In search of more efficient braking systems

Developed over almost two centuries, railway braking systems must be reliable, safe and perform well. Brakes are safety critical, so their design and homologation are highly regulated and standardised. However, this means that railway operators often hesitate to introduce new braking systems.

In the 21^st century, several braking innovations are gaining traction. They focus on new electro-mechanical brakes and adhesion management, but these technologies still have some way to go before they are fully trusted and certified on Europe’s railways.

The R&I work addressed five specific challenges for braking system innovations:

• • • Introducing high SIL (safety integrity level) electronics.
    • Electro-mechanical braking system.
    • Adhesion management solutions.
    • Friction pairing solutions.
    • Virtual validation and certification of brake systems.

Under EU-Rail, building on the work of its predecessor programme Shift2Rail, new innovations for braking systems were developed through several demos:

Demo: Electronic Brake control demonstrator with High Safety Integrity functions

• Solution: A new electronic EP brake control device with SIL4 functions for emergency brake control.
• TRL: 7 (system prototype demonstration in an operational environment).

Discussion

Wabtec Faiveley developed a new system, Metroflexx, a demo incorporating all brake functions in a fully electronical control unit, up to the highest safety integrity (SIL4). With this control unit, hardwired control circuits could be replaced, and the pneumatic equipment could be simplified. The whole equipment was successfully tested in the lab before being installed on a 3-car commuter train of the operator Euskotrain in Spain. The work included a safety analysis of Metroflexx, according to applicable European Standards and UIC leaflets – static tests and a validation campaign with an operational train: brake performance, stopping distances under various adhesion conditions, and the efficiency of the adhesion control system.

Key findings

Static tests demonstrated that Metroflexx can easily be integrated into an existing brake architecture, and has the same performance as the existing system it replaces. Other key findings:

• Braking capacity of the HSIL system is 5% better than a conventional system.
• The innovative adhesion management solution performs better than the existing system (stopping distance, resilience to low and very low adhesion – so no ‘wheel flats’, and faster set-up time).
• Euskotrain now uses the adaptive adhesion management system in regular passenger service.

Electro-mechanical brake system

Two demonstrators of electro-mechanical brake (EMB) systems were developed, built and tested, one by Wabtec Faiveley, the other one by Knorr Bremse.

Demo: Electro-mechanical brake demonstrator (electro-mechanical brake actuators)

• Solution: A closed-loop control of the clamping force of the brake calipers.
• TRL: 4 (technology validated in a lab).

Discussion

Wabtec Faiveley built a test bench with an electro-mechanical brake concept demonstrator, followed by a full prototype of an electro-mechanical brake system. Lab tests checked if the system could match a conventional electro-pneumatic system in terms of service brake, wheel slide protection, holding and parking brake, and load-weighted emergency brake.

Key findings

The proposed concept, which allows a more accurate braking force control than achieved with the electro-pneumatic brakes, was successfully tested in the lab. All brake and wheel slide protection functions could be achieved with sufficient accuracy. A next step will be investigating integration of the EMB system in a train.

Demo: Electro-mechanical brake system (electro-mechanical brake actuators)

• Solution: Electro-mechanical brake actuators, which can directly replace existing electro-pneumatic brake calipers.
• TRL: 7 (system prototype demonstration in an operational environment)

Discussion

Knorr-Bremse developed and built electro-mechanical (EM) brake actuators which can be used to directly replace existing electro-pneumatic (EP) brake calipers. Prototypes of the EM brake calipers were tested in the lab to confirm compliance with the standards for railway equipment, e.g. shock and vibration, dust, dirt, humidity, electromagnetic compatibility, high and low temperatures. For the functional tests on different adhesion conditions, Knorr-Bremse used its ATLAS test bench. This was followed by full-scale tests on a test vehicle of MAV in Hungary. A total of 180 brake cycles were performed and analysed on a 45km long test track at speeds up to 160 km/h.

Key findings

These tests led to the successful validation of the electro-mechanical brake actuators under real operational and environmental conditions. Other key findings:

• The concept of the electric power supply on the train must be adapted to the needs of electro-mechanical brakes.
• The control equipment in the driver’s cab can remain unchanged.
• The EM brakes can provide more accurate information about the health of the brake equipment than what is possible with EP brakes. EM brakes also require less installation volume on the train, because bulky pneumatic hardware is replaced with compact electrical equipment.
• EM brakes alone will not necessarily lead to elimination of the complete air system on trains.
• Standardisation efforts will be needed to achieve a similar level of interoperability between brake systems of different suppliers.
• The safety requirements specified in the TSI LOC PAS can be fulfilled by the EM brakes with their redundant architecture, as is the case with EP brakes.
• The TSI (Technical Specification for Interoperability) requirement, specifying an extremely low permitted frequency of failures with catastrophic consequences, will call for the development of electronic brake control equipment with high safety integrity.

Adhesion management system

Brake performance (deceleration, stopping distance) depends on the coefficient of adhesion between wheel and rail. Low adhesion can lead to blocked wheels and hence ‘wheel flats’, as well as excessive stopping distances. Adhesion management systems aim to make best use of the available adhesion, in order to prevent damage and minimise stopping distances. This leads to improved performance, high safety, and low maintenance costs.

The work on adhesion management included:

• Contributions to the adhesion catalogue.
• Contributions to standards.
• Development and testing of technology demonstrators.

Contributions to the adhesion catalogue

One goal of the adhesion catalogue is to better understand how specific parameters affect the adhesion coefficient between wheel and rail. Adhesion knowledge, collected in the catalogue, can be used to develop and optimise adhesion management systems. Two companies (Wabtec Faiveley, Knorr-Bremse) investigated how the load impacts adhesion in lab tests.

Company	Discussion	Key findings
Wabtec Faiveley	Lab tests on a multi-axle roller rig, with four driven wheelsets exerting a defined (and variable) load on the inner surface of a hollow cylinder representing the rail. The roller rig represents a reduced scale model of the wheel-rail contact. Various loads and low-adhesion conditions were tested, with water and soap solutions in front of the first wheelset. Tests were done in two different configurations: ●        Only one wheel in contact with the roller surface (i.e. the simulated rail) to establish the adhesion limit for one axle. ●        All wheels in contact with the roller surface to determine the multi-axle adhesion limit and to consider possible cleaning effects by the other wheels.	Test results showed a marginal increase of the adhesion limit with increasing load, but not sufficient to justify a dedicated model for this effect.
Knorr-Bremse	Lab tests on the ATLAS full-scale roller rig. Investigated systematically the effect of axle load on wheel-rail adhesion in the presence of different contaminants, e.g. soap-water and oil. Tests were repeated with three different axle loads.	The effect of the axle load on adhesion was within the limits of the reproducibility of test results, with a slight tendency of reduced adhesion with increasing axle load. It was therefore concluded that the effect of wheel load can be neglected in the modelling of adhesion.
Cleaning effect of magnetic track brakes (MTB)	Earlier work showed that the cleaning effect of the MTB leads to a 50% increase of the available adhesion for the axle after the MTB, although this was based on just one field test with activated MTB. So further analysis was done to improve the modelling with additional test data. Alstom provided test reports and test results from field tests with Bombardier M7 trains.	The data were analysed by Wabtec Faiveley. It concluded that the model needs to be refined, by interpolating the cleaning coefficient as a function of the brake entry speed.

 

Contributions to standards

Automated rail traffic systems and signalling systems need to know trains’ braking capacity, in order to calculate stopping distances. Braking capacity depends on current adhesion conditions. If these conditions can be predicted more accurately, safety margins in the calculation of stopping distances can be reduced. This would result in an increased line capacity.

Technically speaking, two topics must be solved:

1. Determination of the current wheel/rail adhesion conditions during operation. Standardised, solution-independent requirements for the adhesion determination system must be specified.
2. Qualification of the technical equipment which influences the braking curve. So far only the wheel slide protection system is qualified (in accordance with EN 15595). If other functions (e.g. smart sanding systems) are to be considered in future, the necessary qualification criteria and test procedures will have to be defined.

Discussions with the standardisation and regulation bodies (e.g. EECT, ERA) took place. The aim was to look more closely at wheel/rail adhesion in the braking curves for ETCS (European Train Control System) operation. The resulting proposal is to update in the long term the formula for the braking curves, independently of any specific solutions. The formula should be unambiguous and independent of any national technical rules.

Exchanged of views and information with the ATO (Automatic Train Operation) working group. One topic was how adhesion management systems can be considered in the calculation of braking curves in ATO systems. Today, the train driver plays a key role when considering adhesion, but new concepts should be defined to allow more automation. The collaboration with the various standardisation and regulation bodies, and with the relevant working groups, has continued.

Demo: Contributions to standards, in particular related to brake curves for ETCS and ATO applications

• Solution: Adaptive wheel-slide protection and the adhesion improving sander.
• TRL: 7 (system prototype demonstration in an operational environment).

Discussion

The adaptive wheel slide protection and the adhesion improving sander, both developed by Wabtec Faiveley, were integrated and installed on train 955 of Euskotrain in Spain, together with the Metroflexx braking system described above. The train consists of two motor cars and one trailer car. It has a maximum speed of 90 km/h. The adaptive wheel-slide protection (WSP) algorithm was implemented on each of the train’s three wheel-slide protection systems, covering the wheelsets of motored axles and trailer axles. The adhesion improving dispenser equipment was installed close to two wheelsets.

The adaptive WSP algorithm can self-adapt its behaviour to a wide range of environmental (and hence adhesion) conditions. This concept enables shorter system tuning, shorter stopping distances, less air consumption, and the prevention of wheel flats. Mechanical supports for installing the demonstrators on train 955 were be designed and tested, to ensure their robustness in regular service, using finite element models and the analysis of static/dynamic load cases. The design’s suitability was confirmed by lab tests, e.g. vibration, shock and endurance tests.

The adaptive WSP algorithm and the adhesion improving dispenser were tested on the Euskotrain track in October and November 2021. Besides the full service-brake applications from the maximum speed of 90 km/h, other braking regimes were also tested, e.g. full service brake with electrodynamic braking, and emergency braking.

Key findings

Test results showed that, with the adaptive WSP system, the extension of the stopping distance under bad adhesion conditions could be reduced by 34%, compared with the current wheel-slide protection system. This is less than half of the extension permitted by the standard EN 15595 and by the UIC (International Union of Railways) technical documentation. With the combination of the adaptive WSP with the variable sanding device, the stopping distance under low adhesion conditions was nearly the same as the nominal stopping distance with dry rail. All test criteria specified by EN 15595 were fulfilled. No wheel flats were observed. The demonstrator of the adaptive WSP was accepted by the train operator and it continues to operate in normal passenger service.

Demo: Adhesion management demonstrator

• Solution: Optimised adhesion through sanding system and software to adapt the wheel-slide protection (WSP algorithm to prevailing conditions.
• TRL: 6/7 (technology demonstrated in a relevant environment / system prototype demonstration in an operational environment).

Discussion

Knorr-Bremse developed two functions to optimise adhesion:

1. A system to activate the sanding function after detection of low adhesion conditions in braking.
2. Software to adapt the wheel-slide protection algorithm to the prevailing rail conditions, resulting in an optimised utilisation of the available adhesion in braking.

These technology demonstrators were installed on the Advanced Train Lab of DB, a diesel-electric high-speed train equipped with sophisticated measurement equipment. The WSP demonstrator and the special sanding system were installed in parallel to the existing systems on the train.

The adhesion conditions were adjusted by using water-soap solution sprayed on the rails, and rails contaminated with oil and with paper, mimicking leaves on the rails. Tests were performed at different speeds between 40 km/h and 120 km/h. Different WSP algorithms were tested, the sanding control function was varied, and the magnetic track brakes were applied to test the influence of the cleaning effect of the magnetic track brakes.

A total of 333 test data sets were taken and analysed. A first set of tests was carried out in line with the European Standard EN 15595, to confirm that the advanced WSP system and the algorithm comply with the applicable standard.

Key findings

Test results with water-soap solutions sprayed on the rails, as specified in the standards, confirmed that the advanced WSP algorithm achieves similar or better results compared with current state-of-the-art WSP equipment. Under extremely low adhesion conditions, caused by oil contamination or paper on the rails, the new WSP algorithm of the demo shows a significantly better performance than the existing systems, achieving up to 25% higher mean train deceleration.

Demo: Adhesion management demonstrator

• Solution: Optimised use of adhesion in traction and braking, and detection and prevention of torsional vibrations and instability in traction.
• TRL: 7 (system prototype demonstration in an operational environment).

Discussion

The adhesion management solution developed by CAF in this demo focused on optimised use of adhesion in traction and braking, and on the detection and prevention of torsional vibrations and instability in traction. Two demos were developed, tested in the lab and finally implemented on trains and tested on track:

1. Detection of torsional vibrations, mainly occurring in traction, and the algorithms in the traction control to prevent and eliminate these instabilities.
2. Brake blending concept to optimise the interaction between the electro-dynamic brake and the friction brake under difficult adhesion conditions.

The torsional vibration detection and prevention system was implemented on a four-car EMU (Electric Multiple Unit) of Euskotrain and tested on track in an extensive test and measurement campaign. Strain gauges were mounted on the axle of the test train, to observe and measure the torsional vibrations of the wheelset without/with the detection and prevention algorithm activated. Adhesion conditions were influenced by spraying water or a soap-water solution on the track. The influence of different parameters on the generation and prevention of torsional vibrations was studied.

Simulations were carried out to define an optimised blending strategy. Some tests were performed on a tramway platform. The concept was further developed and integrated on a real system, a Metro-type EMU, with this advanced blending strategy tested on Brussels Metro. The strategy included applying some friction braking on trailer bogies, while reducing the dynamic brake effort on motor bogies to prevent sliding under bad adhesion conditions.

Key findings

First demo (detection of torsional vibrations): with the new detection and prevention algorithm, torsional vibrations could be reduced by c. 45%.

Second demo (brake blending concept), the stopping distances could be significantly shortened and the number of ED (electro-dynamic) brake trips under difficult adhesion conditions could be reduced.

Demo: Adhesion management and brake blending demonstrator

• Solution: Optimised brake blending strategy.
• TRL: 7 (system prototype demonstration in an operational environment).

Discussion

Prototypes of potentially suitable pairs of brake pads and brake discs were manufactured/tested on a dynamometer test rig in the lab of Wabtec Faiveley. All relevant parameters were measured and analysed, e.g. the amount and particle size of dust produced, noise, wear of brake pads and brake discs, and the brake force achieved under different conditions. Possible signs of local overheating or damage (cracks, material chipping) were detected with visual inspections.

The most promising pair of brake pads and brake disks was selected and installed on a commercial commuter train of Trenord in Italy. In the dynamometer tests, the selected pair of brake pad and brake disk prototypes achieved a significant reduction of brake dust, compared with the commercial equipment in use today.

The prototypes were tested for two months in commercial service on the heavily loaded commuter network of Milan. Three visual inspections were carried out to observe the behaviour of the prototypes and these inspections did not reveal any cracks, hotspots or other defects on brake pads or brake discs.

Key findings

This innovative prototype of friction pairing demonstrated its fitness for purpose in commercial operation, and its eco-friendly low wear and low noise behaviour.

Virtual Certification of Brakes

It takes time and is expensive to do the acceptance process on the braking system of new trains. This process can be speeded up by doing simulations of validated models, of the train and braking system, to replace some of the tests on track.

Demo: Friction Pairing prototype

• Solution: Optimum combination of friction brake pads and brake disks, for eco-friendly braking.
• TRL: 6 (technology demonstrated in a relevant environment).

Demo: Simulator architecture and models for virtual testing of brake performance

• Solution: Simulations with validated models of the train and braking system.
• TRL: 6 (technology demonstrated in a relevant environment).

Discussion

Knorr-Bremse proposed a modular architecture for the train and brake model: components for the train-wide mechanical model, for the train-wide pneumatic model, and for the pneumatic model of the bogie equipment. It is essential that the different model components can be exchanged between different stakeholders, such as the brake manufacturer and the vehicle builder (integrator). For integration of the individual model components in an overall system model, the interfaces must meet open exchange standards, e.g. FMI (Functional Mock-up Interface) and SSP (System Structure and Parameterisation).

This modular simulation concept was implemented to demonstrate its feasibility. The exchange of components of a railway vehicle simulation model for braking performance calculation was done in line with the simulator specification. The overall simulation framework was developed and integrated by Virtual Vehicle (ViF), while development of the simulation models plus the verification/validation activities were done by Knorr-Bremse.

The vehicle model components included models of the following: brake actuation, brake control, car body, energy storage, Organic Base “FA” brake pads, main control, pneumatic converter, and track. Integration of these stand-alone components allowed output data to be generated on vehicle level, which would not have been possible otherwise.

To confirm the correctness of the simulation results, a comparison between two different approaches was conducted. ViF generated a set of SSP files, representing the simulator architecture for simulation models, interface definitions (exchanged variables, data types, time steps etc.), and interface connections between the simulation models using the interface definitions. VIF and Knorr-Bremse used two different commercial tools for the integration process. The previously generated SSP files were used as templates to represent the FMUs (Functional Mock-up Units), i.e. the software components.

Validation of the simulation models is crucial, to ensure confidence in the models and the results. The two train operators provided test data from their commissioning data bases. Commissioning test data of a 10-car EMU was used to validate the simulation model and the results generated with the simulator. Some of the model parameters had to be calibrated.

Key findings

During development of the VVC Process and Simulator Specifications, it became evident that braking distance and any other outputs calculated by the simulator must be represented by a numerical model and a statistical model. Data from the statistical model must be generated on the basis of variation of the numerical model’s parameters. The abovementioned specifications can also serve as a blueprint for implementation of virtual testing in other safety critical domains of the European railway system.

For the implementation into TSI LOC & PAS, VVC will be proposed as an innovative solution in a Technical Document. The technical details can be used for the specification and application of virtual testing for braking performance, until relevant standards have been adapted.

Conclusion

Braking is a safety-critical aspect of trains, which have relied for many decades on traditional and long-trusted systems such as pneumatic and spring-assisted brakes.

New braking solutions have great potential to reduce braking distances and to reduce wheel and track wear, independent of track conditions and train loads. These solutions include more use of electronics, electro-mechanical braking systems, adhesion management systems, innovative friction pairing, as well as virtual validation and certification of braking systems.

Next steps

• Further work on the improved formula for ETCS and ATO braking curves is currently ongoing in the EU-Rail project FP2-R2DATO, in the ATO and ETCS working groups, and in the EU-Rail System Pillar.
• Risk analysis of the use of virtual testing in the brake authorisation process should continue, in the EU-Rail project FP2-R2DATO.