Noise and Vibration Mitigation for Rail Transportation Systems

Railways are a sustainable means of transport. Nonetheless, railways do have an influence on the environment. The most important effect is noise, especially the noise emitted from freight trains.

European Union policy supports noise reduction and has addressed the issue in interoperability directives and corresponding technical specifications. The Environmental Noise Directive (END) requires member states to submit noise maps and action plans. The EU is mostly responsible for noise creation aspects, while member states may additionally enact specific legislation for noise reception. Numerous studies have considered the economics of railway noise control, comparing the costs and benefits of different noise control possibilities. Based on these studies, the railways have adopted the following noise control strategy: 1) Reduce the noise of all new freight vehicles by introducing TSI limit values. 2) Promote the retrofitting of existing freight vehicles with composite brake blocks. 3) Build noise barriers and install insulated windows. 4) Pursue further solutions in special cases.

Noise differentiated track access charges (NDTAC) have been proposed as the main incentive for retrofitting the rolling stock by the EU and several European countries such as Switzerland.

Although the railways have made considerable progress in railway noise reduction, several problematic trends may be observed: 1) There is a tendency towards protecting capital instead of people, for example by introducing compensation for home owners based on property values. 2) Whole system optimizations are rare and infrastructure measures may counteract noise reduction efforts. 3) There are exaggerated expectations from certain new technologies. 4) Often the overall picture is not considered, such as the trade off between noise control and the modal split between road and rail. 5) Simplifications may lead to wrong conclusions, for example the noise reduction potential of a given measure often depends on local conditions and generalizations are not possible.

In 2012 The Netherlands Parliament agreed to a revision of the national Noise Legislation, including the introduction of Noise Production Ceilings for national motorways and railways [1]. Ceiling values are derived from the current calculated noise levels at a large number of reference points along the road or track. Ceiling values are open to the general public. The infrastructure manager is responsible that the ceiling values are respected at all times, i.e. including due to traffic growth. Within the ceiling values he then has an autonomous right to implement mitigation measures or even to apply changes to the track, without further legal procedure. To residents the new system offers more certainty with respect to their future noise situation. To the infrastructure manager it offers flexibility to tune the railway’s capacity to the transport demand without lengthy legal procedures. The system suffers from low credibility both with the general public and with politicians. To a certain extent this is due to the complexity of the first implementation of the system. The low credibility leads to a strong wish to check calculated levels by means of measurements. The next few years, after the first implementation problems have been solved, will decide if and how measurements can solve the credibility problem.

The question ‘What are bearable limits for environmental railway noise?’ is discussed regularly in different forums on a national scale and on a European level. A systematic evaluation of all aspects in what ‘bearable’ could consist of was always missing. The UIC research Project ‘Bearable limits and emission ceilings’ [1, 2] has brought UIC in the position to propose for the first time a well-balanced limit for noise reception. This noise reception limit is a trade-off between the disturbing impact of noise for line side residents and realistic possibilities for viable railways. Findings are based on an extensive study that was commissioned by the UIC and carried out by dBvision in the Netherlands.

The current full revision of the TSI Noise includes consideration of a second step of noise limit values for railway cars, taking into account the developments in technology.

The state-of-the-art of noise emissions of European rail vehicles was determined on the basis of a collection of TSI Noise data of railway cars. From these data the state-of-the-art of the noise emission of European railway cars was derived.

The results indicate that the most of the current limit values for stationary noise and starting noise are significantly above the state-of-the-art noise emission performance of European railway cars. With regard to pass-by noise, the state-of-the-art and the current limit values are closer together. However, in the case of freight wagons and electric multiple units, the current limit values are approximately 5 dB above the state-of-the-art.

It would therefore be reasonable to adjust the limit values to the state-of-the-art within the scope of the TSI Noise revision.

As a result of the revision process, the scope of the TSI is extended to a much wider network, a continuous limit curve for freight wagons is defined in the TSI and new requirements for stationary noise were introduced, which shall help to handle problems with stationary vehicles. For most vehicle categories the redefined limit values are still significantly above the state-of-the-art.

This paper will propose to take a next step to find a consensus method and separation of vehicle noise from track noise to be used for legislative purposes and this method may in addition reduce or eliminate the need for detailed track quality characterization in TSI certification of rolling stock.

Wheel and rail dampers are well known mitigation measures against rolling noise. A combination of laboratory measurements and computations seems the most efficient way to determine their effect. The DEUFRAKO project STARDAMP had the aim of supporting the transfer of wheel and rail dampers from the research phase to their regular application. One goal of the project was the development of a prediction tool that is dedicated to the estimation of the efficiency of wheel and rail dampers. The input data relies on laboratory measurements that are relatively easy to perform. This paper focuses on the wheel damper part. Rail dampers are addressed in a companion paper.

Rail dampers are designed to reduce the rail component of rolling noise by increasing the attenuation with distance along the rail (decay rate, DR). There is no standardized method to assess the performance of rail dampers. The method described here, developed during the Franco-German STARDAMP project, uses laboratory tests and computer simulation to avoid the need for expensive and time-consuming field trials. The premise of the method is that the DRs of a damped track can be found from summing the DRs of a short-section of damped ‘freely supported’ rail and the DRs of an undamped track. Reasonable predictions of the decay rates of a test track have been made using this method. Software has been produced that implements TWINS-like predictions of rolling noise with and without rail dampers to predict the damper effect. The effect of rail pad stiffness on the effectiveness of rail dampers has been considered for track constructions typical in the UK and a regional train travelling at 120 km/h. For track fitted with ‘soft’ 120 MN/m rail pads, the dampers are predicted to reduce the total level by 2.5 dB(A) while with the ‘stiff’ 800 MN/m pads a 0.7 dB(A) reduction is expected.

When a train passes over discontinuities on a rail, e.g. rail joints, impact noise due to the discontinuities is generated impulsively. In this paper, an attempt is made to understand the impact noise generation mechanism experimentally and theoretically. By performing static and running tests, vibratory properties of the track and wheel are investigated first. Second, on the basis of the measurements in the tests, a theoretical model to predict impact noise is developed.

Through the static test for the track, it is found that two facing rails at a rail joint no longer move together above 1000 Hz. In the running tests, for both the wheel and rail vibrations, the measured A-weighted levels at three rail joints show an increase of about 9 dB for a doubling of train speed. Also, by using the measured results, the separate contribution of noise from wheel, rail and sleeper to the total impact noise at one rail joint is estimated quantitatively. The results show that the sleeper has the greater contribution below 630 Hz, whilst the wheel is predominant above 2000 Hz. By using the prediction model, the overall trends in noise are well predicted. Also, the model gives an estimate of the contributions of wheel, rail and sleeper to the total impact noise at the rail joint, and the predictions show good agreement with the results estimated by the measurements.

This paper presents a full finite element (FE) interaction model of wheel-track to study the wheel-rail impact noise caused by a squat. The wheel, the rail and some other track components are modeled with finite elements in three dimensions, where necessary and appropriate. Realistic contact geometry, including geometric irregularity (squat) in the contact surfaces is considered. The integration is performed in the time domain with an explicit central difference scheme. For convergence, the Courant time step condition is enforced, which, together with the detailed modeling of the structure and continuum of the wheel-track system, effectively guarantees that vibration frequency of 10 kHz or higher is reproduced. By making use of the calculated velocities and pressures on the vibrating surfaces, the boundary element method (BEM) based on Helmholtz equation is adopted to transform the vibrations of the track into acoustic signals.

Since 2008, the European Commission recommends Member States to establish noise emission ceilings along their railway corridors. These ceilings should prevent railway noise from increasing due to growing freight transport. This policy instrument is just a rough idea now and it is not clear if (and when) it will be enforced through a directive. However, as the impact of noise emission ceilings on future railway operations could be large, it would be wise for stakeholders to prepare themselves. This paper discusses the pros and cons of ceilings from various points of view, based on experience with Swiss and Dutch ceiling legislation that is in force already.

The social cost of noise is normally used when performing cost-benefit analysis while planning infrastructure projects or noise mitigation measures. In this study it is used together with the official noise prediction methods to estimate the acoustic and monetary impact of transporting 1 ton of cargo through two separate transportation corridors from Gothenburg to Stockholm in Sweden, one via railway and the other via truck transport. The noise cost per transported ton is approximately twice as high for rail transport due to higher noise emissions per ton and more exposed dwellings close to the railway.

A survey of freight locomotive passby noise emissions has been undertaken across five locations on the RailCorp network, including identification of “subjective outlier” events. This paper describes the objectives of the survey, the measurement locations and methodology. Results are presented in the form of statistical parameters for each measurement location, as well as parameters categorised by operating condition (up steep grades, down steep grades and on straight and level track).

It can be seen that high L

Amax

passby levels do not correspond to events identified as subjective outlier passbys. The L

Cmax

parameter was observed to correspond reasonably well with subjectively assessed outlier events, in both the uphill and downhill directions at all locations.

Since 1 July 2012, ProRail has to work with a new legal system of noise production ceilings (NPCs) to control the noise impact of the railway traffic on its network. The new system requires that ProRail computes the noise impact of the yearly traffic on about 60 000 reference points. In addition, ProRail has to demonstrate that the noise does not exceed the NPC at any point.

ProRail has to compute the noise level at each reference point using the Dutch computation method SRMII (which is the recommended European interim computation method for railway noise). This model has to be updated yearly to reflect all changes in the network and surroundings. The data for the model comes from different sources. It is possible that small changes in the model are introduced by small inaccuracies in the data collection process. A source of small changes is the collection of geographic data with photogrammetric measurements with its inherent measurement and processing (in)accuracy.

To investigate the sensitivity of the calculations to small data collection inaccuracies, we have done an extensive parameter study. We performed parameter variations in the national-scale model and analyzed the resulting change in noise level on a statistical basis. This paper presents the results of this study and shows that the accuracy of the height information is crucial.

A number of noise reduction measures for railway infrastructure are available on the market. The manufacturers declare their efficiency, often attenuation in dB. Sometimes the declared efficiency including service life is overestimated.

The main aim of our contribution is to show some results from the noise measurement of different noise reduction measures in normal operating condition. The presented results come from outputs of the above-mentioned project granted by TAČR (Technology Agency of the Czech Republic).

To investigate railway impact noise caused by discrete rail or wheel irregularities, such as wheel flats, rail joints, switches and crossings, a time-domain wheel/rail interaction model is needed. However, time-domain models are normally time consuming to solve and this makes it difficult to carry out parametric studies and gain insight into physical behaviour. A simple mass-spring equivalent track model is developed here to gain insight into the impact vibration induced by a wheel flat irregularity. With the track system simplified to only three degrees-of-freedom, the calculation times of the wheel/rail interaction simulation are reduced by a factor of 10 compared to a corresponding finite element track model. Using this very efficient time-domain wheel/rail interaction model, the characteristics of impact noise in the audio frequency range due to a wheel flat are studied.

Due to the complex geometry of a railway wheel, the shape of its modes may have to be properly considered to understand and predict the sound radiation directivity around it. A straight-web metro wheel is studied in this paper with a radial excitation. First, the near-field sound radiation at the natural frequencies is measured in experiments performed in a semi-anechoic room. The natural frequencies here were determined by modal test. The corresponding mode shape is calculated from a finite element (FE) model. A good agreement is found between measurements and simulations in terms of natural frequencies. Further, the relationship between mode shape and sound field near the wheel is determined. Second, a rotating semi-circular frame of radius 2 m centered at the center of the wheel is employed to measure the directivity pattern. The directivities of A-weighted total level, one-third octave bands and modes are recorded at a total of 667 measuring points. A radial mode (r, 2) and an axial mode (1, 2) are taken as examples to derive empirical equations of directivity pattern. A model developed by the boundary element (BE) method is found to simulate the directivity pattern more precisely than the derived equations.

One microphone recording of train pass-by noise contains the contributions of all noise sources, i.e. traction noise, rolling noise, aerodynamic noise, and less important sources such as train body vibration noise. As rolling noise can now be predicted accurately and traction noise is negligible at high speed, one microphone recordings of high-speed train noise become useful for empirically modeling railway aerodynamic noise.

In this work, rolling noise of Swedish X2 high-speed trains was described using the total transfer function and the total roughness, which were determined using the indirect roughness method based on measurements made in 2004 by the SP Technical Research Institute of Sweden (SP). One microphone data of X2 train pass-by noise were collected by SP in 1994; these data cover a wide speed range from 70 to 270 km/h and are therefore useful both for empirically modeling aerodynamic noise and for validating the modeling of rolling noise.

This work also confirms that railway aerodynamic noise generally has dipole characteristics. However, it was discovered that at 250 Hz and below monopole sources dominate railway aerodynamic noise.

This paper describes a pass-by measurement technique that has been developed for localization and visualization of noise sources on moving rail vehicles using beamforming. Based on measurements with an array of microphones, while also measuring the position of the vehicle, the technique calculates the contribution of noise, and visualizes it as a contour plot on top of a picture of the train. Deconvolution is applied in addition to traditional beamforming in order to get an improved spatial resolution in the noise map. A set of measurements was made on two different types of regional trains on the Danish railway: the Oresund trains and the IR4 trains. The speed of all the trains was approximately 120 km/h. The results show that deconvolution is efficient for identifying wind noise on the pantograph of the Oresund trains. The IR4 trains turned out to have a strong source at the very front part of the train for frequencies around 600 Hz - 800 Hz with a radiated sound power that was approximately 5 dB above the noise radiated by the noisiest bogies. The cause of this noise is yet unknown, but a potential explanation could be an aerodynamic phenomenon at the front.

The work presented in this paper is part of the SNCF project BPS (

noise from singular points

), in which the noise emitted by steel bridges, joints and switches are investigated. Switches and crossings are numerous in urban inhabited areas. From a dynamic point of view, the switch is a complex track component where many wheel/track interaction phenomena are combined. In this complex physical context, the objective of the measurement campaign presented in this paper is to identify and characterize the sources distributed along a specific switch that contribute to the radiated noise. The measurements were performed in December 2012 on a switch (and also on a reference section) crossed by mixed traffic, equipped with accelerometers (for track and ground-borne vibration) and microphones. Compared to the reference section, the track characterisation by impact hammer excitation shows a lower stiffness under the crossing nose and a higher stiffness in the 150-500 Hz frequency range under the switch rail. The amplification due to the wooden sleepers is responsible for a rolling noise increase in the same frequency range. The crossing nose and the insulated joint are impulsive sources with a given power. For the pass-bys generating a high noise level at the reference site, their respective noise contributions are negligible.

The purpose of this paper is to test the validity of a pseudo binary sequence based method which is spectrally flat, and which is called MLS (for Maximum Length Sequences) to assess the differences in the level of vibration between different accelerometers, and evaluate key parameters that can be changed to improve the results. Before checking the validity of the MLS method, a description of the site where the test took place and a comparison of three different methods to setup accelerometers in soil are given. Measurements are also used to compute the improvement of signal to noise ratio as a function of the number of hits. Comparison between the impact hammer and MLS method is then analyzed. The main aspects of the signal processing used to analyze MLS and impact hammer signals are presented. Some information on the type of device used in the study and the primary characteristics of measured impulse responses are also presented. A comparison of mitigation at different distances between the MLS method (fixed parameters) and impact hammer method is then given. A parametric study is then performed on each MLS key parameter: MLS order, number of means, number of impulse responses analyzed, influence of background noise and linear behavior of the electrodynamic shaker. Finally, the variation of soil properties measured with the MLS and impact hammer methods over five days is given.

The sonRAIL web tool represents a public application using the Swiss method of railway noise emission calculation for vehicles and track sections. It is intended to be used by engineers, planners and administrators for the evaluation of specific situations or noise mitigation measures as well as for exchange and enlargement of knowledge about railway noise in the community. The sonRAIL method is the property of the Federal Office of Environment of Switzerland (FOEN) and is hosted at https://sonrail.empa.ch. Access to the web tool can be requested at sonrail@empa.ch.

In high-speed trains aeroacoustic source excitation is one of the most important sources. While the exterior aeroacoustic noise can be measured in wind tunnels the transmission from aeroacoustic noise sources into the passenger cabin is a difficult task. The concept for measuring aeroacoustic noise transmission will show how to determine the source, the response of the body shell and the noise radiation into the cabin of an aircraft. It has already been used on a DLR flight-test campaign [1] and can be applied to high-speed trains, as well.

The aeroacoustic sources were obtained with a two-dimensional array of pressure transducers which allows distinguishing between aeroacoustic and hydrodynamic sources underneath the turbulent boundary layer (TBL). The excitation and transmission of vibro-acoustic energy were measured with accelerometers on the fuselage, the cabin and the aircraft structure while the sound reception inside the cabin was measured with microphones and an artificial head. The data from simultaneous measurements were used to analyse the transfer of the acoustic energy on the basis of correlation methods and to validate empirical and numerical source and transfer models. The measurement concept and results from the flight test will be presented, whereupon the potential for applying this concept to high-speed train measurements will be discussed.

In the course of a large, government-financed project

Konjunkturprogramm II

, infrastructure related innovative measures for the reduction of railway noise and vibrations were installed at numerous hotspots in Germany. The demonstrated effects will be the basis for the formal recognition of these technologies. On behalf of DB Netz (infrastructure), DB Systemtechnik developed a measurement concept and supported the measurement activities.

The results for measures directly at the rail (rail dampers and rail shielding) show effects of up to 3 dB. The application of friction modifiers to reduce curve squeal and rail brake squeal (in the shunting yard) have high effects, both regarding a reduction in maximum noise level as well as occurrence frequency. Low noise barriers may reduce noise by up to 7 dB, but this depends on the height of the barrier, the distance to the noise source and the location of the receiver point. Applied primarily for maintenance of the rail surface, high speed grinding has a very positive side effect on the noise emission. Under-ballast mats and padded under sleeper pads help to tackle vibration problems.

In the framework of the economic stimulus plan “Konjunkturprogramm II” of the German government new measures for noise and vibration mitigation have been tested. The analyses of the measurements on 82 measurement sites shows as result, that in the case of rail dampers, a noise mitigation of about 2 dB(A) dependent on the type of train is possible. The effect of rail shields showed a little bit better results of about 3 dB(A) in summary. The effect of low noise barriers was modeled with regard to the German calculation method with the position of the noise source of rolling noise on the top of the rail. In this case, the shielding - effect of the low noise barrier is reduced comparable with a reduction of the height of the barrier of 30 %.

The paper deals with one of the indirect methods of rail roughness measurement – the axle-box acceleration measurement. This method has been developed by British Rail Research in the 1980s [1]. VUKV is adapting this method to the VUKV test car and to Czech railway conditions. First of all, the influence of out-of-roundness of the axle bearing on the axle-box acceleration was examined. After that, acceleration measurements were performed on different types of tracks. The acceleration was converted to deflection and compared with the direct measurement of acoustic roughness.

A measurement method for combined roughness, track decay rates and transfer functions derived from rail vibration during a train pass-by was initially developed in the late nineties [1]. This method was later implemented in software tools [2] and applied in several countries for various purposes [3, 4, 5]. Due to the broad application potential and the need for a common approach, standardisation work has been undertaken since 2011 within the framework of CEN /TC256/WG3. This resulted in a draft CEN Technical Report [6] in 2013, describing the method and providing background information on benchmarks that could serve as a basis for a new standard. In this paper, the scope and the main elements of the method are presented and performance aspects are discussed.

The paper presents a measurement setup capable of collecting wheel/rail contact noise and vibration signals from a passenger train. A data analysis method based on machine learning is developed for detecting events from the acquired data and classifying them according to relevant railway track components and noise phenomena. A classification rate higher than 84 % is achieved.

New railway vehicles in Europe have to comply with noise limits as defined in the Technical Specification for Interoperability (TSI). The pass-by test has to be carried out at a site that fulfills the TSI requirements. The need has arisen to be able to perform TSI pass-by type testing at non-compliant sites. In the ACOUTRAIN project, procedures have been proposed for the transposition of pass-by measurement results for vehicle and track. This allows for transposition of pass-by data obtained at a non-compliant site to a TSI compliant track, or to another vehicle on the same track. Transposition of track or vehicle requires source separation of the rolling noise contributions of vehicle and track to the total pass-by level. Several methods will be discussed for this purpose, based on combinations of measurement results and calculations. A transposition method is illustrated and validated by a practical case, in which both the reference and the transposed situations have actually been measured.

ACOUTRAIN, a research project within the 7th framework programme co-funded by the European Commission, aims at developing virtual testing processes for TSI Noise certification purposes. Leaded by UNIFE, it started in September 2011 for 3 years of work of 13 partners (rolling stocks manufacturers such as ATSA, Talgo or Bombardier T. and providers such as ABB; railway operators such as DB or SNCF; academic such as ISVR, ECL and TNO and scientific consultant such as VTC, D2S and Cidaut). The project will develop and gather methods (noise source characterization process for example) and tools (numerical software) for proposing a virtual testing as reliable as the real testing used in the present TSI [1], for some test configurations. Even though numerical simulations are largely used at the design and development stage for new vehicles, the use of such tools in a certification process requires a robust, transparent and validated procedure: in ACOUTRAIN, a Verification and Validation approach has allowed this procedure to be defined. It is based on numerical tool certification and validation of a model of rolling stock, called a virtual vehicle. Once the methods and tools well defined, a frame for the use of Virtual Testing shall be given, detailing for which cases VT could be used, partially replacing or supporting real testing.

The number of microphones and the configuration of the microphone array together determine the properties and ability to localize and quantify, as well as determine the source strengths of high-speed train noise sources, which is the goal of this work. The microphone array should perform well in most frequency regions of interest, inclusive of their low-frequency behavior. In order to improve the performance levels of these arrays, microphone configurations are studied to optimize the performance for the identification of the acoustical characteristics of train noise sources and to realize the best configuration to increase in-situ measurement knowledge.

This paper describes free-field tests carried out to test a procedure for the investigation of vibration mitigation measures. The test procedure consists of (1) train passing measurements, (2) measurement of the subsoil transfer-mobility by impact tests and (3) vibration measurement with an artificial, force-controlled excitation by means of a newly-developed force-controlled harmonic shaker including the procedure to simulate the sprung and un-sprung masses of trains.

It is shown that the mitigation effect based on the dynamic spring-mass effect of vibration isolation for resilient elements can be well described by this method. This component of mitigation effect can also be verified by means of the artificial excitation.

On the other hand, the results of the train pass-by measurements show additional reduction potential of under-sleeper pads (USP) due to the mechanical effect of a homogenized track stiffness.

The project Railway Induced Vibration Abatement Solutions (RIVAS) is carried out under the EU’s Seventh Framework Programme for Research and aims at reducing the environmental impact of ground-borne vibration from railway traffic.

Within the frame of the RIVAS project, an experimental procedure has been established to assess the efficiency of different mitigation measures for railway induced vibration. As an alternative to the experimental assessment, the performance of mitigation measures is often assessed relying on numerical simulations, which requires an accurate prediction of the vibration transfer. Benchmark tests are therefore carried out to investigate to what extent the vibration transfer can be predicted based on preliminary site investigation. In this paper, the results are presented for a site in El Realengo (Spain). First, the free field transfer functions are predicted based on a soil profile identified from geophysical tests. After updating of the material damping ratio, good prediction accuracy is obtained. Next, the track is included in the analysis. The track parameters strongly influence the track receptance and are therefore updated based on the measured track receptance under loaded conditions. The track – free field transfer functions are only slightly affected by the track parameters and are predicted accurately.

The aim of this paper is to provide a comprehensive overview of the state of the art on railway-induced ground vibration. The governing physical mechanisms, prediction methods, and mitigation measures of ground-borne vibration are discussed, with focus on low frequency feelable vibration and the case of railway traffic at grade. In order to clarify the importance of quasi-static and dynamic excitation, the basic problems of a moving load with constant magnitude and harmonic magnitude are discussed first. Dynamic excitation due to wheel and track unevenness and parametric excitation is shown to be the dominant source of environmental vibration in most cases. Next, an overview of prediction methods for ground-borne vibration is given. The advantages and limitations of numerical methods, based on physical or mechanical models, and empirical models, derived from measured data, are discussed. Finally, the mitigation of railway-induced ground vibration is considered, where the focus goes to mitigation measures at source (wheel and rail unevenness, rolling stock, track) and measures on the transmission path (trenches and barriers, wave impeding blocks, subgrade stiffening, and heavy masses next to the track). In conclusion, a number of open points requiring further research is given.

This paper presents the prediction model used in the European project RIVAS to evaluate the performances of the mitigation measures in reducing railway induced vibration and ground borne noise inside buildings. A robust empirical model (Vibra-2 from SBB) has been used, with the help of (i) a ground structure calculation model (MEFISSTO from CSTB) to estimate the effect of ground and building foundation changes on vibration immission, and (ii) building acoustics theory to estimate ground borne noise from floor vibration. The paper also presents the exposure descriptors and the associated exposure-response relationships chosen to evaluate the mitigation measures inside buildings in terms of attenuation of vibration and ground borne noise exposure and corresponding decrease of annoyance. Four exposure descriptors have been used: maximum values and equivalent values of both W

m

-weighted vibration and A-weighted ground borne noise; an idealized exposure-response curve, the same for all descriptors has been retained, with target values deduced from existing exposure-response curves.

Current empirical approaches to predicting groundborne railway noise and vibration in buildings exhibit large uncertainties, and are inflexible to design changes which might be suggested by project engineers. For improving this approach, finite element (FE) modelling of buildings is considered for parametric study. FE models of two multi-storey concrete frame buildings have been developed, showing some agreement with measured vibration levels.

Measured and predicted results for inter-storey level difference in both buildings differ from conventional guidance, showing that at frequencies below around 40 Hz

amplification

is observed with height up the building, with only limited attenuation at higher frequencies. Mid-span amplification of suspended floor slabs is roughly in line with the guidance, with the exception of some distinct peaks in the spectrum that exceed the suggested upper range values.

The results support the application of the FE modelling technique in the proposed context, particularly when considering relative vibration levels. Further models for parametric study are to be developed.

The adverse effects that noise and vibration from railway systems in residential environments can have on people are key obstacles for the development of new rail systems and the operation of existing lines. Recent years have seen an increase in public sensitivity towards noise and vibration from rail systems and the success of legislation to control noise levels around railway lines has resulted in an increase in the number of complaints about railway-induced vibration. Costly mitigation measures coupled with unclear or non-existent assessment methods mean that there is a need in industry and consultancy for clear guidance on the assessment of groundborne vibration from rail systems with respect to human response.

The current EU FP7 project

CargoVibes

is to publish a good practice guide on the assessment of the human response to railway induced vibration in residential environments. The aim of the guidance will be to provide end users with a set of practical tools to assess the human impact of “steady state” railway vibration primarily in terms of annoyance and sleep disturbance. Encompassing the current state of knowledge regarding the human response to vibration in residential environments alongside the practical outputs of the

CargoVibes

project, this document is intended to promote policy and standard development in this field.

The current paper will present the preliminary contents of the guidance, which have been shaped by a workshop held at the University of Salford. This paper is intended to promote debate and enable contributions from the IWRN community to ensure that the guidance is relevant to the current needs of legislators, rail and infrastructure operators, consultants, and local authorities.

Preface.

This paper was presented to the IWRN11 in Uddevalla, Sweden in 2013 as a forum for discussion on the contents of a good practice guide for the evaluation of railway vibration that was being developed as part of the FP7 project CargoVibes. Since that time, the guidance document has been completed and can be downloaded at http://usir.salford.ac.uk/30855/.

Sound Transit is constructing a light rail transit system with twin bored tunnels through the University of Washington (UW) campus in Seattle, Washington, USA. Ground vibration impacts on the UW campus were extensively studied with long range vibration propagation modeling and testing, vehicle force density measurements, and dynamic modeling of the vehicle and track isolation system. The vibration level design criteria were specified in third octave band velocities ranging from 500 nm/s to 2,500 nm/s at frequencies ranging from 3.16 to 100 Hz, placing substantial demands on vibration isolation design. Vehicle vibration force density levels were measured at ballasted track and resilient direct fixation fastener track. Vibration reductions of the order of 10 to 15 dB at 40 to 80 Hz were observed after profile grinding of the rails at direct fixation track. Vibration isolation provisions include discontinuous floating slabs with design resonance frequency of 5 Hz and high compliance direct fixation fasteners. Limits for rail undulation were specified to control low frequency vibration. Rail grinding and wheel truing combined with vibration monitoring and wheel flat detection complete the vibration control design. A major design complication is the provision of a direct current traction power cable beneath the slabs to control magnetic fields induced on campus by the trains.

The Pipe-in-Pipe (PiP) model is a fast-running model that calculates vibration levels from an underground railway within a layered halfspace. This model contains many simplifying assumptions, but its fast computation time makes it useful as an early design tool.

Recent developments in PiP have focused on quantifying the effect of various features of the underground environment that are commonly disregarded in railway vibration models. We present a summary of the work done on seven such features: soil inhomogeneity; inclined soil layers; irregular contact at the tunnel-soil interface; track with discontinuous slabs; the presence of piled foundations; second (twin) tunnels; and sources of track roughness. All of these features can introduce inaccuracy of at least 5 dB into vibration predictions.

Two examples of other uses of PiP are presented: a theoretical study of coupling effects between railways and buildings; and a manufacturer’s investigation into the effects of track resilience. Current and future research directions are also discussed.

This paper examines the combination of finite element method and multibody modelling to simulate the generation and propagation of ground vibration in the vicinity of railway networks. Based on the assumption that the source of vibrations lies at the wheel/rail contact, a multibody model of the vehicle is built using minimal coordinates, which leads to a system of pure ordinary differential equations, without constraint equations. Track and foundation dynamic equations are coupled to the vehicle’s equation of motion, using non-linear Hertzian theory. From these results, the ballast reaction on the subgrade is used in a second subproblem where free field ground response is computed using the finite element software ABAQUS.

A transfer path analysis was carried out on the Siemens Combino-Plus tram in Almada – Seixal (south of Lisbon) to provide measurement data for structure-borne noise. The measurements allowed ranking of the different noise paths from the bogie to the car-body of the existing tram and therefore helped to predict interior noise for new vehicles in the design phase. The measurement was validated by a direct measurement of the transfer stiffness of a damper.

When a high-speed train quickly passes along the sound barriers installed on the two sides of a viaduct, the fluctuating wind pressure acting on the sound barriers causes an increase in the aerodynamic loading acting on the viaduct. It not only affects the train operation safety adversely, but also speeds up the sound barrier damage rate and shortens their service life. Thus, high-speed railway noise barriers should not only satisfy the requirement that they have a noise reduction function but also that they have a deloading function due the fluctuating train air-load acting on them. Hence, the design and use of deloading noise barriers serve this purpose. This paper investigates the acoustic performance of different deloading sound barriers numerically. In the investigation, the sound insulation calculation model is established and verified by using an existing effective mass model. The propagation characteristics of the sound transmitted through the barriers are analyzed and the sound diffracted over the barrier top is considered according to the field tests. The results are useful to the efficient acoustic optimization design of high-speed railway deloading sound barriers.

At railroad yards passenger trains are parked overnight and prepared for the next day and/or freight trains are shunted. These activities are diverse and commonly processed ad-hoc. Noise permits for railroad yards are in contrast inflexible. More activities take place at the railroad yards as the intensity on the tracks increases. Noise production is increased and the permit becomes more and more restrictive.

ProRail is the railway infrastructure manager and responsible for complying with the community noise limits from the environmental laws. ProRail is also responsible for the yearly capacity allocation process. A noise propagation model is used for the calculation of the noise levels. Building such a model is time-intensive so the noise impact of only one mode of operation of the yard is investigated, leaving other modes of operations to be unknown.

M+P developed a software program “Dynamic Noise Model” (DNM) on behalf of ProRail. Many possible modes of operation can be evaluated with just a push of a button with this calculation tool. This makes it easier to find alternative modes of operations which produce less noise. Optimization of the use of the yard within noise limits is possible. A DNM also can be used to make informed decisions about taking noise measures or applying changes to the activities in the yard.

The DNM uses a database with two datasets. The first set has the attenuations between standardized processes and receivers. These attenuations are calculated in advance with a propagation model. The second dataset contains all sound power levels (SPL) by train type. This is combined with a list of all activities taking place in the railroad yard, resulting in the total SPL for all receiver positions.

ProRail is going to use the DNM for a large number of railroad yards for testing for compliance with environmental laws during the capacity allocation process.

Some kinds of short pitch rail corrugation are associated with the standing waves of rail vibration. These standing waves can be suppressed by rail vibration dampers, although they were originally developed to reduce the railway rolling noise via attenuation of the wave propagation in the rail. In this study the track models with rail dampers applied are introduced. Two kinds of wheel-track interaction, the single and multiple wheel-rail interaction, are studied to investigate the effects of the rail damper on the suppression of short pitch rail corrugation from the point of view of wheel-rail interaction force.

This paper investigates a significant reduction in environmental noise attributable to changes in the Network Rail grinding strategy employed within Great Britain (GB). Acoustic Track Quality (ATQ) is a measure of the surface roughness of the running rails. Rail roughness has a major influence on wayside rolling noise levels during the passage of a train. Levels of ATQ have been determined for approximately 1,100 km of track on GB’s East and West Coast mainlines from wayside noise measurements and data collected by an under-floor microphone system fitted to a train. The results show a substantial apparent reduction in rail surface roughness and associated wayside noise levels since a similar study was undertaken in 2004, and demonstrate how a maintenance rail grinding strategy can potentially reduce wayside noise levels across large parts of a railway network.

A special track system used to divert a train to other directions or other tracks is generally called a ‘railway turnout’. A traditional turnout system includes rails, switches, crossings (special track components), steel plates, fasteners, screw spikes, timber bearers, ballast and formation. The wheel/rail contact over the crossing transfer zone has a dip-like shape and can often cause detrimental impact loads on the railway track and its components. The large impact also emits disturbing noises (either impact or ground-borne noise) to railway neighbors.

In a brown-field railway track where an existing aged infrastructure requires renewal or maintenance, some physical constraints and construction complexities may dominate the choice of track forms or certain components. With the difficulty to obtain high-quality timbers, a methodology to replace aged timber bearers in harsh dynamic environments is to adopt a suitable material that could mimic responses and characteristics of timber in both static and dynamic situations. A critical review has suggested a field trial of an alternative material called Fibre-reinforced foamed urethane (FFU) because of its comparable characteristics to timber, high-impact attenuation, high damping property, and longer service life.

After the review of laboratory tests, a field trial of the FFU material has been implemented at an urban turnout junction in RailCorp’s suburban rail network. The effectiveness of such a method has then been evaluated using integrated numerical simulations, axle box acceleration and ride quality data obtained from the calibrated track inspection vehicle “AK Car”, and operational pass-by measurements of noise and vibration. The field trial demonstrates that using the FFU bearers in an urban turnout is effective in retaining the level of impact vibration and passenger ride comfort. It is also found that the FFU material responds to operational actions in a similar manner as timber. The material can well suppress the high-frequency impact vibration at the crossings. However, it is important to note that, in addition to lateral stability consideration, the vertical stiffness transition along the track is recommended in order to mitigate the damage of track components due to rigid body modes of track vibration.

This paper describes studies associated with planning for a new Metro Line in close vicinity to heritage buildings in the historic heart of Beijing requiring strict compliance with vibration criteria. To minimize uncertainties, extensive measurements were performed, including borehole tests to assess local soil properties. Furthermore, the influence of several key design factors on the final results is analyzed. The importance of long-term design and installation experience of highly efficient FST is illustrated

This paper describes track refurbishment works undertaken by the Network Rail Thameslink Programme to control groundborne noise and vibration transmission to sensitive buildings located above the existing King’s Cross railway tunnel in London, England. Measurements of noise, vibration and rail roughness have been undertaken to assess the immediate benefit of replacing the existing corrugated rails and continuous rail supports with new rails and the Pandrol Vanguard baseplate system. The effectiveness of these works over a period of 18 months has also been assessed.

The ground vibrations, which are generated by trains on different tracks, have been calculated by finite-element boundary-element models. The ballasted track is modelled in detail by the finite element method. The infinite soil is modelled by the boundary element method as a homogeneous or layered half-space. The track-soil system is coupled to a simple rigid mass model of the vehicle so that the vehicle-track interaction is completely included. Transfer functions are calculated in frequency domain without and with vehicle-track interaction, the compliance of the track and the mobilities of the soil at different distances from the track. Finally, the ratios between the ground vibration amplitudes with and without mitigation measures are calculated to quantify the effectiveness of the mitigation measures.

Tracks with under-sleeper pads have been investigated in a wide parameter study for the RIVAS project. The main parameters that influence the reduction of ground vibration are the stiffness of the under-sleeper pad, the mass and the width of the sleeper. The softest sleeper pad yields the best reduction of the ground vibration. The influence of the sleeper mass is not so strong, as the characteristic frequency is ruled by the mass of the sleeper and the mass of the wheelset as well.

In the course of a large, government-financed project, infrastructure-related innovative measures for the reduction of railway noise and vibrations were installed and tested at numerous hotspots. The demonstrated effects will be the basis for the formal recognition of these technologies. On behalf of DB Netz (infrastructure) DB Systemtechnik set up a measurement concept and supported the acoustic investigations. This paper focusses on those measures to reduce the noise of steel bridges. For the first time, the investigations carried out with resilient rail fastening systems on steel bridges with directly fastened track show an effect in the lower frequency range. The measures used in addition to the resilient rail fasteners, such as the elastically mounted covers, the filling of the bridge railing as a sound barrier and the use of rail dampers can further reduce the noise next to the bridge. For steel bridges with ballasted tracks, the effect of homologated under-sleeper pads in the lower frequency range is not comparable to that of under-ballast mats. If combined with bridge dampers however, the effect is significant. The use of bridge dampers alone requires careful tuning to the bridge resonance frequencies. Since they only affect the rolling noise, rail dampers have only a small influence on the overall bridge noise.

Railway vehicles negotiating tight curves may emit an intense high-pitch noise. The underlying mechanisms of this squeal noise are still a subject of research. Simulation models are complex since they have to consider the non-linear, transient and high-frequency interaction between wheel and rail. Often simplified models are used for wheel and rail to reduce computational effort, which involves the risk of over-simplifications. This paper focuses on the importance to include a rotating wheel instead of a stationary wheel in the simulation models. Two formulations for a rotating wheel are implemented in a previously published wheel/rail interaction model: a realistic model based on an Eulerian modal coordinate approach and a simplified model based on a rotating load and moving Green’s functions. The simulation results for different friction coefficients and values of lateral creepage are compared with results obtained for the stationary wheel. Both approaches for the rotating wheel give almost identical results for the rolling speed considered. Furthermore, it can be concluded that a model of a stationary flexible wheel is sufficient for both capturing the tendency to squeal and predicting the resulting wheel/rail contact forces.

In this paper we show that falling friction poses a principal difficulty for the FASTSIM algorithm. Using velocity-dependent

μ

leads to discontinuities in the tractions, displacements and slip velocity. These discontinuities are due to the local and instantaneous relations that are used. They are circumvented in different, unsatisfactory ways in the extensions of FASTSIM that are presented in the literature. We describe a new FASTSIM algorithm that deals with falling friction properly, using the concept of friction memory.

Curve squeal is a strong tonal noise that may arise when a railway vehicle negotiates a curve. The wheel/rail contact model is the central part of prediction models, describing the frictional instability occurring in the contact during squeal. A previously developed time-domain squeal model considers the wheel and rail dynamics, and the wheel/rail contact is solved using Kalker’s nonlinear transient CONTACT algorithm with Coulomb friction. In this paper, contact models with different degree of simplification are compared to CONTACT within the previously developed squeal model in order to determine a suitable contact algorithm for an engineering curve squeal model. Kalker’s steady-state FASTSIM is evaluated, and, without further modification, shows unsatisfying results. An alternative transient single-point contact algorithm named SPOINT is formulated with the friction model derived from CONTACT. Compared to the original model results, the SPOINT implementation results are promising and similar to results from CONTACT.

Experience has shown that curve noise issues can become more severe when curved track is upgraded from timber to concrete sleepers. This suggests that changing the dynamics of the track structure, for instance using a softer rail fastening system and/or sleeper, could provide ways to reduce curve squeal. The problem, however, is that the changes in dynamic characteristics between timber and concrete sleepers that explain the difference in curve noise behavior are not currently understood. Observations and anecdotes have hinted at the effect of different rail pads, fasteners, rail dampers and gauge relief on curve noise, but have failed to definitively identify either a successful mitigation strategy or a quantitative relationship with curve noise generation. This paper presents the results of field trials on curved track, conducted before and after an upgrade from timber sleepers to concrete sleepers, aimed at improving this understanding.

This paper presents the results of a comprehensive trial of friction management techniques carried out over a 7 month period at a curve in Sydney. The trial tested various top-of-rail friction modification and gauge face lubrication products, both in isolation and in combination, applied at different volumes and locations around the curve. Noise and vibration transducers were situated at three locations around the curve, one upstream of the treatment and the other two downstream. At the upstream location, rolling stock data was also available from a permanent wayside geometry monitoring system.

The findings provide a number of valuable new insights into the performance of gauge face lubrication and top-of-rail friction modification for curve noise mitigation. The most significant finding was that gauge face lubrication of the outer rail provided a substantial reduction of severe tonal wheel squeal at the test curve. This is at odds with conventional theory that wheel squeal is controlled by friction characteristics at the top of rail / wheel-tread interface (rather than the gauge corner / wheel flange interface).

High speed trains have been running for almost 50 years in many countries. High speed railway systems differ from each other in terms of rolling stock, track, commercial speed, operating conditions, maintenance. Then, the environmental noise varies according to all these parameters. To improve the insertion of high speed railway systems in the environment, noise mitigation measures have been developed. Each one provides a noise reduction for the considered system. Combinations of the noise reduction solutions are also very efficient and cost effective.

A state of the art of pass-by noise measurement results from several high-speed railway systems all around the world is presented and discussed. The main noise sources of high speed railway system, the rolling noise, the aerodynamic noise and the equipment noise, are described. The measurement and calculation methods to characterise the sources are presented. Each mitigation measure and its efficiency are described. As a conclusion, some indications for future research topics in high speed railway noise are proposed.

The aerodynamic noise from a high-speed train becomes predominant at sufficiently high train speeds. Pantographs and bogies have been identified as significant aerodynamic noise sources in high-speed trains. The aerodynamic noise generated by pantographs is produced by a vortex shedding process while the aerodynamic noise from bogies is mainly generated by complex circulation of turbulent flow around its components. A semi-empirical component-based method for fast prediction of aerodynamic noise from a high-speed train pantograph is developed on the basis of similar models applied to landing gear noise prediction. Model predictions give good agreement with existing wind tunnel tests showing the approach to be promising. For the bogie, wind tunnel tests have been designed using a 1/10

th

scale bogie mock up. The results obtained are in agreement with existing data from train pass-by tests, showing that these experimental data can be used for model calibration and validation.

As one of the main aerodynamic noise sources of high-speed trains, the bogie is a complex structure containing many components and the flow around it is extremely dynamic with high-level turbulence. Flow around a simplified bogie at scale 1:10 is studied numerically using computational fluid dynamics for comparison with experimental measurements. The upstream inlet flow is represented as a steady uniform flow of low turbulence level in the simulations. Following a rigorous grid refinement study, multi-block fully structured meshes are generated for all cases to improve numerical efficiency and accuracy. The aerodynamic and aeroacoustic behaviour of the flow past an isolated wheelset, tandem wheelsets and a simplified bogie are investigated.

High Speed 2 (HS2) is a planned high-speed railway in the UK connecting London to Manchester and Leeds, via Birmingham. A key concern for many is the environmental impact of the scheme, particularly noise. Trains on HS2 are specified to operate at a maximum speed of 360 km/h. There are currently no legislative passby sound level limits available in Europe for trains running at these speeds. Published measurements for higher speeds are typically for older generation trains, or for emission levels from sub-systems.

This paper discusses the methodology used to derive the sound emission source terms used for the HS2 noise impact assessment. At speeds in excess of 300 km/h, aerodynamic sound contributes to the overall passby sound levels, and cannot be neglected. Therefore source terms have been derived for five source types: power/traction, rolling sound, body aerodynamic sound, pantograph recess, and raised pantograph. The source term magnitudes for traction/power, rolling, and body aerodynamic sound were derived from limits defined in the technical specification for interoperability (TSI) for high-speed rail. Corrections were applied to allow for in service track and growth of wheel and rail roughness. The source terms for the pantograph recess and raised pantograph were derived from a range of measured and simulated data in published literature.

This paper also considers reductions for each of the five sources that have been demonstrated through published literature and which could therefore provide the opportunity for lower train sound emissions by the time that HS2 rolling stock is procured.

This paper deals with the first and comprehensive application of a new prediction and assessment method on the new Katzenbergtunnel. The tunnel follows a new safety concept with two single track tubes and is strongly affected by the micro-pressure wave (MPW) phenomenon. Thus, appropriate MPW countermeasures were designed in the early construction phase. At this stage, the effects could only be predicted by model scale tests and simulations. During the final homologation tests, the values were measured in full scale and compared to the predicted ones. There was a good correlation so the tunnel could be opened for operation in December 2012 without any further modifications.

In this paper the implemented countermeasures are presented. Also, some comparisons between predicted and measured values are shown. A detailed description of the assessment method which was adopted in a guideline of DB Netz AG is given in a separate paper.

To reduce aerodynamic noise generated by a pantograph panhead, a flow control method is applied to the panhead. First, a plasma actuator is applied to the panhead surface to verify that flow control can control flow separation and reduce turbulence intensity behind the panhead. Second, based on the flow control mechanism derived from the experiments applying the plasma actuator, a more practical flow control method is proposed using air suction as an alternative. Experimental results show that air suction near the separation point can reduce narrow band aerodynamic noise from the panhead.

The influence of vehicle design, wheel out-of-roundness (OOR), track irregularity and properties of the layered ground on the generation of train induced ground-borne vibration is investigated through simulations with the TRAFFIC software. By mitigating severe forms of OOR, such as high levels of eccentricity (order 1) and higher orders of wheel polygonalisation, there is a potential of reducing the vibration level in the ground with up to 15 dB in particular 1/3-octave frequency bands and around 5 dB on the overall level. A similar overall reduction is achievable by reducing the unsprung mass whereas the primary suspension stiffness is of minor influence. The performance of the vehicle in terms of induced vibration levels is strongly dependent on the status of the track as defined by track irregularity level and dynamic properties of the ground. The current study was carried out within the 7

th

framework research project RIVAS (Railway Induced Vibration Abatement Solutions).

The goal of the EU project RIVAS (Railway Induced Vibration Abatement Solutions) is to take ground-borne vibration mitigation on open tracks an essential step forward. In this context, measures to be taken on the rolling stock are investigated in Work Package 5 (WP5). The influence of rolling stock was tested in comprehensive ground vibration measurements of regular trains carried out in the SBB railway network.

The vibration analyses were correlated with vehicle-specific data and out-of-roundness measurements. The analyses show that of all railway vehicle parameters, wheel condition and unsprung mass have the major influence on vibration emissions. These analyses result in vibration mitigation measures on the rolling stock with the goal of sustainably improving wheel condition and reducing unsprung mass. These measures have the potential of considerably reducing vibrations and will be further developed.

The research leading to these results has received funding from the European Union Seventh Framework Programme under grant agreement no 265754.

This paper studies the efficiency of stiff wave barriers for the mitigation of railway induced vibrations. Coupled finite element–boundary element models developed at KU Leuven and ISVR are employed; these models have been cross–validated within the EU FP7 project RIVAS (Railway Induced Vibration Abatement Solutions). A first mitigation measure consists of a block of stiffened soil embedded in a halfspace that acts as a wave impeding barrier. The existence of a critical frequency from which this mitigation measure starts to be effective, as well as a critical angle delimiting the area where the vibration levels are reduced, is demonstrated. Next, a sheet piling wall is considered, accounting for the orthotropic behaviour of this wall. Calculations show that the reduction of vibration levels is entirely due to the relatively high axial and bending stiffness in the vertical direction (along the profiles

)

, while the bending stiffness for bending waves traveling in the longitudinal direction (perpendicular to the profiles) is too low to affect the transmission of vibrations. Field tests are being carried out in Spain and Sweden to confirm the conclusions of these numerical computations.

A state-of-the-art study on ground-borne vibration induced by railway turnouts (switches & crossings) is presented. Vibration generating mechanisms and possibilities for cost-effective mitigation measures are discussed. A brief literature survey and examples of results from vibration measurements performed by the Schweizerische Bundesbahnen (SBB) are presented. The present work was performed within the EU FP7 project RIVAS (Railway Induced Vibration Abatement Solutions).

Mitigation measures for turnouts described in the literature mainly aim at reducing the dynamic wheel-rail contact force (impact load) in the crossing panel. This is supported by field observations of amplified ground-borne vibration at turnouts, indicating that there is a need to improve the design of these track components. A smooth wheel transition between wing rail and crossing nose is important to achieve low energy impacts on the crossing. Measures to mitigate the vibrations include the use of soft or stiff Under Sleeper Pads (USP) for better and more stable turnout geometry over time. The use of USP may also reduce the magnitude of vibrations transferred from turnout to soil. Other measures include improving the design of the crossing panel (material and geometry of crossing nose and wing rails) and the use of more resilient rail pads.

Results from a Swiss test campaign at two different sites (Rubigen and Le Landeron), using stiff USP in four turnouts, are presented and compared to results from four nominal turnouts without USP. However, the vibrations measured at Le Landeron, and even more so for Rubigen, seem to be significantly influenced by the turnout and soil conditions. It is important to accurately determine the condition of each turnout to enable a meaningful comparison between the turnouts and to draw conclusions on the influence of turnout mitigation measures on vibration emission.

The best approach to reduce the amplification of vibration is a system approach where different parameters (such as rail profiles, rail material, resilient layers) are designed to interact in a harmonious and robust way. It is thus important to identify the combination of parameters that defines the design of an optimum turnout. Modelling is required to optimize the wheel transition over the crossing panel. Further tests are needed to define the relevant geometric parameters influencing the wheel trajectory and the amplified vibrations. If an improvement of the dynamic wheel-rail interaction (wheel transition) can be accomplished that reduces impact loads on the crossing panel, both vibration magnitudes and the need for crossing panel maintenance can be reduced.

To reduce railway induced low frequency vibration, two mitigation measures - open trenches and buried soft wall barriers have been studied in this paper by using coupled finite element-boundary element models. These models were developed at KU Leuven and ISVR, and have been cross-validated within the EU FP7 project RIVAS (Railway Induced Vibration Abatement Solutions). Variations in the width, depth, location of trench and properties of soft barrier material are considered under various soil conditions. Results show that in all ground conditions, the notional rectangular open trench performs better than the other constructions. The width of an open trench has little influence on its performance, whereas increasing the width of a filled trench reduces the stiffness of the barrier, improving the performance of the trench. Likewise, fill materials with lower Young’s modulus give higher insertion losses.

For high-speed trains, the interior noise of the pantograph area is usually higher than other areas. A lot of achievements have been made to reduce the aerodynamic noise of the pantograph and its facilities. Yet, the exterior noise and vibration of the pantograph area is still much higher while the train operates at 300 km/h or more. So research on the vehicle structure is necessary.

Aiming at reducing the interior noise, operating tests were carried out to get the noise and vibration of the pantograph area and the force on the pantograph mounting. Spectrum characteristics and statistical laws all prove that the interior noise is closely related to the ceiling vibration.

In order to reduce the vibration, an absorber structure is proposed for the pantograph mounting. A specially designed real-car-simulating test bench is used to evaluate the application performance of the absorber. The test results show that the noise and vibration are significantly reduced after adopting the absorber.

The research of this article proposes a new thinking and method for interior noise reduction of the pantograph area from the aspect of vibration attenuation.

With increasing train speeds, the emission of micro-pressure waves (MPW) from tunnel portals can contribute significantly to the annoyance of residents close to a high-speed railway line. Up to now there are no established assessment guidelines or standards in Europe, with the specific purpose of assessing this effect.

In order to develop efficient measures for the reduction of noise emission due to MPW it is necessary to join the planning of a new tunnel at an early stage. This requires reliable prediction methods as well as upper limits for noise levels. As micro-pressure waves are audible as “sonic boom”, the generally used A-weighting is not considered to be a suitable indicator for the evaluation of this effect.

During the years 2007 to 2009, Deutsche Bahn performed a study how to predict the micro-pressure effects for new tunnel constructions as well as how to assess and to limit the micro-pressure effects. The main issues were the protection against hearing damage in the close vicinity of the tunnel entrances (e.g. railway workers), the conspicuity in the portal region, the annoyance in residential areas in the surrounding of the railway line as well as micro-pressure effects on the structures of nearby buildings and secondary acoustical effects by structural vibration excitation.

As a result of this study, a method for prediction and assessment of MPW-immissions was proposed to the German national railway authority. The final procedure, which has been approved by the German national railway authority (EBA) and the German Federal Environment Agency (UBA), includes following requirements: (I) The C-weighted sound peak level

L

pC,peak

should not exceed 115 dB(C) at a distance of 25 m from the tunnel portal, additionally (II) threshold levels for the C-weighted sound exposure level depending on the utilization of the area should not be exceeded at buildings in the neighborhood of tunnel portals. (III) Above all the A-weighted sound exposure level of MPW in the neighborhood should be evaluated according to the German traffic noise protection directive.

The prediction and assessment procedure is adopted into the regulation “Ril 853” of DB Netz AG and came into effect as of Feb, 1

st

2013. It is the basis for all current German tunnel projects either in planning or under construction.

This paper gives a brief overview of the generation mechanism and the acoustical properties of micro-pressure waves. The prediction method that was developed by DB Systemtechnik will be summarized. It also describes the developed assessment criteria and their justification in regard to the relevant noise legislation. An overview about existing measures for reducing MPW emissions is shown.

Based on an equation given by Cremer and Heckl, describing the noise radiation of plates of limited size, a method of predicting the efficiency of noise mitigation measures of ballast-less steel bridges is presented.

Whereas in most studies, computationally demanding numerical modeling is used, this paper describes a simplified approach for investigating three ways of reducing the vibration of bridge structures and hence reducing the noise emission of the bridge.

These mitigation measures, intended in conjunction with maintenance work (not reconstruction), are designed to reduce the impact of excitation forces arising from the wheel rail contact and moving masses, by increasing the internal losses of the bridge structure and reducing the vibration level by increasing the effective mass of the bridge structure itself.

The predicted results are compared to measured data at bridges where these noise mitigation measures were installed. The objective of the method is that it will simplify the process of selecting the best mitigation measure for a specific steel bridge.

Existing curve squeal theory is often contradicted by field observations such as the generation of squeal from the outer wheel (including wheel flange contact), squeal occurring at various wheel natural frequencies, coupled rail vibrations when squealing wheels pass, and the obvious influence of trackform on squeal occurrence and severity. This paper discusses the deficiencies of existing theory and explores an alternative mechanism based on the concept of mode coupling instability which shows a better match with field observations from some sites.

Improved methods for measuring curve squeal and braking noise have been proposed within the ACOUTRAIN project, intended as input to future standards. In this paper an outline of these methods is given and the underlying considerations are explained. The EU TSI regulation limiting noise emission of new rail vehicles does not yet specify limits for curve squeal or braking noise, which are both relevant noise sources but not easy to characterise reliably due to their variability. No ISO or EN standard exists for curve squeal. Braking noise is included in the current EN ISO 3095 noise type testing standard, but the procedure requires improvement in relation to measured quantities. For curve squeal, two methods are proposed, an occurrence test using on board measurements along a route with various curves, and a trackside test for curves or switches. The existing UIC trackside test is taken as basis using the percentage exceedance indicator L

10

.

For braking noise, improvements to the existing trackside measurement standard are proposed, taking the variety of potential noise sources and their variable position and operating condition into account. The measurement quantity, positioning and braking operation are the main topics. Without proposing limit values, some considerations are put forward that may be taken into account when applying the methods for compliance testing for curve squeal and braking noise.

A real case of a city tramcar generating very high curve squeal noise levels has motivated research in this area considering the simultaneous presence of two contact points. In this paper, available measurements are summarised in a qualitative manner in order to highlight the most important frequencies involved and a theoretical model in the frequency domain is developed with the aim of predicting curve squeal tones. Good matching is found between numerically predicted and measured unstable frequencies and a peculiar shift toward higher frequencies is found both in measurements and predictions.

Measurements have been made for ground vibration at four different sites where the existing track form was replaced by one with a much lower vertical stiffness. Other parameters, such as the rail roughness level, were controlled such that these measurements allow a direct comparison to be made between the ground vibration levels obtained for the two different track forms. It is found that the more resilient track form does bring about a significant reduction in ground vibration in the part of the frequency range expected from computer models. The performance of the more resilient track form differs from expectations based on the results of the models in the lower part of the frequency range of interest at some sites. The potential causes of this difference are discussed with reference to the measurements.

Floating slab tracks (FST) are mass-spring systems that are used to reduce groundborne vibration generated by trains. The effectiveness of FSTs can be estimated using the single degree-of-freedom (SDOF) model. More detailed study of the modes and shapes of the FST are sometimes performed using Finite Element Method (FEM) models. This paper also explores the suitability of a natural rubber (NR) based elastomeric spring for FST designs that require relatively high deflection and good performance at low service temperatures (-40 °C). The paper also evaluates the utility of simple SDOF models as a prediction tool, and the potential factors that would result in higher damping at the resonance frequencies (f

0

).

The collaborative work presented in this paper is part of the European project RIVAS, dedicated to the mitigation of ground-borne vibrations from rail traffic. In this paper, focus is made on innovative low stiffness fastening systems designed for ballasted track. These kinds of systems reduce ground-borne vibration by the filtering effect provided by their low stiffness in comparison to the global track stiffness. Numerical simulations performed as a first step confirmed this expected behaviour: ground vibration is strongly attenuated above a cut-off frequency at which resonance occurs. Following these results, two fastening systems designed by Pandrol (member of the RIVAS consortium) were chosen for testing, first in laboratory and then in a commercial track. The laboratory tests mainly consisted in the evaluation of the systems stiffness in realistic conditions (single sleeper in a ballast box), using an appropriate excitation. Combined with acceleration measurements, those tests confirmed the main results expected from simulations but also raised specific issues for the testing of low stiffness fastening systems. In order to be tested in a commercial track in France, some preliminary work was required on these systems; it is also presented in this paper.

This paper focuses on the insertion of resilient layers in the track as mitigation measures for ground-borne noise and vibration. Soft rail pads, under-sleeper pads or ballast mats are considered as mitigation measures.

Common train-track interaction models using unevenness excitation predict reduction of ground vibrations around and above the natural frequency of the rolling stock unsprung mass on the track stiffness, in agreement with experimental observations. In the lower frequency range, i.e. below this natural frequency, computations show an increase of vibration that is not always confirmed by experimental results recorded on ballasted tracks.

The excitation mechanisms play a role in the measured efficiency. In this paper the insertion losses computed with an unevenness and a parametric excitation model are compared. The results show that, depending on the train speed, the low frequency range insertion loss, related to resilient layer insertion in the track, is not always negative for parametric excitation.

Measurements are presented of irregularities on rails in the wavelength range 6.3-5000 mm. Two types of equipment have been used, a hand-held measuring trolley (the so-called CAT) and vehicle-mounted RCA equipment that is typically used for routine quality assurance of rail reprofiling work. There is close correlation of one-third octave spectra from the two types of equipment for most of the wavelength range of interest. The RCA is a more repeatable instrument than the CAT for wavelengths of greater than a metre. Use of such equipment offers a relatively simple, readily available means of controlling residual irregularities from reprofiling and thereby reducing noise. Control of irregularities over a wide wavelength range is within the capabilities of modern reprofiling equipment, but only if an appropriate specification is rigorously monitored.

Wheel roughness measurements available from several different campaigns are presented in terms of average levels and dispersion. The dependence on factors such as brake type and whether the wheel is powered or trailing is also addressed. A method to decide how many wheels from a train are to be measured is then presented. Finally, the main outcomes are described from a round robin test aimed at assessing the effect on wheel roughness measurements of adopting different equipment, used independently by different teams.

The development of rail corrugation (so called rutting corrugation) on a 120 m radius curve on the metro of Stockholm Public Transport was studied by field measurements, laboratory measurements and numerical simulations. The corrugation develops exclusively on the (inner) low rail with wavelengths of about 5 cm and 8 cm. A time-domain model for prediction of long-term roughness growth on small radius curves is developed and validated against measured data. The wavelength-fixing mechanisms of the corrugation are bending eigenmodes of the leading wheelsets. The application of a friction modifier effectively mitigates the problem.

This paper describes track vibration levels and rail roughness growth measurements carried out on a number of tracks with different values of track stiffness and rail damping at a number of metro systems in China. At each test site, measurements of rail surface roughness were repeated several times over a period of time. The results provide useful data on the progress of rail roughness growth, and of change in track vibration levels in the medium term. Rail roughness increased noticeably on the stiffer track, whereas it decreased on the track with tuned dampers attached to the rail. Comparing with the standard baseplate track, the vibration levels measured on the rails in both vertical and lateral directions are all lower on the track with lower stiffness support and tuned vibration dampers.

This paper presents a prediction of the acoustic wall pressure levels on a rolling stock vehicle from wheel-track interaction. This study has taken into account two different approaches: beam-tracing and boundary element calculations. The two types of calculations are benchmarked against results measured on an existing train.

In order to reduce noise along railway lines, sound barriers have been widely used. However, the effects of these barriers on railway noise have not been understood thoroughly. In this paper, the relative acoustical performances of a number of sound barriers in typical Shinkansen railway situations are first investigated through laboratory measurements using scale models. In the scale model studies, the effects of various types and shapes of the sound barriers on Shinkansen noise are investigated. The results show that the most effective barrier is Y-profile type. This is because the Y-profile barrier has noticeable effects on both mitigating the noise increase due to the multiple reflections of sound between the barrier and the vehicle surface and widening the shadow zone behind the barrier because of double diffractive effects. Secondly, the effect of the Y-profile barriers is examined by field tests with running Shinkansen vehicles. The results show similar trends to those in the laboratory measurements.

Field experiments were carried out to investigate abnormal interior noise generation in a coach used as a dining car of a high-speed train, including field measurements to investigate the interior noise conditions of another passenger coach under the same operating conditions. Sound quality indexes at each interior noise measuring point were analyzed to determine the characteristics of the abnormal interior noise. The differences of vibration and noise in the two coaches were then compared to show the effect of the interior structure of the coach on the abnormal interior noise. In addition, the differences of vibration and noise in the dining coach before and after the wheel re-profiling were compared to examine the effect of the wheel roughness on the abnormal interior noise. The present work has identified mechanisms of abnormal interior noise generation in the dining car.

Interior noise in rolling stock is an important index in the evaluation of passenger ride comfort. In this paper, the hybrid FE-SEA method was used to predict the interior noise of high-speed trains. In this process, the hybrid model of the car body was established based on FE models. Then according to the characteristic of the car body, the beams and stiffeners were described by a FE subsystem and thin, light plates and panels were divided into a SEA subsystem. As for the special structure of hollow aluminium alloy extrusion which makes up the car body profile, the equivalent model of a General Laminate was proposed to simplify this structure in the paper. Through theoretical calculation and test, subsystem parameters and external excitations were obtained, and then imported into the prediction model. Finally noise at the central section of the high-speed train was predicted by using the hybrid FE-SEA model. Compared with the test results, in the frequency range of 100 ~ 1000 Hz, the predicted and test noise spectra followed basically the same trend, so the proposed Hybrid FE-SEA method can be efficiently applied to the railway field.

A study of annoying sounds and vibrations generated by train interiors is reported. Different types of annoying sounds are discussed with respect to the effects they have on the passengers and a notation for distinguishing annoying sounds on trains is defined.

A survey of disturbing sounds and interior vibrations on Swedish intercity trains is also reported. A large majority of the annoying sounds found in the survey were of tapping or rattling character, originating from components like ceiling panels, light covers, cabinet doors, interior sliding doors and foldable tables. For some vehicles the number of annoying sounds and vibration issues related to interiors is substantial also for cars with only a few years in operation. This observation underlines the need for systematic abatement procedures and proactive measures from the manufacturers and operators to ensure comfortable train journeys.

A review of processes applied in the automotive sector to control annoying noises is also summarized and opportunities for knowledge transfer to the railway industry are discussed. Finally, best practice design solutions to reduce interior vibrations and annoying sounds from train interiors are summarized in view of the results of the present survey and previous vehicle design experience.