface metrology at Huddersfield has impacted on the global manufacturing industry through international standardisation of areal (3D) surface characterisation (ISO 25178 part 2, 3, and 71). These enabled the consistent interpretation of surface data and aided cooperation across industrial supply chains for many manufactured components worldwide. Today, products from all surface metrology instrument manufacturers are compliant with these ISO standards. A leading metrology software vendor, Digital Surf, has incorporated the research findings directly into its products and 80% of surface measurement instruments in industry globally use their software. The resulting interoperability has hugely benefited end-user companies and consumers by enabling instrument sales, speeding up new product and new applications development.
2. Underpinning research
Surface metrology is the science of measuring small-scale geometrical features (roughness) on surfaces - the surface topography. This is important because, for instance, the control of surface texture allows automobile engines to have reduced running-in times, operate more efficiently and emit reduced emissions. Similarly, orthopaedic implants last longer when their surface topography is optimised. Traditional surface measurement techniques were 2D (so-called profile) and relied on moving a stylus across the surface to be characterised. The topography was measured but only along the exact contact line. In the late 1990s, new quantitative 3D (so-called areal) techniques, such as optical interferometers and electron microscopies, were developed. These collected more detail with greater fidelity. However, each instrument manufacturer developed its own protocols to interpret the measurements and therefore results from different instruments were incompatible. This disrupted the adoption of the new areal techniques because the members of a supply chain could not communicate this new generation of surface measurement in a reproducible and consistent manner.
To resolve such chaos in industry, the representative underpinning research conducted by the Huddersfield team was as follows:
A Surface Assessment Framework was created to develop technologies for areal analytics under the EU SurfStand Project, with Liam Blunt as the PI/coordinator and with 3 academic and 7 industry partners. The framework included surface characterisation, instrumentation and a series of case studies in automotive production, sheet steel manufacture and orthopaedic implant manufacture [1].
A Set of “Field” Numerical Parameters were systemically developed based on statistics calculated from the surface, to classify: deviations, extremes and specific features from a scale-limited topography. As a result, a field surface parameter set was formulated, that were stable, reliable and had functional correlations, describing electrical, thermal, sealing and running-in wear properties (Chapter 2 of [1]).
Novel Segmentation methods were created to decompose a surface into stable features using pioneering fundamental mathematics [2], enabling the analysis of surface features and structures (attributes and their relationships), allowing new approaches to the analytics of surface function (lubrication, paintability, illumination and defects, etc.) in Chapter 3 of [1].
First Generation of Areal Filters were originated with the EU Surfstand project; and followed by two EPSRC Filtration Projects with Jane Jiang as PI, including Wavelet filtration allowing multi-functional surface analytics [3] and generic Gaussian filters from linear to non-linear with fast and robust algorithms [4]. These allowed the establishment of robust reference and residual surfaces at any scale (from micro to sub-nanometers) with fidelity for further parametric assessment, originally discussed in Chapters 5 and 9 of [1].
Application Technologies were explored (with refined and optimised algorithms) and applied to a wide range of surfaces, from automotive, bio-engineering to semiconductor sectors. These were tested and validated by the project partners in real industrial surface quantification scenarios. For example, Volvo undertook ‘blind’ engine tests, analysing the emissions performance of a series of engines and found excellent correlation between the emissions measured and those predicted by the parameter values. The findings were summarized in a series of case studies [1], which clearly showed that measuring and characterising surface texture in areal mode provided massive advantages in understanding surface functions such as wear, paintability, formability and bio-tribology.
This underpinning research received worldwide recognition. In January 2002, ISO/TC 213 invited the Huddersfield team to present their initial findings and technologies, leading to a surface texture taskforce being set-up in June 2002 (as recorded in ISO/TC213 *Resolution 483, Madrid, Spain). In 2003 a new Working Group WG16 was formulated to develop areal surface texture standards (ISO 25178 series, recorded in *Resolution 512, St Petersburg, USA), where *Paul Scott led the surface parameters standards and *Jane Jiang led the software measurement standards. The mathematical results of the research [1-4] were hard-coded into software tools and released as a bespoke non-commercial software package, “SURFSTAND”. In 2007 and 2012, two papers [5-6] formalized the paradigm shift (both of scientific and technological aspects) that had taken place in surface metrology from profile to areal measurement. This research also resulted in the training of 8 PhDs and 2 Postdocs (3 females). Note: *Resolution 483 and *Resolution 512 see Page 1 & 2 [E1]; *Paul Scott and *Jane Jiang see Page 3 & 4 [E1]
3. References to the research
The following outputs provide reference to the body of research and are 2* or higher, being published in high quality peer-reviewed journal articles or well cited books.
Blunt, L., & Jiang, X. (2003). Advanced techniques for assessment surface topography: development of a basis for 3D surface texture standards "SURFSTAND". 355 pages, 1st edn, London Penton Press, Elsevier. ISBN 978-1-903996-11-9, 2003. https://www.sciencedirect.com/book/9781903996119/advanced-techniques-for-assessment-surface-topography (Cited 411). [can be supplied on request]
Scott, P. J. (2004). Pattern analysis and metrology: the extraction of stable features from observable measurements. Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, 460(2050), 2845-2864. DOI: https://doi.org/10.1098/rspa.2004.1291 (Cited 87). [can be supplied on request]
Jiang, X., & Blunt, L. (2004). Third generation wavelet for the extraction of morphological features from micro and nano scalar surfaces. Wear, 257(12), 1235-1240. DOI: https://doi.org/10.1016/j.wear.2004.06.006 (Cited 65).
Zeng, W., Jiang, X., & Scott, P. J. (2010). Fast algorithm of the robust Gaussian regression filter for areal surface analysis. Measurement Science and Technology, 21(5), 055108. DOI: https://doi.org/10.1088/0957-0233/21/5/055108 (Cited 47). [can be supplied on request]
Jiang, X., Scott, P. J., Whitehouse, D. J., & Blunt, L. (2007). Paradigm shifts in surface metrology. Part II. The current shift. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 463(2085), 2071-2099. DOI: https://doi.org/10.1098/rspa.2007.1873 (Cited 265).
Jiang, X. J., & Whitehouse, D. J. (2012). Technological shifts in surface metrology. CIRP annals – Manufacturing Technology, 61(2), 815-836. DOI: https://doi.org/10.1016/j.cirp.2012.05.009 (Cited 234).
Note: In [5] and [6], *D.J. Whitehouse Emeritus Professor Warwick University, Visiting Professor University of Huddersfield 2005-present
4. Details of the impact
Published Areal Surface Standards (ISO 25178 - Part 2, Part 3, Part-71)
Three parts of the standards on areal surface characterisation [ISO 25178 series] resulted from the research [1-6]. Two were drafted by Paul Scott and one on areal surface software standards was drafted by Jane Jiang: The Draft International Standards (DIS) were voted by country representatives by 18/18 (Part 2, *Doc N1015, 2008), 18/19 (Part 3, *Doc N1129, 2008) and 17/18 (Part 71, Doc *N1182, 2010) Countries; and the Final Draft International Standards (FDIS) were voted by country representatives by 21/21 (Part 2, *Doc 1368, 2011)), 19 /19 (Part 3, *Doc N1480,2011) and 18/18 (Part 71, *Doc N1539, 2012) Countries [E1]. As a result, ISO 25178-2 Geometrical Product Specification (GPS) – Surface Texture: Areal - Part 2: Terms, Definitions and Surface Texture Parameters (2012); Part 3: Operators (2012); and Part 71: Software measurement Standards (2012) were published by the International Organization for Standardization, Geneva [E2, E3].
The new standards were adopted from 2014 onwards by all OEM manufacturers of ISO compliant surface measurement equipment, influencing metrology across a wide range of manufacturing industries. The Lead Standards Development Manager at BSI wrote: “The work from the Huddersfield team has been fundamental in the wide acceptance of ISO standards worldwide and has facilitated the impact of the GPS standard ethos across wide swathes of manufacturing engineering” [E4]. Up to now, ISO 25178 Part 2 and Part 3 [E2] have had more than 4000 citations (Google Scholar). Based on the standards, interoperability was enabled along engineering supply chains that needed to share consistent surface metrology protocols. This accelerated, and lowered the cost of innovation, and resulted in commercial benefits for (i) businesses that provide measurement technology, (ii) companies that manufacture products that rely on consistency in surface functionality and (iii) consumers.
Note: *Doc N1015, *Doc N1129 and *Doc N1182 see Page 7,12, 17 of [E1]; *Doc N1368, *Doc N1480 and Doc *N1539 see Page 5,10,15 of [E1].
Commercial Impact for Suppliers of Metrology Instrumentation
During the course of the research and subsequent standards development, the Huddersfield team worked with leading suppliers of surface metrology instruments and software developers to ensure the outputs were industrially applicable. The ISO standard has superseded any proprietary software they may have developed historically.
Digital Surf
Digital Surf specialised in manufacturing 3D non-contact laser profilometers. From 2014 it began to use the output of the SURFSTAND project and the ISO standards to develop a surface analysis software package, MountainsMap®, and since then it has incorporated developing ISO standards into it. Today, Digital Surf’s Mountains® software provides visualization, analysis and reporting tools for data output to a wide range of commercial profilometers and microscopes. Since 2016 the company has outsourced the development of many of its updated algorithms to the University of Huddersfield (UoH). Following the success of Digital Surf, many metrology instrument companies (including larger ones such as Bruker, Zeiss, Nikon and Taylor Hobson) no longer develop “in house” areal software but instead use the Mountains® package. It is now embedded in the equipment of the majority of profilometer (40+) and scanning probe / scanning electron microscope manufacturers (10) and it provides software for 80% of the world market in the field. MountainsMap® has an installed base of 12,000+ licenses worldwide and fully supports the ISO 25178 standards. The VP of Digital Surf confirmed, “The embedding of this research into our products, [..], was a key factor for the Company to develop market leadership and achieve an extremely diverse user-base worldwide, in the profilometry market. Digital Surf is doing around 4 million euros of sales per year, 75% of which is on the profilometry market, which is the main market concerned by ISO parameters and filtration” [E5].
Since 2017 the Huddersfield team has supported the development of new software modules for Digital Surf’s next generation software, MountainsMap® 8. Since the release of the new software version in 2019, sales have gone well (licenses sold per annum (>1,200)). Digitalsurf’s mid-range software retailed for approximately €15,000 in 2020 and the company estimated the increase in turnover to be €20M linked to the underpinning research. In 2019 Digital Surf developed a new version of MountainsMap® moving from a grid representation of surface to a triangular mesh (based on book produced at UoH ISBN 9780128218150 and papers, e.g. https://doi.org/10.1016/j.measurement.2017.05.028\). This has opened new markets, such as in the area of x-ray computer tomography (https://doi.org/10.1016/j.precisioneng.2016.12.008\). As a result Huddersfield has expanded its formal research contract with Digital Surf to develop new modules for MountainsMap® (MESHSURF) under the University of Huddersfield and Digitalsurf Agreements: Phase I (signed 20th December 2019) & Phase II (signed 11th November 2020). The Technical Director at Digital Surf stated, “We recognise Huddersfield’s continued world-leading position in surface metrology research and the value that we as a business can gain from maintaining a close and mutually beneficial collaboration” [E5].
Bruker Alicona
Bruker Alicona provides optical industrial measurement technology for the quality assurance of complex components used in the automotive industry. Its core competence is the measurement of dimension and surface topography; this is based on focus variation. Since 2008 the company has developed its software to incorporate algorithms extracted from the ISO standards into its products. The Head of R&D said “around 15% of our customers use surface roughness with their instruments…a core element of our analysis capability within our software makes use of the research carried out by the Huddersfield team on areal roughness parameters” [E6].
Commercial Impact for Industry End-Users
The end users of the research output cover a wide range of industries. Examples are given below:
Volvo: used the research in its development of IC engines with reduced emissions. The Company Surface Engineering Specialist for Cylinder Liners stated, “in this area, the use of areal roughness parameters has been a core tool in understanding engine performance, engine wear and consequent emissions [..] we have now written their application into our company standards” [E7].
Sarclad: is the world’s leading producer of Electrical Discharge Roll Texturing (EDT) machines, which are used in the automotive industry to make sheet metal easier to form and paint. The Head of R&D wrote, “Textured sheet metal using EDT is specified by all major automotive OEMs as a means of improving formability of sheet [..] the continued success of our Rolltex product range is helped significantly by the ISO 25178 areal surface measurement characterisation and the underpinning work of the Huddersfield team” [E8].
DePuy Synthes: is a world leading bio-implant manufacturer. Surface roughness is a key determinant of how successfully a surgical implant (e.g. replacement hip joint) bonds/integrates with the patient’s bone and also the life of the joint bearing surfaces. The DePuy Testing Laboratory Manager commented, “Areal surface roughness parameters developed at [UoH] are of great importance to our industry and the development of new orthopaedic implants. [..]. The research conducted at the University of Huddersfield has helped shape ISO 25178 that is widely used in our industry” [E9].
Digital Metrology: has provided independent consultancy on surface metrology for over 30 years, mainly in the USA. Its President confirmed, “Without the underpinning work on parameters at Huddersfield, the field of areal surface metrology would have been significantly delayed and would certainly have been disparate and where adopted, would have been utilized within individual silos minimizing its impact and usefulness in innovation and international trade” [E10].
5. Sources to corroborate the impact
ISO/TC 213 Resolutions regarding outcome of Huddersfield’s underpinning research; set-up ISO/TC 213 working Group and different stage voting results for ISO 25178 Part 2, Part 3 and Part 71.
ISO 25178 Geometrical Product Specification (GPS) - Surface Texture: Areal - Part 2: Terms, Definitions and Surface Texture Parameters (Page 1-54), and Part 3: Operators, International Organization for Standardization (Page 55-80), Geneva (2012)
ISO 25178 Geometrical Product Specification (GPS) - Surface Texture: Areal - Part 71: Software measurement Standards, International Organization for Standardization, Geneva (2012)
Letter from Lead Standards Development Manager, BSI, 25th March 2020.
Letter from VP, Digitalsurf (Software Vendor), 17th April 2020
Letter from Head of R&D, Alicona (Instrument Manufacturer), 22nd April 2020
Volvo - Value of UoH-based ISO (End user - Automotive), 25th October 2020
Sarclad - Value of UoH-based ISO (End user - Steel), 2nd November, 2020
DePuy - Value of UoH-based ISO (End user - Bio Implants), 1st November 2020
Michigan Metrology (USA Metrology Consultant and Stds Committee),16th December 2020
- Submitting institution
- The University of Huddersfield
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Technological
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
Machine tools are used widely in manufacturing across many sectors (aerospace, automotive, automation, green energy, consumer goods, etc.). Poor machine accuracy leads to waste in production time and energy, and scrapped components.
Research by the University of Huddersfield created new ways to measure and interpret errors, thus enabling the elimination of predictable failures. The research led to the development of new measurement products and the opening of new markets for machine tools. The processes developed were embedded in blue-chip companies, such as aerospace manufacturers. Two SMEs commercialized elements of the research, which are now sold globally. Further benefits, to second and third-tier users, are estimated to exceed £90m.
2. Underpinning research
CNC precision machine tools are used for cutting materials such as metal or carbon fibre for use in precision-engineering industries, such as automotive and aerospace, and consumer goods such as mobile phones. The components are manufactured to extremely high tolerances. This degree of precision can extend the life of a component (by reducing friction), increase the fuel efficiency of an engine or make components simpler to assemble.
A key challenge was maintaining the accuracy and repeatability of the machine tools that produced components to micron-level accuracy. It should be possible to manufacture a new piece within specification simply by using the nominal values indicated on the CNC control panel, but systematic errors often meant the first piece produced was out of tolerance and had to be discarded. This “sacrificial part” informed adjustments that were then made to the machine program. As a result, manufacture became relatively cheaper if many copies of a component were made (assuming that the machine tool maintained its accuracy in use), but “batch-size-one” was very costly.
Best practice was to calibrate regularly, by checking the alignment and accuracy of the machine using a laser interferometer and adjusting to correct any errors. This was a complex and time-consuming process that could take a machine out of production for several days. As a result, very few machines were regularly calibrated, with businesses having to correct when things went wrong.
The traditional calibration technique used a laser or other device to check the errors of each of the machine axes individually (at least 21 errors on a 3-axis milling machine, rising to 36 on a 5-axis machine). This was a long process because the machine had to stop at several points along the axis for each error measurement, which usually meant that the bare minimum (often only three of the 21 errors) were checked. Therefore, machine-tool users rarely knew the true accuracy of their machines. Further variability was introduced due to thermal effects during machining. These caused expansion and bending that are not normally detected during calibration (where the machine is not under thermal load).
As a result, machines either produced components outside specification, or productive time was lost while performing remedial calibration. “Work-around” solutions that helped maintain quality (e.g. manual offsets and “warm-up” cycles), reduced productivity and only delayed the onset of quality issues. Even machine-tool suppliers only specified performance tolerances using a small number of calibration points. Hence manufacturers, using a machine to make a product for their customer, could purchase one that did not perform at the level of precision required, or lost sales because they could not guarantee the precision needed by their customers. This case study describes how these challenges were tackled.
The research was carried out at the University of Huddersfield (UoH) by Prof. Andrew Longstaff (Professor of Machine Tool Technology, at UoH since 2001) and Dr Simon Fletcher (Principal Enterprise Fellow, at UoH since 2000). There were four key industrial partners: Rolls-Royce (aerospace engine manufacturer), Renishaw (manufacturer of machine-tool calibration equipment and software), MTT (provider of services to the users of machine tools) and Dapatech (which commercially exploits new machine-tool technologies).
Research (with MTT) to characterize the accuracy and maintenance process, showed how the calibration and control problem can be broken down using a lean manufacturing approach [R1] (2010). It established accuracy as a key performance indicator in the regular machine maintenance schedule. New simulations of the effects of errors [R2] (2014) were created to determine their impact on the final accuracy of a machine tool when producing a part. It was clearly demonstrated that existing practices omitted too many factors and were not fit for purpose. The limiting factors were shown to be the complexity of measurement (it was perceived as too specialist) and the machine downtime required to gather sufficiently comprehensive data. The engineers who performed the calibrations chose the order in which the measurements were taken almost at random – convenience was the overwhelming factor. Later research [R3] (2015) showed that the order in which measurements are taken affects the uncertainty of the calibration. A software solution was developed using artificial-intelligence (AI) planning to balance quality with the time needed for calibration.
Using these findings the UoH team worked with industrial partners to develop new hardware and processes to enable comprehensive data to be captured more rapidly. UoH was a key contributor to Rolls-Royce’s SAMULET (Strategic Affordable Manufacturing in the UK with Leading Environmental Technology) project (2011–2013). Rapid calibration strategies using new instrumentation for machine tools were first developed and tested in laboratory settings. They dramatically reduced the average calibration time (from several days to a few hours). Rigorous validation showed that the errors of each linear axis could be measured (i) more rapidly, (ii) while the machine was in motion and (iii) with improved accuracy [R4] (2016).
Next the team investigated the approach in several different industrial settings. Specialist knowledge transfer then enabled the industry partners to implement it independently. This demonstrated that it could be applied successfully across multiple types of machine tool, across many sites with differing ways of working. The team then focused on minimizing the machine downtime. They designed new hardware and refined the algorithms. This enabled engineers to balance the level of precision of the calibration (number of errors calibrated and level of accuracy obtained) against the downtime (cost).
The team then created a comprehensive cost model. It demonstrated the balance between measurement cost and the consequential cost of non-conforming machines and let users calculate the financial benefit of adopting the techniques. Data collected from the industrial partners provided the knowledge base for a new AI decision support system [R5] (2018). This characterized the impact of machine accuracy on productivity, to determine where regular calibration has a positive impact and where “run to failure” is more economical.
Temperature effects were a limiting factor. Even a well-calibrated machine is susceptible to variation from machine warm-up, coolant, draughts, etc. The new calibration methods allowed rapid measurement of thermal errors, but the time to create models remained long. AI was used to optimize sensor location and model the effect of thermal changes in the machine on the accuracy of the manufactured parts. Using an ANFIS (adaptive neuro-fuzzy inference systems) technique, multiple error sources were simulated (including thermal effects). This speeded up prediction of the final accuracy of manufactured components [R6] (2015). Manufacturers do not often provide a detailed model of thermal behaviour of their machines. The UoH-developed approach is a reliable method of predicting thermal effects within a reasonable time. The team then worked with DAPATECH to develop the intelligent thermal error compensation (iTEC) system to reduce the errors.
3. References to the research
These references are considered to be 2* or higher. [1] is from the MATADOR conference (double peer-reviewed, running since 1959). [2–6] are from prestigious journals, with [3, 4, 6] Q1 and [5] Q2 in this field (source: Scimago).
Willoughby, P., Verma, M., Longstaff, Andrew P. and Fletcher, Simon (2010) “A Holistic Approach to Quantifying and Controlling the Accuracy, Performance and Availability of Machine Tools” Proceedings of the 36th International MATADOR Conference. Springer, London, UK, pp. 313–316. ISBN 978-1-84996-431-9 https://doi.org/10.1007/978-1-84996-432-6_71 [can be supplied on request]
Longstaff, A., Fletcher, S., Parkinson, S. and Myers, A. (2014) “The Role of Measurement and Modelling of Machine Tools in Improving Product Quality” International Journal of Metrology and Quality Engineering, 4 (03), pp. 177–184. ISSN 2107-6839 https://doi.org/10.1051/ijmqe/2013054
Parkinson, S. and Longstaff, A. (2015) “Multi-objective optimization of machine tool error mapping using automated planning” Expert Systems With Applications, 42 (6), pp. 3005–3015. ISSN 0957-4174 https://doi.org/10.1016/j.eswa.2014.11.066
Miller, J.E., Longstaff, A.P., Parkinson, S. and Fletcher, S. (2016) “Improved machine tool linear axis calibration through continuous motion data capture” Precision Engineering, vol. 47, pp. 249–260 https://doi.org/10.1016/j.precisioneng.2016.08.010
Shagluf, A., Parkinson, S., Longstaff, A. & Fletcher, S. (2018) “Adaptive Decision Support for Suggesting a Machine Tool Maintenance Strategy: From Reactive to Preventative” Journal of Quality in Maintenance Engineering. 24, 3, p. 376–399 https://doi.org/10.1108/JQME-02-2017-0008
Abdulshahed, A.M., Longstaff, A.P., Fletcher, S. and Myers, A. (2015) “Thermal error modelling of machine tools based on ANFIS with fuzzy c-means clustering using a thermal imaging camera” Applied Mathematical Modelling, 39(7):1837–1852. https://doi.org/10.1016/j.apm.2014.10.016
4. Details of the impact
The research findings benefited businesses and users across the machine-tool supply chain. They enabled the introduction of new products and services and changed professional practice through better and cheaper calibration methodologies. They delivered over £90m of benefit to industry.
The impacts can be summarized under three headings:
Influence on new product development
New revenue streams for providers of services to machine-tool users
Changing professional practice for users of machine tools
Influence on New Product Development
Renishaw is a manufacturer of products for the calibration and regular checking of machine tools, enabling users to establish and maintain their performance. The research [R1] proved that a laser system, developed as a Quality Assurance device for internal use, would work as part of a new solution for the calibration of machine tools. An understanding of the benefits of rapid calibration [R 2,3] led the company to commercialize it as the XM60 multi-axis calibrator.
UoH research strongly influenced the development of the product. The Business Manager in the Laser and Calibration Products Division wrote, “You [..] show[ed] that a three-axis machine tool can have all its linear axes calibrated within one hour, highlighting a competitive edge of our new product. You also devised scientific methods [..] in order to help our own exploitation plan”. The product was launched in 2016 and in 2020 he confirmed that: “Global sales of the XM60 now make up a significant proportion of the revenue generated from calibration product range” [a].
Dapatech Systems was established in 2011 to commercialize new technologies for machine-tool accuracy, compensation and monitoring. It used the research findings to develop a low-cost, mass-producible thermal compensation product (the iTEC). UoH researchers identified where to locate the temperature sensors and how to connect them to the machine tools [R6]. By using iTEC to remove the need for machine warm-up cycles (that can “cost” 30 mins to two hours of productive machining time per day), customers could save significant costs. The MD of Dapatech wrote, “For customers with many hundreds (or even thousands) of machines, using iTEC in this way will save tens of thousands of machining hours each year” [c] (2020). The company has demonstrated the product to customers in the UK, North America, South America and South East Asia and has generated interest from large volume manufacturers of consumer electronics and for high-precision applications in the aerospace sector. The MD added, “This opens up a huge market opportunity for the company” [c]. Although the Covid-19 pandemic slowed progress, he reported, “In 2019 we had negotiated an initial order with one partner for the first 1,000 systems, with the potential for this to rapidly grow to orders of tens of thousands of systems over a 2-3 year period [..] Currently we are negotiating a contract for the delivery of 1500 iTEC Systems to an OEM” [c].
New Revenue Streams for Providers of Services to Machine-Tool Users
MTT (Machine Tool Technologies) is a high-end provider of calibration and other services to users of precision machine tools. It worked closely with the UoH team to develop the early research [R1] into a high-end maintenance offering. MTT used the measurement toolkit created during the SAMULET project, including the Renishaw XM60 and the error modelling methods [R2, 4, 6], to provide rapid calibration services (2014–20). The MD of MTT stated, “[..] in the period 2014–2020 our relationship has facilitated an additional £7m of sales for MTT, as well as creating four directly related jobs”. MTT created training courses in collaboration with the researchers. These were delivered to 12 of their engineers and 30 of their customers. The MD added, “This [..] led to these “intelligent customers” improving their processes, which generated revenue [..] of another £1.5m” [b] (2020).
Accudyne Europe Ltd, in conjunction with their USA based parent company, Accudyne
Systems, Inc.is a manufacturer of Automated Fibre Placement (AFP) heads that are used to lay composite materials, such as carbon fibre. The head must be mounted on a carrier machine which, for challenging applications where materials are laid with extreme curvature, had to be extremely rigid, and was consequently expensive to buy and run. Working with MTT (2016) on instruments and processes developed using the research [R1,2], the company was able to move to a lightweight and cost-effective carrier. Crucially, the new measurement methods let Accudyne prove to potential customers that their new machine was fit for purpose. The methods also provided a reliable solution for regular machine maintenance with reduced downtime. This led to “a new product line, which has already realized turnover of £7.5M and cemented [Accudyne’s] competitive foothold in a strategically important customer base” [d]. It estimated that customers are saving £250k on every machine by using their more cost-effective solution and £50k pa from additional uptime [d] (2020).
Changing Professional Practice for Users of Machine-Tools
The SAMULET program targeted the development of a high-speed machine-tool calibration and verification system. Rolls-Royce (RR) commissioned UoH to define requirements in terms of measurement accuracy, data quality, protocol optimization and evaluation of uncertainty. This led to an anticipated downtime reduction from 2.5 days to a single eight-hour shift, is estimated to save £150k per annum at one site [e], and verifiably improved quality. One RR site adopted the method on 100 machines and the Process Owner at RR confirmed that related “[..] operations implemented to date, demonstrated 100% conforming parts. Thus we have had zero rework and salvages" [e] (2020).
The impact has extended to users of the Renishaw XM60 product, who have been able to reduce the costs of their calibration activities. Renishaw’s Business Manager stated, “Users [..] can also perform full 3D characterisation of the linear axes in one fifth of the time previously required. This significantly reduces the machine downtime and increases the capability of maintenance programs. Combined, this allows the equipment to pay for itself in under 12 months when compared with traditional systems” [a] (2020). This is in line with UoH research [R5].
MTT estimated that the improved calibration services they now deliver have provided significant benefits to their customers, including lower machine-tool downtime and more products manufactured within specification. MTT’s MD wrote that its customers have enjoyed benefits amounting to “c. £70m in the period 2014-2020” in sectors such as aerospace, Formula 1 racing and manufacturing for the MoD [b] (2020).
Since 2013, the UoH team have shared the nature of the research findings and best practice for their application with companies across precision engineering. MTT, Rolls Royce, Dapatech and the Manufacturing Technologies Association (MTA) all cite the knowledge transfer as a key enabler of improvements in professional practice.
The CEO of the MTA commented that the “upskilling employees and improving the quality of the high precision end of machine tools” that has been made possible by the UoH research has been vital to the development of the UK precision-engineering sector and his own organization. He added: “Your work also led to advanced training courses, which two members of our technical team undertook in early 2019. This was invaluable in upskilling our employees, who have as a result been able to offer a more in-depth [..] service to our members” [f] (2020).
The MTA noted that the impact of the research has led to the UoH team being instrumental in the UK’s contribution to international standards in machine-tool accuracy. In its 2018 report on the impact of manufacturing, authored by Oxford Economics, it estimated that GVA per job is £65,000 in the engineering subsector, which is 15% higher than the UK average. It added, “Your research has directly led to more efficient working in this high-value sector” [f].
5. Sources to corroborate the impact
Testimonial from Business Manager (Laser & Calibration Products Manager) Renishaw Plc.
Testimonial from Managing Director, MTT Ltd.
Testimonial from Managing Director, Dapatech SARL.
Testimonial from Technical Director, Accudyne Europe Ltd.
Testimonial from Process Owner for Rolls-Royce.
Testimonial from CEO of the Manufacturing Technologies Association (MTA).
- Submitting institution
- The University of Huddersfield
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Technological
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
Safety is a major concern on railway networks. Traditionally operators had more data available to them than they were able to analyse and interpret, and this led to risks being left unattended, putting passengers and railway staff potentially at risk of injuries and fatality.
The impact of research by the University of Huddersfield (UoH) has been a significant improvement in the effectiveness of railway safety management systems, allowing new safety insights to be extracted from unstructured data, new visualization techniques to be adopted and new tools to be developed that provide rapid analysis of operational data to identify changes in safety risk.
Railway undertakings in the UK were able to improve their safety management (including Network Rail and LNER), benefitting passengers and policy makers (Rail Safety and Standards Board in the UK and the European Railway Agency), plus rail operators in mainland Europe (e.g. RENFE in Spain).
2. Underpinning research
Railways rely on safety management systems to capture data (both structured and unstructured), which is then used by safety experts to better understand and model railway system safety risk. This understanding enables the development of strategies to improve safety, thus reducing accidents, injuries and fatalities.
In 2012 the Rail Safety and Standards Board (RSSB) identified that the volume of data produced by train operators meant that it could not all be analysed using “traditional” techniques and this led to missed opportunities to improve safety. The RSSB Rail Technology Strategy (2012), required the industry to extract more value from the data it had available. University of Huddersfield (UoH) research identified how to apply emerging digital data analysis techniques, to move industry practices from risk-based strategies to ones based on prediction.
The research was carried out in the Institute of Railway Research (IRR), based at UoH. It was performed by Dr Coen Van Gulijk (Professor of Railway Safety and Risk at UoH since 2014), Peter Hughes (Principal Researcher since 2014), Miguel Figueres and Rawia El Rashidy (Research Assistants since 2014) and Julian Stow (Director IRR at UoH since 2012).
The research focused on three key areas:
Text Analysis: As part of a strategy to identify “hidden” risks, railway employees across the UK produce over 300,000 near-miss reports each year. A near-miss report is a free-text report containing details of something that could have caused an accident. Processing this volume of text-based records by hand required a large team of individuals to extract the information that led to safety lessons.
Research by Figueres [R1] and Hughes [R2,3] (2018) showed different approaches of automated text analysis, referred to as Natural Language Processing (NLP), could be a used for the analysis of safety reports. Standard NLP techniques do not perform well with railway jargon or poor spelling (both commonplace in safety reports). A novel NLP approach was developed that dramatically improved the outputs from NLP on this unstructured data. The solution utilized naïve ontology word nets to manually create “clouds” of safety-related words. These were used to analyse the near-miss reports and the word clouds were refined so that individual risk types could be extracted. The team used a state-of-the-art Graph database, which was infinitely scalable and able to process thousands of documents in a matter of seconds. The database also had a graphical user interface that displayed the results of the analyses as bubble diagrams, which were easy for non-experts to interpret. The move from manual to automated analysis reduced the processing time from days, or even weeks, to hours.
Train and Signal Network Data: A train that fails to react to a red signal is known as a Red Signage Passage (also Signal Passed at Danger, or SPAD) and is the most dangerous safety incident on the railways – when a high-speed train SPADs it can collide with other trains, potentially killing hundreds of people. Risk modelling for SPADs traditionally used statistical methods to analyse exposure to SPAD risk, but this is, by definition, imprecise.
The research [R4] (2016) explored how exposure to SPAD risk could be identified from data already collected by train recorders and other infrastructure. One source of useful data was Network Rail’s (NR) Train Describer data-feed, gathered at signals, to detect trains that came dangerously close to a SPAD. An algorithm was developed to use the data to report when a train failed (or nearly failed) to stop at a red signal. The research findings showed, for the first time, the prevalence of dangerous approaches to red signals across the UK network at the individual signal level. The detection system covered the whole of the digitally enabled rail network, amounting to about 60% of NR’s tracks. A practical, safety-related application was developed for RSSB, and statistical guesswork on the exposure to SPAD risk was much reduced.
In 2017 the research was extended by combining on-train data with the signal data feed [R5] in a transparent Graph database. In 2019, new data from the West Coast mainline was added, in collaboration with Virgin Trains. The data was manually extracted from the on-board recorder at the end of a locomotive’s shift and taken to Huddersfield to be extracted, modified and interpreted. This extra locomotive-specific data meant it was possible to take the true speed of travel into account. Since the braking behaviour of a train differs depending on whether the track is dry or wet (due to rain or leaves), slippery tracks could be detected and near-misses could be identified [R6].
Integration of Data into a Visualization Tool: Risk management is a key task for railway operators. Individual companies employ dedicated teams to deal with risk monitoring, incident investigation, assessing protection measures and reporting. The BowTie was a risk management tool used by the Chemical Industry to visualize risk by creating a graphical representation of the “risk space”. At the start of this study the railway industry was beginning to adopt this as a manual tool, with the BowTie drawn on paper by the analyst.
The transparent database approach [R5] (2014) made it possible to integrate the different railway-specific data-sources, e.g. text analysis and the train and signal network data, into commercial software by using an interface built by the UoH team. This enabled a BowTie risk model to be automatically created and to be updated as new data arrived.
3. References to the research
The papers below introduce data techniques for non-engineering safety solutions in the safety science domain. They are among the very first ones to present practical techniques for Natural Language Programming, and data techniques for risk-scenario detection. The work was published in leading international journals for safety in general and safety in the railways.
Figueres-Esteban M, Hughes P and Van Gulijk C (2016) Visual analytics for text-based railway incident reports. Safety Science 89: 72–76. https://doi.org/10.1016/j.ssci.2016.05.009
Hughes P, Figueres-Esteban M and Van Gulijk C (2017) From free-text to structured safety management: Introduction of a semi-automated classification method of railway hazard reports to elements on a bow-tie diagram. Safety Science 110: 11–19. https://doi.org/10.1016/j.ssci.2018.03.011
Hughes P, Robinson R, Figueres-Esteban M and Van Gulijk C (2019) Extracting safety information from multi-lingual accident reports using an ontology-based approach. Safety Science 118: 228–297. https://doi.org/10.1016/j.ssci.2019.05.029
Zhao Y, Stow J and Harrison C (2016) Estimating the frequency of trains approaching red signals: a case study for improving the understanding of SPAD risk, IET Intelligent infrastructures 10-9: 579–586. https://doi.org/10.1049/iet-its.2015.0052
El Rashidy R, Hughes P, Figueres-Esteban M, Harrison C and Van Gulijk C (2017) A big data modeling approach with graph databases for SPAD risk. Safety Science 110: 75–79. https://doi.org/10.1016/j.ssci.2017.11.019
El Rashidy R, Hughes P & Van Gulijk, C (2019) Detection of high-speed red-aspect approaches using multi-data approach. Safety Science 120: 583–588. https://doi.org/10.1016/j.ssci.2019.07.027
4. Details of the impact
The research led to impact on rail operators and their customers in the UK and across Europe, by improving safety (and punctuality) across the networks. European railway policy was also influenced.
The impacts can be summarized under four headings:
Reduction in missed red signals
Extracting safety lessons from close calls
Preparing for digital risk management
Policy influence
Reduction in Missed Red Signals
The research on modelling the prevalence of SPADs [R4,5] was game-changing. For the first time, it was demonstrated that it is possible to replace statistical models, with measurements of actual train behaviour across the UK network. The Rail Safety and Standards Board (RSSB) used the Institute of Railway Research (UoH-IRR) algorithms to develop its own tool to exploit the data. The Risk Aspect Approaches to Signals (RAATS) tool was published through a web service in 2019 [E6], thus making it possible for anyone working on the railways to benefit from superior insights into “a significant safety risk to the railway” [E6]. The web service has had 200 registered users since it was launched, which is the equivalent of about three registered users from each of the over 60 UK train operating companies.
Railway operators used the tool in two ways. Planners used it to identify pinch points on the network (where a high volume of trains approach a signal when it is red) and then planned these situations out of the timetable. This greatly reduced SPAD risk and improved the overall flow of traffic on the network, thus increasing punctuality [E7]. The punctuality improvement is significant because each “delay minute” costs a rail operator up to £3,000. Safety practitioners used the tool to analyse data collected on the frequency of SPADs, to understand the underlying probability of a potentially dangerous red signal approach and whether this was a contributory factor to the SPAD event.
The addition of new, direct from trains, data to the model [R6] isolated two types of red aspect approaches that are particularly dangerous (i) green-light gambling (assuming that the signal will clear in time) and (ii) slippery conditions. Such data is now used by risk analysts at RSSB, Network Rail and other GB railway operators. RSSB stated that “without the research from the IRR this advancement would not have been possible”’ and that “the benefits in reduction in safety risk, both now and in the future, [..] would not have been realised” [E7] (2020).
Extracting Safety Lessons from Close Calls
The structured investigation [R1] and automated classification of close-call data using Natural Language Processing (NLP) [R2,3] has been used in non-standard risk assessments by RSSB. It reported that the data helped to identify risks that were undervalued by the traditional “triage” process. For example, the RSSB’s Professional Head of Safety and Intelligence highlighted its role in identifying, for the first time, risks in relation to vegetation management on slopes [E7]. The method was also used for monitoring purposes [R2] to classify how safety risks increased or decreased over time. RSSB has used the tool to analyse over a million records, stating it “enables rapid interrogation of the data” and “delivers structured results that can more readily support robust and actionable changes to safety practices” [E7]. Similar work was continued by RSSB, as demonstrated in its 2019 report T1152 [E8].
The chair of the steering committee for the strategic partnership that oversaw the research, worked for Network Rail (NR), and encouraged NLP development because of the improvements in speed and accuracy of risk identification that the techniques delivered. The IRR developed a bespoke text-analysis tool for NR, called C-CAT. It used text analysis to classify 300,000 close-call records against 26 risk categories. NR had a team of 30 full-time employees analysing the reports and by removing the need for humans to read every single report, the analysis time was greatly reduced. Importantly C-CAT also helped re-classify the 30% of entries that were unintentionally reported in the “general” category [E9], hence automating the correction of 300,000 entries (2018).
In Spain, the national railway operator, RENFE, reported, “we are investing [..] EUR 0.25m in the development of a ‘close call’ reporting system that utilises the ontology and data processing techniques based on the IRR research” [E5].
Preparing for digital risk management using the BowTie tool
The research findings on the BowTie tool [R5] demonstrated ways to populate commercial risk BowTies with consistent railway-specific data, gathered from disparate automated sources.
LNER worked with IRR to upgrade their safety management system. They are now recognized as having the leading BowTie experts on the GB railways and this is evidenced by their bow-tie manual [E10], which is used across the sector. LNER’s use of advanced digitally enabled BowTie software for its safety management system has provided it with unprecedented oversight of operational risks.
The French railway company SNCF started a digital programme in France in 2016. Adding their own digital developments to the IRR foundation, they initiated a national safety programme called “Prisme”. The Head of SNCF Railway System Safety stated, “The research from IRR and the development of the bowtie tool allow SNCF to better identify where to invest funds to improve, or maintain, safety of operations”. He added, “Without the research from IRR, this improvement would not have happened!” [E4].
In 2019, RENFE in Spain started building a new safety system based on BowTies. Their Head of Safety said, “IRR research on the use of bowtie analyses in safety management systems has been adopted by RENFE and we are now working [..] on a EUR 1M two-year project to connect our operational processes and procedures with safety controls defined within a bowtie analysis” [E5].
IRR researchers spoke at a number of industry-specific BowTie events, which led to an increase in sales [E10] for the organizations that licensed the technology, such as LNER, RSSB and CGE-Risk (a provider of BowTie software).
Policy Influence
The European Railway Agency (ERA) participated at the BDRA (Big Data Risk Analysis) symposia in Huddersfield (2016 and 2017) [E1] and invited the University of Huddersfield to be the only academic partner in their four workshops on the digitization of railways (2016–18). ERA acknowledged the input from UoH workshops on their policy, in their reports [E2, E3].
5. Sources to corroborate the impact
[E1]: Editorial in SaRS journal that printed the proceedings of the 1st industrial BDRA conference tandfonline.com/doi/full/10.1080/09617353.2016.1252082
[E2]: Slide by ERA identifying the inputs for the CSM-ASLP project reviewing the European incident registration system for ERA.
[E3]: Meeting minutes for 3rd industry workshop on DSD demonstrating industry engagement and impact into the ERA.
[E4]: Testimonial by SNCF supporting our claim for the impact of our work on SNCF (France). https://www.era.europa.eu/sites/default/files/events-news/docs/sncf_frederic_delorme_en.pdf
[E5]: Testimonial by RENFE supporting our claim for the impact of our work on RENFE (Spain).
[E6]: the online RAATS tool, accessible through: https://www.rssb.co.uk/en/safety-and-health/improving-safety-health-and-wellbeing/Rail-Risk-Toolkit/Red-Aspect-Approaches-to-Signals-Toolkit
[E7]: Testimonial by RSSB supporting our claim for improved management of railway safety risk in the UK.
**[E8]:**The cover and executive summary of T1152 report demonstrating that RSSB continued research that was originally designed in this work
[E9]: Delivery note and manual for C-Cat delivered to Network Rail.
[E10]: LNER BowTie risk assessment method.
- Submitting institution
- The University of Huddersfield
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Technological
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
3D sound recording systems have often operated sub-optimally because no theoretical model of the relationship between the set-up of the microphones (the array) and the way that a listener perceived the resulting playback existed.
Research from the University of Huddersfield created such a model and established novel principles that enabled microphone arrays to have a more compact form, while improving the sense of realism and spatial impression for the listener.
The knowledge was commercially exploited as the design basis for an award-winning microphone array by Schoeps, a world-renowned microphone manufacturer based in Germany. The new array was adopted by sound engineers for major broadcast events such as the FIFA World Cup, BBC Proms and the French Open. Other microphone arrays and an array design app that was based on the research became essential tools for 3D sound recording at Abbey Road Studios (UK), Austrian Broadcasting Corporation (Austria), Arizona Public Broadcasting Services (USA) and Tianjin Juilliard School for musicians (China).
2. Underpinning research
Since 2011, use of three-dimensional (3D) and surround sound (such as Dolby Atmos and Auro-3D), has grown rapidly in cinemas, music and broadcasting. The techniques use additional loudspeakers (or they manipulate the stereo outputs) to give the illusion that the listener is ‘within’ the soundscape. Changes in practice for both recording and playback were developed by sound engineers. To record 3D sound they positioned groups of microphones (arrays) that captured the directional information on the sound in a way that meant it could be manipulated to give the listener a realistic experience. On playback additional speakers in the height dimension were employed to give a ‘vertical’ element to the sound. Sound engineering has always been a ‘black art’, with best practice resulting from experience mixed with trial and error. The relationship between how the microphones were positioned and the resultant aural experience for a listener was not scientifically understood. The research described in this case study explored the relationship between the physics of sound recording and reproduction and how humans perceive what they hear – a branch of psychophysics called psychoacoustics. The findings identified principles and algorithms that could be used to configure a microphone array to give consistent and reproducible results, whilst helping the sound engineer to mould their desired sound world.
The research described in this case study was carried out at the University of Huddersfield (UoH) by Dr Hyunkook Lee (Reader at UoH since 2010). In 2013, he established the Applied Psychoacoustics Lab (APL), which is now recognized as one of the world-leading groups in this field. Others who contributed to the research include Lee’s previous PhD students Dr Chris Gribben (Acoustic Engineer at B&O, Denmark), Dr Dale Johnson (Research Fellow at UoH) and Dr Rory Wallis (Analytics Manager at AIMS Smarter, UK).
All the research described below used a similar methodology. Each hypothesis was tested using a panel of 15–20 volunteers (the typical number for hearing-based research since the panellists must have sufficient aural acuity to give reliable feedback), who provided their subjective view of the quality of each aural experience based on a questionnaire. Their responses were statistically analysed. Physical measurements were collected in parallel that were used to define physical parameters and ensure reproducibility.
To test the variables associated with 3D recording and reproduction, the effect of altering a number of elements of common practice were modelled. Typical sound engineer practice was compared to see how it measured up against the optimum listener experience.
(i) Experience had taught sound engineers that, for the most natural-sounding stereo reproduction, two microphones had to be separated in the horizontal plane. By analogy, they had assumed that pairs of microphones needed to be positioned at different heights to give the best results in the vertical plane. Lee conducted the first piece of formal research on this topic [R1] (2014), by testing multiple microphone heights and separations.
He found that having a space between vertically arranged microphones provided no perceptual benefit to the listener – the gap could be as low as zero. This resulted in a paradigm shift in 3D microphone array design. Conventional arrays had a cube-like form factor, but the research showed that a more compact, horizontally spaced and vertically ‘flat’ array, gave listeners a more accurate impression of the actual sound. Psychoacoustically, this can be explained by understanding that a pair of human ears are in the same horizontal plane.
(ii) The accepted best-practice of separating microphones in the vertical plane led to ‘interchannel crosstalk’ because the upper microphone captured the same sound as the lower one, which made the resulting sound ‘image’ more blurry due to interference. Using a similar experimental strategy to (i), Lee [R2] (2017) showed that angling the upper microphone by a fixed amount (so that it no longer directly faced the sound source), meant that it preferentially picked up ambient sound, which led to more accurate 3D sound image reproduction. This so-called psychoacoustic threshold level difference was measured to be 7dB and became the theoretical basis for determining the vertical angle between the microphones for several arrays designed by professional sound engineers.
(iii) To try to improve the perceived ‘spread’ of a vertical sound image, the signal to each speaker was slightly manipulated (decorrelated) by sound engineers because it improved how ‘natural’ a sound appeared when using a conventional speaker set-up. Research [R3] (2017) proved that applying decorrelation in the vertical plane had a negligible impact on the perception of the listener.
(iv) The perceived quality of a recorded sound is always a trade-off between the degree of ‘space’ in the soundscape and accuracy with which the listener can pinpoint the position of a particular sound. This is one element of the ‘art’ of the sound engineer. The conventional approach for positioning the microphones in an array that delivered the most accurate ‘position’ in 4-channel 360° reproduction, was to separate them by 23cm. Testing, using an expert panel of listeners, showed that this was wrong with the best results coming from a 50cm separation [R4] (2019). This led Lee to develop a new microphone array, the ESMA-3D. It provided instructions on how to position existing microphones for recording and was incorporated into a new psychoacoustic model [R6].
(v) Lee further refined the model to enable it to accurately predict the position that a sound source appeared to a listener, based on the physical position of the sound sources and the positions of the microphone arrays relative to them. This is known as phantom image localization [R5] (2013). Different notes from various musical instruments (piano, trumpet, etc.) were recorded and analysed in a real setting. The results were then modelled. This generated the data needed to locate the sound image of a source of sound at a specific position between a set of loudspeakers. The resulting localization prediction model could be used to predict the position of perceived sound images.
(vi) A software app, the Microphone Array Recording and Reproduction Simulator (MARRS) was developed by Lee and Johnson based on the data and model [R5]. It enabled the simulation and automatic configuration of a microphone array [R6] (2017) and gave a more accurate prediction of perceived sound image positions than existing tools. It also provided a more intuitive visualization for the sound engineer; based on the position of the sound sources and other input parameters, such as how ‘spacious’ the sound should be and how accurately it should be possible to locate the sound, the app recommends how to configure the microphone arrays and where to place them.
In summary, the research has codified the techniques that sound engineers used for 3D and surround sound recording and shown that their traditional approaches were sub-optimal. It has produced novel methods that improve the spatial impression of recorded 3D sounds and reduced the microphone array size significantly, compared to conventional methods.
3. References to the research
The references listed below are considered to have a 4*/3*/2* quality. [R1] to [R5] have been published in high quality journals: Applied Sciences (Scimago Q1 in Engineering) or Journal of the Audio Engineering Society (Scimago Q2 in Engineering, Q1 in Music). [R6] has been presented at an AES International Convention, which is one of the most prestigious conferences dedicated to audio.
[R1] Lee, H. and Gribben, C. (2014) ‘Effect of Vertical Microphone Layer Spacing for a 3D Microphone Array’, Journal of the Audio Engineering Society, 62 (12), pp. 870–884. https://doi.org/10.17743/jaes.2014.0045
[R2] Wallis, R. and Lee, H. (2017) ‘The Reduction of Vertical Interchannel Crosstalk: The Analysis of Localisation Thresholds for Natural Sound Sources’, Applied sciences, 7 (3), pp. 278. https://doi.org/10.3390/app7030278
[R3] Gribben, C. and Lee, H. (2017) ‘A Comparison between Horizontal and Vertical Interchannel Decorrelation’, Applied sciences, 7 (11), pp. 1202. https://doi.org/10.3390/app7111202
[R4] Lee, H. (2019) ‘Capturing 360° audio using an Equal segment microphone array (ESMA)’, Journal of the Audio Engineering Society, 65 (9), pp. 733–748 https://doi.org/10.17743/jaes.2018.0068
[R5] Lee, H. and Rumsey, F. (2013) ‘Level and Time Panning of Phantom Images for Musical Sources’, Journal of the Audio Engineering Society, 61 (12), pp. 978–988. http://www.aes.org/e-lib/browse.cfm?elib=17075 [can be supplied on request]
[R6] Lee, H., Johnson, D. and Mironovs, M. (2017) ‘An Interactive and Intelligent Tool for Microphone Array Design’, In proceedings of Audio Engineering Society 143rd Convention. http://www.aes.org/e-lib/browse.cfm?elib=19338 [can be supplied on request]
4. Details of the impact
The research generated novel methods to improve the spatial impression (a sense of the relative positions of sound sources and perceived spaciousness) of recorded 3D sounds, whilst reducing the microphone array size significantly compared to conventional methods. Four different 3D microphone arrays were developed and these were exploited by the sound-recording industry for improved 3D music recording (a radio station and a classical music concert), outdoor recording (a microphone manufacturer and a TV station) and for virtual reality music recording (a recording studio and a VR content production company).
The impacts can be summarized under three headings:
Influencing the design of a commercial 3D audio microphone array.
Enhancing immersive sound recording techniques.
Improving the realism of recorded classical music.
The research findings led to talks and workshops on 3D audio at international conferences, including Audio Engineering Society (AES) Conference on Spatial Reproduction (2018). These resulted in industry interest and subsequent adoption of the new approaches.
Influencing the Design of a Commercial 3D Audio Microphone Array
Schoeps is a best-in-class Germany-based manufacturer of microphones for sound professionals. Their microphone array product, the ORTF-3D, was completely redesigned in 2016 based on consultation with Lee, which applied the main findings of three papers [R1, 2, 3]. Before redesign, the array had a cube-like form factor and its large size made it impractical for sound engineers to carry and install easily. This was a particular problem for outdoor location recording and broadcasting environments, such as music concerts and sports events. The company said UoH research “helped [it] to identify crucial and non-crucial aspects of microphone array geometry and the extent to which certain design parameters influence spatial perception” [S1]. As a result, the research “significantly influenced the design of 3D audio microphone arrays at Schoeps” [S1] and led to a new version of the ORTF-3D that was much more compact. The vertical microphone spacing was reduced to zero, which reduced the size and weight of the unit without affecting its performance. This gave users benefits in terms of portability, set-up time and health and safety. ORTF-3D won the prestigious TEC Award for its innovative design during the 32nd NAMM Show (the world’s largest music products trade fair) in 2017 ( tecawards.org/tec-winners).
The new version of ORTF-3D was rapidly adopted and used for broadcasting a number of major musical and sports events, such as the BBC Proms (2016–2019), FIFA World Cup (2018), and the French Open (2018–2019). The BBC’s Technical Producer, responsible for recording and broadcasting the BBC Proms concerts, stated that ORTF-3D provided “flexibility when mixing and better spatial impression than Ambisonics and other coincident mic arrays” [S2]. The sound engineer who was in charge of capturing the crowd sounds of the 2018 Russia World Cup football matches for HBS (a Swiss sports broadcaster) worldwide broadcasting, testified that ORTF-3D “played a major role in delivering excellent immersiveness and precise localisation for crowd sounds in 3D… which would have not been possible with a conventional microphone array like First-Order Ambisonics”. He said that the microphone array provided “a far better sense of envelopment and a more accurate localisation as well as a larger listening area, which is of paramount importance for TV viewers” [S3].
Enhancing Immersive Sound Recording Techniques
The reputation of the UoH APL 3D audio research led to consultancy on immersive sound recording techniques with Abbey Road Studios (London), one of the most famous recording studios in the world. The UoH-developed ESMA-3D microphone array [R4] was introduced to Abbey Road through a collaborative project in 2017. Since that time it has become the studio’s go-to microphone array for various types of 3D recording [S4]. One example was a large orchestral recording session for the Electronic Arts’ computer game, Star Wars Jedi: Fallen Order, which was performed by London Symphony Orchestra [S4, S5]. The award-winning recording and mixing engineer who worked on the session became “a fan of the array from the initial experiment sessions for the game conducted in March 2019” [S4]. The excellent 360° imaging the array offered, led him to develop a new way to bridge the sonic gap between the strings and the woodwinds, which produced impressive results [S4]. The Head of Audio Product at Abbey Road stated that “of all the arrays I have tried over the last couple of years, ESMA-3D is the most useful and sonically pleasing microphone array with versatile end user applications from stereo to 5.1 and Ambisonics to Dolby Atmos” [S4]. ESMA-3D has also become an essential tool to “glue various performers in the room into a cohesive acoustic fingerprint” in Abbey Road’s pioneering work on six-degrees-of-freedom virtual reality (VR) audio, which adds in a representation of body movement to the standard three degrees used in conventional VR [S4]. ESMA-3D was also “the backbone” of a VR experience created for Highways England (HE) [S6] by MagicBeans, a start-up dedicated to providing highly immersive virtual and augmented reality experiences. It recorded various types of motorway traffic noise using the array and used the recordings to create an immersive training environment for roadside workers using VR (2019). The CEO of MagicBeans said that “the imaging stability and sweet spot with ESMA-3D is so much better and larger than with a comparable Ambisonic recording, and the spaciousness and timbre of the ESMA-3D recordings are far better”. He added, “We would not have been able to provide this large-scale sound experience without this manner of recording being pioneered by Dr. Lee …. [it is] a significant step forward in making high quality 3D captures practical and useful” [S6].
Improving the Realism of Recorded Classical Music
Central Sound (CS), an award-winning audio production service, is part of Arizona PBS (Public Broadcast Service). It specializes in capturing acoustic music performances for broadcast. The PCMA-3D microphone array [R1] helped CS to reach their goal of a “quintessential sound” in their immersive recordings for the Grammy-Award-winning Phoenix Chorale and Grammy-nominated True Concord Voices and Orchestra in 2019 [S7]. Having moved away from the ‘trial and error’ approach, their chief recording engineer and manager said “the test methodology Dr Lee has employed gives us confidence that we are utilising proven techniques” [S7].
The Austrian Broadcasting Corporation (ORF) was the first public broadcaster in Europe to start a 5.1 surround sound service and their senior sound engineer was one of the pioneers in 5.1 surround recording. He explored new techniques for outdoor 3D audio recording for broadcast and, working with Lee, developed a new 3D microphone array (2018) [S8]. Named the Double UFIX it used the novel UoH localization prediction model [R5] and the associated iOS/Android MARRS app [R6]. The ORF engineer said that the UoH research “significantly increased sense of realism or an even more convincing illusion of the resulting sounds” [S8].
The app provided “an intuitive way to derive tailormade microphone arrangements for specific recording situations” [S8]. It helped sound recording engineers achieve their goals for accurate and spacious imaging in recording, by visualizing the virtual sound images resulting from different microphone configurations. Compared to other similar tools, MARRS helped them produce results more quickly, accurately and reproducibly. The app is currently used by 600+ users worldwide. The Director of Recording and Music Technology at The Tianjin Juilliard School for musicians in China has used MARRS as a starting point of microphone placement for many of his professional recording sessions, including Peabody Symphony Orchestra and Tianjin International Chamber Music Festival [S9]. He stated that, “it saves me a lot of time in correctly configuring a microphone array and achieving an accurate stereo imaging, thanks to the accurate sound image visualization and the powerful parameter control options it provides” [S9].
5. Sources to corroborate the impact
[S1] Letter from CEO, Schoeps (world-renowned microphone manufacturer
[S2] Letter from Technical Producer, BBC Radio and BBC R&D
[S3] Letter from Professor/Tonmeister, University of Applied Sciences Darmstadt
[S4] Letter from Head of Audio Products, Abbey Road Studios
[S5] Stiles, M. (2019) Composer Stephen Barton Returns for Spatial Audio Experiments [online] Available at: https://www.abbeyroad.com/news/composer-stephen-barton-returns-for-spatial-audio-experiments-2554 [Accessed: 20 Jan 2021]
[S6] Letter from CEO, MagicBeans (VR training specialist)
[S7] Letter from Manager, Central Sound at Arizona PBS
[S8] Letter from Senior Sound Engineer, Austrian Broadcasting Corporation (ORF)
[S9] Letter from The Director of Recording and Music Technology, The Tianjin Juilliard School
- Submitting institution
- The University of Huddersfield
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Technological
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
The interface between the wheel and rail is critical for train operation, and maintenance of wheels and track is one of the largest areas of cost in the rail sector. A significant challenge is transitioning between tram and heavy rail networks. Standard wheel profiles can cause damage and increase the vehicle derailment risk. The tram-train wheel profile developed through this research has been instrumental in the opening of the award-winning tram-train route in Sheffield, with more than ten other transport authorities planning similar services.
Research into the optimization of wheelset maintenance resulted in significant revisions to national standards and guidance to operators and generated operational and cost benefits. The tools and techniques developed were trialled by Alstom, the French rolling stock manufacturer, increasing wheelset maintenance (re-profiling) intervals by up to 43%. A number of the tools have been incorporated into commercial systems, such as the Alstom ‘Trainscanner’ and MRX Technologies’ ‘Surface Crack Measurement’ device.
2. Underpinning research
The shape of the cross-section of a railway wheel can significantly influence the dynamic performance of the wheelset (consisting of the axle and the wheels), as well as the physical properties such as contract stress, creep forces and wear. An incompatible wheel-rail profile combination can be the source of problems such as high rates of wear, cracking of the wheel-rail surface through excessive cyclic stresses (known as rolling contact fatigue (RCF)) and wheelset instability. These problems can result in reduced wheel-rail life, increased maintenance costs and, in some cases, increased derailment risk. Rail vehicle wheelsets are regularly maintained to ensure their safe operation and prolong their life. This is achieved through measurements to inspect roundness, profile shape and surface damage. If necessary, wheels are re-profiled on a lathe to preserve the optimal wheel profile shape and remove any surface damage.
The research described below was undertaken at the Institute of Railway Research (IRR) based at the University of Huddersfield (UoH). It was led by Adam Bevan (Professor, at IRR since 2012). Other members of the team, who all joined IRR in 2012 and are still there, except where indicated, were Julian Stow (Assistant Director), Principal Research Fellows David Crosbee and Paul Molyneux-Berry (left 2018), and Research Fellows Yousif Muhamedsalih and Antonio Antrado (left 2016).
The research included the development of wheel damage models using a combination of detailed vehicle dynamic simulations and bespoke computer models. These models were then applied to the development and assessment of new wheel profiles and the design of wheelset maintenance optimization activities, with the aim of extending wheel-rail life and reducing whole-life costs.
Development of Wheel Damage Models
Research carried out by Bevan and Molyneux-Berry has developed and validated computer models to predict the rate of wear and RCF damage on railway wheels. These models were used to improve the understanding of the main wheel damage mechanisms, the rates at which they occur and the optimum maintenance regimes [3.1]. Wheel maintenance and observation data was reviewed from over 90% of all UK passenger vehicle fleets to determine the main drivers for maintenance, identify the frequency of the different wheel damage mechanisms and to provide validation data for the computer models that were developed during the research. These models were subsequently incorporated into the industry-recognized Vehicle-Track Interaction Strategic Model (VTISM) to support whole system cost modelling [3.2].
Design of Optimal Wheel Profiles
A common and serious cause of wheel-rail deterioration is RCF damage, which if left unattended can result in premature fracture of the surface material with catastrophic consequences, as observed during the Hatfield train crash in October 2000. To better understand how RCF could be reduced, the computer modelling tools [3.1, 3.2] were used to assess the performance of current and new wheel profiles designed by the researchers. A new ‘track friendly’ wheel profile shape, designed to reduce the wheel-rail forces that cause RCF damage, was developed and has since been adopted as one of the standardized profiles used in the UK.
A tram-train is a vehicle that operates on two very different railway infrastructures – as a tram on light-rail infrastructure and as a conventional train on heavy-rail infrastructure. The wheel profile needed for safe running on a street and a heavy-rail network are different, which causes a problem for rail vehicles that can transition between the two. Existing wheel designs caused a significant risk of derailment of the vehicle and increase in damage to the track. A combination of vehicle dynamic simulations and bespoke software was used to create a wheel profile design for use on a tram-train vehicle which was optimized to provide safe running and minimized wheel-rail damage. This resulted in a new wheel profile for GB tram-train operations that was successful on both railway infrastructures [3.5].
Optimization of Wheelset Maintenance
Research into wheelset maintenance, undertaken by Bevan, provided a better understanding of the factors that influence the degradation of wheels and the scheduling of maintenance activities. The development of computer models [3.1, 3.2] and wheelset management tools, to calculate and interpret wheel condition, supported the optimization of wheelset maintenance and renewal activities and an examination of the cost benefits [3.2, 3.4]. Techniques and guidance (developed by the researchers and published by the Rail Safety and Standards Board) to help fleet maintainers better understand wheel degradation and solutions to extend life and reduce costs, were also developed. In particular, Stow and Muhamedsalih investigated the use of Economic Tyre Turning (ETT), a technique not previously permitted in the UK, that minimizes the amount of metal that is removed from railway wheels during planned maintenance [3.4, 3.6]. The technique was proven, through application of computer models [3.1] and field trials on the Alstom Class 390 fleet, to improve maintenance efficiency and reduce costs without compromising safety.
The review of wheelset maintenance on UK passenger vehicle fleets [3.1] identified that to reduce inspection times and optimize wheel re-profiling, a fast and repeatable method of quantifying damage (both surface and near-surface) on railway wheels was needed. Research into novel automated non-destructive testing (NDT) technologies and tools for monitoring wheel condition was undertaken to address this. This included the use of magnetic flux techniques for the detection of wheel surface damage [3.3]. Tools for calculating and trending wheel profile parameters from laser scanned data were developed.
3. References to the research
Included below is a list of the relevant journal papers which evidence some of the underpinning research described in this case study. Three of the research outputs listed below [3.1, 3.5 and 3.6] have been awarded prizes by the Institution of Mechanical Engineers (IMechE) Railway Division for their contribution and achievements in the field of Railway Engineering.
Bevan, A., Molyneux-Berry, P., Eickhoff, B. and Burstow, M. (2013) ‘ Development and validation of a wheel wear and rolling contact fatigue damage model’, Wear, Vol. 307, pp. 100–111. https://doi.org/10.1016/j.wear.2013.08.004
Bevan, A., Molyneux-Berry, P., Mills, S., Rhodes, A. and Ling, D. (2013) ‘ Optimisation of wheelset maintenance using whole system cost modelling’ Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit, Vol. 227, pp. 504–608, ISSN 0954-4097, https://doi.org/10.1177/0954409713484712
Bevan, A. and Klecha, S. (2016) ‘ Use of magnetic flux techniques to detect wheel tread damage’ Proceedings of the ICE - Transport, Vol. 169, pp. 330–338. ISSN 0965-092X,
https://doi.org/10.1680/jtran.16.00025 [can be supplied on request]
Andrade, A. and Stow, J. (2017) ‘ Assessing the potential cost savings of introducing the maintenance option of Economic Tyre Turning in Great Britain railway wheelsets’ Reliability Engineering and System Safety, Vol. 168, pp. 317–325. ISSN 0951-8320, https://doi.org/10.1016/j.ress.2017.05.033
Crosbee, D., Allen, P. and Carroll, R. (2017) ‘ Analysis of design and performance of tram-train profiles for dual-operation running’, Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit, Vol. 231, pp. 578–597. ISSN 0954-4097, https://doi.org/10.1177/0954409716679448
Muhamedsalih, Y., Stow, J. & Bevan, A. (2018) ‘ **Use of railway wheel wear and damage predictions tools to improve maintenance efficiency through the use of Economic Tyre Turning (ETT)**’, Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit, Vol. 233,1 pp.103-117. https://doi.org/10.1177/0954409718781127
4. Details of the impact
The research findings have led to impact on the UK rail and tram networks for passengers and operators. Government policy on tram-train transportation systems and national railway standards were influenced.
The impacts can be summarized under three main headings:
Improvements to public transport in Sheffield and plans to replicate the model
Enabling the use of a cost-effective wheel maintenance procedure
Commercial exploitation by train operators
Improvements to Public Transport in Sheffield and Plans to Replicate the Model
The research [3.5], undertaken as part of a £125m Department of Transport tram-train pilot study, developed a tram-train wheel profile that was subsequently tested on the Sheffield tram-train scheme operated by Stagecoach. Stagecoach stated that “without this work and the identification of a wheel profile that would allow dual operation of a tram on both a street network and a heavy network, it is highly unlikely that Network Rail would have allowed street trams to operate on its network” [5.1].
The economic benefits from the new scheme have helped to regenerate Sheffield City Centre and reduced congestion on the existing road and rail networks, through increased capacity by utilizing existing infrastructure [5.2, 5.3]. The route was opened in October 2018, running three services an hour between Sheffield and Rotherham Parkgate. Seven new vehicles were built and fitted with the new wheel profiles. In 2019 Stagecoach announced: “the route has been in operation for over 12-months and has delivered more than 1,000,000 passenger journeys” [5.2]. The Sheffield tram-train scheme also achieved a 100% customer satisfaction record in its first year of operation as recorded by Transport Focus (the independent watchdog for transport users) [5.2]. In addition, the Senior Engineer- Light Rail, Network Rail, stated that: “no unusual behaviour of track or vehicle has been experienced” on the network between Tinsley and Rotherham Parkgate, despite the increased usage due to the new tram-train service [5.4].
The research has also informed the wider strategic testing and development of tram-train technology for subsequent roll-out across the UK. Following the success of the Sheffield tram-train scheme, Network Rail have confirmed that further ”tram-train schemes are now being actively pursued by a number of UK Authorities” [5.4], with “over ten transport authorities, including Manchester, Birmingham, Glasgow and Cardiff, seeking to create their own tram-train service” [5.3]. These will improve transport links and reduce congestion on the existing UK road and rail networks.
Enabling the Use of a Cost-Effective Wheel Maintenance Procedure
As a result of this research [3.4, 3.6] Economic Tread Turning (ETT) was incorporated into the Railway Group Standard GMRT2466 issue 4.1 (2019). The Chairman of the Industry Wheelset Management Group and Engineering Manager at Avanti West Coast, confirmed that “of particular significance, is the [UoH] work [..] which has demonstrated quite significant cost savings for the industry”. He also comments that “these changes have been adopted through revision of the Standard GMRT2466 issue 4.1. The research can therefore enable significant potential savings in both track and train maintenance costs” [5.6]. The Principal Vehicle Systems Engineer, Rail Safety and Standards Board (RSSB), added “ETT formed a significant part of the revised content and covers its application, geometry and service limits.” [5.7] The standard (GMRT2466), and accompanying Rail Industry Standard for Wheelsets (RIS-2766-RST, 2017), refers directly to a report [5.7] published by UoH (2017), which, along with the research described in [3.6], established the confidence that a range of flange widths could be selected by operators and maintainers while remaining safe and compatible with the infrastructure. In 2016, the researcher (Muhamedsalih) was awarded the ‘Rail Research UK Association / Institute of Mechanical Engineers’ Young Researcher of the Year Award ( Ground breaking research dispels UK myths around ETT) for the research work into ETT on GB Railways [3.6]. The award panel concluded that “this research provides valuable and strong evidence to support the case that train operators should be allowed to implement ETT policies and will provide the opportunity to exploit the cost savings associated with ETT without a significant detrimental effect on the infrastructure over which they operate”. These cost savings are estimated by RSSB to be between £880k and £5.1m per year across the GB passenger train fleet [5.8].
The research outputs on wheel profiles and ETT [3.1, 3.6] were trialled in 2017 by Alstom on the Class 390 train fleet operating on the West Coast Mainline, in an attempt to extend wheelset re-profiling intervals and reduce the cost of maintenance. As detailed by the Wheelsets and Components Engineer, Alstom have seen a “13% increase in re-profiling intervals, from 310,000 miles to over 350,000 miles, allowing one additional trainset to be released for service per week.” The benefits of releasing an additional trainset into service “should not be underestimated”, it allows the operator “to provide more train services” and allows the maintainer to have “a better understanding of the performance of its wheelsets” and therefore make smarter decisions regarding day-to-day operations [5.9].
Commercial Exploitation by Railway Operators
Automated non-destructive testing techniques developed from the research [3.4, 3.6] have been commercially exploited by two major rail infrastructure manufacturers, Alstom and MRX technologies.
Since 2017, the tools arising from the research for calculating wheel condition parameters [3.1] have been integrated within the Alstom ‘Trainscanner’ and ‘HealthHub’ predictive maintenance systems to automatically calculate and interpret wheelset condition and support maintenance scheduling. The Wheelsets and Components Engineer at Alstom, stated that “using these tools, we have recently been able to operate a wheelset up to 500,000 miles (43% increase) without incident as part of trial, and we are currently working with the [UoH] Institute to realize the benefits of this trial across the entire fleet” [5.9].
The wheel surface crack measurement (SCM) technology (2016), developed in collaboration with MRX Technologies [3.3], enabled rail operators and maintainers to reliably detect and characterize the severity of wheel damage which is difficult to quantify through visual inspection [5.10]. Porterbrook, a railway rolling-stock leasing company, have taken the research on SCM and ETT [3.3, 3.6] and applied it to a fleet of diesel locomotives. The Head of Digital Services at Porterbrook, commented: “the operator has trialled this on one fleet, which moved them from a 600,000 miles wheel life to 780,000 miles wheel life (25% extension)” [5.10].
The models developed during the research [3.1, 3.2] are a key element of the wheelset management model, which was incorporated into the Vehicle Track Interaction Strategic Model (VTISM) in 2014. VTISM is a rail industry tool (developed by UoH in association with Serco) for modelling the cost of track and wheelset deterioration. It supports the optimization of maintenance and renewal regimes, thereby increasing wheelset life and reducing costs. This industry-recognized model is run and managed for the UK rail network by the RSSB and is available to all its members for use [5.7]. There are 28 organizations (and 42 users) registered to use the model, of which five are rail industry suppliers, nine are train operating companies, one is an infrastructure manager and the remainder are research organizations. This represents a significant proportion of the UK rail sector.
5. Sources to corroborate the impact
Testimonial Letter from Stagecoach SuperTram, Head of Engineering, Rolling Stock and Infrastructure, April 2020.
Sheffield City Region, Tram-Train Marks One Million Passenger Journeys, October 2019.
Rail Business Daily, South Yorkshire Blueprint to Benefit Future Tram-Train Schemes, October 2020.
Testimonial Letter from Network Rail, Senior Engineer, Light Rail, May 2020.
Testimonial Letter from Avanti West Coast, Chairman, April 2020.
Testimonial Letter from RSSB, Principal Vehicle Systems Engineer, March 2020.
Economic Tyre Turning – Practical Considerations for Creating Thin Flange Variants of Wheel Profiles, 2017, sparkrail.org/Lists/Records/DispForm.aspx?ID=24866.
Rail Safety and Standards Board, Economic Tyre Turning - Can We Cut Wheelset Costs in an Easy and Safe Way? January 2018.
Testimonial Letter from Alstom, Wheelsets and Components Engineer, December 2020.
Testimonial Letter from Porterbrook, Head of Digital Services, January 2021.
- Submitting institution
- The University of Huddersfield
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Technological
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
The design of valves used in challenging conditions, such as those found in the oil, gas and nuclear industries, has traditionally been based on the experience of the engineer rather than reproducible mathematical modelling. This led to unpredictable and sub-optimal performance. The inability to consistently ensure quality compliance, slowed the development of new technology.
Research at the University of Huddersfield, developed valve design methods that were responsible for reducing design lead times and manufacturing costs for an SME, Weir Valves. The resulting valves were up to 250% more accurately sized and they also out-performed competitor products. This resulted in an increase in sales of 640% and helped the company win a contract worth circa £3m.
Workshops explaining the research findings regarding multiphase valves have influenced engineer behaviour at companies such as Exxon Mobil, resulting in wide acceptance of research outcomes in the engineering practice.
2. Underpinning research
All valves (with sizes varying from a few millimetres to multiple metres), work on the same principle; where a fluid medium (which could be a liquid, gas or slurry) enters via an inlet, a piston controls its rate of flow and it exits through an outlet (see figure 1). A safety critical valve is characterized by its application in a critical infrastructure, and also by the degree of precision and manufacturing robustness to which it must be fabricated. They are designed to be used in very challenging conditions, such as those in the piping systems for nuclear power plants, or where the pressure differential across the valve is very high (for example, several thousand pascals across a valve that has a pressure-reducing element that is only 10cm long).
Figure 1: Valve components
In order to regulate the flow, a component called a trim is inserted within the valve in the path of the fluid. The trim consists of a series of interlocking channels, of differing diameter, geometry and interconnectivity – as the fluid flows through them, they change the pressure and flow velocity. A trim that is designed for a very high change in pressure is primarily used in very critical applications and is called a severe service trim. Trims are complex and, if not properly designed, can cause numerous problems in the valves during their operation (see figure 2). Firstly, if there are a number of flow paths that include an abrupt change in the size of the
channel through which the liquid flows, they can cause the fluid pressure to drop below its critical pressure and vaporize. When the vapour enters a high-pressure part of the trim it collapses, releasing massive amounts of energy in a process known as cavitation. This can cause severe damage to the valve. Secondly, if too much energy is taken out of the fluid by the trim, then the energy requirement of the entire system may increase – this is described as the valve
Figure 2: Cylinder arrangement in a trim disc
being inefficient. Furthermore, if the medium being transported is a slurry, the solid particles can cause erosion within the trim, seriously reducing the life of the valve. It was known that these effects existed, but there was no systematic way to accurately predict the performance of a given valve design in a specific use-case.
Although design standards for valve manufacture exist, they are empirical and focused mostly on simplified valve geometries. Before this research there was no way to mathematically model the performance of a valve relative to its trim (and other features) and valve manufacturers developed new products based on trial and error (using a valve from their “back catalogue” as the starting point for developing one for a new application). The lack of comparability between valves, and inconsistency in documented design protocols, meant the development of new design techniques in the industry had stalled.
The research described in this case study was carried out at the University of Huddersfield (UoH) by Prof. Rakesh Mishra (at UoH since 2000), Dr Taimoor Asim (Research Fellow at UoH from 2013–19) and Carlos Oliveira (KTP Associate 2014 to 2017).
The research, which was undertaken as part of a Knowledge Transfer Programme (KTP) with an SME valve manufacturer (Weir Valves) between 2014 and 2017, evolved from a previous KTP project with the same company (2011-2014) and earlier work carried out under an EPSRC Case Award with Bentley Motors in 2005 [R3]. Both of the above works involved flow mapping and design of complex geometries to enable optimum performance. Weir Valves had identified that they could generate more sales if they understood enough about the flow characteristics within their valves to explain them to their clients, and thus justify the reliability of their safety critical valves. The UoH-based research team approached the challenge in the following way.
The initial stage of the research was to map the valve design process used at Weir and investigate how effective it was in producing valves with the desired performance. This was done by measuring the flow characteristics of actual valves and then building computational models to mimic those characteristics. Next, design equations (mathematical models) were built to define how the flow within a valve was affected by all of its constituent parts and these were tested experimentally, to see whether a valve manufactured using the model behaved as predicted [R5,6]. Specifically, local flow fields were quantified and relationships between the geometrical features to be designed and the likely flow behaviour were established. By following an iterative process, it was possible to develop a modified valve design procedure that was faster and more reliable.
The process was applied in the specific case of a low shear valve [R4]. These are valves where the geometrical complexity of the trim needs to be such that there is only a small change in the physical properties of the flowing fluid. The valve had to be designed for low energy loss and required quantification of the interrelation between the geometrical complexity of the trim and the resultant local flow fields. The research characterized the link between them for the first time, by combining experimental and computational analyses. The project created a successful new product for Weir and as a result, they asked UoH to explore the much more complex case of multiphase flow applications, where they wanted to ensure minimal mixing of the phases in a mixture. A practical example is avoiding the recombination of the oil and water fractions present in partially refined crude oil.
In the second KTP, the same modelling process was used to study the behaviour of multi-phase liquids (such as slurries containing water, oil and sand) in order to predict the behaviour of the liquid through a particular valve. This resulted in a set of modified equations [R1] that could be applied in the design process.
Valves, particularly the trims, are extremely complex to manufacture and the lead time from the start of the design process to when a valve is ready to be manufactured can be several months. To speed up this process, additive manufacturing (also known as 3D printing), was being trialled across the valve sector during this period, with an expectation that it would be faster and more cost effective than traditional methods. However, valve trims provide particular challenges for the process, since the flow paths are very complex and narrow. Research [R2] studied and modelled all aspects of the manufacturing process (e.g. scanning speed, power of the laser, etc.) to derive the best balance between the limitations of the manufacturing process and the optimal specification of the valve. More recently, the sophistication of the mathematical models has meant that it is now possible to modify the flow path geometry to ensure control of the desired energy loss across each stage of the trim. This has enabled further optimization of valve performance by reducing cavitation and erosion and thus extending the life of the valve.
3. References to the research
The following outputs provide reference to the body of research and are predominantly 3* or higher, being peer-reviewed journal articles in Q1 and Q2 journals. Authors at the University of Huddersfield at the time of publication are highlighted in bold:
[R1] D. Singh, A. Aliyu, M. Charlton, R. Mishra, T. Asim, A. Oliveira (2020), Local multiphase flow characteristics of a severe-service control valve, Journal of Petroleum Science and Engineering. 195, 18 p., 107557, ISSN092-4105 https://doi.org/10.1016/j.petrol.2020.107557
[R2] D. Singh, M. Charlton, T. Asim, R. Mishra, A. Townsend and L. Blunt (2020), Quantification of additive manufacturing induced variations in the global and local performance characteristics of a complex multi-stage control valve trim, In: Journal of Petroleum Science and Engineering. 190, 13 p., 107053 https://doi.org/10.1016/j.petrol.2020.107053
[R3] E Palmer, R Mishra and J. Fieldhouse (2009), An optimization study of a multiple-row pin-vented brake disc to promote brake cooling using computational fluid dynamics. In: Proceedings of IMechE, Part D, Journal of Automobile Engineering, 223,7,865-875,11p,ISSN: 0954-4070
https://doi.org/10.1243/09544070JAUTO1053
[R4] T. Asim, R. Mishra, M. Charlton and C. Oliveira (2018) ‘Improved Design of a Multi-Stage Continuous-Resistance Trim for minimum Energy Loss in Control Valves’, Energy, 174, 1 May 2019, pp. 954–971, ISSN: 0360-5442 https://doi.org/10.1016/j.energy.2019.03.041
[R5] T. Asim, R. Mishra, M. Charlton and C. Oliveira (2018) ‘Effects of the geometrical features of flow paths on the flow capacity of a control valve trim’, Journal of Petroleum Science and Engineering, 172, pp.124-138, ISSN092-4105 https://doi.org/10.1016/j.petrol.2018.09.050
[R6] T. Asim, M. Charlton and R. Mishra (2017) ‘CFD based Investigations for the Design of Severe Service Control Valves used in Energy Systems’ Energy Conversion and Management, 153, pp. 288–303, ISSN 0196-8904. https://doi.org/10.1016/j.enconman.2017.10.012
4. Details of the impact
The research was carried out within a Knowledge Transfer Programme (KTP) and the findings have been exploited by the sponsoring company, Weir Valves (now part of Trillium Flow Technologies) to develop a new product, and thus increase their turnover. Lessons in the practical application of the new design approach have been shared with companies operating across the flow dynamics industry.
The impacts can be summarized under three headings:
A. Commercial benefits for Weir Valves
B. Improvements in safety-critical valve design
C. Better design practice across the critical valve industry
Commercial Benefits for Weir Valves
The research findings enabled Weir to increase their sales in a number of ways. The mathematical models [R1,4,5] applied to multiphase valve design, enabled the company to design and build valves that met the specifications of their customers more reliably. The demonstrable mathematical rigour behind the valve design, which resulted in an “astonishing improvement in size calculations” [E5] significantly increased credibility in the market and convinced customers that Weir were the partner of choice. The MD of Trillium said that their “detailed understanding of the performance of [their] valves” means “Customers now see us as knowledgeable in this area. We now use this as a sales tool” [E1].
Sales of the X-treme valve had increased by 640% by 2018, according to the KTP final report produced at the end of the project. The project was vital to the award of a £3m contract for an oil and gas industry customer in the North Sea [E5]. Sales volumes have continued to grow and the MD of Trillium stated, “The KTP was a great success [..] and has improved how competitive we are [..]. Our market share of severe service valves has increased significantly” [E1].
The new valve design method developed, also revealed that the existing design methods resulted in oversized valves for a given capacity. The UoH-developed design process increased the accuracy of the valve sizing by up to 250% [E5] which led to the creation of a smaller valve that gave the same performance envelope as a bigger valve. This gave Weir a cost advantage (because less raw materials are required) and meant they were able to reduce the price of the finished product to their customers. Confirming this, the MD of Trillium wrote that the KTP had allowed them to “be more competitive in smaller valve offerings” [E1]. Indeed, the company were able to reduce the size of all the valves, not just multiphase valves [E5]. Smaller valves provided customers with other advantages such as fewer operating issues and associated physical handling problems [E2,3].
Improvements in Safety-Critical Valve Design
The research findings resulted in a set of mathematical models [R5,6] that enabled Weir to improve the performance of their safety-critical valves, both in terms of their performance against specification and their longevity. The MD of Trillium stated that the research findings help the company to “specify valves more accurately, to select the appropriate valve and [..] choose the optimal solution [for] better operational performance” [E1]. Weir also used the models as the starting point for their valve designs. They then iteratively applied the correction factors the models indicated, which produced the optimum design more quickly.
The advantage of this approach, as opposed to the old, trial and error method based on the industry-set empirical standards, is that the performance of valves designed this way is more predictable and repeatable. This resulted in a superior finalized valve design [E1]. This approach enabled Weir to redesign and improve their X-treme valve for safety-critical applications.
The research on additive manufacturing processes [R2] enabled Weir to use this technique to produce valves to a high precision. Additive manufacturing led to a reduction in costs because it uses less material than traditional methods (it builds a structure from scratch, rather than removing metal from block) and it is less labour intensive. The X-treme valve is now manufactured using the new processes.
Better Design Practice Across the Critical Valve Industry
Weir have embedded the knowledge in their business and used it to train their engineers, which has resulted in improved understanding of the product, faster design and improved sales. UoH extensively interacted with Weir through the KTP and one-to-one meetings and explained how they combined the mathematical models with the empirical industry standards to create novel valve designs. Weir have shared this knowledge with many businesses in the multiphase critical valve industry, including VWS Westgarth (who make water treatment plants for oil and gas installations), SBM (a service provider for the oil and gas industry based in the Netherlands) and the following firms based in the global oil capital of Houston, Texas: Wood, Worley Parsons, Exxon Mobil, Williams and Bechtel. This activity increased understanding of multiphase sizing techniques at these companies, resulting in wider incorporation.
The MD of Weir confirmed that they regularly give presentations, which are “providing the industry with more detailed knowledge of valve performance” to “leading major oil and gas contractors globally” [E1]. This has led to better awareness of the new design approach in companies such as Exxon Mobil, Equinor (a global energy company) and Worley (a service provider to the chemical industry). Feedback from a participant at one of the CPD events in 2019 indicated how they planned to change their internal processes. The Technical Director from Koso Kent Introl (an industrial valve manufacturer, formerly at Weir) said, “attending this CPD is considered [as valuable training for those working] towards attaining registered professional engineer status” [E6].
Awareness of the research findings led to interest from other flow handling companies and resulted in four further knowledge transfer partnerships with SME businesses (Koso Kent Introl (valve manufacturer), Woodcock and Wilson (centrifugal fan manufacturer), Trust Electric Heating (who make electric radiators), plus one more at Weir [E7]. The total value of these contracts to the University of Huddersfield is £955,000. It is anticipated that the total value to the businesses at the end of the KTP will be £500,000 in increased net profits and after five years will be a total of £9.65 million (based on the companies’ own calculations) [E7]. The research methodology developed when working on the first two KTP projects with Weir has been used to optimize the performance of their existing products and to develop new ones. A UoH-hosted web-based portal was launched in February 2020 and received 14,000 hits by the end of the year [E4]. All the above resulted in Hull-based Shipham Valves, one of the longest-established industrial valve manufacturers in the world, negotiating development of a KTP project.
5. Sources to corroborate the impact
[E1] Testimonial letter from MD of Trillium flow technologies formerly called Weir about the impact of the KTP.
[E2] https://fluidhandlingmag.com/news/smaller-valve-sizes-ok-for-multiphase-applications/
[E4] http://inflowsem.hud.ac.uk/ - Website used by businesses to secure online help from the UoH researchers.
[E5] KTP Final Report - Demonstrating significant benefits of the programme to the company sales and reputation.
[E6] Private email from the Technical Director, Koso Kent Introl (formally at Weir/Trillium) indicating that the approach developed by the research is now used for engineers training for professional status.
[E7] Testimonial from Knowledge Transfer Advisor, Innovate UK, on impact of the research carried out on the further funding and effects on industries in the region.
- Submitting institution
- The University of Huddersfield
- Unit of assessment
- 12 - Engineering
- Summary impact type
- Technological
- Is this case study continued from a case study submitted in 2014?
- No
1. Summary of the impact
The outputs of this University of Huddersfield research have enabled the UK rail freight industry to reduce the derailment risk for freight wagons by up to 50% via the provision of new metrics. The metrics utilised existing railway sensor infrastructure and were relevant for more than 90% of the wagons in the UK.
The research informed industry action to assuage safety concerns raised by the regulator, which was then able to remove “risk of freight derailment” from its critical issue list. Findings were included in a national standard to prevent future risk from new vehicle designs.
A new derailment risk for was identified for repurposed wagons. Otherwise redundant coal wagons are now to be modified to carry aggregates, with improved safety, savings of up to £175m over replacement rolling stock and greater network utilisation. Beneficiaries include freight operating companies (such as Freightliner), the Rail Safety and Standards Board (RSSB) and Network Rail.
2. Underpinning research
The loading of a rail freight vehicle invariably leads to a degree of imbalance. This can result in increased derailment risk, and in recent years has been identified as a causal or contributory factor in a number of derailments. The imbalance is a consequence of factors such as unevenly loaded or distributed shipping containers, loading practices for bulk products, or uneven unloading, such as discharging ballast from one side of a maintenance wagon. The number of derailment incidents, combined with the risk of future imbalance-driven derailments, caused the Office of Rail and Road (ORR) to take regulatory action and call for an industry-wide response to address the issue in 2014 [5.1]. The research described in this case study was key to enabling the industry response.
The work described below was carried out by the Institute of Railway Research (IRR) at the University of Huddersfield (UoH). It was led by Dr Phil Shackleton (Principal Enterprise Fellow at UoH since 2012), who worked with Simon Iwnicki (Director since 2012), Paul Allen (Associate Director since 2012), Yann Bezin (Head of Research since 2012), Paul Molyneux-Berry (Principal Research Fellow, 2012 to 2018), David Crosbee (Principal Enterprise Fellow since 2012) and Aniruddha Kaushal (Research Assistant, 2013 to 2016). The impact was underpinned by the contribution of UoH to three collaborative European Union research projects, SUSTRAIL [3.1, 3.2], Spectrum [3.3] and D-Rail [3.4]. In these projects, the IRR team investigated the performance of a range of freight vehicle suspension systems. Their research modelled different suspension mechanisms and simulated freight vehicle–track interactions. It focused on:
Incremental improvements of conventional running gear (wheel, bogie, suspension and brake combinations) to permit higher speed freight (2013–15) [3.1, 3.2]
The development of novel running gear concepts to exploit opportunities in the shipment of low density, high value goods (not carried by rail at the time) (2016) [3.3]
The most significant factors contributing to the likelihood of derailment and the derailment resistance of freight vehicles when exposed to those factors (2013) [3.4]
The research outputs were used to develop an improved understanding of freight suspension behaviour and the factors that influence it, and to develop a modelling methodology to investigate and quantify variation in risk performance.
Multiple research projects were undertaken for the Rail Safety and Standards Board (RSSB) (2017–19). These created a holistic approach to managing the risk posed by derailment mechanisms caused by the “imbalanced loading” of freight wagons [3.5, 3.6], across a system with multiple duty-holders. The methodology described above was implemented to study examples representative of the most common wagon types in the GB fleet, covering well over 90% of those in use, and ran 6,000 use cases to maximise its relevance and utility.
The outcomes from the research can be summarised as follows:
[a] An in-depth understanding of how imbalanced loading affects derailment resistance. This redefined operational limits for imbalanced loading. Previously, imbalances in the longitudinal and lateral directions were considered in isolation, while the revised limit more accurately accounts for the combined influence of longitudinal and lateral imbalances. This provided the justification for more optimal controls, with greater operational flexibility where possible, and more stringent limits where necessary.
[b] A metric was developed that replaced the traditional pass/fail limit with an analogue quantification of the derailment risk posed by a given vehicle imbalance. This enabled the identification of the most “at-risk” wagons on the network and for the efficacy of the industry’s risk mitigations to be quantified and trended over time.
[c] Fundamental relationships between derailment metrics were proven. This gave confidence in using convenient proxies, such as established laboratory tests to assess derailment risk, as opposed to more direct but complex methods, such as on-track tests.
[d] Some freight vehicle types were found to have a low tolerance to imbalanced loading. This meant they could pose an increased risk if “re-purposed” for a load, other than that for which they were originally intended. For example, a coal wagon partially filled with an equal (maximum) weight of aggregate can generate significant imbalance, whereas the same wagon filled to the top with coal cannot.
3. References to the research
The following outputs provide reference to the body of research and are predominantly 2* or higher, being either peer-reviewed journal articles, high-quality peer-reviewed conference papers or high-quality technical reports endorsed by industry. Authors at the University of Huddersfield at the time of publication are highlighted in bold.
[3.1] S.D. Iwnicki, S. Stichel, A. Orlova & M. Hecht (2015) Dynamics of railway freight vehicles, Vehicle System Dynamics, 53:7, 995–1033, https://doi.org/10.1080/00423114.2015.1037773
[3.2] S.D. Iwnicki, Y. Bezin, A. Orlova, P-A. Johnsson, S. Stichel, H. Schelle (2013) The ‘SUSTRAIL’ high speed freight vehicle: Simulation of novel running gear design, 23rd Symposium of the International Association for Vehicle System Dynamics (IAVSD 2013): Qingdao, China [can be supplied on request]
[3.3] P. Shackleton, Y. Bezin, D. Crosbee, P. Molyneux-Berry, & A. Kaushal (2016) Development of a new running gear for the Spectrum intermodal vehicle, 24th Symposium of the International Association for Vehicle System Dynamics, IAVSD 2015 (pp. 1461–1470). CRC Press/Balkema. https://doi.org/10.1201/b21185-153 [can be supplied on request]
[3.4] P. Allen, P. Shackleton, D. Crosbee (2013) Influence of vehicle parameters on derailment In D-RAIL D3.2 - Analysis and mitigation of derailment, assessment and commercial impact (pp. 19–72), D-RAIL Consortium, http://d-rail-project.eu/IMG/pdf/DR-D32-F3-Analysis_mitigation_derailment-assessment_commercial_impact.pdf
[3.5] P. Shackleton (2018) Simulating Offset Loading of Container Wagons on Twisted Track (Report), RSSB, https://www.sparkrail.org/Lists/Records/DispForm.aspx?ID=25630 [can be supplied on request]
[3.6] P. Shackleton (2019) Imbalanced loading of bulk wagons (Report), RSSB/University of Huddersfield Strategic Partnership, IRR/110/192/iss2 https://www.sparkrail.org/Lists/Records/DispForm.aspx?ID=26257 [can be supplied on request]
4. Details of the impact
Additional cross references [a, b, etc.] refer to the relevant outcomes in Section 2.
The impact presented in this case study contributed to an improved understanding of the derailment risk posed by imbalanced wagons by the rail industry, and led to a consequent improvement in risk management. The key beneficiary groups were freight operating companies, Her Majesty’s Inspector of Railways and the Rail Safety and Standards Board.
The impact can be divided into four subheadings:
Enabling industry response to a critical issue
Reduction in derailment risk
Commercial benefits
Contribution to standards
Enabling Industry Response to a Critical Issue
In 2014 the Office of Rail and Road (ORR) took regulatory action and called for an industry-wide response [5.1] to address the issue of freight derailments, which reached an all-time high of 14 that year. A cross-industry group was created, which commissioned research [3.5, 3.6] described in this case study and then applied the findings.
Collectively, the research findings [3.5, 3.6] were critical in helping industry provide a satisfactory response, by informing and justifying stakeholder mitigations. Such mitigations included re-orientating imbalanced containers to minimise vehicle imbalance when loaded [5.3], or providing targets so loading operators of bulk products knew where to “aim” the load, such as a painted line down the middle of a flat wagon or ensuring the peak of loaded aggregate “does not exceed 250mm from the [wagon] midline” [5.9]. As a result of risk-mitigation processes introduced across the industry, the incidence of freight imbalances dropped dramatically, with “monthly longitudinal imbalances reducing from 70 to 3 events” (2018–20) [5.4]. As a result, in May 2020 ORR removed freight derailments from their top five concerns [5.8]. The ORR confirmed the significance of the changes: “Phil Shackleton’s work [..] provided [..] the incontrovertible evidence, a measurement system and thresholds to allow the freight industry to act on this matter”, and “This has had a great positive impact. It has led to a change [in] the way in which the industry addresses this concern and has facilitated a step-change in the industry’s attitude to derailment risk”, adding “It has been an exemplar of how academia can work hand in hand with industry stakeholders to solve a complex problem” HM Chief Inspector of Railways, ORR [5.6].
Reduction in Derailment Risk
Derailment risk from imbalanced loading was reduced by up to 50% [5.2, 5.3] as the industry applied the relationship between imbalance and derailment risk [a] identified by the research [3.5, 3.6]. Freight operating companies have been able to make informed changes to their operating policies and procedures and thus substantially reduce derailment risk from imbalance, as industry representatives testify: “For Freightliner, we believe this has allowed us to take action to reduce the risk of derailment by between 30-50%, and managing this risk using the outputs and tools of Phil Shackleton’s work is now in our DNA” Professional Head of Traction & Rolling Stock Engineering, Freightliner [5.2] (2020).
“The industry’s approach to freight loading has transformed as a direct result of the work that Phil’s team did to formally identify and for the first time quantify the relationship between derailment risk and combined lateral and longitudinal load imbalances for rail freight wagons, and enabled a derailment risk reduction of 49%, despite a traffic uplift of at least 25% since 2018” Chair, Cross Industry Freight Derailment Prevention Group [5.3].
The research utilised existing, in-situ sensor and data acquisition technology to quantify the risk factor from offset loads for any freight wagon, train or wagon contents [5.4] using UoH metrics [3.5, 3.6] to quantify acceptable levels [b]. The data generated by this approach has enabled freight operators to monitor the ongoing effects of the resulting risk-mitigation strategies, using regular reports from Network Rail that quantify the risk of derailment and how it has trended over time [5.4]. Although the risk cannot be completely eliminated, recorded instances of excessive imbalance have reduced by 95% since 2018, when records were first made available to the industry [5.3, 5.4]. The historic average cost of a derailment was estimated at over £800k and the justified annual safety spend for the industry was £1m per year. There is an expectation from the regulator that, following the mitigations enabled by the IRR research and undertaken by the industry, these costs will have fallen, but at this time a new calculation has not been made. A Senior Engineer at Network Rail indicated that:
“Without the work that Phil Shackleton at the University of Huddersfield did to confirm the link between longitudinal and transverse offset loads and derailment risk, it is highly likely that freight operators would still lack the information they needed to confirm the need for change” [5.4].
Commercial Benefits
The research also led to the prevention of a new derailment risk. It was planned to repurpose redundant coal wagons to carry denser aggregates. The research identified that the wagon type was sensitive to imbalanced loading, which would not normally be possible when fully loaded ‘to the top’ with coal. However, more dense aggregates, which would not fill the wagon volume, could potentially introduce significant imbalance [d]. The research indicated that shortening the wagons to reduce the load volume to match the new payload density [3.6] would avoid the need to scrap them. The modification minimised the imbalance and hence, the derailment risk. The shorter wagons also meant there could be more wagons in a train, bringing increased operational efficiency [5.3] (as freight trains are length limited in the UK). The Cross-Industry Freight Derailment Prevention Group (XIFDPG) commented on activity since 2018:
“Put simply, this has […] saved costs and extended asset life - by finding a use for approximately 2500 redundant coal wagons (the market for coal has collapsed). It costs up to £50k to repurpose each of these coal wagons and this avoids having to buy a new aggregates wagon at £120k. If the whole fleet ends up getting repurposed, this alone will have saved the industry nearly £175m. This process has started, with 500 already complete or underway to my knowledge” [5.3]. To date the industry has expended £25m on the 500 wagons, producing assets worth £60m at a net saving of £35m.
The research findings were exploited in a number of commercial projects for freight operators including GB Railfreight, Freightliner, VTG, Network Rail, DRS and DB Cargo, as a result of their membership of the XIFDPG [5.7]. The projects [3.5, 3.6] allowed the operators to better understand the imbalances from specific payloads and then develop strategies and operational rules to reduce imbalance-driven derailment risk [5.2, 5.3]. The letter from XIFDPG confirms: “[The] model has proven equally effective in both container and bulk wagon loading and has provided the owners of the UK’s freight wagons with the confidence to make financial decisions about their rolling stock…” [5.3]. Where imbalance could not be avoided, one operator “re-allocated a fleet of aggregate wagons… to a problematic flow based on their superior predicted tolerance to imbalanced loads, using Phil’s model” [5.3].
Contribution to Industry Standards and Best Practice
The imbalance limit [a] [3.5] was used to define new acceptance test conditions in the relevant national standard (GMRT2141 i4 **[5.5]**) owned by the Rail Safety and Standards Board (RSSB); instead of measuring only the Tare (unloaded) and fully loaded behaviour of the vehicles, a number of representative use cases must now be considered. This ensures that new vehicle designs do not introduce greater risk to the network.
“Phil Shackleton’s quantification of the problem was vital to us being able to publish effective guidance for designing wagon suspension to handle the impact of offset loads and to minimise the likelihood of derailment” Director of Standards, RSSB (2020) [5.5].
The National Freight Safety Group produced a new Code of Practice (CoP), which is now used as a reference point for its 11 members [5.9] [3.6] and is freely available to download for wider use. A General Manager at Aggregate Industries wrote: “The guidelines in these sections were informed by the findings [..] and provide operators with practical rules which may be applied on a daily basis as part of ongoing, industry wide, management of derailment risk”. He continued, “This work has addressed risk areas which had not previously been clearly qualified and practices to mitigate them had not been identified” [5.10] (2020).
The CoP also led to a change in practice across the industry. The same General Manager said: “Further to the publication of the CoP a number of toolbox (local training) talks were produced by Aggregate Industries to support the communication of those key practices to relevant operators. These [were] made available within RSSB's SPARK [cross-industry system]”. The CoP and toolbox talks were shared with 90%+ of the construction materials sector and “widely communicated” to relevant staff. He concluded, “As a consequence of the CoP, and supporting research, day to day loading practices are better informed, and the derailment risks from imbalanced bulk wagons has been reduced” [5.10].
5. Sources to corroborate the impact
[5.1] Health and safety regulatory action following recent freight derailments. Letter from HM Chief Inspector of Railways, ORR to rail freight stakeholders, Evidence of Need. http://hud.ac/ict
[5.2] Risk reduction, industry change of behaviour. Source: Professional Head of T&RS Engineering, Freightliner, Letter of Support
[5.3] Derailment risk reduction and repurposing wagons. Source: Chair XIFDWG, Letter of Support
[5.4] Using UoH research to measure offset freight loads Gotcha reports and trending. Source: Senior Engineer ME, Network Rail, Letter of Support
[5.5] Importance of the research in enabling effective guidance on wagon loading Source: Director of Standards, RSSB
[5.6] Importance of research on enabling industry response to ORR. Source: HM Chief Inspector of Railways, ORR, Letter of Support
[5.7] Membership of the XIFDPG http://hud.ac/h9e
[5.8] Derailment Project – Movement to Business as Usual Report, National Freight Safety Group (NFSG), May 2020 http://hud.ac/h9d
[5.9] National Rail Safety Group – Code of Practice. Huddersfield is mentioned on pp.15/16 http://hud.acv/h9c
[5.10] Aggregate Industries – Evidence of changes in industry practice. Source: General Manager, Aggregate Industries