Study, analysis and improvement of the comfort index levels on railways
Abstract
Os futuros meios de transporte baseiam-se em soluções amigas do ambiente. A redução do impacto ambiental, congestionamento do tráfego, e a promoção de baixas emissões de CO2, são passos cruciais para alcançar a neutralidade carbónica. A indústria ferroviária encontra-se classificada como a solução mais proeminente neste objetivo. Especialmente nos países europeus e asiáticos, a maioria dos serviços ferroviários é operada por comboios eletrificados. Deste modo, as suas emissões de CO2 são residuais ao nível do utilizador final. Para manter a atual tendência e aumentar continuamente o número de passageiros dos comboios, é fundamental proporcionar viagens agradáveis. A sua avaliação baseia-se numa análise extremamente complexa da segurança e do conforto, dependendo da interação de múltiplos fatores. Independentemente da dificuldade mencionada, os fatores de conforto dependentes do funcionamento do comboio são o foco desta tese. Para quantificar adequadamente o conforto, é necessário quantificar parâmetros estáticos e dinâmicos. A análise da influência de parâmetros dinâmicos no conforto, levou à caracterização do mesmo nas viagens de longas distâncias realizadas nos comboios portugueses através da aplicação da norma ISO 2631. Com base nessa avaliação, foi sugerida uma nova metodologia para identificar anomalias na infraestrutura ferroviária e detetar possíveis necessidades de manutenção da mesma. Deste estudo resultou a identificação de doze zonas críticas. Além disso, com base numa abordagem inovadora, a identificação modal permitiu distinguir as vibrações estruturais dos assentos dos movimentos que dependem das frequências naturais da estrutura do veículo. A identificação dessas frequências está dependente do tipo de banco estudado. Foi ainda avaliada a influência da espessura da espuma na transmissão de vibrações, sendo que, espumas mais espessas são mais eficazes na mitigação de vibrações. Relativamente ao conforto estático, este foi caracterizado por metodologias subjetivas e objetivas. A influência das propriedades mecânicas da espuma, assim como, da pressão e temperaturas induzidas na interface passageiro-banco, conduziu à identificação dos elementos mais adequados para aplicações ferroviárias. O sistema identificado como a solução mas promissora para elevar os níveis de conforto em viagens de comboio é composto por uma espuma de poliuretano (80 kg/m3) e uma capa de tecido. Os resultados do presente estudo podem ajudar na conceção de novos bancos capazes de oferecer níveis de conforto elevados e, assim, promover a utilização do comboio como o principal meio de transporte.
Full text
Universidade do Minho Escola de Engenharia Patrícia Filipa Pinheiro da Silva Study, analysis and improvement of the confort index levels on railways abril de 2023 UMinho | 2023 Patrícia Filipa Pinheiro da Silva Study, analysis and improvement of the confort index levels on railways
Patrícia Filipa Pinheiro da Silva Study, analysis and improvement of the comfort index levels on railways Doctoral Thesis Doctoral Program in Mechanical Engineer Work developed under the supervision of: Prof. Dr. Eurico Augusto Rodrigues de Seabra Prof. Dr. Joaquim Gabriel Magalhães Mendes abril de 2023
ii DIREITOS DE AUTOR E CONDIÇÕES DE UTILIZAÇÃO DO TRABALHO POR TERCEIROS Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Atribuição-NãoComercial CC BY-NC https://creativecommons.org/licenses/by-nc/4.0/
iii "Science is not only a disciple of reason but also one of romance and passion." - Stephen Hawking During my PhD journey, I had the opportunity to be surrounded by people highly passionate about science and the rail industry. First, I would like to thank my supervisors, Prof. Dr. Eurico Seabra and Prof. Dr. Joaquim Mendes. None of this work would be possible without their full guidance, support and motivation. Their knowledge and passion for discovering inspired me multiple times and led me through the good and hard times. Thank you for everything. I was also blessed by all support and motivation from two railway enthusiasts, Prof. Dr. Diogo Ribeiro (ISEP) and Prof. Dr. Rui Calçada (FEUP). Thank you very much to Eng. Nuno Pinto (LESE) and Prof. Dr. Octavian Postolache (ISCTE) for all experience shared regarding experimental setups and to all Comboios de Portugal employees for their extraordinary support through the experimental campaigns. A special acknowledgement to all my laboratory colleagues and friends for their support and shared visions, especially Joana Cerqueira and Pedro Pratas. Most importantly, to my extraordinary family (particularly my Mom and Sister) and my amazing boyfriend, the most important people in my life, who never let me break down, giving me the love and strength necessary to complete this challenge. Lastly, not only an acknowledgement but also a dedication, to my Dad, who, even from the stars, guided me and was my main motivation source! This research was supported by the project iRail – Innovation in Railway Systems and Technologies Doctoral Programme funds through FCT – Fundação para a Ciência e a Tecnologia and was developed on the aim of the Doctoral grant PD/BD/143161/2019. Acknowledgements
iv STATEMENT OF INTEGRITY I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho.
v Estudo, análise e melhoramento dos níveis de conforto de comboios de passageiros Os futuros meios de transporte baseiam-se em soluções amigas do ambiente. A redução do impacto ambiental, congestionamento do tráfego, e a promoção de baixas emissões de CO2, são passos cruciais para alcançar a neutralidade carbónica. A indústria ferroviária encontra-se classificada como a solução mais proeminente neste objetivo. Especialmente nos países europeus e asiáticos, a maioria dos serviços ferroviários é operada por comboios eletrificados. Deste modo, as suas emissões de CO2 são residuais ao nível do utilizador final. Para manter a atual tendência e aumentar continuamente o número de passageiros dos comboios, é fundamental proporcionar viagens agradáveis. A sua avaliação baseia-se numa análise extremamente complexa da segurança e do conforto, dependendo da interação de múltiplos fatores. Independentemente da dificuldade mencionada, os fatores de conforto dependentes do funcionamento do comboio são o foco desta tese. Para quantificar adequadamente o conforto, é necessário quantificar parâmetros estáticos e dinâmicos. A análise da influência de parâmetros dinâmicos no conforto, levou à caracterização do mesmo nas viagens de longas distâncias realizadas nos comboios portugueses através da aplicação da norma ISO 2631. Com base nessa avaliação, foi sugerida uma nova metodologia para identificar anomalias na infraestrutura ferroviária e detetar possíveis necessidades de manutenção da mesma. Deste estudo resultou a identificação de doze zonas criticas. Além disso, com base numa abordagem inovadora, a identificação modal permitiu distinguir as vibrações estruturais dos assentos dos movimentos que dependem das frequências naturais da estrutura do veículo. A identificação dessas frequências está dependente do tipo de banco estudado. Foi ainda avaliada a influência da espessura da espuma na transmissão de vibrações, sendo que, espumas mais espessas são mais eficazes na mitigação de vibrações. Relativamente ao conforto estático, este foi caracterizado por metodologias subjetivas e objetivas. A influência das propriedades mecânicas da espuma, assim como, da pressão e temperaturas induzidas na interface passageirobanco, conduziu à identificação dos elementos mais adequados para aplicações ferroviárias. O sistema identificado como a solução mas promissora para elevar os níveis de conforto em viagens de comboio é composto por uma espuma de poliuretano (80 kg/m3) e uma capa de tecido. Os resultados do presente estudo podem ajudar na conceção de novos bancos capazes de oferecer níveis de conforto elevados e, assim, promover a utilização do comboio como o principal meio de transporte. Palavras-chave: Comboios de passageiros; Conforto; Conforto dinâmico; Conforto estático; Manutenção. Resumo
vi Study, analysis and improvement of the comfort index levels on railways Future transportation will be based on environmentally friendly solutions, reducing environmental impact and traffic congestion, and promoting low CO2 emissions targeting to achieve carbon neutrality. The railway industry is positioned as the most prominent solution to this goal. Especially in European and Asiatic countries, a predominant proportion of rail services are operated by electrified trains. Thus, its CO2 emissions are residual at the end user. To keep the positive trend and continuously increase rail passengers, it is fundamental to deliver comfortable journeys, which requires a rather complex analysis based on safety and comfort, and depends on the interaction of multiple factors. Regardless of the mentioned difficulty, the comfort factors dependent on the rail motion are the focus of this thesis. To properly quantify comfort, dynamic and static parameters need to be accessed. The analysis of the influence of dynamic parameters when sitting allowed to characterise the ride comfort of the Portuguese rail train journeys, applying ISO 2631. Based on that evaluation, a novel methodology was suggested to identify abnormalities in the rail track infrastructure and detect possible maintenance requirements. A total of twelve critical sections were identified. Moreover, based on an innovative experimental setup and analysis methodology, a modal identification may properly distinguish seat structural vibrations from those dependent on carbody natural frequencies. The identification of these frequencies depends on the seat type, being that of Alfa Pendular and Intercity trains different. The influence of foam thickness in vibration transmission was also evaluated, with thicker foams being more effective in mitigating vibrations. Static comfort was accessed by both, subjective and objective, methodologies. Foam mechanical properties, induced interface pressure, and temperature at the cover-passenger interface led to the identification of the most suitable seating system for railway applications. That identified as the most promising solution to raise comfort levels on train journeys consists of a polyurethane foam (80 kg/m3) and a fabric cover. The results of the present study may help design new railway seats capable of offering increased comfort levels and, thus, promoting rail usage. Keywords: Comfort; Dynamic comfort; Maintenance; Railways passengers; Static comfort. Abstract
vii Resumo .................................................................................................................... v Abstract .................................................................................................................. vi Table of Contents ................................................................................................... vii List of Figures .......................................................................................................... x List of Tables ......................................................................................................... xiii List of Acronyms .....................................................................................................xv Chapter I - Introduction ............................................................................................ 1 1.1 Motivation and context ..................................................................................................... 2 1.2 Research questions ......................................................................................................... 3 1.3 Thesis organization .......................................................................................................... 4 Chapter references ................................................................................................................ 6 Chapter II - Railways passenger's comfort evaluation through motion parameters ... 8 2.1 Introduction ..................................................................................................................... 9 2.2 Materials and Methods .................................................................................................. 10 2.2.1 Information sources .................................................................................... 10 2.2.2 Eligibility Criteria and Screening ................................................................... 10 2.3 Results .......................................................................................................................... 11 2.3.1 Dynamic comfort evaluation ........................................................................ 12 2.3.2 Static evaluation .......................................................................................... 28 2.4 Discussion .................................................................................................................... 33 2.5 Conclusions .................................................................................................................. 36 Chapter references .............................................................................................................. 38 Chapter III - Ride comfort evaluation of the Portuguese Railways Northern Line .... 42 3.1 Introduction ................................................................................................................... 43 3.2 WBV alternative measurement equipment validation ...................................................... 45 Table of Contents
xiv Table 6.12. Maximum and mean pressure and contact area regarding the current and older AP standard seat for subject S1. ................................................................................................ 134 Table 6.13. Temperature records regarding the different seating covers................................. 135 Table 7.1. Compilation of scientific production. ..................................................................... 147
xv 𝑎𝑃95 𝑤 95th percentile of the weighted accelerations (𝑚/𝑠2) AP Alfa pendular 𝐴0−75% Area under the loading compression cycle (𝑚𝑚2) 𝐴75−0% Area under the unloading compression cycle (𝑚𝑚2) 𝑃𝑎𝑣𝑔 Average pressure (𝑚𝑚𝐻𝑔) 𝑓 Average value of each natural frequency BMI Body mass index 𝐹25% Compressive load at 25% deformation (𝑁) 𝐹65% Compressive load at 65% deformation (𝑁) CBM Condition-based maintenance CF Crest factor DAQ Data acquisition DOF Degree-of-freedom 𝐺𝑖 Double-side square acceleration (𝑐𝑚/𝑠2)2 EFDD Enhanced frequency domain decomposition FEA Finite element analysis FSR Force sensitive resistor FDD Frequency domain decomposition 𝐵𝑖 Frequency weighting curve GPS Global positioning system HAV Hand-arm vibration 𝑓𝑚𝑎𝑥 Highest frequency componente (𝐻𝑧) ILD Indentation load deflection (𝑁) IP Infraestruturas de Portugal 𝐺𝑖(𝑓) Input acceleration spectrum at the floor level (𝐻𝑧) IC Intercities IFFT Inverse fourier transform JND Just Noticeable Difference scale 𝑃𝑚𝑎𝑥 Maximum peak pressure (𝑚𝑚𝐻𝑔) 𝑁𝑀𝑉 Mean comfort index MAC Model assurance criterion 𝑘 Multiplying factor NRS Numerical rating scale 𝑓𝑁𝑦𝑞𝑢 Nyquist frequency (𝐻𝑧) 𝐺𝑜(𝑓) Output acceleration spectrum at the seat–user interface (𝐻𝑧) PU Polyurethane PSD Power spectral density PRISMA Preferred Reporting Items for Systematic reviews and Meta-Analyses 𝑝𝑖 Pressure on its cells (𝑚𝑚𝐻𝑔) 𝑎𝑖 Rms accelerations (𝑚/𝑠2) SEAT Seat effective amplitude transmissibility SPD % Seat pressure distribution List of Acronyms
xvi SVD Singular value decomposition 𝑊𝑧 Sperling's ride index SAG factor Supportive factor 𝑆𝑦𝑦(𝑤) SVD decomposition of the PSD matrix 𝑝𝑚 The mean pressure of the 𝑛 elements (𝑚𝑚𝐻𝑔) 𝑛 Total number of nonzero cell elements 𝑎𝑣 Total vibration (𝑚/𝑠2) 𝐻(𝑓) Transmissibility function (𝐻𝑧) 𝜙𝑗∗ Vector containing the coordinates from the comfort seat mode j 𝜙𝑖∗ Vector containing the standard seat information of mode i . VDV Vibration dose value (𝑚/𝑠1.75) 𝑎𝑤 Weighted acceleration (𝑚/𝑠2) 𝑊𝑖 Weighting curves WBV Whole-body vibration
Chapter I 1 Chapter I - Introduction This document presents the thesis for the Doctoral Program in Mechanical Engineering entitled “Study, analysis and improvement of the comfort index levels on railways”. It aims to address the passenger’s comfort levels and the main influence parameters on railway journeys. The developed concepts are envisioned to be submitted as proof of the candidate's suitability to be considered for a Doctor of Philosophy degree. The work was designed and developed primarily on three research groups, namely the Mechanical Engineering and Resource Sustainability Center (MEtRICs) from the University of Minho, the Institute of Science and Innovation in Mechanical and Industrial Engineering (INEGI) from the University of Porto and the CONSTRUCT-LESE from the University of Porto. Experimental tests were performed on-board Pendulino and Intercities trains, in the maintenance facilities of Comboios de Portugal (CP) and at FEUP laboratories.
Chapter I 2 1.1 Motivation and context Transportation modes are pointed out as one of the most significant challenges for future generations. Due to an era of climate change, urban modernisation and its evolution are mostly defined by the development and bid of public transportation. Increased traffic congestion and, consequently, high CO2 release and its impact on environmental pollution led multiple governments to implement sustainable transport modes [1,2]. In this context, the European Union has established a goal to achieve climate neutrality by 2050. Thus, a reduction of carbon emissions in transportation modes by 60% is required [3]. Railways are one of the most efficient and widely used mass transportation systems for mid-range distances, also pointed out as the best strategy to reach European Union decarbonisation goals. Compared with air and car transportation, railways present superior efficiency, transportation capacity and safety, and lower environmental impact [4–8]. Moreover, since its development and service introduction, the railway industry has been linked with social development in both social and economic aspects [9]. The railway transportation importance has been sustained by a steady increase in passengers during the last decade, which will continue in the next future [7,8,10–12]. The quality of public transportation services influences the traveller's choices. Passengers with previous good experiences will probably use the same travel transportation mode again, whereas those that experienced service problems may change to a different transportation mode for their next trip. Thus, to keep the increasing trend that leads passengers to shift to railway transportation, it is crucial to identify the currently dissuading factors, raise the trains' attractivity, and provide safer and more comfortable journeys [13,14]. Journey quality is defined based on safety, comfort and user conditions [7,8]. The analysis of the passengers' comfort is rather complex, as it involves the interaction of multiple factors, such as vibration, noise, temperature, visual stimuli, illumination, smell, humidity, seat design and materials [15–20]. Although all parameters are equally essential and significantly impact the user's comfort levels due to their high complexity and the large number of tests required for their evaluation, the present work focuses on evaluating discomfort parameters induced by the vibration resulting from the train motion. Both static and dynamic properties that interfere with the perceived comfort will be analysed: ride comfort, seat effective amplitude transmissibility (SEAT) and seat transmissibility, interface pressure and foam properties, such as supportive factor (SAG factor) and hysteresis loss [8,21].
Chapter I 3 Passengers spend most of their time seated; thus, the seat affects passengers' comfort among other interior facilities. A combination of optimum dynamic properties (to minimise unwanted vibration) and the finest static performance (to decrease interface pressure) define the ideal seat [8,22]. Therefore, passengers' comfort is only properly quantified by performing static and dynamic evaluations. This thesis was motivated by the need and challenge to properly quantify and improve railway passengers' comfort levels in order to promote its use as the main daily public transportation mode. Inspired by the presently applied methodologies, new approaches were developed to fulfil the current research gaps and limitations and provide new insights on this topic. 1.2 Research questions As highlighted in sub-chapter 1.1, nowadays, when travelling by public transportation, passengers do not only expect to move from one to place to another but also demand comfortable journeys. Therefore, balancing transportation's primary function and high comfort levels is needed [23]. Derived from the train motion and the railway track-wheel interaction, vibration is particularly problematic and the leading cause of passengers' discomfort. Human vibration is defined as the physical factor effect on the human body induced by mechanical energy transmission from oscillation sources [24,25]. Regarding a train journey, vibration is transmitted to the users due to contact with the seat and floor [8,26,27]. Besides affecting comfort, vibration reduces the passenger's ability to perform simple tasks, like reading or watching a movie; therefore, investigating the seat's ability to mitigate unwanted vibration is fundamental [28–30]. Moreover, dynamic discomfort may be raised by unproperly railway track infrastructure conditions or maintenance requirements, as those may increase vibration levels at the rail vehicle carbody [31–35]. Static parameters mainly influence perceived comfort in the absence of vibration. Those parameters are primarily related to the user's feelings and the surface quality [21,22,36]. Static and dynamic comfort are firmly linked, as the influence of one depends on the other. This way, it is essential to consider both comfort approaches when evaluating comfort levels. Based on the referenced trend, some questions are imposed: - What are the current comfort levels on Portuguese railway trains? - How do dynamic and static factors influence the current comfort levels? - Can we predict maintenance needs based on comfort indexes?
Chapter I 4 These are stated as the broad research questions of this study. To achieve their answers, a study on static and dynamic parameters influencing discomfort in railway journeys was performed. Multiple trains (Alfa Pendular (AP) and Intercities (IC)) were instrumented to rate ride comfort, SEAT, seat transmissibility and modal identification, foam properties and interface pressure. Moreover, a methodology capable of identifying railway track infrastructure maintenance needs was developed based on ride comfort evaluation. This way, methodologies were developed and applied to assess both railway vehicles and infrastructures' current condition. 1.3 Thesis organization The present thesis outlines the phases taken to approach the main research questions. Each chapter is presented in research paper format and addresses the description of the developed study. Therefore, it is organised into seven chapters accordingly: - Chapter I presented a global introduction of this thesis, introducing the theme, the motivation and the research questions. - Chapter II details a systematic review regarding the evaluation of motion parameters influencing passengers' comfort and discomfort. - Chapter III intends to characterise the comfort of Portuguese AP and IC trains passengers. Initially, a vibration measurement system was validated. Then, a case study is presented regarding the ride comfort evaluation of Portuguese rail users at the Northern Line in both AP and IC trains. Moreover, the SEAT is also calculated as a complementary comfort evaluation. ISO 2631 standard methodology was employed, assessing the passenger's comfort levels in the time domain. - Chapter IV shows a railway track infrastructure maintenance requirements identification methodology based on comfort levels. As stated in Chapter III, multiple trains were instrumented. Using floor vibration measurements and the global positioning system (GPS) location of those trains, it was possible to investigate the railway track infrastructure maintenance needs due to its influence on passengers' comfort levels. - Chapter V evidences a modal identification of comfort and standard seats of AP and IC trains, corresponding to the analysis of seating behaviour in the frequency domain. That was based on output-only techniques, namely transmissibility and enhanced frequency domain decomposition (EFDD). Results highlighted the influence of seat structural frame movements, those derived from carbody natural frequencies, and the importance of foam's ability to absorb vibration.
Chapter I 5 - Chapter VI describes the analysis of static comfort by combining the influence of interface pressure and foam properties, i.e. SAG factor and hysteresis loss. Moreover, the effect of interface pressure distribution and temperature on the seat foam was estimated by employing an array of force sensitive resistor (FSR) sensors and a thermographic analysis. This way, a complete analysis of dynamic and static factors influencing passengers' comfort was performed. - Chapter VII presents the overall study conclusions and intends to answer the research questions mentioned in Chapter I. Future work and disseminated scientific articles and oral presentations are also introduced.
Chapter I 6 Chapter references [1] dell’Olio L, Ibeas A, Cecin P. The quality of service desired by public transport users. Transp Policy (Oxf) 2011;18:217–27. https://doi.org/10.1016/j.tranpol.2010.08.005. [2] Han Y, Li W, Wei S, Zhang T. Research on Passenger's Travel Mode Choice Behavior Waiting at Bus Station Based on SEM-Logit Integration Model. Sustainability 2018;10:1996. https://doi.org/10.3390/su10061996. [3] Mega VP. Sustainable Energy and Transport Systems. Conscious Coastal Cities, Cham: Springer International Publishing; 2016, p. 107–46. https://doi.org/10.1007/978-3-319-20218-1_4. [4] European Union. EU Transport in Figures. Luxembourg: 2018. [5] Grob L, Craven N, Union I. Analysis of Regional Differences in Global Rail Projects by Cost, Length and Project stage. Int Union Railw 2018:1–18. [6] De Gruyter C, Currie G, Rose G. Sustainability Measures of Urban Public Transport in Cities: A World Review and Focus on the Asia/Middle East Region. Sustainability 2016;9:43. https://doi.org/10.3390/su9010043. [7] Jiang Y, Chen BK, Thompson C. A comparison study of ride comfort indices between Sperling's method and EN 12299. International Journal of Rail Transportation 2019:1–18. https://doi.org/10.1080/23248378.2019.1616329. [8] Kim YG, Kwon HB, Kim SW, Park CK, Park TW. Correlation of ride comfort evaluation methods for railway vehicles. Proc Inst Mech Eng F J Rail Rapid Transit 2002;217:73–88. https://doi.org/10.1243/095440903765762823. [9] European Commission. European year of rail 2021. [10] Railway passenger transport statistics - quarterly and annual data. Eurostat - Statistics Explained 2021. [11] Statistics Explained. Railway passenger transport statistics-quarterly and annual data. Eurostat 2020. [12] Office of Rail and Road. Passenger Rail Usage 2020-21 Quarter 3. 2021. [13] Eboli L, Mazzulla G. A New Customer Satisfaction Index for Evaluating Transit Service Quality. J Public Trans 2009;12:21–37. https://doi.org/10.5038/2375-0901.12.3.2. [14] Lin C-Y, Chen L-J, Chen Y-Y, Lee W-C. A comfort measuring system for public transportation systems using participatory phone sensing. Proceedings of PhoneSense 2010. [15] Wibowo RKK, Soekarno S, Puspitasari I. Analysis of Train Passenger Seat Using Ergonomic Function Deployment Method. International Journal of Transport and Vehicle Engineering 2017;11:1747–51. https://doi.org/https://doi.org/10.5281/zenodo.1132711. [16] Bosso N, Gugliotta A, Zampieri N. Design and simulation of a railway vehicle for the transport of people with reduced mobility. Shock and Vibration 2018. https://doi.org/10.1155/2018/9207639. [17] Zhang L, Helander MG, Drury CG. Identifying factors of comfort and discomfort in sitting. Hum Factors 1996;38:377–89. https://doi.org/10.1518/001872096778701962. [18] Groenesteijn L, Hiemstra-van Mastrigt S, Gallais C, Blok M, Kuijt-Evers L, Vink P. Activities, postures and comfort perception of train passengers as input for train seat design. Ergonomics 2014;57:1154– 65. https://doi.org/10.1080/00140139.2014.914577. [19] Silva P, Ribeiro D, Mendes J, Seabra EAR, Postolache O. Railways Passengers Comfort Evaluation through Motion Parameters: A Systematic Review. Machines 2023;11:465. https://doi.org/10.3390/machines11040465. [20] Silva P, Mendes J. Passengers Comfort Perception and Demands on Railway Vehicles: A Review. KnE Engineering 2020. https://doi.org/10.18502/keg.v5i6.7039.
Chapter I 7 [21] Ebe K, Griffing M. Factors affecting static seat cushion comfort. Ergonomics 2001;44:901–21. https://doi.org/https://doi.org/10.1080/00140130110064685. [22] Ebe K, Griffin M. Quantitative prediction of overall seat discomfort. Ergonomics 2000;43:791–806. [23] Cucu L, Simion I, Carzola M, Stoica G, Cananau S. Ergonomics and accessibility in passenger trains. UPB Scientific Bulletin, Series D: Mechanical Engineering 2016;78. [24] Sinik V, Despotovis Z, Jankovic S. Exposure to whole-body vibration, tools for calculating daily exposures and measurement. V International Conference - Ecology of Urban Areas 2016. [25] Issever H, Aksoy C, Sabuncu H, Karan A. Vibration and its effects on the body. Medical Principles and Practice 2003;12:34–8. https://doi.org/10.1159/000068155. [26] Gong W, Griffin MJ. Measuring, evaluating and assessing the transmission of vibration through the seats of railway vehicles. Proc Inst Mech Eng F J Rail Rapid Transit 2018;232:384–95. https://doi.org/10.1177/0954409716671547. [27] Năstac S, Picu M. Evaluating Methods of Whole-Body-Vibration Exposure in Trains. The Annals of “Dunarea de Jos” 2010. [28] Basri B, Griffin MJ. The application of SEAT values for predicting how compliant seats with backrests influence vibration discomfort. Appl Ergon 2014;45:1461–74. https://doi.org/https://doi.org/10.1016/j.apergo.2014.04.004. [29] Patelli G, Griffin MJ. Effects of seating on the discomfort caused by mechanical shocks: Measurement and prediction of SEAT values. Appl Ergon 2019;74:134–44. https://doi.org/10.1016/j.apergo.2018.08.003. [30] Silva P, Ribeiro D, Mendes J, Seabra E. Modal Identification of Train Passenger Seats Based on Dynamic Tests and Output-Only Techniques. Applied Sciences 2023;13:2277. https://doi.org/10.3390/app13042277. [31] Falamarzi A, Moridpour S, Nazem M. A Review on Existing Sensors and Devices for Inspecting Railway Infrastructure. Jurnal Kejuruteraan 2019;31:1–10. https://doi.org/10.17576/jkukm-201931(1)-01. [32] Ji A, Woo WL, Wong EWL, Quek YT. Rail track condition monitoring: a review on deep learning approaches. Intelligence & Robotics 2021. https://doi.org/10.20517/ir.2021.14. [33] Ferrante C, Bianchini Ciampoli L, Benedetto A, Alani AM, Tosti F. Non-destructive technologies for sustainable assessment and monitoring of railway infrastructure: a focus on GPR and InSAR methods. Environ Earth Sci 2021;80:806. https://doi.org/10.1007/s12665-021-10068-z. [34] Kaewunruen S, Remennikov A. Dynamic properties of railway track and its components : a state-ofthe-art review. In: Weiss B, editor. New Research in Acoustics, Nova Science Publishers; 2008, p. 197– 220. [35] Fesharaki M, Wang T-L. The Effect of Rail Defects on Track Impact Factors. Civil Engineering Journal 2016;2:458–73. https://doi.org/10.28991/cej-2016-00000049. [36] Park SJ, Min SN, Lee H, Subramaniyam M, Suh WS. Express Train Seat Discomfort Evaluation using Body Pressure and Anthropometric Data. Journal of the Ergonomics Society of Korea 2014;33:215–27. https://doi.org/10.5143/jesk.2014.33.3.215.
Chapter II 14 Figure 2.2. Acceleration measurement locations defined by ISO 2631 and respective axis orientation. (Adapted from [24]). The root-mean-square (rms) acceleration is calculated for each axis, and the corresponding weighting curve is applied [2,19–21,24]. The weighting process is calculated following Equation (2.1): 𝑎𝑤= [∑(𝑊𝑖𝑎𝑖)2]12 ⁄ (2.1) where 𝑊𝑖 represents the weighting frequencies and 𝑎𝑖 the rms accelerations, weighting curves application depends on the measurement location and purpose. The total vibration (𝑎𝑣) is calculated according to Equation (2.2): 𝑎𝑣= (𝑘𝑥 2𝑎𝑤𝑥 2+ 𝑘𝑦 2𝑎𝑤𝑦 2 + 𝑘𝑧 2𝑎𝑤𝑧 2)12 ⁄ (2.2) where 𝑎𝑤 are the rms accelerations for each axis, and 𝑘 represents the multiplying factor dependent on the measuring position, presented in Table 2.1 regarding the comfort approach. Table 2.1. Weighting curves and multiplying factors applied following the comfort approach for a seated passenger defined by ISO 2631. X - axis Y - axis Z - axis Floor Wk and kx = 0.25 Wk and ky = 0.25 Wk and kz = 0.40 Seat surface Wd and kx = 1.0 Wd and ky = 1.0 Wk and kz = 1.0 Seatback Wc and kx = 0.80 Wd and ky = 0.50 Wd and kz = 0.40 Finally, based on 𝑎𝑣, the discomfort is assessed by a defined scale (see Table 2.2) where accelerations higher than 0.315 m/s2 are ranked as uncomfortable.
Chapter II 15 Table 2.2. ISO 2631 comfort evaluation scale. 𝒂𝒗(𝒎/𝒔𝟐) Ride comfort ≤ 0.315 Not uncomfortable 0.5 – 0.63 Little uncomfortable 0.63 – 0.8 Little uncomfortable to fairly uncomfortable 0.8 – 1.0 Fairly uncomfortable to uncomfortable 1.0 – 1.25 Uncomfortable 1.25 – 1.6 Uncomfortable to very uncomfortable 1.6 – 2.0 Very uncomfortable 2.0 – 2.5 Very uncomfortable to extremely uncomfortable ≥ 2.5 Extremely uncomfortable Large fluctuations in acceleration and frequency levels may arise throughout a train ride due to the track section's condition, rail irregularities, braking or speed restrictions. ISO 2631 is suitable for evaluating WBV when minor frequency-level variations occur. To overcome the standard major limitation, the vibration dose value (VDV) was introduced. This method uses the fourth power of acceleration instead of the second power; thus, it emphasises acceleration peaks [1,2,24]. The VDV is calculated as follows in Equation (2.3): 𝑉𝐷𝑉 = [∫ [𝑎𝑤(𝑡)]4 𝑑𝑡 𝑡2 𝑡1 ]14 ⁄ (2.3) The statistical analysis introduced by the EN 12299 standard was developed based on the rms methodology. EN 12299 standard defines ride comfort based on two methods, the standard and the complete. The former considers only floor vibration (three directions). In contrast, the latter uses both floor (vertical acceleration) and seat locations (vertical and lateral at the seat surface and longitudinal at the backrest). Thus, the standard method can be classified as a simplification of the complete one [25]. Moreover, although this standard presents several methods to calculate different passenger comfort indices, the Mean Comfort Index (𝑁𝑀𝑉) is the one often used. 𝑁𝑀𝑉 quantifies the passenger mean comfort during a continuous 5-minute run. This way, the measurement duration shall be a multiple of five, and a minimum of four travelled zones at constant speed must be accomplished to apply the method [1,26,27]. In contrast to the ISO 2631 method, weighting curves are applied initially. Then the rms acceleration over 5 seconds is calculated for each direction (longitudinal, lateral, and vertical) throughout the tested track. Finally, the 95th percentile (i.e. the 4th highest value) is determined for periods of 5 minutes, allowing the achievement of the index for each axis [1,26]. The partial 𝑁𝑀𝑉 are calculated as follows,
Chapter II 16 𝑁𝑀𝑉𝑋 =6 (𝑎𝑋𝑃95 𝑊𝑑) (Longitudinal) (2.4) 𝑁𝑀𝑉𝑌 =6 (𝑎𝑌𝑃95 𝑊𝑑) (Lateral) (2.5) 𝑁𝑀𝑉𝑍 =6 (𝑎𝑍𝑃95 𝑊𝑘) (Vertical) (2.6) where, 𝑎𝑃95 𝑤 represents the 95th percentile of the weighted accelerations in the three directions, x, y and z. By combining the acceleration in the three directions, a global comfort index is obtained, as follows in Equation (2.7), 𝑁𝑀𝑉 =6√(𝑎𝑋𝑃95 𝑊𝑑)2 + (𝑎𝑌𝑃95 𝑊𝑑)2 + (𝑎𝑍𝑃95 𝑊𝑘)2 (2.7) The evaluation of 𝑁𝑀𝑉 is defined by an established scale, Table 2.3, which considers values between 1 and 5. Values under 1 are considered a "very comfortable ride", and for results above 5 the journey is ranked as "very uncomfortable ride" [27]. Table 2.3. EN 12299 evaluation scale. 𝑁𝑀𝑉 Ride comfort ≤ 1 Very comfortable 1 – 2 Comfortable 2 – 4 Medium 4 – 5 Uncomfortable ≥ 5 Very uncomfortable The main advantage of this method is that it avoids sensitivity to artefactual extremes. On the other hand, the use of the 95th percentile may lead to data exclusion (loss of information) and some doubtful analysis [1]. Kufver et al. [27] considered three different hypothetical 5-minute vibration patterns. In opposition to what was expected, the three patterns were demonstrated to be equally comfortable. Moreover, the lack of possibility to correspond the track irregularity's location with the 𝑁𝑀𝑉 values (the highest 𝑁𝑀𝑉 values can occur during three different 5 seconds time intervals) represents another significant limitation [27]. Finally, the test defines that it should be executed during four blocks of 5 minutes while keeping a constant speed. These conditions are difficult to achieve during passenger service and do not reflect the natural in-service circumstances [25,27]. Sperling proposed an alternative ride comfort evaluation method, fundamentally distinct from the methods based on rms analysis. The Sperling's ride index (𝑊𝑧) is evaluated individually for longitudinal, lateral and vertical directions. Moreover, 𝑊𝑧 is calculated for defined time intervals or track sections [1,22] according to the following Equation (2.8): 𝑊𝑍𝑖 = [∫ 𝐺𝑖(𝑓) 30 0.5 𝐵𝑖2(𝑓) 𝑑𝑓]16.67 ⁄ (2.8)
Chapter II 17 where 𝐺𝑖 corresponds to the double-side square acceleration [(cm/s2)2] and 𝐵𝑖 represents the frequency weighting curve. As with the previous methods, passenger comfort is evaluated by a scale, Table 2.4, where passengers will not feel discomfort for values below 3 and will feel extreme discomfort for results higher than 3.5 [2]. Table 2.4. Sperling's method comfort evaluation scale. 𝑾𝒛 Ride comfort 1 Just noticeable 2 Clearly noticeable 2.5 More pronounced but not unpleasant 3 Strong, irregular but still tolerable 3.25 Very irregular 3.5 Extremely irregular, unpleasant, annoying; prolonged exposure intolerable 4 Extremely unpleasant; prolonged exposure harmful The particularity of being evaluated based on a fixed number makes Sperling's method the most convenient for comparing different conditions or trains. Multiple authors agree with that statement and highlight that when compared with ISO 2631 methodology, Sperling's method is easier to apply and understand. However, the ISO methodology is highly superior in accuracy and results quality. Additionally, this method assesses vibration individually for all directions, being generally employed to evaluate the carriage's vibration level instead of the passenger's comfort. Sperling's method's primary limitation is that vibration influences in distinct frequency bands and directions in sitting comfort are neglected [22,23,28,29]. Multiple authors apply the three ride comfort analysis methodologies around the world. Liu et al. [25] applied ISO 2631 to investigate the effect of train speed and track geometry on ride comfort in highspeed trains (speed up to 400km/h). Initially, the authors take real measurements at a high-speed train (steady speed of about 210 km/h) on the floor, seat surface and backrest. The data showed a total acceleration of 0.358 m/s2, corresponding to a "little uncomfortable" ride. Then, using the threedimensional multi-body vehicle model developed by Wickens and Huang, Liu et al. expanded its conclusion for speed up to 400 km/h. The authors stated that the track's vertical irregularities are the predominant source of vibration discomfort, even though the lateral irregularities also develop to be essential as speed increases. Moreover, it was concluded that as speed increases from 200 to 300 km/h, the total acceleration increases by a factor of 2. That factor rises to 3.5 when the speed rises from 200 km/h to 400 km/h [25]. Similar conclusions were drawn by Peng et al. [23] when studying the ride
Chapter II 18 comfort on high-speed (300 km/h) railways connecting Shanghai-Kunming in China. The authors applied three-axial accelerometers on the seat surface and seatback. Maximum total accelerations of 0.12 m/s2 were obtained. Results demonstrated higher discomfort provoked by the vertical acceleration on both seat surface and seatback. Therefore, vertical vibration had the most significant impact among the three directions (vertical, lateral, and longitudinal). Furthermore, Peng et al. [23] also demonstrated the significant influence of tunnels and running speed on passengers' ride comfort [23]. In the research developed by Johanning et al. [30], when evaluating the WBV exposure of locomotive engineers under normal operating conditions, the same conclusions as Liu et al. [25] and Peng et al. [23] were obtained. By placing three-axial accelerometers at both floor and seat surface, higher weighted accelerations were found for the vertical axis, again demonstrating this direction's significant influence on ride comfort [30]. Cheng and Hsu [31] numerically evaluated the effect of the vehicle speed and tilting angles based on a tilting railway vehicle modelled by a 27-degree-of-freedom (DOF) car system. Regarding the titling vehicle, as the speed rises, the ride comfort also rises. On the contrary, the non-tilting train ride comfort initially rises, decreasing subsequently when vehicle speed increases. Concerning the tilting angles, higher discomfort was found for higher angles. Ride comfort also appears to be sensitive to rail irregularities, as it increases when those are introduced [31]. To ascertain the vibration limit levels that affect sedentary activities, Khan and Sundstrom [22] instrumented second-class seats of three different Inter-regional Swedish trains, particularly the IR-train, Y2-train and X50-train. The authors conducted a survey questionnaire and calculated both 𝑎𝑣 and 𝑊𝑧. While 𝑎𝑣 results noticed "not uncomfortable" rides, the same measurements but evaluated with 𝑊𝑧 ranked the journeys as "more pronounced but not unpleasant". The results highlighted the differences between both standards' applications. Regarding the vibration effect on sedentary activities, even with the low levels of vibration demonstrated by the ride comfort methodologies, 47% of the passengers described that shocks and vibration were their leading causes of disturbances. Moreover, 60% of passengers reported moderate difficulties in performing a short writing test, and 12% noticed great difficulty in performing that test. Those results show that even low vibration levels can cause substantial limitations for rail users [22]. Jiang et al. [1] tried to establish a relationship between 𝑁𝑀𝑉 and 𝑊𝑍 indices. For that purpose, the authors installed three-axial accelerometers on the floor of Australian N-series passenger carriages. As in the studies mentioned above, both methods agreed on the vertical direction vibration effect on passengers' comfort caused by track irregularities. However, the vibration ranking did not agree as the vibration evaluated based on 𝑊𝑍 revealed higher discomfort levels than those of 𝑁𝑀𝑉 [1]. Dimitriu and Leu [32] compare similarly to Jiang et al. [1] but on a mechanical model travelling on a track with
Chapter II 19 vertical irregularities. The vehicle consisted of a rigid-flexible coupled model, including an elastic beam for the carbody and six rigid bodies for bogies and axles. Initially, the authors observed that ride comfort indices increase in non-uniform growth when speed increases. The non-uniform growth is justified by the geometric filtering effect, which is a critical feature regarding the vertical vibration behaviour in railway vehicles. Then, the critical comfort points were identified. Indeed, at high speed (250 km/h), both bogies and carbody centre are comparable when applying EN 12299 standard, demonstrating results of 1.91 𝑁𝑀𝑉 at the carbody centre, 1.82 𝑁𝑀𝑉 above bogie 1 (front), and 1.99 𝑁𝑀𝑉 above bogie 2 (rear). As in the other mentioned studies, Sperling's method results are slightly higher than those of Mean Comfort, presenting 2.0 𝑊𝑍 at the carbody centre, 2.7 𝑊𝑍 above bogie 1 and 3.1 𝑊𝑍 above bogie 2 [32]. Munawir et al. [33] also developed a study to compare Sperling's method and the EN 12299 ride comfort indices on the Ampang line route in Malaysia. Similar evaluations were found for both EN 12299 and Sperling's method. Mostly, the journey was considered as comfortable. The highest discomfort levels were 3.34 for 𝑁𝑀𝑉 and 2.63 for 𝑊𝑍, respectively [33]. Also, investigating the comparison between Sperling's method and Mean Comfort, Dimitriu and Stanica [34] predicted, through a numerical simulation, the effects of speed and train suspension on the ride comfort index. The vehicle model consisted of a rigid-flexible coupled model, Figure 2.3. The carbody was represented by an equivalent Euler-Bernoulli beam, and the bogies and wheelsets were modelled through rigid bodies. Utilising a speed interval between 0-240 km/h, the authors found a relationship between increasing speed and low comfort levels, exposing lower comfort levels when speed increases. Additionally, when analysing the three reference points (one at the carbody centre and two above each bogie) for speed until 80 km/h, the three points demonstrated similar comfort levels on both methods. As speed increases above this threshold, the discomfort levels are higher at bogies (1.72 𝑁𝑀𝑉 above bogie 1 and 1.86 𝑁𝑀𝑉 above bogie 2 for 240 km/h) than at the carbody centre (1.68 𝑁𝑀𝑉 at 240 km/h). Under the same running conditions, higher 𝑊𝑍 were noticed, as the carbody centre demonstrated 1.9 𝑊𝑍 discomfort, bogie 1 discomfort equal to 3.0, and 2.7 𝑊𝑍 regarding bogie 2. Therefore, significant differences between both ride comfort assessment methods were reported, being that of 𝑊𝑍 producing a weaker ride comfort than the 𝑁𝑀𝑉. For the suspension effect, employing values of damping ratio within the range of 0.10 to 0.40, it was found that lower levels of damping will result in a lower ride comfort index, corresponding to higher comfort levels [34].
Chapter II 20 Figure 2.3. Vehicle model used by Dimitriu and Stanica. (Adapted from [34]). Kim et al. [2] tried to compare ISO 2631, EN 12299 and Sperling's method comfort levels. To achieve this goal, the authors recorded acceleration measurements on Korean high-speed trains, which were used as vibration models of railway vehicles. Based on those models, equations of fitted curves between two ride comfort indices were obtained, namely 𝑊𝑍− 𝑎𝑣, 𝑊𝑍− 𝑁𝑀𝑉 and 𝑁𝑀𝑉 −𝑎𝑣. However, the authors did not report values of ride comfort indices [2]. Although all three ride comfort analysis methodologies present some advantages and limitations, the ISO 2631 standard seems to be the most employed one. The methods share the application of frequency-dependent filters and acceleration records at the passenger-seat-car body interfaces. That approach involves a set of dedicated accelerometers installed in multiple locations, which may be challenging. Indeed, the passengers' involuntary movements may modify the accelerometers' positioning, compromising the measurements' quality. Nevertheless, when performed adequately, this approach constitutes the more adequate one to evaluate passenger comfort because it considers a more realistic model of seat structure dynamics [35–37]. Tables 2.5, 2.6 and 2.7, respectively, resume the 𝑎𝑣, 𝑁𝑀𝑉 and 𝑊𝑧 results reported in the aforementioned studies.
Chapter II 21 Table 2.5. ISO 2631 results resumed from the retrieved literature, with respective measurement position (floor, seat surface and seatback), total acceleration and discomfort level and the type of experiment (experimental or numerical simulation). Ref. Measurement position 𝒂𝒗(𝒎/𝒔𝟐) Discomfort level Type of experiment [17] Floor 0.06 Not uncomfortable Experimental (IR-train) Seat surface 0.24 Not uncomfortable Seatback 0.18 Not uncomfortable Floor 0.16 Not uncomfortable Experimental (Y2-train) Seat surface 0.41 Little uncomfortable Seatback 0.28 Not uncomfortable Floor 0.07 Not uncomfortable Experimental (X50-train) Seat surface 0.28 Not uncomfortable Seatbac\k 0.17 Not uncomfortable [23] Seat surface 0.12 Not uncomfortable Experimental (high-speed train) Seatback 0.12 Not uncomfortable [25] Seat surface 0.358 Little uncomfortable Experimental (high-speed train) [30] Seat surface 0.59 Little uncomfortable Experimental (Locomotives) Table 2.6. EN 12299 yielded literature resume, with respective measurement position (floor, seat surface and seatback or carbody centre, front bogie and rear bogie), total acceleration and discomfort level and the type of experiment (experimental or numerical simulation). Ref. Measurement position 𝑵𝑴𝑽 Discomfort level Type of experiment [1] Floor 0.5 Very comfortable Experimental (N-series passenger carriage) [32] Carbody centre 1.91 Comfortable Numerical simulation Front bogie 1.82 Comfortable Rear bogie 1.99 Comfortable [33] Seat surface 3.34 Medium Experimental (Light rapid train) [34] Carbody centre 1.68 Comfortable Numerical simulation Front bogie 1.72 Comfortable Rear bogie 1.86 Comfortable
Chapter II 22 Table 2.7. Sperling’s method presented literature resume, with respective measurement position (floor, seat surface and seatback or carbody centre, front bogie and rear bogie), total acceleration and discomfort level and the type of experiment (experimental or numerical simulation). Ref. Measurement position 𝑾𝒁 Discomfort level Type of experiment [1] Floor 1.5 Clearly noticeable Experimental (N-series passenger carriage) [32] Carbody centre 2.0 Clearly noticeable Numerical simulation Front bogie 2.7 Strong, irregular but still tolerable Rear bogie 3.1 Strong, irregular but still tolerable [33] Seat surface 2.63 Strong, irregular but still tolerable Experimental (Light rapid train) [34] Carbody centre 1.9 Clearly noticeable Numerical simulation Front bogie 3.0 Strong, irregular but still tolerable Rear bogie 2.7 Strong, irregular but still tolerable 2.3.1.2 Seat Effective Amplitude Transmissibility On a rail journey, passengers spend the majority of their time seated. Thus, the vibration felt by users depends on the seat's dynamic behaviour, namely its vibration isolation performance. Seats with poor dynamic behaviour will lead to unpleasant ride comfort [38]. SEAT value is a complementary method used to estimate dynamic seat comfort. This indicator of seat isolation efficiency reflects the extent to which a seat amplifies or attenuates vibration transmission [39]. The extent to which a seat can provide vibration attenuation relies on three factors: the spectrum of vibration (characterised by the vehicle motion), how the seat modifies the vibration spectrum (seat vibration transmissibility) and the sensitivity of the human body to the vibration range causing discomfort. This way, SEAT reports how those three parameters fluctuate with the frequency and vibration direction to estimate the seat's vibration isolation [38,40,41]. SEAT compares the vibration discomfort when sitting on a rigid seat to discomfort on a non-rigid seat [41]. Thus, SEAT is specified as the ratio between the VDV measured on the seat and the VDV assessed on a rigid support beneath the seat surface, according to Equation 2.9: 𝑆𝐸𝐴𝑇 % = 𝑉𝐷𝑉𝑠𝑒𝑎𝑡 𝑉𝐷𝑉𝑓𝑙𝑜𝑜𝑟 × 100 (2.9)
Chapter II 23 where 𝑉𝐷𝑉𝑠𝑒𝑎𝑡 represents the vibration dose value measured at the seat, and 𝑉𝐷𝑉𝑓𝑙𝑜𝑜𝑟 means those vibration dose values obtained at the floor. Depending on vibration characteristics, it is possible to calculate SEAT by either VDV or weighted rms acceleration (𝑎𝑤). The vibration Crest Factor (CF) determines how impulsive a measure is by calculating the ratio of the peak weighted acceleration (max |𝑎𝑤|) to its rms value (𝑎𝑤), as in Equation (2.10): 𝐶𝐹= 𝑚𝑎𝑥|𝑎𝑤| 𝑎𝑤 (2.10) Data with crest factors under nine (low crest factors) can be evaluated with 𝑎𝑤, while data with higher crest factors must be evaluated based on VDV. Nevertheless, SEAT evaluation is rather performed using VDV once this highlights the vibration peaks effect [42]. A SEAT value greater than 100% implies that the seat magnifies vibration transmission; thus, the discomfort the user feels on the seat is higher than that of a rigid seat. On the other side, SEAT values lower than 100% indicate vibration mitigation by the seat, and therefore, the seat provides proper vibration attenuation and increases passengers' comfort. The seat does not influence vibration transmission if a SEAT result equals 100%. Due to the foam cushion's inability to absorb low-frequency vibrations, typically dominant in the vertical direction of railways, the vertical SEAT of those vehicle seats is generally higher than 100% [38]. Seats are designed to have the lowest SEAT value, yet this must fit other constraints. Reducing the SEAT value may lead to a too-soft or hard seat, resulting in poor static comfort for the passengers [38]. Therefore, the optimum SEAT value will not eliminate the vibration transmission to the user but will minimise vibration exposure [38,40]. SEAT values can be achieved by direct acceleration measurements at rail seats and floor, laboratory experiments where the railway environment is mimicked, or even by estimation using vibration records and transmissibility functions of the seat/user combination [39,42]. Multiple authors applied SEAT to measure or predict dynamic seat behaviour. Gong and Griffin [38] used SEAT to understand vibration transmission through railway passenger seats, specifically in each direction. Based on a laboratory experiment, a double train seat unit (similar to subway seats) was fixed to a six-axis motion simulator and instrumented with two three-axial accelerometers at both the seat surface and seatback. Three types of stimuli were applied: simulated train vibration, three-axial random vibration, and single random vibration. The authors calculated, measured, and predicted SEAT values for each direction (longitudinal, lateral and fore-and-aft). Results were similar for the three axes when applying three-axial and single-axis random vibration. For the same stimulus, the SEAT at the backrest
Chapter II 30 lower pressure magnitudes (maximum and mean pressures) and lower contact areas. In opposition, the obese subject generates the highest contact area. Results demonstrated that contact area would linearly increase when subjects' weight and height increased. Due to its larger contact area, the obese subject also had the lowest SPD % of all subjects. Therefore, larger contact areas represent a more distributed pressure and, consequently, lower discomfort levels [60]. Akgunduz et al. [66] examined the influence of knee angle on seat interface pressure. Results suggested that increasing the angle from 95° to 115° and up to 135° decreases the pressure around ischial tuberosities and increases pressure beneath the thighs. This way, pressure is distributed along the seat surface, reducing pressure magnitudes. The increased contact area also supports this fact [66]. Xu et al. [59] discussed the association between local comfort and whole-body comfort. Four body parts were defined: back, waist, hips, and thighs. Pressure measurements and a subjective ranking assessed discomfort. Results highlighted the hips as the most significant body part affecting whole-body discomfort, followed by the back, waist and thighs, Figure 2.7 (a) [59]. A similar study by Peng et al. [62] revealed the passengers sitting local and overall comfort degradation mechanism in high-speed railways. Pressure map sensors measure the interface pressure between the user and the seat on a second-class Chinese high-speed rail seat, Figure 2.7 (b). At every 10 minutes of the experiment, passengers evaluated discomfort based on a subjective ranking. Local comfort was divided into eight parts: shoulders, mid-back, side back, waist, buttocks, upper, side and lower thigh. Results demonstrated that the comfort of the shoulders, waist and buttocks mainly influence overall comfort. In contrast, the side back, upper thigh and lower thigh noticed the lowest correlation with overall comfort. Comfort degradation was defined based on sitting time. During the initial 20 minutes of sitting, the seat negatively influences the comfort degradation rate, with fatigue appearing after this time [62]. Yuxie et al. [3] also analysed sitting duration's influence on overall comfort on high-speed railways. By evaluating comfort with both objective (Tekscan body pressure measurement system) and subjective (numerical rating scale (NRS)) methods, the authors concluded that higher sitting duration leads to higher discomfort levels [3].
Chapter II 31 (a) (b) Figure 2.7. Interface pressure experiments (a) typical body pressure distribution (adapted from [59]); (b) experimental setup employed by Peng et al. on a Chinese high-speed rail 2nd class seat (adapted from [62]). Regarding foam characteristics, the hardness, hysteresis loss, SAG factor and foam shape may interfere with pressure distribution and are generally presented as static comfort properties. Seat hardness seems to be the factor that mainly affects pressure distribution comfort. Multiple authors stated that softer seats lead to higher comfort levels. Indeed, a soft seat contact area is considerably larger than that on a rigid seat; this way, pressure peaks are minimised by the soft seat. Moreover, a more rigid foam is less flexible than a soft one, which induces higher pressure around the ischial tuberosities. Therefore, while on a rigid seat, the contact area is caused by the subject's buttocks' deformation; on a soft seat, both seat and buttocks deformations contribute to enlarging the contact area [9, 60,66].
Chapter II 32 Commonly, the foam hardness is obtained following ISO 2439 standard specifications, particularly the 25% indentation load deflection (ILD) hardness. That standard also defines the calculation of the SAG factor, while ISO 3386 defines the foam hysteresis loss. Seat materials require a high SAG factor and low hysteresis values [9,57,67]. The SAG factor represents comfort while seated and is determined as the ratio of the stress value at 65% strain with that of 25%. A higher than 2.8 SAG factor is proposed for seating applications. At lower values, bottoming may occur as the passengers feel the seat pad bottom [9,57]. Hysteresis loss is calculated as the percentage ratio in energies between loading (0% - 75%) and unloading (75% - 0%) compression cycles. Energy is obtained based on the area under the stress-strain curve. This way, hysteresis loss represents the passengers' comfort during de-seating [57]. This property is highly influenced by the cellular foam structure, particularly the cell wall area ratio, once energy loss occurs mainly due to the buckling affecting the unrecovered cell walls and struts. Therefore, low cell wall area ratios are required to obtain low hysteresis values [57]. Figure 2.8 illustrates the influence of crosslinking agent found by Choi and Kim [57] on hysteresis loss evaluation, where the BHMTA agent promoted the lowest cell wall area ratio. Figure 2.8. Influence of crosslinking agent on hysteresis loss evaluation. (Adapted from [57]). Ebe and Griffin [9] related multiple static physical foam characteristics to subjects' impressions of sitting comfort. Based on four foams with equal hardness but different densities (45, 52, 55 and 65 kg/m3), the influence of density, hysteresis loss and SAG factor on sitting comfort was investigated. The low-density foam (45 kg/m3) was judged as the least comfortable foam, whereas a high-durability foam (55 kg/m3) was ranked as the most comfortable seat. The same foams presented 34.3% and 29% hysteresis loss, respectively. This way, as expected, it was observed that the foam with the lowest
Chapter II 33 hysteresis loss was classified as the most comfortable. Moreover, from the investigated factors, the hysteresis loss revealed the lowest correlation with comfort evaluations, whereas a higher correlation between seat comfort and SAG factor was obtained [9]. Tang et al. [68] developed an experiment to indicate which seat cushion foam shape model is preferred based on discomfort analysis. Three surface shape models were produced with equal thickness and material: (1) bilateral protruding cushion, (2) front protruding cushion and (3) flat cushion. A pressure sensor system and a subjective body region discomfort questionnaire quantified discomfort. Static pressure distribution among the three cushions showed a significant difference, where foam (2) presented the highest-pressure peaks, followed by (1), whereas cushion (3) presented the lowest peaks. On the subjective questionnaire, the highest overall comfort was achieved by foam (3) and the lowest by (1). Thus, the flat cushion shape was considered the most suitable foam shape for high-speed railways seats [68]. 2.4 Discussion Multiple factors influence railways passengers' comfort; being those dependent on the vehicle and its motion the primary research concern. Passengers' comfort relies on the static and dynamic vehicle and seat characteristics, depending on the presence or absence of vibration. The influence of both comfort types depends on the environmental conditions. For low vibration magnitudes, comfort relies on static characteristics, while in opposite conditions, when vibration magnitude increases, discomfort is dominated by dynamic characteristics [10]. Dynamic comfort is evaluated by ride comfort indexes, SEAT and transmissibility, whereas static comfort is mainly assessed by measuring the seat-user interface pressure, which highly depends on the user's anthropometric characteristics and foam properties. Ride comfort defines the human tolerance to vibration exposure over time. Besides several countries applying their own comfort evaluation standards, those were developed based on three principal methodologies/standards specially developed for that purpose, ISO 2631, EN 12299 and Sperling's method. Evaluations following the different methods agree that vertical direction induces higher discomfort than longitudinal or lateral accelerations. Moreover, the track's vertical irregularities are the dominant cause of vibration discomfort [23,25,30]. However, besides sharing the same assumptions, frequency dependence and goals, each method has its calculation formulation and evaluation scale. Thus, one method cannot be directly converted into another without performing a complete analysis [1]. Although several authors tried to establish a comparison methodology between the ride comfort methods,
Chapter II 34 that goal was still unachieved. Experimental campaigns must be performed and evaluated based on the three methods to fill this research gap. Moreover, the lack of a consensus methodology capable of being applied worldwide and in multiple conditions leads to misevaluation when comparing results. Based on the ride comfort analysis, notably the ISO 2631 methodology, SEAT values should be calculated as a complementary comfort examination for an accurate passenger comfort evaluation. The SEAT reflects seat isolation efficiency [39]. Due to the low-frequency vibrations, dominant in the vertical direction of railways, and foam's incapacity to absorb that, a vertical SEAT higher than 100% is expected [38]. Currently, SEAT evaluations are primarily performed in laboratory experiments, which do not correctly traduce the seat behaviour under service conditions. Gong and Griffin [38] concluded that SEAT values should be obtained based on real vibration conditions; otherwise, the conclusions are not valid for service conditions. Seat transmissibility quantifies seat dynamics and verifies its efficiency in handling vibration discomfort. Seat efficiency depends on human body biodynamics, vibration excitation and foam properties. The human body produces nonlinear physiological responses to vibration. Thus, vibration frequency range influences more some body regions than others; for example, frequencies comprehended between 4-8 Hz led to resonance in the abdominal region [20,41,44]. Moreover, once the biodynamic response of the human body is cross-axis-coupled, vibration transmissibility presents the same characteristic. The cross-axis-coupled transmissibility is a crucial factor once, this way, vibration induced in one axis induces responses in other axes [39]. Seat users' anthropometric characteristics do not influence vibration transmissibility [48–52]. Additionally to being cross-axis-coupled, transmissibility differs in direction and location. Therefore, transmissibility peaks in vertical, longitudinal, and lateral directions in both seat surface and seatback present different trends. Whereas the vertical seat surface transmissibility reveals a peak between 4-6 Hz, that transmissibility increases to approximately 25 Hz for both lateral and longitudinal directions. The seatback transmissibility peaks are a rotation of those of the seat surface. This way, while the vertical and lateral transmissibilities have peaked around 25 Hz, the longitudinal direction ones are comprehended between 4-6 Hz [38]. The most referenced bibliography studies' major limitation concerns its performance on a seat simplification instead of a train seat. Seat dimensions, support points, foam coupling, real train vibration inducing and the influence of carbody natural frequencies on seat transmissibility are crucial parameters not evaluated in that studies. Transmissibility tests should be conducted on real train environments instead of laboratory analysis. This way, tests can be adapted to perform a modal identification and analysis of the seat structure. Particularly, natural frequencies related to the seat frame structure can be distinguished from those dependent on
Chapter II 35 the carbody, increasing the ability to perform a more robust comfort evaluation. Regarding foam properties, the thickness was demonstrated to influence seat transmissibility. Doubling the foam thickness approximately halves its stiffness, increasing transmissibility and discomfort [41,46,48,55,56]. Foam hardness, composition and density do not significantly affect seat transmissibility [9]. The biomechanical load the human body's musculoskeletal structure imposes on a seat surface defines its interface pressure. That depends on the users' anatomical characteristics and foam properties, such as hardness, hysteresis loss and SAG factor. The human buttocks are not flat; this way, local pressure varies over the seating contact area, achieving higher values under the ischial tuberosities. Those local pressure peaks are highly related to discomfort and should be avoided [3,9,59,60]. The maximum interface pressure threshold is not consensus in the literature. Some authors defend that a 32 mmHg pressure should not be exceeded once it represents the capillary pressure value, while others accept a higher 37.75 mmHg pressure. However, although the lack of consensus defines a maximum pressure threshold, it seems unanimous that the ideal seat will avoid pressure peaks, leading to a widely dispersed distribution [58,60,62–66,69]. Regarding passengers' anthropometric characteristics influencing interface pressure, subjects' weight induces the most substantial impact. Heavier subjects report higher contact areas with the seat surface, leading to lower pressure and increased comfort. In opposition, thinner subjects tend to produce a higher pressure around the ischial tuberosities, conducting higher discomfort levels [60]. Additionally, the waist and buttocks are the most remarkable body parts influencing whole-body discomfort [59,62]. Currently, to the author's knowledge and research, only experiments conducted with adults were performed. Since the interface pressure depends on the subject's weight, conducting an interface pressure experiment in children would be interesting since they are typically shorter and lighter than adults. Foam properties, i.e., hardness, hysteresis loss and SAG factor, also significantly impact pressure distribution. Softer seats reach higher comfort levels due to increased contact area than rigid seats [9,60,66]. The SAG factor, which represents comfort while seated, should present higher than 2.8 values for seating applications. Lower values induce bottoming and, consequently, passenger discomfort [9,57]. In its turn, hysteresis loss represents the passenger's comfort during de-seating. Thus, theoretically, low hysteresis loss values are required to increase passengers' comfort [9,57]. Moreover, Ebe and Griffin [9] concluded that the SAG factor highly correlates with passenger comfort. In resume, the ideal foam presents low hysteresis loss, high SAG factor and low vibration transmissibility.
Chapter II 36 To the author's knowledge and appraised research, there is a lack of interface pressure tests conducted on real train seats under static and dynamic conditions. To accurately characterise static foam comfort, foams evaluation should be performed in a complement between interface pressure analysis and foam properties laboratory analysis, namely density, hardness, hysteresis loss and SAG factor. This way, the foam behaviour under static conditions could be thoroughly evaluated. Most authors agree on the performance of combined static and dynamic comfort evaluation. However, to the authors' knowledge, only four published studies evaluate both comfort types simultaneously, representing an investigation gap. 2.5 Conclusions Railways are currently one of the most used public transportation systems. Therefore, its comfort performance is now a central theme regarding rail design. Depending on multiple factors, rail comfort is evaluated by both static and dynamic behaviour of motion parameters. Interface pressure, density, hysteresis loss and SAG factor define static comfort, whereas dynamic comfort is assessed employing ride comfort, SEAT and seat transmissibility. Ride comfort is evaluated by three main methods, ISO 2631, EN 12299 and Sperling's method. Besides sharing similar assumptions and goals, the methodologies vary in comfort calculations, and no correlation has been found between methods, invalidating a universal analysis methodology. It also should be highlighted that the lack of a consensus methodology to evaluate passengers' ride comfort possibly leads to misevaluation when comparing results. This way, a research gap can be identified, and more efforts must be made to accomplish a global and consensus ride comfort evaluation methodology. SEAT and transmissibility agree in determining the importance of vertical vibrations in passengers' discomfort. Nevertheless, the main bibliographic research conducted on these topics is performed under laboratory conditions, which do not consider the influence of the seat frame structure and carbody natural frequencies. Therefore, those experiments may not correctly reproduce the railway in service conditions and how those parameters affect passengers' comfort. When applied together, interface pressure analysis and foam properties accurately define the foam's static behaviour. However, only one study reported the combined use of static comfort analysis methods, and more work needs to be performed. Furthermore, multiple studies demonstrated that the subject's weight influences the interface pressure. Since children are typically shorter and lighter than adults, conducting experimental research on children's interface pressure in rail seats would be relevant.
Chapter II 37 A generalised lack of experiments conducted on real rail vehicles and seats exists. Laboratory experiments allow comfort predictions but do not represent actual in-service conditions. Moreover, comfort is only thoroughly evaluated if static and dynamic evaluations are performed. The present research revealed an investigation gap once only four studies were obtained regarding that type of evaluation. More work is needed to evaluate and increase passengers' comfort and promote rail usage as a daily transportation system.
Chapter II 38 Chapter references [1] Jiang Y, Chen BK, Thompson C. A comparison study of ride comfort indices between Sperling's method and EN 12299. International Journal of Rail Transportation 2019:1–18. https://doi.org/10.1080/23248378.2019.1616329. [2] Kim YG, Kwon HB, Kim SW, Park CK, Park TW. Correlation of ride comfort evaluation methods for railway vehicles. Proc Inst Mech Eng F J Rail Rapid Transit 2002;217:73–88. https://doi.org/10.1243/095440903765762823. [3] Yuxue B, Bingchen G, Jianjie C, Wenzhe C, Hang Z, Chen C. Sitting comfort analysis and prediction for high-speed rail passengers based on statistical analysis and machine learning. Build Environ 2022;225. https://doi.org/10.1016/j.buildenv.2022.109589. [4] Vink P, Hallbeck S. Editorial: Comfort and discomfort studies demonstrate the need for a new model. Appl Ergon 2012;43:271–6. https://doi.org/10.1016/j.apergo.2011.06.001. [5] L. Z, M.G. H. Identifying factors of comfort and discomfort: A multidisciplinary approach 1992:395– 402. [6] Zhang L, Helander MG, Drury CG. Identifying factors of comfort and discomfort in sitting. Hum Factors 1996;38:377–89. https://doi.org/10.1518/001872096778701962. [7] de Looze MP, Kuijt-Evers LFM, van Dieën J. Sitting comfort and discomfort and the relationships with objective measures. Ergonomics 2003;46:985–97. https://doi.org/10.1080/0014013031000121977. [8] Hiemstra-van Mastrigt S, Groenesteijn L, Vink P, Kuijt-Evers LFM. Predicting passenger seat comfort and discomfort on the basis of human, context and seat characteristics: a literature review. Ergonomics 2016;60:889–911. https://doi.org/10.1080/00140139.2016.1233356. [9] Ebe K, Griffing M. Factors affecting static seat cushion comfort. Ergonomics 2001;44:901–21. https://doi.org/https://doi.org/10.1080/00140130110064685. [10] Ebe K, Griffin M. Quantitative prediction of overall seat discomfort. Ergonomics 2000;43:791–806. https://doi.org/10.1080/001401300404751. [11] Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 2009;6:e1000097. https://doi.org/10.1371/journal.pmed.1000097. [12] Hostens I. Analysis of Seating during Low Frequency Vibration Exposure. Katholieke Universiteit Leuven, 2004. [13] Park SJ, Subramaniyam M. Evaluating Methods of Vibration Exposure and Ride Comfort in Car. Journal of the Ergonomics Society of Korea 2013;32:381–7. https://doi.org/10.5143/jesk.2013.32.4.381. [14] British Standards Institution, International Standard Organization. BS ISO 2631 - Mechanical vibration and shock. Evaluation of human exposure to whole-body vibration. 2001. [15] British Standards Institution, European Committee for Standardization. BS EN 12299:2009 Railway applications - ride comfort for passengers - measurement and evaluation. British Standards Policy and Strategy Committee; 2009. [16] Railway Technical Research Institute (Ed.). Outline of design standards for railway structures and commentary (displacement limits). Tokyo: 2006. [17] American Society of Civil Engineers. ANSI/ASCE/T&DI 21-13 Automated people mover standards. Virginia: 2013. [18] People's Republic of China National Railway Administration (CNRA-PRC). GB/T 5599-2019 Specification for dynamic performance assessment and testing verification of rolling stock. Beijing: 2019. [19] Bovenzi M. Health effects of mechanical vibration. G Ital Med Lav Erg 2005;27:58–64.
Chapter II 39 [20] Năstac S, Picu M. Evaluating Methods of Whole-Body-Vibration Exposure in Trains. The Annals of “Dunarea de Jos”, University of Galați: Galați, Romania , 2010. [21] Griffin MJ. Handbook of Human Vibration. London: Academic Press; 1990. https://doi.org/10.1016/C2009-0-02730-5. [22] Khan S, Sundstrom J. Effects of vibration on sedentary activities in passenger trains. Journal of Low Frequency Noise, Vibration and Active Control 2007;26:43–55. https://doi.org/10.1260/026309207781487448. [23] Peng Y, Wu Z, Fan C, Zhou J, Yi S, Peng Y, et al. Assessment of passenger long-term vibration discomfort: a field study in high-speed train environments. Ergonomics 2021;65:659–71. https://doi.org/10.1080/00140139.2021.1980113. [24] International Standard Organization. ISO 2631 - Mechanical vibration and shock - Evaluation of human exposure to whole-body vibration 2001. [25] Liu C, Thompson D, Griffin MJ, Entezami M. Effect of train speed and track geometry on the ride comfort in high-speed railways based on ISO 2631-1. Proc Inst Mech Eng F J Rail Rapid Transit 2019;0:1–14. https://doi.org/10.1177/0954409719868050. [26] European Committee for Standardization. EN 12299 - Railway applications — Ride comfort for passengers — Measurement and evaluation 1999. [27] Kufver B, Persson R, Wingren J. Certain aspects of the CEN standard for the evaluation of ride comfort for rail passengers. WIT Transactions on the Built Environment 2010;114:605–14. https://doi.org/10.2495/CR100561. [28] Darlton AO, Marinov M. Suitability of Tilting Technology to the Tyne and Wear Metro System. Urban Rail Transit 2015;1:47–68. https://doi.org/10.1007/s40864-015-0007-8. [29] Karakasis K, Skarlatos D, Zakinthinos T. A factorial analysis for the determination of an optimal train speed with a desired ride comfort. Applied Acoustics 2005;66:1121–34. https://doi.org/10.1016/j.apacoust.2005.02.006. [30] Johanning E, Fischer S, Christ E, Göres B, Landsbergis P. Whole-body vibration exposure study in u.S. railroad locomotives—an ergonomic risk assessment. Am Ind Hyg Assoc J 2002;63:439–46. https://doi.org/10.1080/15428110208984732. [31] Cheng YC, Hsu C te. Parametric Analysis of Ride Comfort for Tilting Railway Vehicles Running on Irregular Curved Tracks. International Journal of Structural Stability and Dynamics 2016;16. https://doi.org/10.1142/S021945541550056X. [32] Dumitriu M, Leu M. Correlation between Ride Comfort Index and Sperling's Index for Evaluation Ride Comfort in Railway Vehicles. Applied Mechanics and Materials 2018;880:201–6. https://doi.org/10.4028/www.scientific.net/amm.880.201. [33] Munawir TIT, Samah AAA, Rosle MAA, Azlis-Sani J, Hasnan K, Sabri SM, et al. A Comparison Study on the Assessment of Ride Comfort for LRT Passengers. IOP Conf Ser Mater Sci Eng 2017;226. https://doi.org/10.1088/1757-899X/226/1/012039. [34] Dumitriu M, Stănică DI. Study on the evaluation methods of the vertical ride comfort of railway vehicle—mean comfort method and sperling's method. Applied Sciences (Switzerland) 2021;11. https://doi.org/10.3390/app11093953. [35] Schust M, Blüthner R, Seidel H. Examination of perceptions (intensity, seat comfort, effort) and reaction times (brake and accelerator) during low-frequency vibration in xor y-direction and biaxial (xy-) vibration of driver seats with activated and deactivated suspension. J Sound Vib 2006;298:606–26. https://doi.org/10.1016/j.jsv.2006.06.029. [36] Hong K-T, Hwang S-H, Hong K-S. Automotive Ride-Comfort Improvement with an Air Cushion Seat. SICE Annual Conference, Fukui University, Japan: 2003. [37] Bonin G, Cantisani G., Loprencipe G, Sbrolli M. Ride quality evaluation: 8 D.O.F. vehicle model calibration. 4th INTERNATIONAL SIIV CONGRESS, 2007.
Chapter III 46 𝑓𝑁𝑦𝑞𝑢 ≥2𝑓𝑚𝑎𝑥 (3.1) 𝑓𝑁𝑦𝑞𝑢 ≥2×80≥160 𝐻𝑧 where 𝑓𝑁𝑦𝑞𝑢 represents the Nyquist frequency and 𝑓𝑚𝑎𝑥 indicates the highest frequency component of the signal [17]. Moreover, due to the need to access vibration on three axes, only 3-axial accelerometers were considered [6,8]. PCE Instruments company, a current project partner, kindly provided the PCE-VDL-24I accelerometer. This 3-axes accelerometer presented the following technical characteristics, scale range: ±16 g, sample rating between 0 – 2400 Hz, precision: ±0.24 g, resolution: 0.004 g, autonomous recording (32 GB µSD – at 200 Hz sample rate records during 1d08h32min), Li-Ion 3.7 V/500 mAh integrated battery and, allows post-analysis due to data export (CSV file). The accelerometer was programmed and calibrated, being selected as the measurement unit [15]. Moreover, a metallic box was designed and validated to set the accelerometer and protect it from external disturbances [6]. Two pieces were thought to form a metallic box. Also provided by PCE Instruments, a mounting plate guaranteed the accelerometer's rigid fixation. Figure 3.1 illustrates the accelerometer and its coupling with the mounting plate. (a) (b) Figure 3.1. 3-axial accelerometer attached on the mounting plate where (a) top view and (b) bottom view. The top of the protector metal case needs to support and sustain a user's weight without damaging the accelerometer or influencing the vibration measurements. The proposed solution involved the development of an aluminium 6061 box, reinforced at the top and laterally, as represented in Figure 3.2.
Chapter III 47 Figure 3.2. Multiple views of the box designed in Autodesk Inventor [19]. To validate the top box element, a pressure test was conducted using a finite element analysis (FEA) on Autodesk Inventor software [18]. According to SAE J1384 [19], a 75 kg mass should be used in place of a human when testing vibration equipment. To obtain a more elevated test margin and the correspondent representation of people with a higher weight, the pressure test was performed with 1kPa, representing the exerted pressure of a 100 kg person seated on the box. Results, presented in Figure 3.3, showed a maximum Von Misses stress of 0.08 MPa, a maximum displacement of 8.94x10-2 μm, and 15 safety factor, demonstrating the box's reliability and security [20]. (a) (b) Figure 3.3. FEA pressure tests results where (a) Von Misses stress and (b) displacement analysis.
Chapter III 48 Finally, a silicone disc having a 235 mm diameter was designed. A fit with the metal box dimensions was withdrawn from the centre of the disc to couple the metal box. Figure 3.4 shows the developed system, composed of the metallic box housed in the silicone disc. Figure 3.4. Developed system, metallic box coupled with a silicone rubber 80 Shore A disc. 3.2.4 System elements validation As described, three elements compose the alternative WBV measurement system. Although the accelerometer was factory calibrated, it was necessary to experimentally ascertain the minimum or noninfluence of the new elements in vibration measurements. A vibration measurement test was performed using a Brovind vibration equipment connected to a CB controller [21–23]. Initially, the equipment was fixed on the vibrating container; then, the vibration was set at level 0 for 75 seconds, increased to level 1 for another 75 seconds and, finally, returned to level 0 for the last 75 seconds. The accelerometer axes (x, y, and z) were fixed in the same position for all experiments. Each experiment was repeated three times for each configuration (accelerometer, mounting plate, metallic box, and complete system). Figure 3.5 shows the experiments and axis configuration.
Chapter III 49 (a) (b) (c) (d) Figure 3.5. Experimental setup: (a) accelerometer; (b) mounting plate; (c) metallic box; (s) rubber disc/complete equipment. The results were obtained and analysed using MATLAB software [24]. Although the accelerometer measurements are in g's, they were posteriorly converted in m/s2. An ANOVA test was used to validate the complete equipment. This test determines the probability of whether the means of the different datasets are equal. Four ANOVA tests were performed. The influence of each piece of equipment was observed when compared with the accelerometer, namely the accelerometer/mounting plate, accelerometer/metallic box, and accelerometer/complete equipment, which studied the influence of the three components as a complete WBV measurement equipment. The null hypothesis was that the equipment's presented the same mean. If the p-value returns higher than 0.05, the means are equal; otherwise, they are considered different [25–27]. The results exhibited the highest acceleration values for the accelerometer, with a mean of 3.136 m/s2, a maximum of 18.013 m/s2, and a minimum value of -3.373 m/s2. The other configurations presented similar results, and no significant differences were reported. The outcomes are presented in Table 3.2.
Chapter III 50 Table 3.2. Experimental results according to system configuration. Accelerometer Mounting plate Metallic box Complete equipment Mean (m/s2) 3.136 3.134 3.133 3.134 Maximum (m/s2) 18.013 18.009 18.012 18.012 Minimum (m/s2) -3.373 -3.374 -3.374 -3.374 An ANOVA statistical analysis was performed for the interaction between the primary element (accelerometer) and the other configurations. This way, differences in mean values for each test were statistically tested and further analysed (see Table 3.3). Table 3.3. ANOVA statistical analysis results. Comparison p-value Accelerometer/ Mounting plate 0.3295 Accelerometer/ Metallic box 0.652 Accelerometer/ Complete equipment 0.8025 Likewise, an analysis comprehended the comparison between all configurations and the accelerometer was performed. As the results were all above 0.05, it can be assumed that tey have similar average outputs. Thus, the null hypothesis was not rejected, and therefore, the results showed significance. Once the accelerometer, main element of the measurement system, was previously calibrated, these were the expected outcomes, representing the system's suitability to be used as an alternative WBV measurement system. 3.3 Case study: Portuguese Railways Northern Line ride comfort Vibration is the critical factor affecting passengers' comfort, performance and health, and it can limit the user's ability to carry out simple tasks, such as reading and writing [28–30]. Therefore, assessing the harmful consequences of vibration on train users is crucial. That can be accomplished following ISO 2631 standard comfort approach recommendations [6]. In order to characterise the passenger's comfort levels at the Portuguese Railways Northern Line, experimental tests were conducted on several AP and IC trains. Then, the standard methodology was applied to properly quantify users' discomfort levels. Moreover, vibration transmission can be mitigated or amplified by the seat and, consequently, vibration discomfort may decrease or increase. In order to quantify that influence, the SEAT calculation was applied as a complementary analysis method [31].
Chapter III 51 3.3.1 Procedures and equipment ISO 2631 approach was followed to perform the WBV evaluation during Porto (Campanhã) – Lisbon (Oriente) connection on AP and IC trains. As mentioned, according to this standard, 3-axes acceleration measurements should occur at the interface surfaces where vibration is transmitted to the user, namely: floor, seat surface and seatback. Therefore, the previously developed measurement equipment was attached to those interface surfaces at both AP and IC seats. A sample rate of 200 Hz, preventing aliasing and obeying the Nyquist theorem, was set. Moreover, it should be highlighted that the human feeling while exposed to vibration depends on those frequencies. To quantify the effect of sensitive vibration ranges, the ISO 2631 weighting curves amplify the effect of those frequencies most influencing passenger's discomfort and reduce the impact of those frequencies less affecting passenger's discomfort [3]. Experiments run at the Portuguese CPA 4000 series Pendolino, traditionally named AP. This train achieves a maximum 220 km/h speed and is operated as a single unit, divided into six cars. Those are organised according to two classes and bar facilities. The first and second cars are classified as comfort class, whereas the others are named as standard. Moreover, the bar is placed on the third car, reducing the available seats. Belonging to different classes, comfort and standard seats are visually and dimensionally different. Those of the comfort class present higher dimensions than the standard class [32,33]. According to the train motion, the first car corresponds to the begging of the train; consequently, the sixth car is placed at the opposite extremity, meaning the end of the train. The procedure at AP trains comprised nine complete journeys following the location and class; thus, three journeys took place on the first car, three on the fourth car, representing the middle of the train, and three on the sixth car. Figure 3.6 presents the AP measurement location and the experimental setup.
Chapter III 52 (a) (b) Figure 3.6. (a) AP interior layout and measurement points and (b) experimental setup, right side standard class and left side standard class with the subject. IC train service, introduced in 1980 in Portugal, runs at a maximum 200 km/h speed. This train is run by 5600 series locomotives with Corail coaches, renovated in 2002. The electric locomotive trails five coaches, the first and second representing the comfort class and the bar, and the third, fourth, and fifth classified as a standard class. Depending on the class, the seats show different dimensions and foam thicknesses. Furthermore, Corail coaches are characterised by having two seats per row. However, for the comfort class, besides matching this parameter, the seats are arranged individually; thus, the seat support frame is not shared by two seats. In opposition, the touristic seat frame is shared by two seats, representing a more similar structure to AP seats [34]. IC instrumented seats were localised at the first, third and fifth carriages, corresponding to the train start, mid and end points. Figure 3.7 illustrates the comfort class seat measurement location and experimental setup.
Chapter III 53 (a) (b) Figure 3.7. (a) IC train interior layout and measurement points; (b) experimental setup, right side comfort class and left side comfort class setup with the subject. As on the AP train, according to the train motion, the first car corresponds to the start of the train; consequently, the sixth car, placed at the opposite extremity, means the end of the train. Regarding the Northern Line, it presents a total length of 275 km. When travelled by AP, it takes around 2h50m to complete, comprising five station stops. For the IC train, approximately 3h15m are needed to complete the journey, been that divided into twelve train station breaks. All journeys run under regular operation conditions and passenger transportation. 3.3.2 Ride comfort analysis: results and discussion Following the ISO 2631 recommendations, weighted acceleration (𝑎𝑤) per axis and total acceleration (𝑎𝑣) were calculated to evaluate AP and IC passengers' discomfort levels. Table 3.4 presents the results regarding the nine AP journeys, demonstrating the measurement car (1, 4 or 6) and location (floor, seat surface and seatback), as well as the train identification (Train ID). The train identification number was replaced to keep the train's ID anonymous.
Chapter III 54 Table 3.4. Weighted acceleration per axis and total acceleration results regarding the AP trains. Car Train ID Measurement location 𝒂𝒘(𝒎/𝒔𝟐) 𝒂𝒗 (𝒎/𝒔𝟐) Discomfort level 𝑥 𝑦 𝑧 1 A Floor 0.07 0.08 0.15 0.06 Not uncomfortable Seat surface 0.06 0.09 0.26 0.28 Not uncomfortable Seatback 0.16 0.11 0.15 0.15 Not uncomfortable B Floor 0.07 0.08 0.14 0.06 Not uncomfortable Seat surface 0.06 0.10 0.24 0.26 Not uncomfortable Seatback 0.21 0.11 0.15 0.19 Not uncomfortable B Floor 0.07 0.09 0.15 0.07 Not uncomfortable Seat surface 0.06 0.10 0.23 0.26 Not uncomfortable Seatback 0.19 0.11 0.16 0.17 Not uncomfortable 4 C Floor 0.07 0.07 0.12 0.06 Not uncomfortable Seat surface 0.05 0.09 0.26 0.28 Not uncomfortable Seatback 0.19 0.09 0.15 0.17 Not uncomfortable D Floor 0.05 0.06 0.12 0.05 Not uncomfortable Seat surface 0.05 0.08 0.24 0.26 Not uncomfortable Seatback 0.14 0.08 0.13 0.13 Not uncomfortable A Floor 0.08 0.07 0.13 0.06 Not uncomfortable Seat surface 0.05 0.08 0.21 0.23 Not uncomfortable Seatback 0.21 0.09 0.15 0.19 Not uncomfortable 6 C Floor 0.07 0.07 0.13 0.06 Not uncomfortable Seat surface 0.05 0.11 0.23 0.26 Not uncomfortable Seatback 0.16 0.11 0.15 0.15 Not uncomfortable C Floor 0.08 0.09 0.14 0.06 Not uncomfortable Seat surface 0.05 0.13 0.27 0.30 Not uncomfortable Seatback 0.18 0.14 0.15 0.17 Not uncomfortable E Floor 0.07 0.08 0.14 0.06 Not uncomfortable Seat surface 0.05 0.10 0.25 0.27 Not uncomfortable Seatback 0.15 0.11 0.15 0.15 Not uncomfortable Observing the results, it was possible to conclude that regarding the floor and seat surface 𝑎𝑤 the vertical axis is the most prominent influencing discomfort, whereas, on the seatback, the x and z axes present a similar influence complying with Peng et al. [9] and Johanning et al. [10]. Moreover, as in the investigation developed by Khan and Sundstrom [11], when observing the vibration evolution from the floor to the seat surface and seatback, an increased discomfort level was noticed for the seat surface. These results highlight the seating capacity to modify vibration transmission, being in the present case
Chapter III 55 amplified [26,35,36]. Figure 3.8 illustrates and highlights the vibration transmission from the floor throughout the seat by comparing 𝑎𝑤 at the floor, seat surface and seatback. Figure 3.8. Vibration discomfort levels evolution from the floor throughout the seat of AP trains. Additionally, based on Figure 3.8, the seat's tendency to increase vibration peaks recorded on the floor can be observed. The seating capacity to amplify or mitigate vibration shocks will be further analysed through SEAT calculations. Moreover, the location of the train stations is also identified based on the low 𝑎𝑤 presented at those locations. Focusing on each measurement location, independently of the car, comparable results were found for the floor vibrations. The same trend was noticed regarding the seat surface and seatback results, where both 𝑎𝑤 and 𝑎𝑣 have demonstrated similar values in the three cars. This way, results may indicate that seat discomfort induced by vibration does not depend on the seat location nor on the seat type. Globally comparing the AP discomfort results with those reported by Liu et al. [8] and Peng et al. [9] on Chinese high-speed railways, AP trains presented an intermedium value for the seat surface measurements. Revealing better performance than the seats evaluated by Liu et al. [8] and worst than those instrumented by Peng et al. [9]. Regarding seatback, only one comparison can be conducted; the AP seats have a poor dynamic behaviour than those evidenced in the seats evaluated by Peng et at. [9]. Nevertheless, all AP journeys were ranked as "Not uncomfortable". Furthermore, it should be highlighted that the AP is not a high-speed train but the most similar train owned by Portugal.
Chapter III 62 [19] Huston DR, Johnson CC, Zhao XD. A human analog for testing vibration attenuating seating. J Sound Vib 1998;214:195–200. [20] Stefanello J, Herbert F, Gomes M. Evaluation of the floor-seat transmissibility (SEAT) in riding vehicles and verification of vibration levels regarding health and comfort in WBV. 24th ABCM International Congress of Mechanical Engineering; December 3-8, Curitiba, PR, Brazil 2017. [21] Vibratori B. Brovind vibratori vibration plate. BrovindvibratoriIt/Catalogo/Basi-Vibranti n.d. [22] Vibratori B. Brovind vibratori vibration container. BrovindvibratoriIt/Catalogo/Contenitori-Scalini n.d. [23] Vibratori B. Brovind vibratori controller. BrovindvibratoriIt/Controller n.d. [24] The Mathworks I. MATLAB R2020a 2020. [25] Box P, Cox DR. An analysis of transformations. Journal of the Royal Statistical Society 1964;Series B:211–52. [26] Smith MJ. Statistical Analysis Handbook. 1st ed. Statsref; 2018. [27] Kutner M, Nachtsheim C, Neter J, Li W. Applied Linear Statistical Models. Homewood: McGraw-Hill; 2004. [28] Hostens I. Analysis of Seating during Low Frequency Vibration Exposure. Katholieke Universiteit Leuven, 2004. [29] Picu L, Picu M. An analysis of whole-body vibration and hand-arm vibration exposure on the Danube ship crew. J Phys Conf Ser 2019. [30] Park SJ, Subramaniyam M. Evaluating Methods of Vibration Exposure and Ride Comfort in Car. Journal of the Ergonomics Society of Korea 2013;32:381–7. https://doi.org/10.5143/jesk.2013.32.4.381. [31] Sapena J, Caminal R. Psychoacoustic evaluation of noises generated by passenger seats for high speed trains. Notes on Numerical Fluid Mechanics and Multidisciplinary Design 2018;139:439–50. https://doi.org/10.1007/978-3-319-73411-8_34. [32] Comboios de Portugal. CP - Comboio Elétrico Pendular, Séries 4000 2017. [33] Comboios de Portugal. CP - Comboio Elétrico Pendular, Séries 4000 1998. [34] Comboios de Portugal. CP - Carruagem Corail 2002. [35] Gong W, Griffin MJ. Measuring, evaluating and assessing the transmission of vibration through the seats of railway vehicles. Proc Inst Mech Eng F J Rail Rapid Transit 2018;232:384–95. https://doi.org/10.1177/0954409716671547. [36] van der Westhuizen A, van Niekerk JL. Verification of seat effective amplitude transmissibility (SEAT) value as a reliable metric to predict dynamic seat comfort. J Sound Vib 2006;295:1060–75. https://doi.org/10.1016/j.jsv.2006.02.010. [37] Patelli G, Griffin MJ. Effects of seating on the discomfort caused by mechanical shocks: Measurement and prediction of SEAT values. Appl Ergon 2019;74:134–44. https://doi.org/10.1016/j.apergo.2018.08.003. [38] Van Niekerk JL, Pielemeier WJ, Greenberg JA. The use of seat effective amplitude transmissibility (SEAT) values to predict dynamic seat comfort. J Sound Vib 2003;260:867–88. https://doi.org/10.1016/S0022-460X(02)00934-3. [39] Li T, Liang H, Zhang J, Zhang J. Numerical study on aerodynamic resistance reduction of high-speed train using vortex generator. Engineering Applications of Computational Fluid Mechanics 2023;17. https://doi.org/10.1080/19942060.2022.2153925. [40] Liang H, Sun Y, Li T, Zhang J. Influence of Marshalling Length on Aerodynamic Characteristics of Urban Emus under Crosswind. Journal of Applied Fluid Mechanics 2023;16. https://doi.org/10.47176/jafm.16.01.1338.
Chapter IV 63 Chapter IV - Indirect assessment of railway infrastructure anomalies based on passenger comfort criteria Railways are among the most efficient and widely used mass transportation systems for mid-range distances. To improve the attractiveness of this type of transport, it is necessary to improve the comfort, which is much influenced by the vibration derived from the train motion and wheel-track interaction, making railway track infrastructure condition and maintenance a main concern. Based on the discomfort level, a methodology capable of detecting railway track infrastructure failures is proposed. During the regular passenger service, acceleration and GPS measurements were taken on AP and IC trains between Porto (Campanhã) and Lisbon (Oriente) stations. ISO 2631 methodology was used to calculate instantaneous floor discomfort levels. By matching results for both trains, using GPS coordinates, twelve track section locations were found to require preventive maintenance actions. The methodology was validated by comparing these results with those obtained by the EM 120 track inspection vehicle for which similar locations were found. The developed system is a complementary condition-based maintenance tool that presents the advantage of being low-cost and not disturbing the regular train operation.
Chapter IV 64 4.1 Introduction High safety levels are guaranteed through adequate maintenance actions based on corrective or preventive strategies. The corrective action occurs after a fault is recognised and intends to recover the normal working state [1–4]. In contrast, proactive measures intend to prevent and minimise eventual faults that did not yet cause the system to fail, thus anticipating the problems. Preventive maintenance's main goal is to preserve the system functions and to prevent rail system failure, being performed according to a defined scheduled time. On the other hand, condition-based maintenance (CBM), the typical railway track infrastructure condition monitoring system, is a preventive maintenance strategy that acts when there is evidence of failure. Railway track infrastructure monitoring actions can be performed in two ways, either by human or automatic means, depending on the type and extent of work [5–7]. Human inspection is performed by well-trained inspectors who periodically walk along railway track infrastructure to detect defects. The inspectors can be aided by portable equipment for the auscultation and measurement of geometric parameters of the track infrastructure. However, this maintenance technique results in high costs and can be potentially hazardous for inspectors. Moreover, the human inspection can only be applied by stopping or restraining traffic, and its results highly depend on the observer's capability to detect anomalies and recognise critical situations [1,8]. Due to the problems associated with human inspection and advances in technology, automatic inspection methods have been developed, such as inspection vehicles. These dedicated vehicles can detect railway track infrastructure defects and evaluate the infrastructure performance. Generally, an inspection vehicle uses optical and inertial sensors linked to a GPS, which increases the method's efficiency and decreases the required time [2,3,9,10]. This inspection is periodically performed along the railway track infrastructure. However, in addition to being expensive, these vehicles introduce traffic disruptions during the inspection, affecting the regular service operation. Different types of inspection vehicles are used worldwide. In Portugal, the EM 120 inspection vehicle, owned by Infraestruturas de Portugal (IP), is responsible for identifying maintenance needs and railway track infrastructure failures. After detecting a failure, the defect is fixed, and ideally, the system is restored to its initial state [4,6]. More recently, cargo vehicles are also being used as instrumented inspection vehicles, particularly in the axle box [11–15]. However, implementing the experimental setup is problematic for that approach, limiting its use in passenger trains. These vehicles have the advantage of continuously surveying the track without any traffic interference, thus providing information on a more regular base at a reduced cost. Moreover, railway track infrastructure maintenance also improves the interaction performance of the rail vehicle and the overhead
Chapter IV 65 infrastructures, such as the pantograph-catenary interaction, which leads to a safer, more stable and comfortable journey [16]. It is known that there is a significant difference in the vibration level when comparing a healthy with a defective railway track infrastructure, being the latter characterised by higher peak values. Thus, besides affecting safety, isolated and continuous infrastructure defects lead to poor passenger comfort due to increased vibration levels [3]. Once comfort is mainly affected by vibration, which mainly derives from the train motion and rail track infrastructure irregularities, passengers are subjected to it throughout the journey due to contact with the seat and floor. Therefore, vibration is transmitted through all the passenger-seat and passenger-floor contact surfaces, generally defined as WBV. Besides causing discomfort, this can also lead to fatigue and, in some extreme cases, diseases. Due to its adverse consequences, it is vital to classify vibration levels in the rail environment. One of the reference international documents is ISO 2631 standard, which precisely quantifies WBV regarding comfort, human health, and motion sickness [17–20]. Up to this date, passenger comfort levels have not been used for possible damage detection. Based on the close connection between railway track infrastructure condition, induced vibration, and passengers' discomfort levels, it was possible to develop a new CBM methodology to identify critical railway track infrastructure locations. The goal was to provide a low-cost solution to improve maintenance needs detection and increase infrastructure availability. The present method overcomes the limitations of the traditional railway infrastructure maintenance detection methods, as it does not cause disturbance in the railway operation. This study considers that the railway track infrastructure requires maintenance actions if multiple trains with different dynamic characteristics present floor discomfort on the same GPS location. Based on the limitations of the previous experiments, this study aims to give clear contributions about some aspects that presently, according to the authors' knowledge, are not sufficiently addressed in the existing literature, particularly: - Development of methodology capable of detecting rail track infrastructure abnormalities or damages based on measurements on in-service trains. This way, the service runs under regular operation without disruption or interference. - Development of an easy-to-adapt and implement methodology, capable of being applied in any in-service rail vehicle. This way, overcoming the limitations of the accelerometer installation in the axle box, which imposes difficulties on the processes of installation, maintenance and dismounting.
Chapter IV 66 - The results of the present methodology are not dependent on the vehicle type. Thus, it can be applied to passenger trains with different characteristics. This was stated based on the results of a large extent of measurement campaigns proving the methodology's accuracy and precision. 4.2 Ballasted track Worldwide, as well as in Portugal, the most common railway system is the ballasted track system, which has lower construction costs and adequate response to static and dynamic forces [21]. This track type is grouped into two main categories: superstructure and substructure. The former consists of rail, sleepers, ballast, and components that connect these elements. In contrast, the latter is associated with the geotechnical system and comprises sub-ballast, embankments and subgrade layers [14,15,22,23]. Figure 4.1 presents a schematic representation of the railway ballasted track system. Figure 4.1. Schematic illustration of the railway ballasted infrastructure. Track integrity problems are dependent on the superstructure and substructure deformation. Both are related, but the health of the track substructure defines the structural performance [15,23]. The main substructure issues are related to sub-ballast and subgrade deterioration, which can induce poor drainage and track settlements. Those issues may also reduce track stability, capacity and safety, leading to misalignments and increased wear [15]. Superstructure track faults can be grouped into three main conditions, those related to geometry (cross-level, alignment, longitudinal levelling, twist and gauge), those dependent on rail surface faults (surface, corrugation, fatigue cracking, squat and creep) and, lastly, those dependent on the ballast condition (fouled ballast, ballast pockets and poor ballast drainage) [14,21]. Track defects and irregularities negatively affect track performance and safety. Wavelengths between 10 – 120 m mainly influence passengers' comfort. EN 13848 [24] defines three wavelength intervals for
Chapter IV 67 the evaluation of vertical and lateral track geometry conditions, D1 (3-25 m), D2 (25-70 m) and D3 (70150 m). The D1 wavelength range irregularities are mainly associated with running safety conditions, whereas D2 and D3 ranges strongly affect ride comfort [24,25]. Lower wavelengths define rail surface faults; for example, rail corrugation can appear between 0.03-0.08 m wavelength, whereas rail squat comprehends 0.02-0.04 m wavelengths. A discontinuity in the rail caused by a weld or a joint can present 0.01-0.02 m wavelengths [26,27]. It should be highlighted that the aforementioned defects' wavelengths are the most typical ones. Nevertheless, depending on its severity and dimension, the same type of defect can produce higher or lower wavelengths. Figure 4.2 presents the rail track geometry conditions evaluation defined by EN 13848, along with rail surface faults (rail corrugation, squats and discontinuities) and passenger's comfort wavelengths. Figure 4.2. Typical rail surface faults and EN 13848 track geometry evaluation wavelengths and those capable of affecting passengers' comfort. Rail defects and increased speed further amplify the rail vehicle-track dynamic interaction forces [13,14,23,28]. When a rail vehicle and track do not present defects or abnormalities, the vehicle exerts low-frequency forces (under 20 Hz). However, that frequency increases in the presence of faults, leading to dynamic impact wheel-rail forces. Moreover, dynamic loads, which are connected to track irregularities, amplify the rail deterioration rate. Vehicle vibration magnitudes are strongly linked with track irregularities and can be used to assess the general track condition effectively [13,22,23]. Karoumi et al. [29] and Norris [30] noticed that railway track defects could be identified based on acceleration records according to their peak values, as illustrated in the example of Figure 4.3. Figure 4.3. Detection of rail irregularities based on acceleration measurements. (Adapted from [29]).
Chapter IV 68 Besides affecting comfort, track irregularities can lead to a derailment, which may have significant consequences [12]. Therefore, promoting scheduled railway infrastructure maintenance is crucial to maintain high safety levels and the comfort of passengers. When a railway vehicle passes over a railway track, it causes static (weight) and dynamic (inertia and impact) loads. That dynamic load leads the railway track to vibrate for a certain period and, consequently, to deform. Moreover, those loads need to be transmitted to the track subgrade. The dynamic load exerted on the track is the critical parameter causing track deterioration, whereas wear and fatigue are the main consequences. Generally, rail defects are caused by loads applied to the rails in longitudinal, transverse and vertical directions. Additionally, the weather also plays a significant role in railway infrastructure deterioration as it can promote corrosion, thermal expansion and buckling. This way, the railway track experiences different conditions at different points along the track, and, consequently, different track sections show different degradation behaviours and require different maintenance plans [15]. 4.3 Suspension systems The interaction between the wheels and railway track infrastructure, healthy or with defects, generates vibration with varying amplitude and frequency. This vibration can cause damage to both trains and railway track infrastructure; thus, it must be controlled. The train suspension system, constituted by primary and secondary suspension, fulfils this function by enabling the filtering of the vibrations derived from the vehicle-track dynamic interaction (promoting ride comfort) and controlling the kinematic modes of the bogie (promoting stability). Suspension systems are natural low-pass filters, preventing the transmission of high frequencies from the rail track and wheel to the carbody and seats [31]. The primary suspension's main goals are to secure stability and guidance, reduce track forces and wear, and improve curve performance. Regarding the secondary suspensions, it suppresses the vibration transmission from the bogie to the carbody and attenuate vehicle vibrations derived from railway track irregularities, improving ride comfort and controlling the quasi-static motion [31–33]. Although with the same goal, different trains present different suspension systems, which interfere with the filtered ranges of frequencies. The present research conducts experimental tests on AP and IC trains. Due to their model series and tilting mechanisms, these trains present different suspension systems. Therefore, characterising those trains' suspension systems is essential.
Chapter IV 69 4.3.1 Alfa Pendular vehicles suspension AP train, composed of six cars, has an active tilting system, which reduces the lateral acceleration perceived by the passengers and, consequently, allows the performance of curves at higher speeds than the balanced one while maintaining high passenger comfort levels [31]. Two bogies support each car distanced by 19 m, and are 6 m far from the bogies of consecutive cars. Figure 4.4 (a) illustrates the primary AP suspension, while Figure 4.4 (b) shows its secondary suspension system. Primary suspension is composed of four helicoidal springs (sets of two plus two) (1) and a vertical damper (2), both filtering the vibration derived from the vehicle-track interaction in each wheel. Thus, each bogie of the AP tilting train has twelve flexi-coil springs combined with six dampers acting as secondary suspension and sixteen springs coupled to four vertical hydraulic dampers performing as primary suspension [34–36]. The mechanical elements of the secondary suspension ensure the carbody-bogie connection. The secondary suspension (see Figure 4.4 (b)) is constituted of twelve flexi-coil springs grouped within four units of three (3). Between spring units, on each side of the bogie, there are a vertical (4), a transversal (5) and an anti-yaw (6) hydraulic damper. (a) (b) Figure 4.4. AP train bogie: (a) primary suspension; (b) secondary suspension. 4.3.2 Intercity vehicles suspension IC train service is currently run by 5600 series locomotives with hauled Corail coaches. The electric locomotive trails five Corail coaches, each with a length of 26.4 m. As on the AP train, each carriage has two bogies, separated by 18.4 m distance within the same vehicle and 8 m between consecutive carriages.
Chapter IV 70 Figure 4.5 presents both, primary and secondary, suspension systems of IC vehicles [37]. The primary suspension (see Figure 4.5 (a)) comprises eight helicoidal springs and four hydraulic dampers, resulting in two springs (1) and one damper (2) per wheel. The bogie-carbody connection, partially performed by the secondary suspension (Figure 4.5 (b)), comprises two grouped helicoidal springs (3) aided by two vertical (4) and two transversal (5) dampers, one on each side of the bogie. (a) (b) Figure 4.5. IC bogie: (a) primary suspension; (b) secondary suspension. Table 4.1 resumes the primary and secondary elements of both types of trains (AP and IC). Table 4.1. Primary and secondary elements of AP and IC trains suspension systems. Train Primary suspension Secondary suspension AP 16 helicoidal springs + 4 vertical hydraulic dampers 12 flexi-coil springs + 6 hydraulic dampers (2 vertical, 2 transversals, 2 anti-yaw) IC 8 helicoidal springs + 4 vertical hydraulic dampers 4 helicoidal springs + 4 hydraulic dampers (2 vertical, 2 transversal) 4.4 Indirect method for infrastructure condition assessment based on comfort criteria Railway track infrastructure maintenance interventions are commonly decided based on measurements obtained by inspection vehicles and not on the dynamic response of in-service railway vehicles.
Chapter IV 71 Passenger comfort is affected by the vibration that the carbody experiences due to motion and railway track infrastructure irregularities. As aforementioned, railway track irregularities can be identified based on acceleration records according to visible peaks in those measurements. Passenger comfort levels are also assessed through acceleration measurements. Thus, a strong link between isolated railway track irregularities and passenger comfort levels can be identified. The present method relies on that connection. Based on the natural excitation created by the passage of railway vehicles over the track and the fact that a defective railway track system induces higher vibration, it is expected that high discomfort levels will be accomplished when in the presence of an abnormality on the railway track infrastructure. Thus, it was hypothesised that the railway track infrastructure needed maintenance if multiple trains with different suspension mechanisms reported instantaneous floor discomfort levels at the same geographic location. Therefore, a CBM identification methodology was developed to investigate the defined hypothesis. The proposed methodology includes two stages, the experimental data measurement and acquisition and the data analysis. Acceleration measurements were performed based on a 3-axial accelerometer aided by a GPS to determine the vehicle's location. ISO 2631 standard provided the reference evaluation method for assessing passenger comfort. Then, a MATLAB algorithm was developed to match multiple railway vehicles' discomfort locations and identify the maintenance needs locations. It should be highlighted that the methodology was applied on ballasted tracks once, that is, the track type of the Northern Line of the Portuguese railways. Nevertheless, there are no expected assessment differences when applying the present method to ballastless tracks. 4.4.1 Whole-body vibration evaluation WBV evaluation was performed based on the aforementioned ISO 2631 standard, following its requirements and recommendations [38-41]. As mentioned, the standard determines acceleration measurements at interface surfaces where vibration is transmitted to the user. However, it is well known that the seat modifies the vibration transmission [42–45]. Additionally, the floor represents the closer location between the users and the carbody. This way, the calculation of instantaneous floor discomfort by performing the standard analysis per second and applying the recommendations concerning floor measurement location ( 𝑊𝑘 as weighting curve and multiplying factors 𝑘𝑥=𝑘𝑦=0.25 and 𝑘𝑧=0.40) was performed [41].
Chapter IV 78 Figure 4.11. Nothern Line railway track infrastructure maintenance needs identification; detailed zones (left side and right side) are at the laterals. According to Figure 4.11, twelve maintenance zones were identified. For each zone, the start and end coordinates and their associated line kilometre were registered in Table 4.3, which also defines the matching with the IP-identified track sections. Due to the map resolution, some close zones appear to be just one. Blue and green markers define two sections that are, indeed, divided into small zones, 1 and 2 for the blue marker and 3, 4 and 5 for the green section. The discomfort location obtained at the Porto (Campanhã) train station is not numbered, as it is due to the acceleration to start the motion and not to a railway infrastructure abnormality.
Chapter IV 79 Table 4.3. Maintenance needs identification segments and comparison with EM 120 inspection vehicle results. Zone number Initial coordinates Final coordinates Line km Track segment IP Latitude Longitude Latitude Longitude 1 40.9776 -8.6374 40.9753 -8.6368 315 D 2 40.9609 -8.6354 40.9589 -8.6353 313 D 3 40.9005 -8.6221 40.8948 -8.6212 306 D 4 40.8805 -8.6189 40.8789 -8.6187 304 D 5 40.8725 -8.6178 40.8708 -8.6176 303 D 6 40.7044 -8.5727 40.7038 -8.5735 283 C 7 40.5196 -8.5108 40.5180 -8.5087 255 C 8 40.4198 -8.4564 40.4172 -8.4566 242 C 9 39.8820 -8.6489 39.8807 -8.6507 166 B 10 39.0651 -8.8734 39.0639 -8.8758 46 - 11 39.0081 -8.9518 39.0047 -8.9542 37 - 12 38.9159 -9.0155 38.9149 -9.0160 26 A Ten listed locations match those the Portuguese infrastructure manager company (IP) pointed out to be subject to maintenance, identified by the EM 120 inspection vehicle [50]. Thus, it confirms that the obtained locations are critical points of the railway track infrastructure, therefore validating this methodology. Moreover, two extra zones were detected, zones 10 and 11. In addition to its proven accuracy, the model demonstrates the ability to be considered a viable alternative to the current methods. It should be highlighted that it is a low-cost method capable of simultaneously identifying railway track infrastructure isolated irregularities and analysing the passenger's comfort levels. It should be noted that the proposed methodology cannot characterise the abnormality types present on the railway track infrastructure, but it is focused on detecting the critical track sections. Additionally, the present study does not consider the effect of aerodynamic forces. That has a more significant effect on high-speed trains, namely those achieving speeds higher than 300 km/h, which is not the case in this experimental research [51]. Nevertheless, it should be highlighted that since AP and IC trains have different structural designs, their aerodynamic behaviours are different, which may have a distinct influence on their vibration levels [52]. In the future, more work is intended to be developed regarding abnormality identification using more advanced methodologies, such as supervised machine learning procedures. Implementing these procedures will require vehicles to pass through track sections with well-known and characterised abnormalities. This way, it will be possible to learn and identify the typical dynamic response patterns of each abnormality [53].
Chapter IV 80 4.6 Conclusions Railways are currently one of the most used mass transportation systems worldwide. The increase in railway transportation demands more velocity and load. These two factors and the weather conditions accelerate the railway track infrastructure degradation, which must be assessed more frequently. The interaction between rail infrastructure, wheels and vehicle motion creates a complex vibration environment, which is filtered by two suspension systems, primary and secondary. The former intends to induce stability and reduce the vibration transmission from the track-wheels interaction, while the latter concerns the bogie-carbody transmission. These systems decrease the vibration transmission to the passenger and thus increase comfort. The railway track infrastructure, main component of the rail industry, affects both safety and comfort; thus, it needs continuous maintenance. Nowadays, railway track infrastructure evaluation is mainly accomplished by inspection vehicles. However, these vehicles are expensive, and their passage disrupts the programmed timetable. Based on these limitations, a low-cost CBM system was developed to identify railway track infrastructure maintenance needs based on comfort measurements without timetable disruptions. AP and IC trains ran multiple journeys at the Portuguese Railways Northern Line, where accelerations and GPS measurements were recorded. ISO 2631 standard methodology was followed to obtain floor discomfort levels. Matching the instantaneous floor discomfort for both types of trains, the railway track infrastructure maintenance sections were found. The identified geographic locations are similar to those obtained by the EM 120 inspection vehicle. Therefore, the system was validated and proved its precision. Moreover, compared with previous studies where acceleration measurements were acquired at the axle box, the application of comfort levels to assess maintenance requirements is a novelty. The developed system provides a complementary, low-cost, CBM railway track infrastructure analysis capable of detecting abnormalities (although not the type of), using in-service passenger trains to get measurements, thus avoiding any service disruption. A future goal of the present research is to identify the abnormality type. Therefore, machine learning procedures will be developed and applied.
Chapter IV 81 Chapter references [1] Fontul S, Fortunato E, de Chiara F, Burrinha R, Baldeiras M. Railways Track Characterization Using Ground Penetrating Radar. Procedia Eng, Elsevier Ltd 2016;143:1193–200. https://doi.org/10.1016/j.proeng.2016.06.120. [2] Al-Douri YK, Tretten P, Karim R. Improvement of railway performance: a study of Swedish railway infrastructure. Journal of Modern Transportation 2016;24:22–37. https://doi.org/10.1007/s40534015-0092-0. [3] Falamarzi A, Moridpour S, Nazem M. A Review on Existing Sensors and Devices for Inspecting Railway Infrastructure. Jurnal Kejuruteraan 2019;31:1–10. https://doi.org/10.17576/jkukm-2019-31(1)-01. [4] Bressi S, Santos J, Losa M. Optimization of maintenance strategies for railway track-bed considering probabilistic degradation models and different reliability levels. Reliab Eng Syst Saf 2021;207. https://doi.org/10.1016/j.ress.2020.107359. [5] Kim SS, Park C, Kim YG, Park C. Parameter characteristics of rail inspection measurement system of HSR-350x. Journal of Mechanical Science and Technology 2009;23:1019–22. https://doi.org/10.1007/s12206-009-0332-5. [6] Nijland F, Gkiotsalitis K, van Berkum EC. Improving railway maintenance schedules by considering hindrance and capacity constraints. Transp Res Part C Emerg Technol 2021;126. https://doi.org/10.1016/j.trc.2021.103108. [7] Sharma S, Cui Y, He Q, Mohammadi R, Li Z. Data-driven optimization of railway maintenance for track geometry. Transp Res Part C Emerg Technol 2018;90:34–58. https://doi.org/10.1016/j.trc.2018.02.019. [8] de Ruvo P, de Ruvo G, Distante A, Nitti M, Stella E, Marino F. A Visual Inspection System for Rail Detection & Tracking in Real Time Railway Maintenance. The Open Cybernetics and Systemics Journal 2008;2:57–67. [9] Zhao Y, Liu Z, Yi D, Yu X, Sha X, Li L, et al. A Review on Rail Defect Detection Systems Based on Wireless Sensors. Sensors 2022;22. https://doi.org/10.3390/s22176409. [10] Malekjafarian A, OBrien E, Quirke P, Bowe C. Railway track monitoring using train measurements: An experimental case study. Applied Sciences (Switzerland) 2019;9. https://doi.org/10.3390/app9224859. [11] Chudzikiewicz A, Bogacz R, Kostrzewski M, Konowrocki R. Condition monitoring of railway track systems by using acceleration signals on wheelset axle-boxes. Transport 2017;33:555–66. https://doi.org/10.3846/16484142.2017.1342101. [12] Tsunashima H, Mori H, Ogino M, Asano A. Development of Track Condition Monitoring System Using Onboard Sensing Device. Railway Research - Selected Topics on Development, Safety and Technology, InTech; 2015. https://doi.org/10.5772/61077. [13] Ji A, Woo WL, Wong EWL, Quek YT. Rail track condition monitoring: a review on deep learning approaches. Intelligence & Robotics 2021. https://doi.org/10.20517/ir.2021.14. [14] Ferrante C, Bianchini Ciampoli L, Benedetto A, Alani AM, Tosti F. Non-destructive technologies for sustainable assessment and monitoring of railway infrastructure: a focus on GPR and InSAR methods. Environ Earth Sci 2021;80:806. https://doi.org/10.1007/s12665-021-10068-z. [15] Artagan SS, Bianchini Ciampoli L, D’Amico F, Calvi A, Tosti F. Non-destructive Assessment and Health Monitoring of Railway Infrastructures. Surv Geophys 2020;41:447–83. https://doi.org/10.1007/s10712-019-09544-w. [16] Song Y, Wang Z, Liu Z, Wang R. A spatial coupling model to study dynamic performance of pantograph-catenary with vehicle-track excitation. Mech Syst Signal Process 2021;151:107336. https://doi.org/10.1016/j.ymssp.2020.107336.
Chapter IV 82 [17] Hostens I. Analysis of Seating during Low Frequency Vibration Exposure. Katholieke Universiteit Leuven, 2004. [18] Picu L, Picu M. An analysis of whole-body vibration and hand-arm vibration exposure on the Danube ship crew. J Phys Conf Ser 2019. [19] Park SJ, Subramaniyam M. Evaluating Methods of Vibration Exposure and Ride Comfort in Car. Journal of the Ergonomics Society of Korea 2013;32:381–7. https://doi.org/10.5143/jesk.2013.32.4.381. [20] Huston DR, Johnson CC, Zhao XD. A human analog for testing vibration attenuating seating. J Sound Vib 1998;214:195–200. [21] Soleimanmeigouni I, Ahmadi A, Kumar U. Track geometry degradation and maintenance modelling: A review. Proc Inst Mech Eng F J Rail Rapid Transit 2018;232:73–102. https://doi.org/10.1177/0954409716657849. [22] Remennikov AM, Kaewunruen S. A review of loading conditions for railway track structures due to train and track vertical interaction. Struct Control Health Monit 2008;15:207–34. https://doi.org/10.1002/stc.227. [23] Kaewunruen S, Remennikov A. Dynamic properties of railway track and its components : a state-ofthe-art review. In: Weiss B, editor. New Research in Acoustics, Nova Science Publishers; 2008;197–220. [24] European Committee for Standardization. EN 13848-1:2019 - Railway applications – Track – Track geometry quality – Part 1: characterization of track geometry 2019. [25] De Rosa A, Kulkarni R, Qazizadeh A, Berg M, Di Gialleonardo E, Facchinetti A, et al. Monitoring of lateral and cross level track geometry irregularities through onboard vehicle dynamics measurements using machine learning classification algorithms. Proc Inst Mech Eng F J Rail Rapid Transit 2021;235:107–20. https://doi.org/10.1177/0954409720906649. [26] Chang W, Cai X, Wang Q, Tang X, Sun J, Yang F. The Influence of Track Irregularity in Front of the Turnout on the Dynamic Performance of Vehicles. Applied Sciences 2022;12:4169. https://doi.org/10.3390/app12094169. [27] Loidolt M, Marschnig S. Evaluating Short-Wave Effects in Railway Track Using the Rail Surface Signal. Applied Sciences 2022;12:2529. https://doi.org/10.3390/app12052529. [28] Fesharaki M, Wang T-L. The Effect of Rail Defects on Track Impact Factors. Civil Engineering Journal 2016;2:458–73. https://doi.org/10.28991/cej-2016-00000049. [29] Karoumi R, Wiberg J, Liljencrantz A. Monitoring traffic loads and dynamic effects using an instrumented railway bridge. Eng Struct 2005;27:1813–9. https://doi.org/10.1016/j.engstruct.2005.04.022. [30] Norris P. Recent advances in the understanding of bridge dynamic behaviour on the West Coast Main Line Route Modernisation Project. Dynamics of High-Speed Railway Bridges. 1st Edition, CRC Press; 2008. [31] Fu B, Giossi RL, Persson R, Stichel S, Bruni S, Goodall R. Active suspension in railway vehicles: a literature survey. Railway Engineering Science 2020;28:3–35. https://doi.org/10.1007/s40534-02000207-w. [32] Sugahara Y, Kazato A, Koganei R, Sampei M, Nakaura S. Suppression of vertical bending and rigidbody-mode vibration in railway vehicle car body by primary and secondary suspension control: Results of simulations and running tests using Shinkansen vehicle. Proc Inst Mech Eng F J Rail Rapid Transit 2009;223:517–31. https://doi.org/10.1243/09544097JRRT265. [33] Kargarnovin MH, Younesian D, Thompson D, Jones C. Ride comfort of high-speed trains travelling over railway bridges. Vehicle System Dynamics 2005;43:173–97. https://doi.org/10.1080/00423110512331335111.
Chapter IV 83 [34] Pombo J, Almeida T, Magalhães H, Antunes P, Ambrósio J. Finite element methodology for flexible track models in railway dynamics applications. International Journal of Vehicle Structures and Systems 2013;5:43–52. https://doi.org/10.4273/ijvss.5.2.01. [35] Ribeiro D, Calçada R, Delgado R, Brehm M, Zabel V. Finite-element model calibration of a railway vehicle based on experimental modal parameters. Vehicle System Dynamics 2013;51:821–56. https://doi.org/10.1080/00423114.2013.778416. [36] Comboios de Portugal. CP - Comboio Elétrico Pendular, Séries 4000 2017. [37] Comboios de Portugal. CP - Carruagem Corail 2002. [38] Bovenzi M. Health effects of mechanical vibration. G Ital Med Lav Erg 2005;27:58–64. [39] Griffin MJ. Handbook of Human Vibration. London: Academic Press; 1990. https://doi.org/10.1016/C2009-0-02730-5. [40] Năstac S, Picu M. Evaluating Methods of Whole-Body-Vibration Exposure in Trains. The Annals of “Dunarea de Jos” 2010. [41] International Standard Organization. ISO 2631 - Mechanical vibration and shock - Evaluation of human exposure to whole-body vibration 2001. [42] Gong W, Griffin MJ. Measuring, evaluating and assessing the transmission of vibration through the seats of railway vehicles. Proc Inst Mech Eng F J Rail Rapid Transit 2018;232:384–95. https://doi.org/10.1177/0954409716671547. [43] van der Westhuizen A, van Niekerk JL. Verification of seat effective amplitude transmissibility (SEAT) value as a reliable metric to predict dynamic seat comfort. J Sound Vib 2006;295:1060–75. https://doi.org/10.1016/j.jsv.2006.02.010. [44] Basri B, Griffin MJ. The application of SEAT values for predicting how compliant seats with backrests influence vibration discomfort. Appl Ergon 2014;45:1461–74. https://doi.org/https://doi.org/10.1016/j.apergo.2014.04.004. [45] Silva P, Ribeiro D, Mendes J, Seabra E. Modal Identification of Train Passenger Seats Based on Dynamic Tests and Output-Only Techniques. Applied Sciences 2023;13:2277. https://doi.org/10.3390/app13042277. [46] SparkFun. SparkFun GPS Logger Shield. https://www.sparkfun.com/products/13750. [47] Sparkfun. SparkFun RedBoard Qwiic. https://www.sparkfun.com/products/15123 [48] The Mathworks I. MATLAB R2020a 2020. [49] Ribeiro D, Calçada R, Brehm M, Zabel V. Train–Track–Bridge Dynamic Interaction on a BowstringArch Railway Bridge: Advanced Modeling and Experimental Validation. Sensors 2022;23:171. https://doi.org/10.3390/s23010171. [50] Infraestruturas de Portugal. Network Statement 2022. 2021. [51] Li T, Liang H, Zhang J, Zhang J. Numerical study on aerodynamic resistance reduction of high-speed train using vortex generator. Engineering Applications of Computational Fluid Mechanics 2023;17. https://doi.org/10.1080/19942060.2022.2153925. [52] Liang H, Sun Y, Li T, Zhang J. Influence of Marshalling Length on Aerodynamic Characteristics of Urban Emus under Crosswind. Journal of Applied Fluid Mechanics 2023;16. https://doi.org/10.47176/jafm.16.01.1338. [53] Meixedo A, Santos J, Ribeiro D, Calçada R, Todd MD. Online unsupervised detection of structural changes using train–induced dynamic responses. Mech Syst Signal Process 2022;165:108268. https://doi.org/10.1016/j.ymssp.2021.108268.
Chapter V 84 Chapter V - Modal Identification of Train Passenger Seats Based on Dynamic Tests and Output-Only Techniques This chapter describes the experimental vibration modal identification of train seats based on a dedicated set of dynamic tests performed on AP and IC trains. This work uses two output-only modal identification techniques: the transmissibility functions and the Enhanced Frequency Domain Decomposition (EFDD) method. The last method allows us to clearly distinguish the seat structural movements, particularly those related to torsion and bending of the seat frame, from the local vertical foam vibrations. The natural frequencies and mode shapes are validated by matching the results derived from the transmissibility functions and EFDD method. The identified modal parameters are particularly relevant to characterise the vibration transmissibility provided by the foams (local transmissibility) and the vibration transmissibility derived from the metallic seat frame (global transmissibility).
Chapter V 85 5.1 Introduction Trains' attractivity is usually achieved based on three main parameters: safety, comfort, and user conditions [1–4]. Resulting of the rail motion and wheel–track interaction, seat vibration affects users' comfort and health, limiting the attractiveness performance. As passengers spend most of their time seated, vibration will be transmitted to the user through interfaces between the human body, the seat structure, and the vehicle [8]. Thus, evaluating vibration transmission in a rail environment is crucial to assess the vibration frequencies and amplitudes transmitted to the user and evaluate passengers' comfort [4–9]. Several reference standards and guidelines, such as EN 12299 [3], UIC 513 [10], and ISO 2631-1 [11], define methodologies for the assessment of passengers' comfort based on both simplified and complete measurements. The simplified measurements are exclusively based on the accelerations measured at the floor of the carbody. This approach has the advantage of using a simpler measurement layout. However, the transposition of the acceleration from the floor to the passenger seat level requires the application of dedicated weighting frequency-dependent filters [3,10,11]. Otherwise, the complete measurements are based on the acceleration recorded on the passenger–seat–carbody interfaces, such as headrest–neck, armrest–upper arms, seat–hip, seat–back, and floor–seat, as well as on the carbody floor, close to the seat. This approach requires a set of dedicated accelerometers installed on several sites, which may be difficult. Inclusively, involuntary movements of the passengers can change the accelerometers' positioning, compromising the measurements' quality. Nevertheless, if performed adequately, it constitutes the most adequate approach to evaluate passenger comfort as it considers more realistically the complexity of the seat structure dynamics [12–15]. Seat dynamics is quantified regarding transmissibility, which verifies seat efficiency in handling vibration discomfort. Train seats are crucial to decrease vibration transmission from the floor to the user. Together, the seat and human body constitute a coupled, complex dynamic system [9,16–18]. A significant discrepancy can be observed when comparing the transmissibility of subjects with identical characteristics, including identical mass. Moreover, the human body has its own natural resonance frequencies that may lead to physiological responses when matched. The range between 5 and 10 Hz leads to resonance in the chest and abdominal region, and a higher frequency range, comprehended between 20 and 30 Hz, affects the head and neck muscles. Lastly, in the interval between 30 and 60 Hz, the ocular system perceives resonance [19–22]. Therefore, seat transmissibility can be significantly influenced by the biodynamics of the human body and its pronounced nonlinearity [23].
Chapter V 86 The biodynamic response of the human body is cross-axis-coupled, i.e., inducing vibration in one axis may lead to a response in another axis. Significant body movements in the fore-and-aft direction are produced during vertical excitation. This way, seat transmissibility is cross-axis-coupled. Additionally, the influence of subjects' physical characteristics, such as weight, on the resonance frequency and seat transmissibility has been evaluated by several authors. Multiple studies developed using subjects with varying mass ranges reported unaffected results. Increasing the loading on the seat surface (by the subject's weight) tends to increase the foam's dynamic stiffness. This foam behaviour explains the absence of the subject's weight's effect on seat transmissibility [24–27]. Transmissibility differs in direction, particularly in the vertical, fore-and-aft, and lateral directions, and location, namely the seat surface and seatback. Several laboratory experiments have demonstrated a vertical transmissibility peak at 4–6 Hz when sitting upright with seatback support [18,27–29]. However, these experiments did not implement transmissibility tests in traditional train seats. Instead, a single rigid seat frame with distinct characteristics from the ones of a current train seat, such as frame dimensions and support points, was considered. In addition, the foam was freely placed on top of the seat surface without any restricting cover [18,27–29]. The constraint conditions of the foam are crucial to correctly address the influence of the foam thickness variation on the transmissibility of the seat, particularly on the evaluation of its dynamic stiffness [29]. Regarding seat cushion properties, changing foam thickness has generally been found to have the most prominent and predictable effects on seat transmissibility. Doubling the foam thickness roughly halves its stiffness, leading to increased transmissibility and discomfort. Patelli and Griffin [18] conducted a study to observe the effect of foam thickness on seat transmissibility. Increasing the thickness from 40 to 80 mm reduced the transmissibility from 4–7 Hz to 3–5 Hz. Resonance frequencies were found to be equal to 3–4 Hz, for 40 mm foam, decreasing to 2–3 Hz for 80 mm foam. The experimented individuals did not have contact with the backrest [18]. A similar tendency was reported by Zhang et al. [30], where a higher transmissibility resonance was observed when the seat surface foam increased from 60 mm to 80 mm, and posteriorly to 100 mm. In turn, changing the foam hardness (composition and density) does not significantly affect seat transmissibility. Due to hysteresis loss, density affects the static comfort instead of the seat transmissibility. Ebe and Griffin [8] examined seat transmissibility by considering four foam cushions with equal dimensions but varying densities (from 45 to 65 kg/m3). Seat transmissibility did not present significative differences, contrary to what was observed with the subjects' comfort judgments. Therefore, those judgment differences were related to the static seat comfort [8].
Chapter V 87 Based on the limitations of the previous experiments, this chapter aims to give clear contributions about some aspects that presently, according to the authors' knowledge, are not sufficiently addressed in the existing literature, particularly: - Development of an experimental test setup on a real train environment. This way, the influence of the flexible seat base, instead of the currently rigid base used in laboratory experiments, as well as the influence of the carbody/bogies resonance movements, are considered on the seat dynamic performance. In addition, this dedicated experimental setup can clearly distinguish the seat frame's structural movements from the foam's local vibrations. - Accurate characterisation of modal parameters of the seat structural frame and surface foam for different types of seats belonging to AP and IC trains, including standard and comfort seats. - Train seat modal parameters validation based on the application of two distinct output-only techniques, one based on transmissibility functions, which is used by most of the authors in the bibliography, and the other based on the Enhanced Frequency Domain Decomposition (EFDD) method. The latter allows for visualising the modal configurations, which is a novelty in relation to the previous research works. 5.2 Modal identification methodologies Vibration measurements have been intensively applied to structural health monitoring applications in civil and transportation engineering [31–33]. The output-only modal analysis, also called operational modal analysis, relies on applying modal identification techniques based exclusively on the structural response measurements. Modal identification techniques are divided into two groups, those based on the frequency domain and those based on the time domain. The present chapter considers only frequency domain techniques [34], mainly the transmissibility function and the EFDD. 5.2.1 Transmissibility Function Seat dynamics is quantified regarding transmissibility, which is an indicator of dynamic comfort [1,2,6,7,16,27]. Transmissibility is defined as the ratio between the response amplitude at a specific seat location under an external forced vibration and the excitation amplitude at the seat base, both expressed in the frequency domain [35]. Therefore, the transmissibility function between the user–seat interface and the floor, 𝐻(𝑓), is a non-dimensional parameter calculated by the following Equation (5.1):
Chapter V 94 (a) (b) Figure 5.6. Experimental test setup: (a) seat pad location and experimental test; (b) accelerometers distribution along seat frame structure. Each number represents the modal node used on ARTeMIS software. Time series were acquired in periods of 5 min, with a sampling frequency of 2048 Hz, posteriorly decimated to a frequency of 64 Hz. The down-sample was conducted based on the assumption that the frequencies of interest regarding passengers' comfort are, generally, ranged between DC and 20 Hz. In addition, the data are presented between DC and 12 Hz, which enables a closer view of the frequencies of interest. The data acquisition system was composed of a National Instruments (NI) cDAQ-9172 equipped with IEPE analogue input modules with 24-bit resolution (NI 9234) connected to a PC to acquire and record data measurements. The vibration was induced by a group of people randomly walking and jumping nearby the seat. The experiments were run with an interior train temperature of around 20 °C. 5.3.2 Results Results are divided according to the modal identification technique used: the transmissibility curves and EFDD method.
Chapter V 95 5.3.2.1 Transmissibility curves Transmissibility curves were evaluated using MATLAB scripts validated in previous works [50]. These curves were obtained based on the information from accelerometers 1 and 6, both at the z-axis. Due to the similarity between experiments and to obtain a cleaner presentation, only one experiment per subject is presented. Figure 5.7 shows the transmissibility curves for the three subjects (S1, S2, and S3), considering the comfort (Figure 5.7 (a)) and standard (Figure 5.7 (b)) seats. It is possible to observe the resonance peaks (marked by red circles) and the antiresonance peaks (marked by blue circles). (a) (b) Figure 5.7. Transmissibility curves: (a) comfort seat; (b) standard seat. Table 5.2 summarises the main resonance and antiresonance peaks for comfort and standard seats. Due to its intensity, it is expected that some of these frequencies cannot induce a different type of movement.
Chapter V 96 Table 5.2. Main resonance and antiresonance frequency values for comfort and standard seats of AP train. Subject Comfort Standard Resonance (Hz) Antiresonance (Hz) Resonance (Hz) Antiresonance (Hz) S1 0.20 0.59 0.20 1.37 1.37 4.49 1.76 3.91 2.54 2.34 4.88 4.69 S2 0.20 0.39 0.20 1.17 1.56 4.69 1.76 3.52 2.54 2.34 4.88 4.30 S3 0.20 0.39 0.20 0.78 1.37 4.30 1.76 4.30 2.54 2.73 4.69 4.69 Comfort seat transmissibilities are characterised by resonance frequencies of 0.20, 1.37–1.56, 2.54, and 4.69–4.88 Hz. It should be pointed out that the higher-amplitude transmissibility peaks occur at low frequencies, namely 0.20 and 1.37–1.56 Hz. The antiresonance values are found between 0.39–0.59 and 4.30–4.69 Hz. Despite the inter-subject variability, no significant differences are found between the transmissibility resonances and antiresonances frequency values when comparing all subjects' results. These conclusions comply with the ones found by Toward and Griffin [27]. Regarding standard seat transmissibility, the results are similar for all analysed subjects. Resonances of 0.20, 1.76, 2.34–2.73, and 4.30–4.69 Hz and antiresonances equal to 0.78–1.37 and 3.52–4.30 Hz characterise this class. Higher transmissibility resonances are obtained for 0.20 and 1.76 Hz. Those transmissibility peaks are higher than the ones of the comfort seat. This result agrees with the conclusions derived from the studies performed by Patelli and Griffin [18] and Zhang et al. [30], where the transmissibility peak values decreased for higher foam thicknesses. As aforementioned, the comfort seat has a seat surface thickness of 190 mm, whereas the standard seat thickness is equal to 130 mm. The previously mentioned studies evaluated the effect of foam thickness on thicker foams, namely 40, 60, 80, and 100 mm. Therefore, the AP seat foam thickness is higher than those of the previous studies. Following the reported trend, comparing those results with the present ones, lower transmissibility frequencies in the present results were observed.
Chapter V 97 The transmissibility frequency resonance peaks presented in this study are lower than those presented by other authors [18,27–29], who found vertical transmissibility frequencies around 4.30 Hz. However, the experiments reported in the bibliography were performed in laboratory conditions instead of using a real railway vehicle. More specifically, those tests considered a single seat with a simplified rigid metallic structure (with a mass of approximately 1000 kg) composed of 4 vertical rigid columns and a free foam on top of the surface without covering it nor restricting its movement. In turn, the train seats instrumented in this study present a leather cover with a large seam hem. The cover significantly modifies the foam vibration transmission by restricting the foam cellular morphology movements. Therefore, the vibration absorption is clearly distinct compared to the laboratory experiments' free foam. Figure 5.7 also identifies a peak around 3 Hz associated with the rigid body movement of the seat frame over the carbody due to the stiffness provided by the flexible connection elements, particularly screws and rubber pads. These elements act as a spring that flexibilize the seat frame movements. 5.3.2.2 EFDD method The modal identification was performed using the EFDD method available in the ARTeMIS software [51]. Once the inter-subject variability did not notice any significant alteration in transmissibility analysis, the EFDD method was performed only for subject S1. The natural frequencies estimated from the EFDD method and transmissibility functions are equal once the latter is the base for identifying the natural modes. Figure 5.8 shows the first two curves of the average normalised singular values of the spectral density matrices of subject S1 considering the comfort seat test. In this figure, it is possible to observe the peaks corresponding to the identified frequencies and mode shapes.
Chapter V 98 Figure 5.8. EFDD method: average normalised singular values of the spectral density matrices regarding AP comfort seat. The seat modal parameters, namely the natural frequencies and modal configurations, are depicted in Figure 5.9, where 𝑓 represents the average value of each natural frequency. The figure represents each modal configuration by the front and lateral views. Moreover, the undeformed structure is represented in blue, while the deformed structure is red-coloured. Figure 5.9. Experimental modal parameters for the comfort seat. As expected, each vibration mode corresponds to a different seat movement. Mode 1, with a frequency of 0.20 Hz is related to the vertical foam local movements. Mode 2, with a frequency of 1.37 Hz, is possibly associated with a global movement of the vehicle carbody, which induces a rigid body motion on the seat structure, with a slight foam movement in the vertical and longitudinal directions. This longitudinal component is associated with distorsional Mode 1 f1 = 0.20 Hz Mode 2 / Mode 3 f2 = 1.37 Hz / f3 = 2.54 Hz Mode 4 f3 = 4.49 Hz Mode 5 f4 = 4.88 Hz
Chapter V 99 movements of the foam layer. Concerning the AP train natural frequencies, Ribeiro et al. [50], based on a forced vibration test of the carbody, estimated the frequency values of the three rigid body modes of the carbody, associated with rolling, bouncing, and pitching movements, ranging between 1.0 Hz and 1.6 Hz, which are in full compliance with Mode 2's physical interpretation. It should be highlighted that the 0.59 Hz antiresonance frequency identified on the transmissibility curve matches the same movements as Mode 2. Mode 3, with a frequency around 3 Hz, is associated with rigid body movements of the seat frame due to the flexibility of the connection elements between the seat and the carbody. This modal configuration has similar movements to the ones of Mode 2. Mode 4 demonstrates a seat movement transition. Characterised by this antiresonance frequency, the seat frame presents a quite well-defined bending movement coupled with a torsion movement, as stated by the longitudinal movement of the seat foam. The bending of the seat frame is naturally mobilised by the externally induced vibration and the mass of the passenger (Figure 5.10). At this frequency value, the foam layer strictly follows the movement of the seat frame without any type of independent vibration, being comparable to the result obtained by Ribeiro [50], 4.23 Hz, conducted in the same environment but on a pre-renovation seat and with a simplified experimental setup. That comprised an accelerometer placed on the floor and one tri-axial accelerometer on the seat surface. Therefore, the author obtained the transmissibility curves but could not perform the modal identification. The new foams and restricting covers justify the slight difference between results. Figure 5.10. Demonstration of seat bending movement. Mode 5, with a frequency equal to 4.88 Hz, is associated with a coupled mode involving structural movement of the carbody, as stated by the quite visible support vertical movement, structural movement of the seat frame, demonstrated by bending movement, and vertical foam movements.
Chapter V 100 In conclusion, resonance modes 1 and 5 are connected with the seat frame and foam natural frequencies and, therefore, amplified movements. Mode 2 and the antiresonance valley, Mode 4, are strictly connected with seat frame movements and the generation of those movements transitions. Figure 5.11 illustrates the first two curves of the average normalised singular values of the spectral density matrices of the S1 standard seat test, where it is possible to observe the peaks corresponding to the identified frequencies and mode shapes. Figure 5.11. EFDD method: average normalised singular values of the spectral density matrices regarding AP standard seat. The seat movements derived from the vibration modes are demonstrated in Figure 5.12, where 𝑓 represents the average value of the natural frequency. Figure 5.12. AP standard seat experimental modal parameters, where the front and lateral views are presented. Mode 1 f1 = 0.20 Hz Mode 2 / Mode 3 f2 = 1.37 Hz / f3 = 2.34 Hz Mode 4 f3 = 3.91 Hz Mode 5 f4 = 4.69 Hz
Chapter V 101 As previously mentioned, standard and comfort seats differ mainly in terms of dimensions. Therefore, it was anticipated that both seats face a similar structural behaviour. Due to the foam thickness difference, a distinct foam vibration range absorption was expected, as observed on transmissibility curves. To confirm the similar structural behaviour of both seats, mode pairing was conducted through the MAC parameter. Mode 2 presents an MAC value of 0.96, followed by Mode 1 (MAC equal to 0.93) and Mode 5 (0.86). Indeed, the modal analysis demonstrated that standard and comfort seats present the same seat frame movements occurring at different frequencies. Mode 1 is the exception, as it occurs on the same frequency for both seats. Standard seat Mode 2 matches Mode 2 of the comfort seat. Modes 4 and 5, characterised by higher deformations, occur at slightly, but not significantly, lower frequencies for the standard seat. The standard seat foam demonstrates the capacity to absorb vibrations within the 1.37–3.91 Hz frequency range. This parameter represents the main difference regarding both seats, as the comfort foam presents a larger vibration range absorption (0.59–4.88 Hz). 5.4 Case study: Intercity train The IC train service was introduced earlier than the AP in 1980 and was renovated in 2002. Figure 5.13 illustrates the IC train in its regular passenger service. This service is run by 5600 series locomotives with five hauled Corail coaches. Generally, coaches 1 and 2 are comfort classes, while the other cars are designed as standard classes. Figure 5.13. Intercity train in operation.
Chapter V 102 Being from different classes, comfort and standard seats are distinct. In opposition to AP train seats, the IC seats differ in terms of dimensions and structural frame. While the comfort seat is a single seat (like those used in laboratory experiments), the standard seat is a classical double seat (Figure 5.14). The figure demonstrates the seat location within the train. As on the AP experimental setup, both seats were positioned in the same location inside vehicles 1 and 5, respectively, for comfort and standard classes, particularly near the rear bogies. (a) (b) Figure 5.14. Intercity train seats: (a) comfort class; (b) standard class. Regarding the dimensions, the comfort seat has a foam thickness of 80 mm, and equal to the seatback. For the standard class, the thickness is 60 mm for both seat surface and seatback. Although presenting different foam dimensions, both seats share the fabric cover type. This way, the foam is placed freely on top of the seat's metallic support structure without any movement restriction. This represents a more similar configuration with the laboratory experiments setup in comparison to the one of AP train. Regarding the structural frame, the comfort seat is composed of a vertical column connected on its base to the floor. The standard seat comprises two circular transversal girders connected at both extremities and linked with two vertical columns located at the centre of each seat. Figure 5.15 illustrates the metallic seat frame concerning both seat types.
Chapter V 103 (a) (b) Figure 5.15. Seat metallic structural frame: (a) comfort seat structural frame; (b) standard seat structural frame. 5.4.1 Dynamic tests The experimental tests were accomplished in the CP-maintenance facilities with the same subjects, accelerometers, seat location, and vibration conditions as the AP experiments. One seat of each class was instrumented on the seat surface and support frame. Due to the metallic frame differences, a different experimental setup was designed for each seating class. Regarding the comfort seat, acceleration measurements took place at the floor connection (Figure 5.16 (b), left), beneath the seat surface, at the seat surface (Figure 5.16 (a) represents this configuration with/without subject), and at the seatback (Figure 5.16 (b), right). For the standard seat, accelerometers were placed at the floor connection (z-axis), beneath both seat surfaces (z-axis), at the seat surface (z and x-axes), and at the seatback (z and x-axes). The experiments were performed with an interior train temperature of approximately 20 °C.
Chapter V 110 Mode 1 is characterised by the foam's incapacity to absorb low-frequency vibrations (<0.20 Hz), while Mode 2 combines that foam incapacity with seatback local bending movements. Mode 3 demonstrates rigid body movements induced by the car-body, associated with seatback local bending movements and foam vertical and longitudinal motion. As on the other analysed seats, the last mode is defined by a combination of the remaining modes. The fore-and-aft seatback transmissibility results show resonance peaks within the modes presenting seatback local bending movements, complying with comfort seat observations. Modes 1 and 2 are related to vibration transmission associated with vertical and longitudinal foam movements, so these can be designated as local transmissibility. Mode 3 relies on structural and foam movements; thus, it can be classified as global and local transmissibility. 5.5 Conclusions This chapter presented a dynamic experimental analysis of typical rail seats under a real train in situ environment. This analysis included the modal identification of the seat based on output-only methodologies, namely the transmissibility functions and the EFDD method. The study included the development of dedicated experimental setups that were applied on real train seats, namely the comfort and standard seats of AP and IC trains. Being employed in a real train environment, the experiments can realistically consider all aspects of the seat dynamics, particularly the effect of the flexibility of the seat surfaces and carbody components. Based on the dynamic testing results, it was possible to draw conclusions about the influence of the seat structural movements and dependence on the foam vibration absorption capacity. Regarding AP train seats, the inter-subject variability did not affect transmissibility resonances and antiresonances. Moreover, the standard seat presented higher transmissibility peaks than the comfort seat. Compared with laboratory experiments, AP seats presented lower transmissibility resonance peaks, which reveals the influence of the seat structural frame and foam leather cover. These factors significantly modify the vibration transmission; particularly, the cover restricts the foam cellular morphology movements, leading to variations in foam absorption capacity, while the seat frame movements may couple with movements of the carbody components. Modal identification depicted seat bending movements and vertical foam movements. By matching transmissibility and modal identification, it was possible to associate the resonance frequencies with seat and foam natural frequencies and the corresponding amplified
Chapter V 111 movements. In contrast, the antiresonance valleys generated those movement changes on the structural support. The IC train's comfort seat is similar to that used in laboratory experiments. The foam is freely placed on the seat surface, covered by a non-restricting fabric cover. Its results are close to those of laboratory conditions. The Intercity seats experimental setup allows for concluding about the influence of the seatback local bending movements on the transmissibility response. Seatback transmissibility resonances matched the local bending movement frequencies identified by modal identification. This way, the applied methodology was validated based on the strong correlation between the transmissibility curve and the EFDD method. The present work demonstrated the influence of seat structural frame movements and foam vibration absorption on user–seat vibration transmissibility. Moreover, this is a pioneer study regarding train seat modal identification, which may be used for further studies concerning seat vibration reduction to increase passengers' comfort.
Chapter V 112 Chapter references [1] Kim, Y.G.; Kwon, H.B.; Kim, S.W.; Park, C.K.; Park, T.W. Correlation of ride comfort evaluation methods for railway vehicles. Proc. Inst. Mech. Eng. Part F J. Rail Rapid Transit 2002, 217, 73–88. https://doi.org/10.1243/095440903765762823. [2] Jiang, Y.; Chen, B.K.; Thompson, C. A comparison study of ride comfort indices between Sperling’s method and EN 12299. Int. J. Rail Transp. 2019, 7, 1–18. https://doi.org/10.1080/23248378.2019.1616329. [3] EN 12299; Railway Applications—Ride Comfort for Passengers—Measurement and Evaluation. European Committee for Standardization: Brussels, Belgium , 2009. [4] Wibowo, R.K.K.; Soekarno, S.; Puspitasari, I. Analysis of Train Passenger Seat Using Ergonomic Function Deployment Method. Int. J. Transp. Veh. Eng. 2017, 11, 1747–1751. https://doi.org/10.5281/zenodo.1132711. [5] Bosso, N.; Gugliotta, A.; Zampieri, N. Design and simulation of a railway vehicle for the transport of people with reduced mobility. Shock. Vib. 2018, 2018, 9207639. https://doi.org/10.1155/2018/9207639. [6] Hostens, I. Analysis of Seating during Low Frequency Vibration Exposure. Ph.D. Thesis Katholieke Universiteit Leuven, Leuven, Belgium, 2004. [7] Park, S.J.; Subramaniyam, M. Evaluating Methods of Vibration Exposure and Ride Comfort in Car. J. Ergon. Soc. Korea 2013, 32, 381–387. https://doi.org/10.5143/jesk.2013.32.4.381. [8] Ebe, K.; Griffing, M. Factors affecting seat cushion comfort. Ergonomics 2001, 44, 901–921. https://doi.org/10.1080/00140130110064685. [9] Montenegro, P.A.; Ribeiro, D.; Ortega, M.; Millanes, F.; Goicolea, J.M.; Zhai, W.; Calçada, R. Impact of the train-track-bridge system characteristics in the runnability of high-speed trains against crosswinds— Part II: Riding comfort. J. Wind. Eng. Ind. Aerodyn. 2022, 224, 104987. https://doi.org/10.1016/j.jweia.2022.104987. [10] UIC 513 (E). Guidelines for Evaluating Passenger Comfort in Relation to Vibration in Railway Vehicles; International Union of Railways (UIC): Paris, France , 1994. [11] ISO 2631; Mechanical Vibration and Shock—Evaluation of Human Exposure to Whole-Body Vibration. International Standard Organization: Geneva, Switzerland , 2001. [12] Schust, M.; Blüthner, R.; Seidel, H. Examination of perceptions (intensity, seat comfort, effort) and reaction times (brake and accelerator) during low-frequency vibration in xor y-direction and biaxial (xy-) vibration of driver seats with activated and deactivated suspension. J. Sound Vib. 2006, 298, 606–626. https://doi.org/10.1016/j.jsv.2006.06.029. [13] Hong, K.-T.; Hwang, S.-H.; Hong, K.-S. Automotive Ride-Comfort Improvement with an Air Cushion Seat. In Proceedings of SICE Annual Conference in Fukui, Fukui, Japan, 4–6 August 2003 ; Fukui University: Fukui, Japan, 2003. [14] Bonin, G.; Cantisani, G.; Loprencipe, G.; Sbrolli, M. Ride quality evaluation: 8 D.O.F. vehicle model calibration. In Proceedings of the 4th International SIIV Congress, Palermo, Italy, 12–14 September 2007. [15] van Niekerk, J.L.; Pielemeier, W.J.; Greenberg, J.A. The use of seat effective amplitude transmissibility (SEAT) values to predict dynamic seat comfort. J. Sound Vib. 2003, 260, 867–888. https://doi.org/10.1016/S0022-460X(02)00934-3. [16] Maier, J.; Zierke, O.; Hoermann, H.J.; Goerke, P. Effects of personal control for thermal comfort in long-distance trains. Energy Build. 2021, 247, 111125. https://doi.org/10.1016/j.enbuild.2021.111125.
Chapter V 113 [17] Năstac, S.; Picu, M. Evaluating Methods of Whole-Body-Vibration Exposure in Trains. In The Annals of “Dunarea de Jos”; University of Galați: Galați, Romania , 2010. [18] Patelli, G.; Griffin, M.J. Effects of seating on the discomfort caused by mechanical shocks: Measurement and prediction of SEAT values. Appl. Ergon. 2019, 74, 134–144. https://doi.org/10.1016/j.apergo.2018.08.003. [19] Harris, C.; Piersol, A. Harris’ Shock and Vibration Handbook; McGraw-Hill: New York, NY, USA, 2002. [20] Chaffin, D.; Anderson, G.; Martin, B. Occupational Biomechanics; John Wiley & Sons: New York, NY, USA, 1999. [21] Bruel & Kjaer. Primer: Human Vibration, Booklet; K. Larsen & Son: Naerum, Denmark, 1989. [22] Griffin, M.J. Handbook of Human Vibration; Academic Press: London, UK, 1990. https://doi.org/10.1016/C2009-0-02730-5. [23] Tufano, S.; Griffin, M. Nonlinearity in the vertical transmissibility of seating: The role of the human body apparent mass and seat dynamic stiffness. Veh. Syst. Dyn. 2013, 51, 122–138. https://doi.org/10.1080/00423114.2012.715652. [24] White, S.W.; Kim, S.K.; Bajaj, A.K.; Davies, P.; Showers, D.K.; Liedtke, P.E. Experimental techniques and identification of nonlinear and viscoelastic properties of flexible polyurethane foam. Nonlinear. Dyn. 2000, 22, 281–313. https://doi.org/10.1023/A:1008302208269. [25] Wei, L.; Griffin, M. The prediction of seat transmissibility from measures of seat impedance. J. Sound Vib. 1998, 214, 121–137. https://doi.org/10.1006/jsvi.1998.1540. [26] Varterasian, J.; Thompson, R. The dynamic characteristics of automobile seats with human occupants. Soc. Automot. Eng. SAE 1977 , 1173–1182. https://doi.org/10.4271/770249. [27] Toward, M.G.R.; Griffin, M.J. The transmission of vertical vibration through seats: Influence of the characteristics of the human body. J. Sound Vib. 2011, 330, 6526–6543. https://doi.org/10.1016/j.jsv.2011.07.033. [28] Gong, W.; Griffin, M.J. Measuring, evaluating and assessing the transmission of vibration through the seats of railway vehicles. Proc. Inst. Mech. Eng. Part F J. Rail Rapid Transit 2018, 232, 384–395. https://doi.org/10.1177/0954409716671547. [29] Basri, B.; Griffin, M.J. The application of SEAT values for predicting how compliant seats with backrests influence vibration discomfort. Appl. Ergon. 2014, 45, 1461–1474. https://doi.org/10.1016/j.apergo.2014.04.004. [30] Zhang, X.; Qiu, Y.; Griffin, M. Transmission of vertical vibration through a seat: Effect of thickness of foam cushions at the seat pan and the backrest. Int. J. Ind. Ergon. 2015, 48, 36–45. https://doi.org/10.1016/j.ergon.2015.03.006. [31] Gharehbaghi, V.R.; Noroozinejad Farsangi, E.; Noori, M.; Yang, T.Y.; Li, S.; Nguyen, A.; MálagaChuquitaype, C.; Gardoni, P.; Mirjalili, S. A Critical Review on Structural Health Monitoring: Definitions, Methods, and Perspectives. Arch. Comput. Methods Eng. 2022, 29, 2209–2235. https://doi.org/10.1007/s11831-021-09665-9. [32] Ghannadi, P.; Kourehli, S.S.; Mirjalili, S. The application of PSO in structural damage detection: An analysis of the previously released publications (2005–2020). Frat. Integrità Strutt. 2022, 16, 460–489. https://doi.org/10.3221/IGF-ESIS.62.32. [33] Yang, Y.; Zhang, Y.; Tan, X. Review on vibration-based structural health monitoring techniques and technical codes. Symmetry 2021, 13, 1998. https://doi.org/10.3390/sym13111998. [34] Dziedziech, K.; Staszewski, W.J.; Mendrok, K.; Basu, B. Wavelet-Based Transmissibility for Structural Damage Detection. Materials 2022, 15, 2722. https://doi.org/10.3390/ma15082722. [35] Kolich, M.; Essenmacher, S.; McEvoy, J. Automotive seating: The effect of foam physical properties on occupied vertical vi-bration transmissibility. J. Sound Vib. 2005, 281, 409–416. https://doi.org/10.1016/j.jsv.2004.03.058.
Chapter V 114 [36] Dong, R.; Welcome, D.; McDowell, T.; Wu, J. Theoretical relationship between vibration transmissibility and driving-point response functions of the human body. J. Sound Vib. 2013, 332, 6193– 6202. https://doi.org/10.1016/j.jsv.2013.07.017. [37] Pozharskiy, D.; Zhang, Y.; Williams, M.O.; McFarland, D.M.; Kevrekidis, P.G.; Vakakis, A.F.; Kevrekidis, I.G. Nonlinear res-onances and antiresonances of a forced sonic vacuum. Phys. Rev. E 2015, 92, 063203. https://doi.org/10.1103/PhysRevE.92.063203. [38] Shi, Y.; Li, S. An inverse modification method for assigning antiresonant frequencies. Appl. Acoust. 2020, 170, 107524. https://doi.org/10.1016/j.apacoust.2020.107524. [39] Brincker, R.; Zhang, L.; Andersen, P. Modal Identification from Ambient Responses using Frequency Domain Decomposition. In Proceedings of the 18th International Modal Analysis Conference (IMAC), San Antonio, TX, USA, 7–10 February 2000 ; pp. 625–630 [40] Allemang, R.; Brown, D. A Correlation Coefficient for Modal Vector Analysis. In Proceedings of the 1st International Modal Analysis Conference (IMAC I), Orlando, FL, USA, November 1982 ; pp. 110– 116. [41] Magalhães, F. Identificação modal estocástica para validação experimental de modelos numéricos [Stochastic modal identifi-cation for experimental validation of numerical models]. Ph.D. Thesis, University of Porto, Porto, Portugal , 2004. (In Portuguese) [42] Brincker, R.; Ventura, C.E.; Andersen, P. Damping estimation by frequency domain decomposition. In Proceedings of the International Modal Analysis Conference—IMAC 2001, Kissimmee, FL, USA, 5–8 February 2001; pp. 698–703. [43] Allemang, R.J. The modal assurance criterion–twenty years of use and abuse. Sound Vib. 2003, 37, 14–23. [44] Ghannadi, P.; Kourehli, S.S. Efficiency of the slime mold algorithm for damage detection of largescale structures. Struct. Des. Tall Spec. Build. 2022, 31, e1967. https://doi.org/10.1002/tal.1967. [45] Pastor, M.; Binda, M.; Harčarik, T. Modal assurance criterion. Procedia Eng. 2012, 48, 543–548. https://doi.org/10.1016/j.proeng.2012.09.551. [46] Ribeiro, D.; Calçada, R.; Brehm, M.; Zabel, V. Train–Track–Bridge Dynamic Interaction on a Bowstring-Arch Railway Bridge: Advanced Modeling and Experimental Validation. Sensors 2022, 23, 171. https://doi.org/10.3390/s23010171. [47] Meixedo, A.; Ribeiro, D.; Calçada, R.; Delgado, R. Global and local dynamic effects on a railway viaduct with precast deck. In Proceedings of the Civil-Comp Proceedings 2014, Naples, Italy, 2–5 September 2014; p. 104. [48] Comboios de Portugal. CP—Comboio Elétrico Pendular, Séries 4000; Comboios de Portugal: Lisbon, Portugal , 1998. [49] Comboios de Portugal. CP—Comboio Elétrico Pendular, Séries 4000; Comboios de Portugal: Lisbon, Portugal , 2017. [50] Ribeiro, D.; Calçada, R.; Delgado, R.; Brehm, M.; Zabel, V. Finite-element model calibration of a railway vehicle based on experimental modal parameters. Veh. Syst. Dyn. 2013, 51, 821–856. https://doi.org/10.1080/00423114.2013.778416. [51] ARTeMIS. ARTeMIS Extractor Pro—Academic License, User’s Manual, SVS 2009; Aalborg; Denmark; ARTeMIS
Chapter VI 115 Chapter VI - Influence of static factors on sitting comfort Static comfort is relevant in low or no-vibration situations. The foam mechanical properties and interface pressure mainly evaluates the static comfort parameters. Another critical parameter affecting perceived comfort is the seat-passenger interface temperature. The present chapter presents a complete static evaluation of different foams and seat covers to assess the most suitable solution for railway seating applications. Based on the evaluation of foam mechanical properties and interface pressure profiles, it was concluded that the foam presenting a higher density (80 kg/m3) is the most prominent solution. Regarding the foam cover, a thermographic assessment demonstrated that the fabric cover induced lower temperatures at passenger interface contact, promoting higher comfort levels. It should be highlighted that experiments were conducted on real train seat cushions and environments with a thermographic camera and pressure map sensor.
Chapter VI 116 6.1 Introduction Public transportation users demand faster, safer, and more comfortable journeys. Passengers spend most of their time seated, so the seat assumes a crucial impact on users' comfort [1]. This way, seats should fulfil their primary support function and provide increased comfort. Perceived comfort is divided into static and dynamic categories, dependent on the presence or absence of vibration. Both are strongly connected, as the influence of one relies on the other. Thus, both comfort categories must be evaluated to properly quantify comfort levels. Without vibration, static comfort dominates the overall comfort evaluation [2,3]. When seated, passengers are in close contact with seat cushions, being those considered one of the most critical factors affecting sitting comfort [4,5]. Therefore, evaluating the foam's mechanical properties and its influence on passengers' comfort is critical. Traditionally, seat cushions are made of polyurethane (PU) foams. Those have been universally used in multiple industries due to their versatile physical properties, such as excellent sound absorption and low thermal conductivity, comfort properties, low cost and easy processing. PU foams are cellular polymers, typically produced by reacting an isocyanate with polyols in the presence of blowing and crosslinking agents and additives [6–8]. As mentioned in Chapter II, there is no consensus on comfort definition; thus, comfort quantitate evaluation is still ambiguous [9–12]. Several authors suggest evaluating seat physical characteristics and its relation with sitting comfort as the first approach to overcome the aforementioned limitation. From those physical properties, the hardness, the support factor (SAG factor) and the hysteresis loss were identified as strongly affecting seating comfort, becoming the main focus of most studies. Hardness corresponds to the foam's firmness, the SAG factor indicates the supporting property and the hysteresis loss correlates with resilience [7]. Understanding the influence of the seat's physical characteristics on passengers' sitting impressions is fundamental for designing the ideal foam, which can be accomplished by changing its mechanical properties to achieve high comfort levels [3,13]. According to De Looze et al. [14] comfort model, which is applied as the base comfort model of the present study, comfort and discomfort resulting from the interaction between the product, the user and the environment — being evaluated by a combination of psychological and physical factors. The users' expectations define psychological factors, being assessed based on questionnaires and rating scales (subjective evaluation). In contrast, the physical parameters depend on the human body weight, shape and anthropometric dimensions, being defined by objective measurements [13,15–17]. The passenger's pressure on the seat cushion strongly affects the seating interaction due to the biomechanical load imposed by the user's musculoskeletal structure. That load may lead to skin and tissue deformation,
Chapter VI 117 affecting blood flow and the musculoskeletal system [18]. The skin has thermoreceptors capable of detecting temperature changes and four mechanoreceptors that provide information about touch, pressure, vibration, and cutaneous tension. The Meissner's corpuscles are exceptionally efficient in transducing information regarding tactile and sensitive changes. The Pacinian corpuscles, responsible for detecting vibration, have a lower response threshold than the Meissner ones acting faster than these. Merkel's disks and Ruffini's corpuscles are slowly adapted mechanoreceptors. The former specialises in detecting pressure changes, and the latter can detect stretch and shear stresses. The thermoreceptors and the mechanoreceptors record the stressed surface changes and send the information to the brain, leading to comfort perception [19–21]. Therefore, although the foam's mechanical properties influence comfort, the seat/skin interface pressure and temperature are also essential factors. The present chapter deals with the evaluation of static parameters and their influence on passengers' discomfort. The foam properties of two samples were initially assessed in the lab. Then, those foams' interface pressure and temperature were evaluated using a pressure map sensor and an infrared thermographic camera. Moreover, it was also possible to draw conclusions about the influence of the seat cover. 6.2 Influence of foam properties in sitting comfort impressions As previously described, foam cushions represent the primary contact between the user and the seat. Thus, some authors studied the relationship between foam's physical characteristics and sitting comfort. This connection was assessed by means of subjective questionnaires and interface pressure measurements. Lei et al. [22] related hysteresis loss with seating comfort once it correlates with resilience. Hysteresis loss is defined as the percentage ratio between the area under the loading (0-75%) and unloading (75-0%) compression cycles. Its calculation proceeds as follows (6.1): 𝐻𝑦𝑠𝑡𝑒𝑟𝑒𝑠𝑖𝑠 𝑙𝑜𝑠𝑠 %= 𝐴0−75% 𝐴75−0% (6.1) Ideally, foams should present low hysteresis loss percentages, this value increases as resilience decreases [23,24]. Hysteresis loss is highly influenced by the cellular foam structure, specifically the cell/wall area ratio. During the compression cycle, the energy loss happens primarily due to the buckling affecting unrecovered cell walls and struts. Thus, to obtain low hysteresis, low cell/wall area ratios are required [25].
Chapter VI 118 The SAG factor is another measure of foam's suitability for sitting applications. This is obtained by the ratio of the compressive load at 65% deformation (𝐹65%) and that load at 25% foam deformation (𝐹25%), as demonstrated in Equation (6.2): 𝑆𝐴𝐺 𝑓𝑎𝑐𝑡𝑜𝑟= 𝐹65% 𝐹25% (6.2) Once the SAG factor provides an indication of the foam-supporting property, high values are associated with higher comfort performance. The supportive factor is influenced by the foam's cell structure and density. Regarding seating applications, a SAG value higher than 2.8 is recommended; otherwise, users may feel the seat pad bottom [13,23–26]. Lee and Ferraiuolo [27] identified foam thickness and hardness as significantly affecting sitting comfort. Hardness is defined by the 25% indentation load deflection (ILD). This relates to foam stiffness and defines the foam's resistance to deformation when applying a load. High values correspond to high firmness, whereas low values are associated with lower firmness [28]. Using foam samples with equal density and hardness but varying thicknesses (50, 75 and 100 mm), Neal [29] concluded that foam thickness does not influence hysteresis loss but influences the 25% load, as thicker foams reported higher forces. Ebe and Griffin [13] conducted a study to relate the influence of foam properties, such as density, SAG factor and hysteresis loss, in sitting comfort impressions. The sitting impressions of twelve subjects were obtained to judge the comfort of four foams with equal hardness (≈205 N) but varying density (4565 kg/m3). The high-durability seat (55 kg/m3) was rated as the most comfortable, whereas the lowestdensity foam (45 kg/m3) was stated as the least comfortable. Moreover, the low-density foam presented the highest hysteresis loss (34%). In contrast, the high-durability foam demonstrated a lower hysteresis (29%). A high correlation was observed between the comfort scores and the SAG factor. A second experiment was performed to evaluate the influence of foam's hardness. The same twelve subjects rated the comfort perception of five foams with 25% ILD hardness between 120-285 N. Results suggested that hardness influences static comfort, but the preferred foam varied between subjects. Additionally, it should be highlighted that harder foams noticed the lowest hysteresis loss. Similar conclusions were reported by Moon et al. [23] when studying the production of PU foams with multiple layers. The foam considered the best for seating applications demonstrated a SAG factor around 3, hardness of 150 N and hysteresis loss of approximately 22.5%. Analysing the aforementioned studies, the ideal foam is characterised by a high SAG factor, low hysteresis loss and intermedium to high density.
Chapter VI 119 6.3. Influence of interface pressure in sitting impressions While sitting, the human body is in direct contact with the seat cushion, exerting pressure due to the biomechanical load imposed by the user's musculoskeletal structure. Following this assumption, multiple authors explored and defined comfort based on the seat-passenger interface pressure. The pressure profile strongly connects with discomfort, as proved by the correlation obtained between objective and subjective evaluations, being capable of evaluating the biomechanic factors of sitting discomfort. Therefore, the seat static comfort evaluation usually begins by considering body pressure distribution [9,13,14,30]. Indeed, when seated, more than 70% of the human body weight is supported by the seat cushion. Since the human buttocks are not flat, the seat-passenger interface pressure varies over the seat surface. That feature leads to a significant pressure concentration in the ischial tuberosity area, resulting in pressure peaks. Those peaks may provoke low cell oxygen content, prompting fatigue, pain, and discomfort. Although defining a maximum pressure threshold is fundamental for providing high comfort levels and reducing the harmful consequences for passengers, no consensus has been obtained [3,30– 32]. Nevertheless, multiple authors stated that 32 mmHg, corresponding to the capillary pressure, should not be exceeded. Higher pressure may obstruct the capillaries, limiting blood circulation and leading to oxygen deprivation in the tissues, causing discomfort [2,16,31,33,34]. In opposition, other authors defended that a seat is still considered comfortable when presenting a maximum pressure of 43.50 mmHg under the ischial tuberosities and 21.75 mmHg elsewhere. However, those authors defined that the maximum mean pressure should never exceed 37.75 mmHg; once the skin capillaries close at this pressure [15,32]. Besides, no consensus on defining the pressure distribution threshold has been reached, it is unanimous that a widely dispersed distribution without local peak concentration represents the ideal foam [2,15,16,32–35]. Pressure distribution profiles provide information regarding multiple parameters. Those more related to discomfort are the contact area, maximum pressure peaks (especially those near the ischial tuberosities) and average pressure [36,37]. Moreover, the pressure profile can be affected by several factors, namely the user's anatomical characteristics (weight, height, BMI, and body part) and foam properties (hardness, SAG factor, hysteresis, thickness, and contour). Several authors reported lower peak pressure concerning heavier subjects. Those tend to have higher contact areas, leading to lower peak pressures but higher mean pressures. In opposition, thinner subjects present higher peak pressure around the ischial tuberosities [15,38]. Li et al. [38] evaluated
Chapter VI 126 Observing the results, the older foams demonstrated a higher SAG factor than the new ones. Its results are closer and superior to 2.8 regarding the standard and comfort cushions, respectively. Current and older seat foams are different in terms of density, but equal in dimensions. Therefore, based on the current results, it may be concluded that density is affecting the foam-supporting capacity. Moreover, the current standard seat has the lowest supporting factor of all foams. When comparing both seat types (comfort and standard) SAG factors, an increase is noticed for the thicker seat, which may indicate the influence of foam thickness on this physical property. The 25% ILD hardness evaluated hardness. To understand the effect of the seat cover, cushions were compressed with and without cover, and the increasing hardness percentage was calculated. Table 6.3 exhibits the 25% ILD cushion hardness with and without seat cover and the increase %. Table 6.3. Foam hardness results with and without a seat cover. Seat cushion 25% ILD hardness (foam) (N) 25% ILD hardness with seat cover (N) Increasing % Current comfort seat 288 360 26 Current standard seat 250 320 28 Older comfort seat 271 354 44 Older standard seat 167 240 30 By analysing the results of Table 6.3 it can be concluded that current foams have a higher hardness than older ones. Moreover, hardness increases as the foam thickness rises, which is demonstrated by comfort foams that show a higher 25% ILD hardness than the standard ones, in agreement with Neal [29]. Additionally, the seat cover highly influences the system's (foam and cover) hardness. While the leather cover promoted a maximum increase of 28%, the fabric cover corresponded to a 44% hardness increase. It should be highlighted that if a cover is stretchable and loosely connected to the foam, its mechanical properties can be more predominant as the foam can freely deform. In opposition, a stiff cover tightly attached to the foam may limit its deformation behaviour and, consequently, influence its performance [18]. Additionally, an extra experiment assessed the foam and cover capacity for absorbing energy. That consisted of releasing a ball (52 mm diameter and weighing 127 g) and measuring its initial and bump heights, Figure 6.3. That was constituted by the seat (with and without cover) (1), a motion sensor CBR 2 (Texas Instruments) (2) [62] capable of measuring and recording distance, velocity, and acceleration, connected with a computer (3) and the ball (4). The calculation of energy absorption %, follows Equation (6.4):
Chapter VI 127 % 𝐸𝑎𝑏𝑠 =𝑣2 𝑣1 ×100=(ℎ2 ℎ1)12 ⁄×100 (6.4) where 𝑣1 and ℎ1are corresponding to the ball's initial speed and height and 𝑣2 and ℎ2 represent the ball's rebound speed and height. It was assumed that the ball acts as a rigid body. This way, the ball did not lose energy, but the seat cushion suffered deformation due to the collision [61]. Figure 6.3. Experimental setup employed to assess the energy absorption. Measurements were recorded at 50 Hz sample frequency. Experiments run five times for each cushion with and without the seat cover. Table 6.4 presents the experiments’ mean energy absorption percentage. Table 6.4. Mean energy absorption percentage with and without seat cover. Seat cushion Energy absorption % (foam) Energy absorption % with cover Current comfort seat 68 82 Current standard seat 66 76 Older comfort seat 71 73 Older standard seat 58 69 The cover increases the energy absorption percentage, especially the leather cover. At the current AP seats, energy absorption increases by 14 % and 10%, respectively, for the comfort and standard seats. That absorption % increase is less noticed for the older seats, as the fabric cover only increased 2 % and 11% of the energy absorption for the comfort and standard cushions, respectively. This way, based on the hardness increase and the energy absorption percentage, the leather cover appears to be more
Chapter VI 128 suitable for seating applications. Nevertheless, more experiments need to be performed to properly evaluate the cover influence. 6.5.2. Experiment 2: the influence of seat-passenger interface pressure and temperature A set of dedicated experimental tests were developed and performed to evaluate interface pressure and temperature influence on sitting comfort. The University of Minho Ethics Committee previously approved the experiment's realisation (Appendix I). Before the beginning of the experiments, the volunteers gave their informed consent (Appendix II). Three volunteers (2 male and 1 female) participated in the experiment. Table 6.5 illustrates the subjects’ anthropometric characteristics. Table 6.5. Volunteers’ anthropometric characteristics. Subject Gender Age (Years) Mass (kg) Height (m) S1 Female 25 58 1.70 S2 Male 26 80 1.87 S3 Male 39 115 1.85 The experimental setup was designed to simultaneously assess pressure and temperature at the seat-passenger interface. The pressure was measured using a CONFORMat, developed by Tekscan Inc [63], presenting a sensitive area of 471.4 mm x 471.4 mm. The pressure map was positioned at the seat surface, where the subjects sat. A FLIR A325sc thermal imaging camera was positioned perpendicularly to the seat to evaluate the seat temperature [64]. Both, pressure map and thermographic camera, were connected to individual computers containing specialised software. Figure 6.4 illustrates the complete experimental setup.
Chapter VI 129 Figure 6.4. Experimental setup developed to assess the pressure and temperature at the seat-passenger interface. The experiment consisted of sitting for 10 minutes, while pressure was measured and recorded during the entire sitting experience at a 1 Hz sample frequency. Maximum peak pressure (𝑃𝑚𝑎𝑥), average pressure (𝑃𝑎𝑣𝑔) and contact area were analysed. Concluded the 10 minutes period, thermal images of the seat were obtained. Moreover, following the aforementioned studies of multiple authors that evidenced the importance of matching objective and subjective static evaluations, volunteers rated their total and local sitting comfort impressions using a 7-point scale (Table 6.6). Local comfort was divided into four body parts, as demonstrated by Figure 6.5. Table 6.6. Subjective 7-point comfort/discomfort rating scale. Ranking Comfort/discomfort level -3 Very uncomfortable -2 Uncomfortable -1 Slightly uncomfortable 0 Neutral 1 Slightly comfortable 2 Comfortable 3 Very comfortable
Chapter VI 130 Figure 6.5. Local body parts distribution. The experiment was performed by the three subjects in the current AP seats to investigate the effects of users' anthropometric characteristics. After assessing the effects, subject S1 experimented with the older seats and fabric cover to compare seats (current and older), matching results with foam's mechanical properties and assessing the best seat configuration (foam and cover). 6.5.2.1 The influence of anthropometric characteristics The three subjects performed the experiments at the current AP train seats (comfort and standard) to evaluate the influence of users' anthropometric characteristics on perceived comfort. Pressure results regarding the comfort seat can be found in Table 6.7. Table 6.7. Maximum and mean pressure and contact area regarding the current comfort AP seat. Sub. 𝑷𝒎𝒂𝒙 (mmHg) 𝑷𝒂𝒗𝒈 (mmHg) Contact area (cm2) R1 R2 L1 L2 R1 R2 L1 L2 S1 37 26 29 20 29 20 23 15 1076 S2 57 18 50 17 46 16 44 14 1060 S3 41 29 65 27 35 22 58 22 1339 By observing the results, both, maximum and mean pressures, demonstrated values above the maximum 32 mmHg. Some authors defined this threshold as the trigger for developing discomfort and increasing the risk of health consequences, particularly for longer sitting exposures [2,16,31,33,34]. Those pressures and contact areas revealed an increasing tendency related to the subject's weight. Moreover, pressure peaks were obtained at the ischial tuberosities regions (R1 and L1). Those measured for subjects S2 (57 mmHg) and S3 (65 mmHg) are even above the pressure threshold (43.50 mmHg) stated for that specific region [15,32]. Additionally, Hu et al. [15] and Peng et al. [32] suggested that the
Chapter VI 131 mean pressure should never exceed 37.75 mmHg elsewhere in the sitting region, which is true for the present case. S1 and S3, lightest and heaviest subject’s pressure profiles are illustrated in Figure 6.6. (a) (b) Figure 6.6. Current AP comfort seat pressure profiles where: (a) S1 subject profile and (b) S3 subject profile. As observed, the two subjects demonstrated different pressure profiles, being the S3 larger and the S1 with a higher length. Comparing both contact areas, as expected, the heaviest subject produced a higher contact area than the lightest individual. Table 6.8 presents the pressure distributions regarding the standard seat. Table 6.8. Maximum and mean pressure and contact area regarding the current standard AP seat. Sub. 𝑷𝒎𝒂𝒙 (mmHg) 𝑷𝒂𝒗𝒈 (mmHg) Contact area (cm2) R1 R2 L1 L2 R1 R2 L1 L2 S1 38 21 35 24 28 16 28 19 1100 S2 51 15 46 34 41 13 38 25 1035 S3 50 49 46 35 37 38 40 29 1350 As on the comfort seat, all subjects presented maximum pressures above 32 mmHg for the standard seat as well. Nevertheless, those pressure peak values are slightly inferior to those of the comfort seat. The same decreasing tendency was observed for the mean pressure. This fact can be justified by the highest contact areas promoted by the standard seat. Regarding the interface temperature, results present the maximum, minimum, and average temperatures, and the difference between maximum and minimum temperatures for the sitting area and for a profile line. Figure 6.7 (a) shows a 3D thermal image in the course of the experiment; Figure 6.7 (b) the region of interest and a profile line, of the seat.
Chapter VI 132 (a) (b) Figure 6.7. Thermographic illustrations where (a) is on course experiment and (b) defined sitting area and analysis line. Results concerning the comfort seat are presented in Table 6.9. Table 6.9. Temperature records regarding the comfort seat. Subject Sitting area Sitting profile line Min. (°𝐶) Max. (°𝐶) Max-Min (°𝐶) Avg. (°𝐶) Min. (°𝐶) Max. (°𝐶) Max-Min (°𝐶) Avg. (°𝐶) S1 28 34.1 6.1 31.8 25.2 32.9 7.7 31.8 S2 27.1 33.1 6.0 31.3 25.0 31.2 6.1 28.5 S3 24.4 30.2 5.8 27.7 23 28.2 5.3 26.4 Regarding the sitting area, similar temperatures were obtained for the difference between maximum and minimum values. This result may indicate that the user's anthropometric characteristics do not influence the temperature increase. The same parameter of the sitting line revealed that subject S1 induced higher temperature differences when compared to the seat's initial temperature. Moreover, the same subject also induced the highest maximum and average temperatures in the sitting area. This is the only female subject, which may justify the results. Therefore, although the passengers’ anthropometric features (weight and height) do not impact the temperature, it may be observed that gender is a cause of influence. Regarding standard seat, results are exhibited in Table 6.10. Table 6.10. Temperature records regarding the standard seat. Subject Sitting area Sitting profile line Min. (°𝐶) Max. (°𝐶) Max-Min (°𝐶) Avg. (°𝐶) Min. (°𝐶) Max. (°𝐶) Max-Min (°𝐶) Avg. (°𝐶) S1 26.2 32.9 6.7 30.5 23.9 31.3 7.4 27.6 S2 25.1 31.7 6.7 29.6 21.8 31 9.2 28.1 S3 24.3 32.1 7.8 29.8 23.3 31.1 7.8 28.5
Chapter VI 133 Results of standard seats revealed a similar tendency to that of comfort, both demonstrated equivalent maximum and average temperatures in the sitting area and profile line. Subjectively, three subjects unanimously ranked the ischial tuberosity areas (R1 and L1) as "Slightly uncomfortable", which in fact, exhibited the highest-pressure peaks. Therefore, complying with the findings of Xu et al. [31] and Peng et al. [32], the ischial tuberosity area seems to be the one most influencing the perception of sitting comfort. Additionally, based on the mentioned assumption, it is also possible to conclude that pressure profiles were strongly linked to discomfort rankings [9,13,14,30]. Regarding temperature, the similarities between experiment results did not allow to conclude about its impact on sitting impressions. That influence will be further analysed in the following sub-section. It should be highlighted that the small number of volunteers is the main limitation of this experiment. 6.5.2.2. The influence of seat and cover features on sitting comfort To verify the influence of the seat and the cover on sitting comfort, S1 subject performed experiments with the current and older AP seats. As aforementioned, these seats have foams with similar dimensions, but different densities. Moreover, their covers are produced from different materials. This way, it will be possible to investigate the influence of foam mechanical properties on interface pressure and the impact of interface temperature promoted by the seat cover. Results regarding the interface pressure of current and older AP comfort seats are depicted in Table 6.11. Table 6.11. Maximum and mean pressure and contact area regarding the current and older AP comfort seat for subject S1. Seat type 𝑷𝒎𝒂𝒙 (mmHg) 𝑷𝒂𝒗𝒈 (mmHg) Contact area (cm2) R1 R2 L1 L2 R1 R2 L1 L2 Current 37 26 29 20 29 20 23 15 1076 Older 33 16 27 17 28 14 21 14 983 Generally, the older seat presented a lower average and peak pressures. Although the main peak pressure occurs at the same region as current seats (R1), it presents a lower 4 mmHg magnitude. Therefore, based on the previously found correlation between discomfort and pressure distribution, it can be concluded that the older foams induce higher comfort levels. This result is also sustained by the lower hardness presented by the older foams [13,15,35]. Pressure profiles (Figure 6.8) highlighted the differences between both foams. The current foam (Figure 6.8 (a)) induced multiple pressure peaks at the ischial tuberosities regions and a generally higher average pressure. Concerning the older foam
Chapter VI 134 (Figure 6.8 (b)), a reduced number of localised pressure peaks around the ischial tuberosities are observed. Those are coupled with a globally distributed pressure, reducing the average pressure. (a) (b) Figure 6.8. Pressure distribution profiles of (a) current AP comfort seats and (b) older AP comfort seats. As described in Table 6.12, similar pressure tendencies were obtained on the standard seats. Table 6.12. Maximum and mean pressure and contact area regarding the current and older AP standard seat for subject S1. Seat type 𝑷𝒎𝒂𝒙 (mmHg) 𝑷𝒂𝒗𝒈 (mmHg) Contact area (cm2) R1 R2 L1 L2 R1 R2 L1 L2 Current 38 21 35 24 28 16 28 19 1100 Older 33 16 19 14 25 13 16 12 981 Standard and comfort seats only differ in terms of dimensions, namely thickness. Comparing the similarity of results of those seat types, it can be concluded that thickness is not influencing pressure distribution. These findings contradict Ebe and Griffin [13] and Kumar et al. [39], which stated that increasing foam thickness reduces interface pressures. Seat cover influence was assessed by the thermographic images. The evaluation methodology was the same as in section 6.5.2.1; a sitting area and a defined profile line were applied. Table 6.13 presents the thermal results of both seat covers.
Chapter VI 135 Table 6.13. Temperature records regarding the different seating covers. Seat type/Cover Sitting area Sitting profile line Min. (°𝐶) Max. (°𝐶) Max-Min (°𝐶) Avg. (°𝐶) Min. (°𝐶) Max. (°𝐶) Max-Min (°𝐶) Avg. (°𝐶) Comfort/Leather 28 34.1 6.1 31.8 25.2 32.9 7.7 31.8 Comfort/Fabric 19 27.1 8.1 24.1 17.1 25.1 7.9 20.9 Standard/Leather 26.2 32.9 6.7 30.5 23.9 31.3 7.4 27.6 Standard/Fabric 18.7 27.5 8.8 24.1 16.7 26.3 9.6 23.6 Analysing Table 6.13 results it was noticed that fabric cover seats have a lower minimum, maximum and average temperatures, independently of the seat type. Similar results were found within the same cover seats, meaning the non-influence of seat thickness. For the leather cover, both comfort and standard seats presented minimum temperatures of approximately 28 °𝐶, whereas the maximum temperature was around 33 °𝐶, see Figure 6.9. (a) (b) Figure 6.9. Thermographic profiles of leather cover seats, where (a) is comfort seat and (b) standard seat.