Making rail transport more sustainable, comfortable and quieter

If rail is to become a more attractive transport option for passengers, there must be greater focus on reducing the energy consumption of rolling stock and rail infrastructure. This will lead to significant cost reductions for both railway operators and users, while bringing environmental benefits. Another major goal for the sector is to reduce the noise and vibration generated by trains and railway lines, such as by modelling and predicting these unwanted disturbances.

Energy and Sustainability

Sustainability is a key parameter when measuring rail transport’s ability to compete with both road and air traffic. In addition, enhancing the energy performance of the railway system makes this form of transport more cost-efficient and more environmentally friendly.

Under the Shift2Rail programme’s Energy work area, industrial partners set out to reduce energy consumption. This will make train travel more cost-efficient and contribute to ensuring the sector boosts its reputation as an environmentally friendly means of transport.

One key goal was to develop a standardised methodology for estimating energy consumption, by calling on simulation and measurement. This will enable a standardised specification of energy-efficient railways, achieve and assess the overall energy reduction of the Shift2Rail programme, and demonstrate cost-effectiveness and energy-saving features. The following activities showcase the main achievements of this technical area:

Activity: Energy-efficient solution and future railway system with respect to energy

Discussion

Early research defined the parameters for the energy calculation of reference trains for four traffic segments: high-speed, regional, urban and freight. This energy baseline was then updated and amended to account for five topics:

• HVAC systems, including innovative systems with heat pumps and natural refrigerants.
• Auxiliary power of the train, excluding Heating, Ventilation, and Air Conditioning (HVAC).
• Battery-Powered Electrical Multiple Units (BEMU), with reference parameters for trains/infrastructure.
• Generic thermal model of the carriage body, with thermal reference parameters.
• Freight trains with updated reference parameters for aerodynamic losses/energy consumption.

To ensure the effect of technological innovations on energy consumption are visible, the models and parameters must take these topics into account when calculating energy.

Key findings

HVAC systems: their energy consumption can be better calculated by taking into account new technologies and assessing the operating conditions (e.g. train in parking position or operating with passengers) more accurately. For calculation of the average yearly energy consumption of the HVAC system, the operational states and their daily duration were taken from the European Standard.

Auxiliary power (for lighting, battery charging, cooling of the traction equipment, air compressors, and control and onboard signalling systems): the reference parameters were updated, enabling use of a more realistic average power consumption.

BEMUs (Battery-Powered Electrical Multiple Units): these are increasingly used as substitutes for diesel engines on non-electrified regional lines. A four-carriage train of this kind was successfully operated with power supply either from the overhead line or from its battery, plus battery charging from stationary converters or while operating on electrified lines.

Generic thermal carriage body: definition of a typical two-carriage BEMU, with energy assumptions for the total area of walls, roof, floor and windows; plus a heat transfer coefficient, etc.

Mainline freight trains: when calculating their energy consumption, the aerodynamic part of the running resistance is important. Aerodynamic effects depend on the shape of wagons and on the composition of the whole freight train.

Building on earlier research, energy reference parameters were updated to differentiate between trains with mixed wagons, and trains for combined traffic with partly loaded and partly empty wagons. A simulation tool was used to calculate the aerodynamic coefficients more accurately for two types of reference freight trains.

 Reducing railways’ carbon dioxide emissions

Railways are one of the most energy-efficient mode of transport, and they have made the biggest progress in energy efficiency since the year 2000: railways have increased efficiency by 37% (buses +3%, cars +8%, trucks and light vehicles +13%). However, this energy efficiency can be improved and researchers have identified decarbonisation of rail transport as the most effective way for the railway sector to react to the challenges of global warming, shortage of fossil energy resources and rising energy prices. After evaluation of their feasibility and effectiveness, eight measures were proposed to achieve a 100% decarbonisation of rail transport:

1. Electrical energy supply from non-fossil resources: prioritising renewable energy sources for DC-supplied railway networks and networks.
2. Electrification of more railway lines: 60% of the European rail network is already electrified, as this is economical for lines with regular traffic. Electrification gaps on longer and non-electrified main routes should be closed, especially on routes used by freight trains.
3. Substitution of diesel trains by electrical, battery or hydrogen trains: battery-powered trains (BEMUs) are increasingly replacing diesel trains on non-electrified regional lines of up to 70 km. For routes with gaps in the electrification, regional BEMUs could be upgraded for long-distance traffic (higher passenger comfort, higher speed). Hydrogen-powered trains are an alternative to battery-powered trains. Freight locomotives with batteries or hydrogen are available and used for shunting and the last mile, e.g. non-electrified terminals or ports.
4. Retrofit of existing diesel trains to battery or hydrogen power: to speed up the decarbonisation of rail traffic, existing diesel trains should be re-fitted wherever possible with battery or fuel cells before they reach the end of their life (typically 30 to 40 years).
5. Conversion of existing diesel trains to synthetic fuel or hydrogen: demonstration projects have confirmed that diesel engines can be powered with fuels (e.g. synthetic fuels or hydrogenated vegetable oil) produced with renewable energy, with few changes to the motors. Diesel engines can operate with hydrogen, but this requires more retrofitting effort. Retrofitted diesel engines cost more to operate, because of the higher costs of these alternative fuels.
6. Infrastructure for battery and hydrogen trains: extensive landside infrastructure is needed for these trains, e.g. fast charging stations, external energy supply stations, and hydrogen fuel stations. Existing diesel fuelling stations must be adapted to use synthetic fuel and hydrogenated vegetable oil.
7. Renewable energy for stations and platforms: the electrical energy comes from the public grid, so any supply of 100% renewable energy must be negotiated with energy providers. Solar panels (roofs of buildings and platforms) and electrically powered heat pumps (to replace gas and oil used for heating and air conditioning of buildings) can play a key role.
8. Reduction of train energy consumption: trains’ energy consumption can be reduced by 10 to 30%, depending on the traffic segment (high-speed, regional, urban or freight). Innovations to achieve that goal most effectively are: increased efficiency of the traction equipment; reducing the train weight; use of clever traffic control strategies and driver assistance systems to avoid unnecessary stops and speed changes; improved aerodynamics (high-speed and freight), and reducing the auxiliary energy consumption (HVAC).

Innovations affecting energy consumption For each innovation, experts looked at the potential percentage of energy saving and the Technology Readiness Level for application. The total energy saving potential of all the programme’s innovations is about 10% for passenger trains and 30% for freight trains. The most effective and promising measures and innovations are: Energy optimised driving with driver advisory systems or Automatic Train Operation (ATO), and with a capable traffic management system. Traction and auxiliary converters with highly efficient silicon-carbide (SiC) power semiconductors. Extended freight wagon market with improved aerodynamics. Optimised container sequence for intermodal freight trains.

 

Activity: Evaluation of energy KPIs (Key Performance Indicators)

Researchers quantitatively assessed the overall energy reduction that could be achieved through the programme’s Technical Demonstrators, with a focus on the rolling stock. The energy baseline for the four theoretical system platform demonstrators – high-speed, regional, urban (metro and tram), and freight – were defined.

Initial analysis of energy KPIs mainly focused on weight reductions and the efficiency improvements achieved with SiC converters. For most other System Platform Demonstrations (SPDs), the weight reduction significantly reduced mechanical brake energy and wear. Weight reduction is a very efficient measure to reduce overall Life Cycle Cost (LCC) of rolling stock, thanks to reduced energy consumption and secondary effects like reduced wear of brakes.

Final evaluation of energy KPI: the Shift2Rail innovations with an impact on energy were analysed and mapped onto the SPDs. The energy-saving potential of some demos is very small, e.g. Smart Metering or bogies for freight locomotives. Other demos have a significant energy-saving potential, but only those with a sufficiently high Technology Readiness Level (TRL) could be considered.

Key findings

SPD HST300 (high-speed trains, max. speed 300 km/h or higher), energy KPI improvement of 3.77% mainly thanks to three technological improvements: 1) a relative weight reduction; 2) synchronous traction motors with permanent magnets without gearbox; and 3) innovative HVAC system with heat pump.

SPD HST250 (high-speed train, max. speed 250 km/h): technical improvements and achieved energy savings are identical to those for HST300 high-speed trains, but the KPI energy reduction (5.71%) is more significant because of the lower maximum service speed of the HST250.

SPD Regional-140 and SPD Regional-160 (identical regional trains but different service profiles, with respective max. speeds of 140 km/h and 160 km/h): technical innovations include 1) weight reduction on train level but a weight increase of the innovative HVAC system; 2) improved efficiency of line and motor converters with SiC; 3) improved efficiency of conventional main transformer thanks to SiC converters; and 4) innovative HVAC system with heat pump. Together, these enable a KPI energy reduction (specific energy consumption per km) of 14.19% for the SPD Regional-140 and 12.35% for the SPD Regional-160.

SPD Metro: reduced energy consumption of 12.52% thanks to 1) weight reduction, but a weight increase of the HVAC; improved efficiency of motor inverters with SiC.

SPD Suburban: the energy KPI can be reduced by 6.46% for this train category,  through 1) weight reduction of the traction system; 2) improved efficiency of the traction system; and 3) improved HVAC system with a more efficient CO₂ refrigerant.

SPD Freight Mainline: energy KPI improvements (between 19.82% and 24.24%, depending on which measures are combined) through 1) improved efficiency with line and motor converters in SiC technology; 2) improved efficiency of the conventional main transformer thanks to SiC converters with higher switching frequencies; 3) reduced energy consumption thanks to the driver assistant system; 4) reduced energy consumption for freight trains with extended market wagons, thanks to better aerodynamics and reduced running resistance; and 5) reduced energy consumption for intermodal trains with market wagons and optimised container sequence.

Conclusion

A massive effort is underway to reduce the energy consumed by rail transport, including passenger and freight trains as well as all relevant infrastructure. This work can be accelerated by calling on a standardised baseline to estimate energy consumption through simulation and measurement, so as to enable a standardised specification of energy-efficient railways.

Research done in the S2R programme, in the overall system level, highlighted a potential energy saving in rail transport of around 9 to 19%, depending on the different service types for passenger services (high-speed 9%, regional 19%, and urban 12%), and of 4 to 5% for freight services.

Next steps

Several energy-saving topics investigated in the Shift2Rail programme are being covered in flagship projects (FP) under the EU-Rail Joint Undertaking, with the aim of achieving a higher TRL. These projects explore the energy-saving potential of innovations, through demonstrators in real operation. They include automatic train operation with energy-optimised driving (FP1, FP2 and FP6); a Traffic Management System allowing smoother traffic and fewer unplanned stops (FP1 & FP6); substitution of diesel trains by battery and hydrogen-powered trains (FP4 and FP6); an airless train with airless brakes, suspension, pantograph and doors (FP4); and HVAC solutions with alternative natural refrigerant with higher Coefficient of Performance (FP4).

Noise and Vibration

For train passengers and people living near railway lines, noise and vibration (N&V) are undesirable disturbances. Shift2Rail programme included work to reduce the annoyance and exposure to N&V, through two main strands. First, by developing simulation methodologies for the exterior noise of a train (at standstill and pass-by), based on existing tools from ongoing and past projects. Second, by developing and improving the prediction of ground-vibrations by passing trains, a key part of evaluating the effectiveness of innovations aimed at reducing these issues. The outcomes of all this N&V work will enhance the acoustic certification process for new trains, e.g. during authorisation, and support impact studies on vibration prediction during the Environmental Impact Assessment for new or upgrading railway lines.

Challenges

• The N&V produced by trains are influenced by the quality and characteristics of the rails, the track bed, the ground and the surroundings (buildings, plants, walls, etc.). For rolling stock’s acceptance tests, it is essential to separate the noise sources and to judge only the contributions of the train.
• EU Technical Specifications and Interoperability (TSI) and standards define test conditions for the noise testing of trains to minimise the unwanted effects of rails, track and surroundings, and to make test results comparable. However, these ideal conditions (test track) are difficult to find. The challenge here is to correct the effects caused by the discrepancies between ideal conditions and real test conditions.
• Objections from neighbouring residents can block the construction of new railway lines. One solution is to use virtual reality, which is more powerful than plans and abstract figures, to present the future situation (visual and acoustic) to the public and experts.
• Railway lines in built-up areas can cause damage to buildings due to vibrations; sophisticated simulation tools can predict these vibrations.

Under the Shift2Rail programme, researchers completed a number of activities linked to reducing the N&V of rail transport:

Activity: Pass-by source separation

Solution: Development of several methods to identify and separate the noise sources from trains passing by a measuring point.

Discussion

Researchers developed several methods to identify and separate the noise sources from trains passing by a measuring point. The goal is to separate the noise generated by the train (mainly the wheels) from the noise generated and propagated by the rail. Although the main noise is produced at the contact between wheels and rail, there are other noise contributors: ventilation of traction equipment, compressors, gears and traction motors, HVAC, and the contact between pantograph and contact wire. At higher speeds, aerodynamic noise becomes more significant, as well as noise from the contact between pantograph and overhead wire.

Two different methods of source separation were tested and validated in field test campaigns:

Pass-by-analysis method (PBA): test train’s sound is measured with a microphone at a defined position from the track, and vibrations of the rails are simultaneously measured with accelerometers on the rails. Measurements from various test runs are analysed to 1) separate the rolling noise from other noise sources, 2) separate the noise caused by the wheels from the noise coming from the track (rails, sleepers).

Microphone array method: test train’s sound is measured with 75 microphones, on an array in parallel to the track at the measuring point. Rail vibrations can also be measured simultaneously with accelerometers. Measurements taken during the test runs with the train running past this microphone array are analysed, to identify the location and the sound power level of the different noise sources. Compared with the PBA method, the microphone array method can identify the noise from all significant sources on the train, including ventilation, traction motors and gears, pantograph, HVAC and other roof-mounted equipment, plus the aerodynamic noise at high speed at the front of the train.

The PBA method was tested and validated with two field test campaigns:

Test ring in Velim (Czech Republic): 34 pass-bys with a regional train of eight coaches were recorded and analysed. Test runs covered the whole speed range: very low speed (5 km/h) to the maximum speed of 200 km/h. PBA measurements: microphones on both sides of the track, accelerometers on both rails.

Tests near Ledanca (high-speed line Madrid-Zaragoza), a high-speed Talgo-train with power heads at both ends and 12 articulated passenger coaches in between passed the measurement location 12 times at different speeds between 80 km/h and 280 km/h. Measurement point: microphones on one side of the track, accelerometers on both rails.

In both measurement campaigns, the sensitivity of various parameters on the accuracy of the results was analysed, e.g. the microphone position, roughness of rails and wheels, and absorption of the ground.

Key findings

In the two field test campaigns, the PBA source-separation method produced plausible results for the regional train and for the high-speed train, for the overall separated levels of rolling noise and other sources. The high-speed train has clear aerodynamic noise sources at high speeds, mainly at the front of the train. The PBS method allows a reasonably accurate prediction of sound pressure levels, if the site transfer function can be determined correctly. Directivity can also be shown with strong noise sources.

The microphone array method was validated in the same two test campaigns (Velim, Ledanca) as the PBA method. With the microphone array measurements, a map of the train showing the sound pressure levels of all significant sources can be produced for different frequency bands. The influence of train speed on the noise levels and the frequency bands of the different sources can be shown. With the microphone array method, sound propagation’s directivity can also be determined.

Conclusion: the two methods (PBA and microphone array) are not fully comparable, each has its benefits and limitations. The PBA method gives straightforward pass-by spectra at each speed for rolling noise and other sources, which can also be used for overall noise levels and to estimate the sound power for the whole pass-by or parts of it. The PBA method does not produce a spatial image of the sound or a sound power estimate for individual sources.

The microphone array method can pinpoint and quantify individual sources over the train’s whole length and height. This is not necessary for type testing, but it is important for identifying and mitigating noise. With this method, it was possible to clearly identify the contribution of the rail, which was a milestone as this was mostly unsuccessful in the past.

The microphone array method is rather complex, both the measurement setup and the specific algorithms. The PBA method (ideally with under-rail accelerometers and extra pass-bys covering various train speeds) can easily be added to the measurement procedure required by the standard EN ISO 3095, for measurement points, running speeds and track conditions.

Who benefits: Infrastructure managers, Railway operators.

Activity: Use of pass-by noise source separation outputs for simulation

Discussion

Today’s European certification process for rolling stock defines maximum pass-by noise levels, for compliance with the TSI Noise. For noise measurements, the applicable standard requires a minimum track quality – defined as roughness and track decay rate. But it is difficult to find test tracks that meet both these requirements. Measurements on different tracks can produce significantly different results, even if the minimum requirements for track quality are met.

One solution is virtual certification, based fully or partly on simulations. This can reduce time and effort by replacing expensive on-track testing. But virtual testing needs reliable and trusted simulation tools, in order to model pass-by noise.

Virtual certification involves a small number of pass-by measurements, with the train to be certified. These test runs are done on a track that does not necessarily meet the standard’s track quality requirements. Measurement data are then analysed with innovative source separation methods, developed in the Shift2Rail programme, to identify the parameters of the train’s different noise sources.

These parameters are entered into a trusted noise simulation model to calculate the train’s exterior noise, without the disturbing effects of rails, track and surroundings. Thanks to this simulation tool, the pass-by test scenarios defined in the standard and in the TSI can be simulated, and the results can be compared with the limits defined by the TSI.

Researchers assessed the different source separation methods, to see how suitable they are for generating the input parameters of the noise simulation tools. In three field test campaigns, noise measurements were processed with different source separation algorithms to produce the input parameters for the simulation tools:

• Test campaign (Spain) with a meter-gauge commuter train: two motor carriages at each end and two intermediate trailer carriages, test runs 40 km/h to 90 km/h.
• Test campaign (Czech Republic) with a regional train: eight-carriage EMU for electrified lines, test runs 40 km/h to 200 km/h.
• Test campaign (Spain, high-speed line Madrid to Zaragoza) with a high-speed train: articulated train with power heads at each end and 12 trailing passenger coaches, test runs 80 km/h to 280 km/h.

Key findings

For each acoustic source, the simulation tools require input from the sound power level and directivity pattern. The pass-by analysis (PBA) was found to be unsuitable to provide the required input, as it provides only the sound power level for complete vehicles or trains. The beamforming method (using data measured with the microphone array)trl successfully identified the strength of individual sources, but it cannot reliably identify the directivity.

Three different noise simulation tools were used to model the trains and to calculate the noise:

• SITARE: a noise simulation tool developed and used by Alstom, to model the train structure and the ground reflection in the noise propagation. All possible noise sources of the train can be simulated. TWINS tool can calculate the power levels for the rolling noise of wheels and rail.
• TraiNoiS: in-house tool of CAF for the simulation of exterior noise, part of a virtual noise suite.
• ProgNoise: in-house tool of Siemens for the simulation of exterior noise, and modelling of the train structure with the noise sources and ground properties.

The results, i.e. the reconstructed noise levels of the pass-by test runs, were compared between the different tools and with the original pass-by measurements. The achieved accuracy of the pass-by noise calculated with the simulation tools, and compared with the measurements, was:

• The overall pass-by level has been calculated with an accuracy of 2 dB or less in 90% of all cases.
• The middle frequency content of the noise spectra has been predicted with an accuracy of 3 dB or less in 75% of all cases.
• The extreme frequency content (low and high) of the noise spectra has been calculated with less than 6 dB deviation from the measurements in 70% of all cases.

Who benefits: Infrastructure managers, Railway operators.

Activity: Uncertainty assessment of exterior noise simulations

Discussion

Simulation models’ results for the noise emissions of trains depend on the accuracy of many parameters – some measured, others estimated or taken from literature. For the virtual certification of vehicles, the calculation uncertainty is relevant, since the simulation results must be compared to limit values. Experts therefore assessed to what extent the simulation results were influenced by the uncertainty of parameters and calculation methods.

First, they conducted an uncertainty assessment of the simulation tool SITARE noise model for: 1) noise emissions of a complete train, 2) noise sources, 3) equipment noise, 4) rolling noise, and 5) sound propagation and ground absorption (ground resistivity).

Next, they assessed the uncertainty of the noise simulation for the test case of a three-carriage unit with two motor bogies, two Jakobs trailer-bogies, two underfloor-mounted traction inverters, two underfloor mounted auxiliary converters, three roof-mounted HVAC units, and one roof-mounted air compressor. The train was fitted with fairings on the roof. For the simulation of the noise with the train at standstill, only the traction inverters, the auxiliary converters and the HVAC units were considered as noise sources.

Another uncertainty investigation focused on the influence of selected parameters on the overall result, concluding that was also investigated. For this sensitivity analysis, uncertainty was studied separately for ground absorption coefficient (parameter 1) and microphone position (parameter 2).

A further uncertainty analysis focused on two cases: diffraction effects for roof-mounted equipment, and reflection effects for equipment installed on the train’s underframe. To quantify the uncertainty, a comparison was made of the diffraction and reflection effects. Calculated with different models, these were the analysis results:

–              The differences of the results calculated with different models are between 0 and 2 dB for rectangular or chamfered roof geometries.

–              For round corners, the differences range from -2 dB to 3 dB.

–              For roofs with fairings, the differences range from -1dB to 4 dB.

To investigate the impact of reflection of underfloor-mounted equipment at standstill, a separate model was developed. The analytical models were validated with measurements. There was good agreement between the two analytical models and with the measurements.

Key findings

For the assessment of the uncertainty of the noise simulation for the test case of a three-carriage unit, in the standstill case, the uncertainty of the results was calculated to be ±2 dBA. This is considered accurate enough to reproduce the test conditions.

For the modelling of pass-by runs at 80 km/h, also traction motors, gear boxes, and the rolling noise are active noise sources, in addition to traction inverters, auxiliary converters and HVAC units. Here, the calculated uncertainty is also ±2 dBA.

Who benefits: Infrastructure managers, Railway operators.

Activity: Feasibility of reference track normalisation

Solution: Transferring pass-by measurements taken on a real track, into results obtained on a normalised reference track.

Discussion

The EU’s TSI-Noise (Technical Specifications for Interoperability) defines limits for the pass-by noise of trains. It refers to ISO 3095, which defines a reference track with minimum acoustic properties, such as rail roughness and decay rates, to limit the track’s influence on the pass-by noise. Any track equal to or better than these requirements can be used for certification noise tests.

It is not easy to find test tracks that fulfil these requirements and a train’s pass-by noise depends on the track properties, so the test results do not show only the vehicle performance. Ideally, pass-by measurements taken on a real track could be transferred into results obtained on a normalised reference track. With this approach, pass-by noise of the trains would be fully comparable and independent of the test track’s properties.

Earlier research looked at methods to separate the noise components of infrastructure and vehicles, as well as ways to transpose measurement results from one measurement site to another, or to a virtual site. This transposition must consider rail roughness, track decay rates (vertical and lateral), radiation properties of rails and sleepers, and ground reflection and absorption properties.

To correct for all differences of track properties, a full simulation would be the most efficient approach. Train manufacturers have used tools to simulate exterior noise propagation (e.g. rolling noise, equipment noise, or aerodynamic noise) for many years, calling on the sound power levels of the different noise sources as input. For any transposition method, it is essential to separate the rolling noise contributions from track and from the vehicle, as rolling noise is usually the main part of the pass-by noise.

Railways experts assessed the suitability of the rolling noise separation methods against various noise and vibration requirements:

–              ATPA method: advanced transfer path analysis is an experimental method to obtain the noise at the receiver (microphone) position, combining the noise contributions from different parts of the system.

–              Pass-By Analysis (PBA) method: uses the measurement results from several pass-by test runs at different speeds to identify and characterise the different noise sources. The measurement setup consists of a few microphones to measure sound levels, and some accelerometers to measure the rail vibrations.

• TWINS-based method: with the given rail roughness, wheel roughness and the track decay rates, this can transpose the rolling noise of the measurement track to a reference track.
• Microphone array method: this can localise the noise sources on the train and estimate their strength. The method is useful to identify and characterise unknown noise sources on a moving train, but it is not suitable for reference track normalisation.

Key findings

Researchers made proposals for future revisions of TSI and standards, to enable future transpositions to a virtual reference track. This would also promote the use and further development of the methods needed to transpose test results to a virtual reference track.

Proposals for future revisions of ISO 3095: 1) introduce background information about the track’s influence on rolling noise, 2) provide data of a typical ISO-track (rather than only worst-case limits), and 3) in the medium term, propose methods for noise separation and normalisation.

Proposal for future revisions of EN 15461: add a chapter about uncertainties of the track decay rate assessment.

Proposal for future revisions of TSI-Noise: 1) allow the transposition for the pass-by noise assessment, 2) TSI-Noise certification will be possible with either the process as defined today, or with transposition for pass-by noise.

Who benefits: Infrastructure managers, Railway operators.

Demo: Simulation Tool for Vibrations

Solution: Validating the vibration prediction tool’s ability to predict vibrations from railway traffic in buildings next to the track.

Discussion

The goal was to develop and validate a hybrid approach to vibration prediction, combining numerical prediction with experimental results, for integration into software for other environmental studies and graphical user interfaces (GIS level). Thanks to this unique software platform, engineers can perform noise and vibration environmental impact studies in the same integrated software environment. In 2022, a prototype of this prediction tool was finalised and validated as the Simulation Tool for Vibrations, which was used for 18 case histories.

In Lincent in Belgium, measured data (transfer functions and train passages) from various test campaigns was used to validate the vibration prediction tool, focusing on the input parameters and the results produced by the tool. Assessments were made of the tool’s plausibility and usability, as well as the uncertainty of the prediction.

Further validation of the tool included extensive measurement campaigns, recording a wide variety of parameters with a possible influence on the vibration. The comparison of data calculated with the tool with measured data (e.g. for urban traffic, mixed traffic and high-speed traffic) was used to validate the tool’s ability to determine vibrations in the buildings, in free field, and at the emission point.

Key findings

The prototype tool’s vibration predictions were good at the Lincent site, after comparison was made between experimental data and numerical predictions. But researchers found some discrepancies in the predictions, due mainly to uncertainty about track and soil properties at the site. More sophisticated models are required to capture accurately these and other factors around and under the rail track, including rail unevenness and wheel roughness. In general, the models tend to overestimate the global vibration levels. However, when comparing the global vibration velocity levels, differences between measurements and predictions are reasonable: 1 dB for the best cases, with an average difference of about 4 dB, but up to 15 dB for the worst case.

The validation of the Simulation Tool for Vibrations underlined how both the core version and the later IMMI version can predict vibrations with high accuracy, especially when measured data can be used for the emission, the propagation transfer functions, and the transfer functions to the building.

Demo: software tools for auralisation and visualisation

• TRL: 5.
• Who benefits: Infrastructure managers, Railway operators.

Discussion

Auralisation is the acoustical counterpart to visualisation, allowing users to audibly experience situations. The auralisation model was enhanced (improvement of wheel and track models, and of noise barrier model, etc.) and the catalogue of scenarios and noise mitigation measures was expanded (wheel and rail dampers, low-height barriers, etc.). More sophisticated virtual reality (VR) scenarios were developed for the demonstrator. Wearing VR glasses and headphones to cover the 3D effects, a user can dive into a scenario, freely rotate their head and dynamically switch between different variants (e.g. with or without noise barriers).

Under the Shift2Raill programme, a software package with an acoustic module for auralisation, and a virtual reality module for visualisation, was developed. These modules were validated through data from two large-scale projects with measured data.

The developed software was validated by comparing basic acoustic quantities and the psychoacoustic quantity loudness. Comparative listening tests were also carried out to validate the auditory impression on people. Two test cases were used for validation, a pass-by of a regional train without any mitigation measure (no mitigation measure) at velocities of 140 to 160 km/h and a second pass-by of a regional train with a standard barrier of 3 m height at speeds of 140 to 160 km/h.

The validation’s quantitative part included comparing calculated and measured acoustic quantities for each case at the observer locations. A group of acoustic experts also compared two audio files for several scenarios and judged whether the files contained measured or synthetic noise, and whether the noise level was higher, lower or equal.

Key findings

The goals of the acoustic validation were to ensure the:

• Quality of the A&V system.
• Correctness of the calculated noise environment.
• Feasibility and applicability of the A&V system for the use case communication of mitigation measures (mandatory) and optional for the use case “perception studies” (depending on final realisation in the project).
• A good correlation between the synthetically generated sound and sound recorded and measured in field tests. This also included auditory perception during extensive listening tests with test persons, to ensure that the physics was correlated and the human perception of observers.

The validation’s quantitative part showed that measured and simulated noise were very similar at both observer positions. Listening tests showed that experienced test persons could generally find differences in the auditory impression of measured and auralised audio files, but they struggled to identify which file was measured and which file was auralised. Thus the tool can be used to communicate mitigation measures.

Who benefits: Infrastructure managers, Railway operators.

Conclusion

Noise and vibration are unavoidable consequences of rail transport. However, there are many innovations to reduce such disturbance, including prediction using real and virtual testing as well as modelling.

Next steps

• The new directivity method applied in the measurements of pass-by noise requires further application and verification.
• To transpose noise measurement results from the real measurement site to an ideal track (compliant with the standard), further work is needed to improve the methods and to define suitable reference methodologies.
• The recommendations made for improving the SILVARSTAR vibration prediction tool should be implemented in further development of the tool. The numerical and empirical approaches could be enhanced with additional measurements to increase the accuracy of the vibration predictions.
• Inclusion of a high-speed train scenario in the auralisation tool to consider the higher share of aerodynamic noise at high speeds.