Traction Inverters - Efficient by Design

Making efficient use of available materials and reducing material dependency is particularly important in the field of electromobility. Semiconductor manufacturer Infineon shows how proper distribution of silicon (Si) and silicon carbide (SiC) within the vehicle will lead to the desired performance as well as help optimize efficiency and material usage.

Power semiconductors such as Si-IGBTs or wide bandgap (WBG) semiconductors feature unique characteristics, including their high carrier mobility, which make them particularly suitable for use in traction inverters [1]. This is not just about being cheaper or downsizing applications. Rather, in conjunction with a deep understanding of electric drivetrains and vehicle applications, these technologies lead to creative solutions that enable better end applications using less material and help overcome supply constraints.

The power flow diagram in Figure 1 highlights where energy is used, dissipated or recuperated in an electric vehicle (EV) drive train: the propulsion cycle on the left and the braking and recuperation cycle on the right. This version of the diagram shows a vehicle design with two electric axles, assuming one on the front and one on the rear, with a single motor each.

Figure 1

Power flow diagram for a two-eAxle EV (© Infineon Technologies)

Functions of a Traction Inverter

To get an EV in motion, in simplified terms, all you need is a charger, a battery, an inverter, and a motor. However, the inverter and motor are not only responsible for propulsion, they also act as a generator that recovers energy during deceleration and feeds it back into the battery. There are some secondary functions to the inverter and motor, like hill hold and battery preconditioning. The system is also linked to related functions such as torque management, steering, and vehicle stability.

For the traction inverter, Silicon Carbide (SiC) is the technology of choice to achieve highest performance and efficiency. Which suggests that the front and rear axles should also be built SiC-based. In view of the increasing numbers of electric vehicles, this requires some further investigation.

Standardized drive cycle vs. peak performance

The so-called WLTP drive cycle (Worldwide Harmonized Light-Vehicles Test Procedure) reflects near-real-world driving. Being standardized, it provides a reference for OEMs and consumers to compare the efficiency of vehicles. For EVs, it is expressed as energy consumption over a given distance, for example 10 kWh per 100 km, or as MPGe (miles per gallon of gasoline-equivalent), which also allows comparison to traditional combustion engine vehicles.

The WLTP drive cycle, Figure 2, similar to any other mission profile, consists of various acceleration, deceleration, and performance periods over a 23.3-km distance with a duration of 1800 seconds. Its suitability for real-world driving is controversial because drivers are different. However, it is suitable as a basis for rating the efficiency of a car. Such a mission profile allows for calculating the required motor performance for a given vehicle and its key parameters, such as weight, aerodynamic drag as well as efficiency in driving, accelerating, and recuperating. For a typical 1500-kg car, the calculation indicates that less than 40 kW would be sufficient to perform the WLTP drive cycle. This is remarkably little power to accelerate, reach peak speed and recuperate an EV for this mission profile. In order to include vehicles with higher weight and aerodynamic drag, and to sketch a car configuration that features reasonable power to be successfully sold, 80 kW is used as a reference value in our further evaluations.

Figure 2

The standardized WLTP drive cycle represents a typical drive profile (© Infineon Technologies)

Build more with less

Decarbonization drives an ever stronger need to replace traditional combustion vehicles by electrified options. The supply chain of electrified vehicles is highly complex, expensive, and at risk. EV makers are moving from global to local sourcing in order to establish simpler, faster, and cheaper supply chains. By 2030, every second car sold is expected to be electrified; this will then mean 40 to 50 million vehicles sold.

An impressive growth that needs to be enabled. In other words, the supply chain needs to be filled with the materials required for motors and batteries, as well as with aluminum, steel, and semiconductors. Making ist as sustainable as possible, a circular economy is also needed where at the end of their life, materials are reused to produce new vehicles. EV batteries in particular are given a second life in industrial or consumer applications.

However, the electrification of mobility is just at the beginning. Currently, the market sector of battery-powered electric vehicles (BEVs) is still being established. So the materials to equip a BEV fleet needs to be procured, Figure 3. Subsequently, when the majority of all cars are battery-powered, the gathered materials will be circulated, and only the lost materials will need to be replaced.

Figure 3

Thoughtful use of raw materials is crucial for a sustainable EV circular economy in this early phase (© Infineon Technologies)

The goal is to build vehicles with high efficiency according to the majority of driving events, while optimizing the usage of available materials. This will enable the electrification rate to be maximized even though the available material remains limited.

SiC for the traction inverter

Concerning electric vehicles for the broader market, cost-effectiveness and a decent power rating at reasonable efficiency are the goals for the traction inverter. This is not as straightforward as it sounds.

If one looks at the car as a whole and deal with motor performance and performance distribution only, at this level the minimum performance for the drive mission profile and the peak performance desired can be evaluated. This allows to decide where and how to best implement power semiconductors - Si or WBG.

In the example above, an 80-kW motor was sufficient to cope with the standardized WLTP drive cycle and consequently the majority of driving events. If more power is implemented in the car with SiC, it will stay untapped for a very high percentage of the car usage. For driving experience as a fun factor, the 80 kW may not be enough. Some Silicon can be added to increase the vehicle's peak performance: for example, with a Si-part capable of delivering additional 160 kW. This would end up in quite a sporty car. Addressing an entry-level electric vehicle, these values could also be scaled down to 40 kW SiC and 80 kW Si to achieve a vehicle power of 120 kW.

As the vehicle designers can decide on how the Si and the SiC chips shall be distributed within the car, a deeper look into options for configuring the electric drivetrain is a must.

There are various ways of harnessing the benefits of the different technologies for the powertrain and in particular for the traction inverter - whether in terms of efficiency, performance, or cost, Figure 4: from simply migrating to the next SiC generation (1), to right-sizing vehicle performance (2) and finding cost-optimized solutions (3), to deciding for Si or SiC for a secondary eAxle (4), or even mixing technologies within one traction inverter (5). This broad solution space makes it difficult to find decisions, considering that not only the drive but also the recuperation has to be efficient.

Figure 4

Feasible technology options for electric drive trains according to target use case (© Infineon Technologies)

All these options should consider the target mission profile, such as the WLTP drive cycle, in order to achieve the targeted vehicle use cases with the customer value desired. OEMs need to decide how to position the vehicle in the market and where to add value to their target customers' typical use case. The WLTP drive cycle can serve as a reference to compare different options.

There are several options for distributing performance within an EV. The most obvious is splitting it across a main and a secondary drive axle. In our example in Figure 5 configurations 1 and 2 use the full SiC share either on the rear or the front axle, which is a typical set-up for current electric vehicles. Configurations 3 and 4 show fusion inverter options as a tradeoff between efficiency and cost: Si and SiC chips operate in parallel within the same inverter. A closer look at different driving scenarios can help understand why such a fusion inverter gains efficiency compared to the alternatives.

Figure 5

Fusion inverters offer efficiency for propulsion, recuperation and peak performance (© Infineon Technologies)

Taking the energy flow diagram from Figure 1 and diving deeper, the energy flow within the fusion inverters can be evaluated, Figure 6. At standard load, the SiC of the rear axle is used to accelerate and the SiC of the front and rear axles is used in a 2:1 ratio to smoothly decelerate the car. For this load situation, which is comparable to most driving and the WLTP test conditions, acceleration and recuperation can be done entirely via SiC and on the desired axles. In order to increase performance to the peak level, additional Si is used for acceleration. Only Si is used for recuperation, as it achieves better efficiency at high loads. To summarize: Fusion inverters optimize the utilization of the capability of the different semiconductor power switches towards efficient and affordable e-mobility.

Figure 6

Efficient drive profile with SiC, and peak performance enabled by Si (© Infineon Technologies)

Is GaN an option for next-generation inverters?

Currently, Si and SiC are the two most important power semiconductors for inverters. In the future, gallium nitride (GaN) may also play a role. Currently, GaN is used in DC/DC converters and on-board chargers, where it offers additional advantages due to its switching frequency, including high du/dt values.

At first glance, GaN does not seem to be particularly suitable for traction inverters: lower power than SiC, switching speeds, which may not be compatible to motor insulation and a price higher than that of Si. But as the industry had learned from the change from Si to SiC, a merely replacing Si by SiC does not provide the desired benefits: The entire commutation loop has to be adapted to a low-inductance design to leverage the complete SiC advantage. These mainly do not occur inside the inverter itself, but become visible at the vehicle level.

The same is true for GaN: pure replacement of SiC by GaN will not result in an attractive solution. Specific ways have to be found for GaN to play to its strengths, which will exceed the expected cost advantage further down the road.

Conclusion

In order to take advantage of power semiconductors in electric vehicles, component distribution, consideration of efficiency requirements, cost considerations, and sustainability in design are key. The choice of power semiconductor technologies for specific functions such as propulsion, braking, and all-wheel drive depends on the respective vehicle's use case and requirements.

Silicon carbide emerges as the technology of choice for traction inverters, offering benchmarking performance and efficiency. In order to optimize costs and efficiency the appropriate SiC content for a vehicle must be determined based on the intended use case. Although gallium nitride may not initially match the power capabilities of SiC, specific solutions can be developed to leverage its advantages as a potential technology for future inverters.

Reference

• [1] Geiger, D.; Bauer, Ch.: Drivetrain and semiconductor technologies in future EVs. In: ATZelectronics worldwide 18 (2023), no. 3-4, pp. 52-56