Towards more efficient, Silicon Carbide-based traction systems

Trams, metros, and regional and high-speed trains are all important components of a public transport system. They also have the potential to reduce transportation’s carbon footprint and, in doing so, help deliver a carbon-free future. This potential lies largely within these vehicles’ electric propulsion systems.

Novel electric propulsion systems offer an opportunity to reduce energy consumption, weight, volume, and maintenance costs. Unfortunately, the electronic technology currently used in traction drives is based on silicon (Si) materials like insulated-gate bipolar transistors (IGBT), which face some significant limitations.

Silicon Carbide (SiC) is emerging as a new alternative to Si materials, one that can provide high-speed switching and lower losses. As such, SiC-based devices are being investigated for their ability to optimise traction systems.

At the forefront of this effort is the Traction System Technical Demonstrator (TD). With the goal of paving the way towards new, more efficient traction system architectures, the TD developed five demonstrators (demos), each of which is based on SiC technologies.

AC traction-system used on dual system tramway

• Solution: Alternate Current (AC) traction system for tramways comprised of a SiC-based line-inverter and a main transformer.
• Technology Readiness Level (TRL): 5 (technology validated in relevant environment)

Discussion

This demo looked at the effect SiC technology has on an AC traction system used for dual system tramways, with a specific focus on how it impacts energy consumption, weight, and volume.

The demonstration was performed first on a simulation, the results of which served as the basis for a new traction system designed to replace the IGBT with a SiC metal-oxide-semiconductor field-effect transistor (MOSFET).

The demonstrator compared the Si- and SiC-based systems in as close to real-world conditions as possible.

Key findings

As to the impact on the traction system, the use of SiC-based converters resulted in twice the power density than a Si-based system. The former also saw a 25 – 30% reduction in losses (depending on the mission profile).

In terms of impact on the transformer, the SiC-based system reduced the number of windings by two and leakage inductance by 75%. Although weight increased by 5%, the SiC MOSFET led to a 15% reduction in mission profile losses and a 20% reduction in transformer losses.

Other key findings include:

• A higher rate of voltage change over time (dV/dT) was exhibited, the result of the higher switching speed of the SiC MOSFET in the case of unshielded cables (dV/dT 5 to 10 kV/µs generated).
• The SiC-based system increased stress in the components and in the insulation of the transformer.
• The use of shielded cables reduces generated dV/dT to 3 to 6 kV/µs.

SiC vs Si

Main parameters of the optimised traction system based on SiC technology compared to a Si-based system (converter and transformer used on tram).

• • • Cost: >100% higher
    • Weight: -20%
    • Volume: -15%
    • Losses: -10 to -15%
    • Acoustic noise: -5 to -10 dBA

Conclusion

SiC MOSFETs have the potential to improve traction systems. Based on the required vehicle performance, power supply data, and electromagnetic compatibility (EMC) requirements, the technical parameters of the line converter and transformer were defined. By carefully optimising the system design and addressing electromagnetic compatibility requirements, this technology can enhance performance while reducing both energy consumption and the overall size of the system.

Next Steps

• Specific issues related to the higher switching speed and resulting higher dV/dT must be addressed to reduce the stress on other components, including the motor.
• To further optimise energy consumption, both the tram line converter and IGBT traction converter must be replaced with a SiC MOSFET converter.

Metro traction system with SiC technology

• Solution: New traction converter for metro application designed and fabricated with SiC MOSFET based power modules.
• TRL: 7 (system prototype demonstration in operational environment)

Discussion (a) – SiC-based traction system

To provide metro rolling stock with more efficient, lighter, and smaller equipment, this demo evaluated a full SiC-based traction system installed on an Euskotren unit. The goal was to demonstrate how the use of SiC-based technology can reduce volume, weight, and maintenance costs while also increasing energy savings and reliability. This was done via three tests:

• The new rolling stock equipment’s compatibility with existing track circuits (within the Euskotren infrastructure and ATP system) was analysed. The tests confirmed that SiC-based traction converters are compatible with such infrastructures.
• Next, efficiency measurements were carried out through laboratory tests for different operating conditions, loads and parameters (e.g. switching frequency, motor flux, catenary voltage). These tests showed that efficiency improvement is highly dependent on operating conditions.

More details

During lab tests, four different load profiles were simulated to compare losses of the SiC-based equipment versus the Si-based equipment. Losses were reduced by between 800 W and 1350 W (depending on load profile). This corresponds to an energy savings of 2.43 to 5.91%. The demo concluded that the combination of the optimal flux strategy to the motors and using SiC technology can achieve a 2900 W improvement over Si-based equipment.

• The equipment was then mounted on the train for field tests.

Key findings

• Results obtained were better than expected, with an improvement of approximately 4% (expected improvement was 2.43%).
• Improvement due to SiC MOSFET was greater at low speeds.
• Combing optimal flux strategy and SiC technology can result in a 2300 W improvement over Si-based equipment.
• A 10.51% saving in energy consumption was achieved over the Si-based Euskotren solution.
• With further optimisation, energy consumption could be reduced by another 10.51% (optimise motor cooling, motor de-fluxing, optimum switching).

Discussion (b) – SiC devices and other powertrain stages

The demo also looked at the benefits of using SiC devices in other stages of the powertrain, including the DC/DC and input AC stages. As to the former, the proposed DC/DC converter for a DC-link voltage of 1800 V using SiC technology cuts weight in half while also reducing volume by 40%. It further achieves a 50% efficiency increase. As to the inductor, improvements include a 50% reduction in weight, 37% decrease in volume, and 14% increase in efficiency. More so, because the SiC-based converter’s cooling system requires lower airflow, lower acoustic noise is emitted by the fan. It also means reduced power consumption (between 53 to 65%, depending on the configuration). A similar analysis was performed for an input stage operating in a 15 kV AC system, with a particular focus on its impact on the transformer design.

Conclusion

Using SiC technology offers significant opportunities to improve efficiency at different stages of the traction system (AC/DC and DC/DC converters, phase inductors, transformer). Furthermore, the use of SiC MOSFETs is fully compatible with current infrastructure and peripherals.

AC traction system with SiC for suburban application

• Solution: Prototype of AC traction system for suburban application designed to address customer requirements.
• TRL: 3 (experimental proof of concept)

Discussion

The solution includes SiC-based converters, a main transformer, and a traction motor. It is configured based on a single axle drive, car motion cooled converter, high-capacity line converter with parallel 800A SiC devices, and a high-capacity motor converter with one 450A SiC module per phase per motor. The traction converter unit consists of two converter front sections, with a middle section between. Each front section contains equipment for individually driving two traction motors (single axle drive). The converters use 3.3 kV / 800A SiC power modules. The demo conducted simulations aimed at optimising the main transformer, as well as the traction motor. As to the latter, the goal was to achieve a design that both minimised energy losses and reduced the motor’s weight compared to the base system. The target was to achieve a 5% weight reduction for the entire traction system.

Key findings

• The optimised system’s total losses, including SiC-based traction inverters, was 24% lower than the base system.
• The prototype’s traction motor achieves 16% losses and 0.95 per unit reduced weight compared to the base system.
• Net energy consumption is reduced to 88% of the base system’s consumption (when performing the suburban operational profile).
• Line Converter Module (LCM) switching frequency must be lowered to 2 kHz when performing thermal dimensioning load cycle (due to the high temperatures of the LCM semiconductors).
• Even though total losses of the SiC-based system are 83% compared to the base system, net energy consumption is lowered to 93% of base system consumption (when performing thermal dimensioning load cycle).

Conclusion

The prototype AC traction system for suburban applications addresses key customer requirements, including reducing maintenance, energy and overall costs; decreasing volume, weight and noise; and increasing reliability.

Full SiC sub-system on regional train SiC prototype

• Solution: Prototype of an improved suburban traction system that meets customer expectations.
• TRL: 7 (system prototype demonstration in operational environment)

Discussion

The prototype consists of newly developed components covering the full traction SiC sub-system – from the pantograph to the wheels – for an AC catenary voltage. The traction system includes:

• Transformer with wind speed cooling catenary voltage AC 2500 V.
• Traction case-power converters (4 QC rectifiers, three-phase inverter based on SiC devices).
• High speed traction motor (three-phase squirrel cage asynchronous, six poles, open motor, maximum speed of 7,500 rpm).
• Improved gear box (reduced drag losses).

The entire traction system was validated on a specific test bench capable of reproducing the French catenary system of 25 kV/50 Hz and 1500 V DC. Specific tests, inspections, calculations, simulations, and mock-ups were performed to assess that the system met all requirements and performances before moving to train validation tests, which took place using an XCD regional train. The train is compatible with 1500V DC and 25 kV/50 Hz contact lines, with a maximum speed (during testing) of 140 km/h + 10%.

Key Findings

• Aeraulic and electrical noise: noise at 80 km/h (train speed) is higher than the simulation results, showing that the system is not fully compliant.
• Energy consumption measurements confirmed the theoretical energy saving of 13% (train).
• The energy saving on a regional route profile was 12.8%.
• The PINTA3 Traction Drive is less noisy than the XCD one, although results depend on the train’s operational mode.
• All the components inside the traction case are under their maximum temperature limits.
• The efficiency of the cooling system depends on cab configuration. In some cases, margins are reduced, but maximum temperatures are still below the 8% limit.
• Test results on the motor did not reveal any issue for temperatures or vibrations regarding its operation in a bogie and train environment, nor did any breakdowns or failures occur.
• The new high-speed gearbox and cooling solution resulted in a 1% improvement in train energy consumption.
• The new gearbox, together with a new lubrication system, reduced drag losses by 12% and has been validated up to maximum speed of 140 km/h.
• The highest temperatures reached on the high-speed gearbox are all below the allowed limits.

Conclusion

The new full traction sub-system based on SiC high-power density converters, high efficiency transformer, and high-speed motors and gear box can improve the performance of the entire traction system.

HST traction system integrated in train based on independent rotating wheels

• Solution: 350 kW traction motor for independent rotating wheel for high-speed trains applications.
• TRL: 6 (technology demonstrated in relevant environment)

DiscussionAfter reviewing different options, ranging from three to eight motorised independent rotating wheels, including six to 16 motors and a per motor power rating ranging from 170 kW to 450 kW, the demo selected a 350 kW option.Called MK5, the Permanent Magnet Synchronous Motor (PMSM) was designed for direct drive configuration with high efficiency. It can fit onto a Talgo train’s single axle running gear, thus enabling a different concept from a motorised conventional bogie. All the motor’s parts (rotor, stator, skewed magnet ring, pole-to-slot ratio) were optimised by taking the best trade-off between electrical and mechanical characteristics. Apart from the motor’s new components, some existing components (wheel, axle, and box) were modified to adapt to the new wheel motor. The solution was tested under different ambient conditions.

Key findings

Three types of wheelsets were compared:

• Talgo 350 with two discs on the wheel and a third on the axle
• Talgo AVRIL with two discs on the wheel
• Motor on wheel (no disc brake)

The motor on wheel solution provided up to 60% of the maximum brake required in emergency, reaching 2.6 kNm during 100 seconds per wheel without sustaining any damage.The solution with four motorised axels combined with a wheelset with a third brake disc provides 7.9% and 15% more braking effort in high and low stages respectively.The results for motor and inverter efficiency showed good agreement with expected values. The exception was at very low torque, where small errors in the calculated inverter losses could give rise to significant errors in the calculated value.

Developed wheel motor parameters

• • • • Continuous power: 350 kW
      • Torque: 1.5 kNm
      • Efficiency: 97.4%
      • Torque ripple: 3.7%
      • EMF constant Ke: 7.43 Vs/rad (measured) / 7.5 Vs/rad (modelled)
      • Torque constant: Kt = 6.13 Nm/A (measured) / Kt = 6.05 Nm/A (peak)
      • Winding temperature (during test): under 100 C

Conclusion

While the motor’s measured efficiency is lower than expected, it nonetheless achieves the desired speed/torque performance. This positions the motor as a potential solution for achieving improved traction system performance at the train level, such as an interoperable high speed of 360 km/h (vs 330 km/h) with increased power at wheels of 10.9 MW (vs 8.8 MW) and based on distribution traction with an independent rotating wheelset.

Next Steps

• Finalising installation on the train and conducting on-track testing.

Paving the way to more efficient traction systems

By developing new, SiC-based traction components and subsystems, the Traction System Technical Demonstrator has successfully paved the way towards more efficient traction system architectures, including for use in metro, regional, and high-speed trains.