The Schaeffler eDrive Platform

Modular and Highly Integrated

Thomas Pfund

​I. Introduction

​II. Modular Approach

​III. Platform Components

​IV. Summary and Outlook

I. Introduction

The development of electric drives goes way beyond the electric motor. Optimal efficiency, range, and system costs can only be achieved through the interaction of the motor, power electronics, sensors, mechanical integration, and control strategy, which is why a system-oriented approach is needed.

Schaeffler already proved back in 2011 that it possesses system expertise with the “Active e-drive” concept vehicle. The combustion engine in a Škoda Octavia was replaced with two electric axle drives, each with 105 kW of nominal power. Even then, these motors were units that had been developed within the company. While the gear ratio of the motor and axle speed was fixed initially, a third generation rear axle drive with a two-speed transmission and nominal power reduced to 65 kW was employed in the same concept vehicle beginning in 2014. The drive now included new power electronics developed within the company. Both generations of the concept vehicle had torque vectoring in each wheel independently. While the power electronics for the tractive drive was still a separate system at this time, it was already installed on the axle for torque vectoring.

Meanwhile, it is now foreseeable that electrified vehicle quantities will reach large proportions during the current decade and that the degree of mechatronic integration will increase considerably. At the same time, very different topologies are being implemented for the powertrain, which can be classified based on its installation position [1]. Four  classes of aggregates are needed for the design implementation of these powertrains:

•  Hybrid modules for integration in the powertrain of the combustion engine

•  Dedicated hybrid transmissions for specifically realizing hybrid and electric driving modes

•  Electric axle drives for dedicated hybrid powertrains and purely electric vehicles

•  Wheel-hub drives for new mobility concepts.

Their complexity is increased by the fact that electric drives can be operated at various voltage levels, ranging from 48 V for initial hybridization to 400 V in purely battery-driven vehicles. The first manufacturers have already launched projects with an 800-volt on-board electric system in order to achieve short battery charging times at acceptable current levels. Accordingly, the power spectrum of electric drives is very broad – from 20 kW to more than 400 kW.

What all future aggregate concepts have in common is that they will integrate the mechanical and electrical/electronic systems to a large degree – Figure 1. However, the requirements for electric motors are very different. For example, the axial installation space is limited both in hybrid modules as well as in wheel-hub motors. Moreover, the speed is directly coupled to the powertrain, meaning that power scaling mainly needs to happen via the torque of the e-machine. While electric axle drives have greater axial installation space available, the diameter is generally limited due to the installation position. Coupling to the axle is always via at least one transmission stage in order to guarantee the required drive-away torque. This makes it possible to increase the capacity of the electric motor over higher speeds, which in turn makes smaller designs and material sa-vings possible. The same applies for most dedicated hybrid transmissions. In contrast, power electronics is largely independent of the powertrain type; it is mainly defined by the voltage level, the amount of current needed for maximum output, and the installation space specified for the powertrain.

Figure 1 Powertrain concepts with an integrated electric drive

The large number of requirements listed necessitates a modular approach for various installation spaces and power classes. This is even more true since – depending on the design – the drive’s electric and electronic components make up to 80 percent of the total added value and are therefore key elements in the total cost.

II. Modular Approach

Developing a Technology Platform

In order to support a large number of vehicles and drive concepts on the one hand and yet minimize the development expenses on the other, Schaeffler has developed a modular technology approach for electric drives. Its three levels – Figure 2 – not only include electrical components, but also the hardware and software needed for controlling them.

Figure 2 Modular technology platform for electric drives

The electric motors form the base of the platform. Due to the different requirements described at the beginning, multiple series need to be defined, each with a scalable output. According to the current estimate, the complete spectrum of future applications can be covered completely with six series. Depending on the necessary power density and other requirements, both permanently-excited motors as well as asynchronous machines will be used for this.

The middle level, to begin with, includes the power electronics with all key components, such as the power switches, capacitor, bus bars, driver stages, and sensors. Another part of the platform concept is a carrier frame that serves to channel the coolant and as a heat sink for all components of the power electronics. Also pertaining to the second level is the hardware for the drive control, the specifications of which not only depend on the electric motor and power electronics, but also on the functions carried out at the vehicle level, thereby requiring it to be implemented specific to the application. A key example of this is the communication network, which may be executed as a FlexRay, CAN, or CAN FD network.

For this reason, it cannot be separated from the third level where the software platform is located, which has likewise been developed according to a function-oriented approach. The software includes a functional library based on AUTOSAR, which also defines the requirements for the associated hardware.

III. Platform Components

Electric Motors

When regarding highly-integrated powertrains such as hybrid modules, electric axles, or wheel-hub drives, it is difficult to identify the electric motor as an independently functioning unit. The complete motor function often cannot be tested until the powertrain has been assembled, since multifunctional modules are also determining factors for the functioning of other subsystems. Examples of this are bearing arrangements that also provide support for an integrated drive, or rotor carriers that are also plate carriers for a clutch. The upshot of this is the high requirements for the test concept du-ring production. It is necessary to utilize the measurement of parameters such as winding resistance, inductance, or magnetic field distribution of the components to assess the quality of the magnetic circuit assembly (stator and rotor) in order to immediately sort out faulty parts and not fail to identify them until during the end-of-line test.

In addition to the main function, that of representing defined speed-torque behavior, other demands need to be handled by this magnetic circuit as well:

•  Ideal cooling-down capacity and a high copper filling ratio to ensure long-lasting continuous performance

•  Minimal use of materials in order to optimize costs

•  Minimized harmonically occurring radial and tangential force effects in the stator for NVH optimization

•  Minimized harmonics, which for their part likewise generate force effects, but also induce eddy-current-driving  voltages and ultimately contribute towards losses and stator/rotor heating

•  Minimal cogging and ripple torques.

And the list goes on and on, as any detailed analysis of the individual phenomena will show. Thus there is an optimization problem, the solution for which must be oriented towards the requirements of the application and the expected load cycle.

Distributed winding has proven to be advantageous for a high torque density along with low harmonics and good heat flow from the current-carrying winding into the stator’s laminated core. Since the coils extend over slots at different angles, the coil ends are large compared to concentrated winding.  Round wire distributed win-ding, known from industrial electrical engineering, has proven to be poorly suited for use in automobiles. More and more solutions are shown that use what is referred to as hairpin or I pin technology, which involves inserting copper bars and welding them to the end faces of the laminated core. The requirements for this production technology are high, since there is a high number of weld points for each stator. Compared to coiled wires, the bars have a much larger cross-section, leading to eddy current and skin effect related losses during operation that greatly increase the more frequent changes in polarity become.

While the fact of larger coil ends compared to concentrated winding is considerably alleviated through hairpin technology, it nevertheless continues to persist. For this reason, there are still applications for electric machines where the advantages of distributed winding no longer outweigh the disadvantages, such as applications with extremely short axial installation spaces, e.g. 48 V hybrid modules, in which the active motor length is far less than 50 mm – Figure 3. At the same time, this is a good example of the strong influence of the available production technology. The smaller the coil ends can be made, the further the application limit of distributed winding will shift in the direction of shorter axial lengths.

Figure 3 Application areas of concentrated and distributed winding, depending on the active length of the electric motor

Schaeffler has investigated whether there are alternatives to hairpin winding that utilize the advantages, yet minimize the disadvantages. One good alternative is wave winding, in which the (distributed) winding is produced in a kind of braiding process and then joined in the stator slots . By allowing certain concessions with regard to the copper filling ratio, it is possible to work with smaller cross-sections. The potential number of slots is thereby increased and the effect of the eddy current losses reduced. Figure 4 shows a qualitative comparison of the three win-ding types.

Figure 4 Application areas of concentrated and distributed winding, depending on the active length of the electric motor

The extent to which the advantages and disadvantages come into force depends on the specific application case. In order to quantify the differences, the comparison between the hairpin and wave winding was depicted using a specific example, namely their use in an electric axle of a purely electrical powertrain in the C segment. In coordination with the downstream gear stage, this resulted in the following specification values:

•  Pmax = 147 kW

•  Mmax = 265 Nm

•  nmax = 18,000/min.

The table in Figure 5 shows the performance data of the e-machine with wave winding at selected, application-specific load points, with a outside stator diameter of 220 mm, 96 slots , and an active length of 110 mm. A second motor with the same installation space requirements was produced using hairpin technology, whereby 72 slots were able to be implemented. Subsequently, the stator and rotor losses at the same operating points were compared with each other – Figure 6. This shows that the stator losses in hairpin winding are somewhat less in the lower speed range than in the motor with the wave winding. In contrast, hairpin winding scores much worse in the upper speed range due to the high frequency losses. Moreover, the rotor losses are less at all operating points compared when wave winding was used, an effect that is mainly due to the lower harmonics.

Figure 5 Calculated level of efficiency in an electric motor with wave winding for various operating points

Figure 6 Comparison of the stator and rotor losses for the electric motor in an electric axle application with wave and hairpin winding

In addition, the greater number of slots for wave winding results in a larger overall surface, which is useful for heat dissipation. This becomes evident in the temperatures that occur in the rotor and stator when the losses occurring in the actual design are considered. Figure 7 shows a comparison at a low load and high load operating point.

Figure 7 Comparison of the stator and rotor temperatures at two operating points in an electric motor with wave and hairpin winding

Finally, the efficiency was compared in the cycle (WLTP). For wave winding, there was an average level of efficiency of 94% for the motor, while the value for the motor in hairpin design added up to an average of 89%. One general advantage of the distributed winding is that the stator can be used for permanently-excited synchronous motors, asynchronous motors, or even separately excited synchronous motors, thereby making it suitable as the basis for a modular system.

Power Electronics

Due to the use of power electronics in automobiles, its integration in powertrain components such as electric axles, hybrid modules, and dedicated hybrid transmissions, and its high quantities, it is also subject to specific requirements that must be carefully borne in mind:

•  High level of robustness, since the variance of the user profiles increases at high quantities and there are greater environment-related requirements in highly-integrated powertrains (vibrations, ambient temperature)

•  High power density in narrow installation spaces, resulting in special requirements for lea-ding away power losses

•  Flexibility of the design, as the installation spaces vary in the different applications

•  Optimal current formation to minimize harmonic losses

•  High torque accuracy under all operating conditions

•  Operational reliability in a wide range of operating voltages in order to respond to specific battery configurations

•  Functional safety.

As with electric motors, a detailed analysis would be able to add to this list indefinitely. In addition, power electronics need to be adapted to the application-specific power requirements. To scale power electronics in various applications in light of the specified requirements, Schaeffler has developed a modular concept, which is explained as follows on the basis of a sample implementation in an 85-kW class hybrid module – Figure 8.

Figure 8 Modular concept for power electronics, taking a hybrid module as an example

To begin with, there must be strict separation between the control module and the power module. The control module consists of the control board, the actual intelligence of the power electronics, and an optional actuator power amplifier. As a rule, the control board can be used to depict two engine control channels. In the sample case, the second channel is provided for controlling the K0 actuator in the P2 hybrid module. As an alternative, it can also be used for a gearshift actuator in a two-speed axle, a parking lock, or a second tractive drive in a power split transmission.

In the power module, it is first necessary to consider the scaling of the power semiconductors, which make up a large portion of the value and entail a lot of the effort needed for the qualification. Schaeffler has decided to use half-bridge modules as the smallest unit, since they also enable the construction of multiphase drive systems (with more than three phases). In order to satisfy the range of requirements, two mechanical sizes and one basic population quantity (chip size and chip number) were defined. The interfaces to the outside for contacting and cooling are always the same. Moreover, for the sake of robustness, special importance was placed on avoiding soldering connections and aluminum bond wires in the mounting and connecting techniques used at critical points in the module. In the current design, the IGBTs (Insulated-Gate Bipolar Transistors) and parallel diodes are made using silicon semiconductors. However, the use of wide-bandgap semiconductors is being considered as an alternative, as it is conceivable in the same module at activation frequencies of up to 20 kHz. Frequencies even higher than this require adaptations to be made to the basic design of the power electronics. Figure 9 illustrates the basic scaling approach.

Figure 9 Power scaling through the size and number of IGTB semiconductors and the number of phases in the electric motor

The consistent use of sinter technology increases the cycle stability of the IGBT modules by a factor of 10 in comparison with conventional aluminum bond wire technology. Other basic components involved in carrying current are the capacitor and the bus bars, which also lead to a loss of power due to the flow of current.

The flexibility of the installation space is controlled by the central carrier frame (Figure 8 “Carrier and cooling frame”). This injectionmolded part controls the flow of coolant, connects the current-carrying components to the heat sink, and arranges the basic components with respect to each other. CFD simulation is used to optimize this component in order to ensure symmetrical warming of the half bridges. Moreover, the electrical connections are optimized with regard to parasitic inductances and capacities. In the specific example, a power density of more than 30 kW/l was able to be represented.

Thanks to the flexibility of the installation space and the modular approach, it is easy to react to deviating requirements. Figure 10 shows an example of a design for use in an electric axle with a maximum output of 150 kW. An EMC filter can be integrated on the DC side as an option and can be designed big enough to enable shielding for the DC supply line to be done without. This represents a good option for optimizing the system costs at the vehicle level.

Figure 10 Power electronics setup for an electric axle drive

It was able to be demonstrated that the basic concept selected for the power electronics has the necessary flexibility for meeting the different requirements of P2 and P4 high-voltage drives with regard to the installation space and electric output.

In a further evolution step, it was checked whether the selected approach can also be used for integration in a coaxial P2 hybrid module with a voltage level of 48 V. Used as power switches in this case are MOSFETs (metal-oxide semiconductor field-effect transistors). The components are attached directly to a ceramic substrate (called “bare dies”). The substrate is connected with the coolant-carrying carrier across its entire surface. This makes for very good heat dissipation and thus very high power density, which is important for the relatively high current levels that can occur in powerful 48-volt drives. In the specific application with 15 kW of nominal power (20-second value), the maximum current that occurs is 650 Arms. The capacitors are attached directly via the MOSFETs in order to keep impedance low.

Since it was only possible to have an active length of 45 mm due to the available installation space, the electric motor was also designed with concentrated single-tooth winding. Despite the differences compared to a high-voltage application, key technology platform components were able to be used in this case as well, thereby allowing for a compact hybrid module structure – Figure 11.

Figure 11 Integration of the power electronics in a P2 hybrid module at the 48-volt level

Software

The approach pursued by Schaeffler involves the development of an extensive function-oriented software library, which also includes the specifications for the required hardware. The following are key elements of the library:

•  Analysis of sensor signals, such as for determining the rotor position, the phase currents, or temperatures at defined points

•  Functions for motor control as a function of the motor type used (PSM, ASM)

•  Functions for current control, such as a field-oriented control system that factors in all relevant influencing variables (e.g. field weakening)

•  Superordinate controllers for functional integration in the powertrain, which can also be integrated as customer modules upon special request

•  Monitoring functions, such as for controlling power derating for thermal reasons and for providing functional safety.

In addition to software modules, the library also contains any necessary hardware circuits, along with their definitions and preferred components. Special rules define the implementation in the final layout to guarantee optimum heat dissipation or electromagnetic compatibility. This layout is prepared specific to the application in order to react to special customer requirements. The various communication network options (CAN, FlexRay) have already been mentioned as examples.

With regard to the architecture, the software strictly follows the AUTOSAR paradigms in order to guarantee a high level of reutilization – Figure 12.

Figure 12 Software architecture

System Development

The system development for a specific drive unit based on the technology platform always depends on the boundary parameters required by the overall powertrain and – in part – the vehicle concept as well. This can be illustrated by a wheel-hub drive design in which disturbing acoustic phenomena occurred during operation. A systematic analysis of all of the components, the software, and the transfer paths that was conducted together with research partners KIT, FAST, and ETI revealed the cause: Magnetic field fluctuations generated longitudinal and transverse forces in the electric motor’s stator, which were transferred to the bodywork via the chassis.

The initial countermeasure that presented itself was a change in the electric motor’s design in order to prevent excitations in critical frequency ranges or orders. When this method reached its limitations, a reduced physical model was able to be used to adjust the engine control system. In the remaining critical frequency ranges, the electric motor worked as a damper by making a slight, targeted change in the motor torque. Figure 13 illustrates the reduction in vibration amplitudes achieved in relation to the vehicle velocity.

Figure 13 Vibration reduction in a wheel-hub drive through targeted torque variation in the e-machine

Another example of the high relevance of a system-oriented development approach is torque accuracy, which is part of the specifications for every e-machine. However, each physical component in the torque generation chain is subject to certain tolerances that can cause the output torque to fluctuate. If the voltage in an analog signal generator for the rotor position varies by only 1 %, this can cause the angle during activation to be incorrect, resulting in a torque deviation of 0.5 Nm. Since the complete chain has been modeled – Figure 14 – it is possible to determine the effect of each individual component on the torque accuracy. It is necessary to estimate the anticipated deviation in relation to the operating point, which will make it possible to initiate any targeted countermeasures that are necessary.

Figure 14 Propagation of torque generation inaccuracies in a typical e-machine

VI. Summary and Outlook

Due to a growing variety of electrified powertrains in ever greater quantities, new solutions are needed for the electric drive. Schaeffler has found an answer to the balancing act between variety and standardization with its scalable technology platform for electric drives.

This platform includes both the electric motor as well as the power electronics and the hardware/software for drive control. On the basis of different applications (P2 hybrid modules with high-voltage and low-voltage technology, P4 axle drives), it has been shown that a targeted approach is able to cover a very broad application spectrum.

The technologies used in the platform correspond to the very highest demands for efficiency, power density, and scalability. It has been developed in a modular fashion so that hardware and software from third-party providers can be seamlessly integrated.

Future generations of electric drives will be characterized by increasing integration of electrical, electronic, and mechanical elements. If the opportunities offered by the Schaeffler technology platform are consistently put to use, it will be possible to obtain considerable improvements at the system level. For example, the current development stage of the axle drive mentioned above at the beginning, with a 15-kg reduction in the overall weight, has doubled the maximum torque, which is now available for 60 seconds instead of only 10 – Figure 15 – while the nominal power has more than doubled, going from 60 to 145 kW.

Figure 15 Comparison of electric axle drives from 2011 and 2017

Literature

[1] Englisch, A.; Pfund, Th.: Schaeffler E-Mobility – With Creativity and System Competence in the Field of Endless Opportunities. 11. Schaeffler Kolloquium, Baden-Baden, 2018

[2] Reitz, D.: One Idea, Many Applications – Further Development of the Schaeffler Hybrid Module. 10. Schaeffler Kolloquium, Baden-Baden, 2014

The Schaeffler eDrive Platform

Modular and Highly Integrated

Thomas Pfund

​I. Introduction

​II. Modular Approach

​III. Platform Components

​IV. Summary and Outlook

I. Introduction

The development of electric drives goes way beyond the electric motor. Optimal efficiency, range, and system costs can only be achieved through the interaction of the motor, power electronics, sensors, mechanical integration, and control strategy, which is why a system-oriented approach is needed.

Schaeffler already proved back in 2011 that it possesses system expertise with the “Active e-drive” concept vehicle. The combustion engine in a Škoda Octavia was replaced with two electric axle drives, each with 105 kW of nominal power. Even then, these motors were units that had been developed within the company. While the gear ratio of the motor and axle speed was fixed initially, a third generation rear axle drive with a two-speed transmission and nominal power reduced to 65 kW was employed in the same concept vehicle beginning in 2014. The drive now included new power electronics developed within the company. Both generations of the concept vehicle had torque vectoring in each wheel independently. While the power electronics for the tractive drive was still a separate system at this time, it was already installed on the axle for torque vectoring.

Meanwhile, it is now foreseeable that electrified vehicle quantities will reach large proportions during the current decade and that the degree of mechatronic integration will increase considerably. At the same time, very different topologies are being implemented for the powertrain, which can be classified based on its installation position [1]. Four  classes of aggregates are needed for the design implementation of these powertrains:

•  Hybrid modules for integration in the powertrain of the combustion engine

•  Dedicated hybrid transmissions for specifically realizing hybrid and electric driving modes

•  Electric axle drives for dedicated hybrid powertrains and purely electric vehicles

•  Wheel-hub drives for new mobility concepts.

Their complexity is increased by the fact that electric drives can be operated at various voltage levels, ranging from 48 V for initial hybridization to 400 V in purely battery-driven vehicles. The first manufacturers have already launched projects with an 800-volt on-board electric system in order to achieve short battery charging times at acceptable current levels. Accordingly, the power spectrum of electric drives is very broad – from 20 kW to more than 400 kW.

What all future aggregate concepts have in common is that they will integrate the mechanical and electrical/electronic systems to a large degree – Figure 1. However, the requirements for electric motors are very different. For example, the axial installation space is limited both in hybrid modules as well as in wheel-hub motors. Moreover, the speed is directly coupled to the powertrain, meaning that power scaling mainly needs to happen via the torque of the e-machine. While electric axle drives have greater axial installation space available, the diameter is generally limited due to the installation position. Coupling to the axle is always via at least one transmission stage in order to guarantee the required drive-away torque. This makes it possible to increase the capacity of the electric motor over higher speeds, which in turn makes smaller designs and material sa-vings possible. The same applies for most dedicated hybrid transmissions. In contrast, power electronics is largely independent of the powertrain type; it is mainly defined by the voltage level, the amount of current needed for maximum output, and the installation space specified for the powertrain.

Figure 1 Powertrain concepts with an integrated electric drive

The large number of requirements listed necessitates a modular approach for various installation spaces and power classes. This is even more true since – depending on the design – the drive’s electric and electronic components make up to 80 percent of the total added value and are therefore key elements in the total cost.

II. Modular Approach

Developing a Technology Platform

In order to support a large number of vehicles and drive concepts on the one hand and yet minimize the development expenses on the other, Schaeffler has developed a modular technology approach for electric drives. Its three levels – Figure 2 – not only include electrical components, but also the hardware and software needed for controlling them.

Figure 2 Modular technology platform for electric drives

The electric motors form the base of the platform. Due to the different requirements described at the beginning, multiple series need to be defined, each with a scalable output. According to the current estimate, the complete spectrum of future applications can be covered completely with six series. Depending on the necessary power density and other requirements, both permanently-excited motors as well as asynchronous machines will be used for this.

The middle level, to begin with, includes the power electronics with all key components, such as the power switches, capacitor, bus bars, driver stages, and sensors. Another part of the platform concept is a carrier frame that serves to channel the coolant and as a heat sink for all components of the power electronics. Also pertaining to the second level is the hardware for the drive control, the specifications of which not only depend on the electric motor and power electronics, but also on the functions carried out at the vehicle level, thereby requiring it to be implemented specific to the application. A key example of this is the communication network, which may be executed as a FlexRay, CAN, or CAN FD network.

For this reason, it cannot be separated from the third level where the software platform is located, which has likewise been developed according to a function-oriented approach. The software includes a functional library based on AUTOSAR, which also defines the requirements for the associated hardware.

III. Platform Components

Electric Motors

When regarding highly-integrated powertrains such as hybrid modules, electric axles, or wheel-hub drives, it is difficult to identify the electric motor as an independently functioning unit. The complete motor function often cannot be tested until the powertrain has been assembled, since multifunctional modules are also determining factors for the functioning of other subsystems. Examples of this are bearing arrangements that also provide support for an integrated drive, or rotor carriers that are also plate carriers for a clutch. The upshot of this is the high requirements for the test concept du-ring production. It is necessary to utilize the measurement of parameters such as winding resistance, inductance, or magnetic field distribution of the components to assess the quality of the magnetic circuit assembly (stator and rotor) in order to immediately sort out faulty parts and not fail to identify them until during the end-of-line test.

In addition to the main function, that of representing defined speed-torque behavior, other demands need to be handled by this magnetic circuit as well:

•  Ideal cooling-down capacity and a high copper filling ratio to ensure long-lasting continuous performance

•  Minimal use of materials in order to optimize costs

•  Minimized harmonically occurring radial and tangential force effects in the stator for NVH optimization

•  Minimized harmonics, which for their part likewise generate force effects, but also induce eddy-current-driving  voltages and ultimately contribute towards losses and stator/rotor heating

•  Minimal cogging and ripple torques.

And the list goes on and on, as any detailed analysis of the individual phenomena will show. Thus there is an optimization problem, the solution for which must be oriented towards the requirements of the application and the expected load cycle.

Distributed winding has proven to be advantageous for a high torque density along with low harmonics and good heat flow from the current-carrying winding into the stator’s laminated core. Since the coils extend over slots at different angles, the coil ends are large compared to concentrated winding.  Round wire distributed win-ding, known from industrial electrical engineering, has proven to be poorly suited for use in automobiles. More and more solutions are shown that use what is referred to as hairpin or I pin technology, which involves inserting copper bars and welding them to the end faces of the laminated core. The requirements for this production technology are high, since there is a high number of weld points for each stator. Compared to coiled wires, the bars have a much larger cross-section, leading to eddy current and skin effect related losses during operation that greatly increase the more frequent changes in polarity become.

While the fact of larger coil ends compared to concentrated winding is considerably alleviated through hairpin technology, it nevertheless continues to persist. For this reason, there are still applications for electric machines where the advantages of distributed winding no longer outweigh the disadvantages, such as applications with extremely short axial installation spaces, e.g. 48 V hybrid modules, in which the active motor length is far less than 50 mm – Figure 3. At the same time, this is a good example of the strong influence of the available production technology. The smaller the coil ends can be made, the further the application limit of distributed winding will shift in the direction of shorter axial lengths.

Figure 3 Application areas of concentrated and distributed winding, depending on the active length of the electric motor

Schaeffler has investigated whether there are alternatives to hairpin winding that utilize the advantages, yet minimize the disadvantages. One good alternative is wave winding, in which the (distributed) winding is produced in a kind of braiding process and then joined in the stator slots . By allowing certain concessions with regard to the copper filling ratio, it is possible to work with smaller cross-sections. The potential number of slots is thereby increased and the effect of the eddy current losses reduced. Figure 4 shows a qualitative comparison of the three win-ding types.

Figure 4 Application areas of concentrated and distributed winding, depending on the active length of the electric motor

The extent to which the advantages and disadvantages come into force depends on the specific application case. In order to quantify the differences, the comparison between the hairpin and wave winding was depicted using a specific example, namely their use in an electric axle of a purely electrical powertrain in the C segment. In coordination with the downstream gear stage, this resulted in the following specification values:

•  Pmax = 147 kW

•  Mmax = 265 Nm

•  nmax = 18,000/min.

The table in Figure 5 shows the performance data of the e-machine with wave winding at selected, application-specific load points, with a outside stator diameter of 220 mm, 96 slots , and an active length of 110 mm. A second motor with the same installation space requirements was produced using hairpin technology, whereby 72 slots were able to be implemented. Subsequently, the stator and rotor losses at the same operating points were compared with each other – Figure 6. This shows that the stator losses in hairpin winding are somewhat less in the lower speed range than in the motor with the wave winding. In contrast, hairpin winding scores much worse in the upper speed range due to the high frequency losses. Moreover, the rotor losses are less at all operating points compared when wave winding was used, an effect that is mainly due to the lower harmonics.

Figure 5 Calculated level of efficiency in an electric motor with wave winding for various operating points

Figure 6 Comparison of the stator and rotor losses for the electric motor in an electric axle application with wave and hairpin winding

In addition, the greater number of slots for wave winding results in a larger overall surface, which is useful for heat dissipation. This becomes evident in the temperatures that occur in the rotor and stator when the losses occurring in the actual design are considered. Figure 7 shows a comparison at a low load and high load operating point.

Figure 7 Comparison of the stator and rotor temperatures at two operating points in an electric motor with wave and hairpin winding

Finally, the efficiency was compared in the cycle (WLTP). For wave winding, there was an average level of efficiency of 94% for the motor, while the value for the motor in hairpin design added up to an average of 89%. One general advantage of the distributed winding is that the stator can be used for permanently-excited synchronous motors, asynchronous motors, or even separately excited synchronous motors, thereby making it suitable as the basis for a modular system.

Power Electronics

Due to the use of power electronics in automobiles, its integration in powertrain components such as electric axles, hybrid modules, and dedicated hybrid transmissions, and its high quantities, it is also subject to specific requirements that must be carefully borne in mind:

•  High level of robustness, since the variance of the user profiles increases at high quantities and there are greater environment-related requirements in highly-integrated powertrains (vibrations, ambient temperature)

•  High power density in narrow installation spaces, resulting in special requirements for lea-ding away power losses

•  Flexibility of the design, as the installation spaces vary in the different applications

•  Optimal current formation to minimize harmonic losses

•  High torque accuracy under all operating conditions

•  Operational reliability in a wide range of operating voltages in order to respond to specific battery configurations

•  Functional safety.

As with electric motors, a detailed analysis would be able to add to this list indefinitely. In addition, power electronics need to be adapted to the application-specific power requirements. To scale power electronics in various applications in light of the specified requirements, Schaeffler has developed a modular concept, which is explained as follows on the basis of a sample implementation in an 85-kW class hybrid module – Figure 8.

Figure 8 Modular concept for power electronics, taking a hybrid module as an example

To begin with, there must be strict separation between the control module and the power module. The control module consists of the control board, the actual intelligence of the power electronics, and an optional actuator power amplifier. As a rule, the control board can be used to depict two engine control channels. In the sample case, the second channel is provided for controlling the K0 actuator in the P2 hybrid module. As an alternative, it can also be used for a gearshift actuator in a two-speed axle, a parking lock, or a second tractive drive in a power split transmission.

In the power module, it is first necessary to consider the scaling of the power semiconductors, which make up a large portion of the value and entail a lot of the effort needed for the qualification. Schaeffler has decided to use half-bridge modules as the smallest unit, since they also enable the construction of multiphase drive systems (with more than three phases). In order to satisfy the range of requirements, two mechanical sizes and one basic population quantity (chip size and chip number) were defined. The interfaces to the outside for contacting and cooling are always the same. Moreover, for the sake of robustness, special importance was placed on avoiding soldering connections and aluminum bond wires in the mounting and connecting techniques used at critical points in the module. In the current design, the IGBTs (Insulated-Gate Bipolar Transistors) and parallel diodes are made using silicon semiconductors. However, the use of wide-bandgap semiconductors is being considered as an alternative, as it is conceivable in the same module at activation frequencies of up to 20 kHz. Frequencies even higher than this require adaptations to be made to the basic design of the power electronics. Figure 9 illustrates the basic scaling approach.

Figure 9 Power scaling through the size and number of IGTB semiconductors and the number of phases in the electric motor

The consistent use of sinter technology increases the cycle stability of the IGBT modules by a factor of 10 in comparison with conventional aluminum bond wire technology. Other basic components involved in carrying current are the capacitor and the bus bars, which also lead to a loss of power due to the flow of current.

The flexibility of the installation space is controlled by the central carrier frame (Figure 8 “Carrier and cooling frame”). This injectionmolded part controls the flow of coolant, connects the current-carrying components to the heat sink, and arranges the basic components with respect to each other. CFD simulation is used to optimize this component in order to ensure symmetrical warming of the half bridges. Moreover, the electrical connections are optimized with regard to parasitic inductances and capacities. In the specific example, a power density of more than 30 kW/l was able to be represented.

Thanks to the flexibility of the installation space and the modular approach, it is easy to react to deviating requirements. Figure 10 shows an example of a design for use in an electric axle with a maximum output of 150 kW. An EMC filter can be integrated on the DC side as an option and can be designed big enough to enable shielding for the DC supply line to be done without. This represents a good option for optimizing the system costs at the vehicle level.

Figure 10 Power electronics setup for an electric axle drive

It was able to be demonstrated that the basic concept selected for the power electronics has the necessary flexibility for meeting the different requirements of P2 and P4 high-voltage drives with regard to the installation space and electric output.

In a further evolution step, it was checked whether the selected approach can also be used for integration in a coaxial P2 hybrid module with a voltage level of 48 V. Used as power switches in this case are MOSFETs (metal-oxide semiconductor field-effect transistors). The components are attached directly to a ceramic substrate (called “bare dies”). The substrate is connected with the coolant-carrying carrier across its entire surface. This makes for very good heat dissipation and thus very high power density, which is important for the relatively high current levels that can occur in powerful 48-volt drives. In the specific application with 15 kW of nominal power (20-second value), the maximum current that occurs is 650 Arms. The capacitors are attached directly via the MOSFETs in order to keep impedance low.

Since it was only possible to have an active length of 45 mm due to the available installation space, the electric motor was also designed with concentrated single-tooth winding. Despite the differences compared to a high-voltage application, key technology platform components were able to be used in this case as well, thereby allowing for a compact hybrid module structure – Figure 11.

Figure 11 Integration of the power electronics in a P2 hybrid module at the 48-volt level

Software

The approach pursued by Schaeffler involves the development of an extensive function-oriented software library, which also includes the specifications for the required hardware. The following are key elements of the library:

•  Analysis of sensor signals, such as for determining the rotor position, the phase currents, or temperatures at defined points

•  Functions for motor control as a function of the motor type used (PSM, ASM)

•  Functions for current control, such as a field-oriented control system that factors in all relevant influencing variables (e.g. field weakening)

•  Superordinate controllers for functional integration in the powertrain, which can also be integrated as customer modules upon special request

•  Monitoring functions, such as for controlling power derating for thermal reasons and for providing functional safety.

In addition to software modules, the library also contains any necessary hardware circuits, along with their definitions and preferred components. Special rules define the implementation in the final layout to guarantee optimum heat dissipation or electromagnetic compatibility. This layout is prepared specific to the application in order to react to special customer requirements. The various communication network options (CAN, FlexRay) have already been mentioned as examples.

With regard to the architecture, the software strictly follows the AUTOSAR paradigms in order to guarantee a high level of reutilization – Figure 12.

Figure 12 Software architecture

System Development

The system development for a specific drive unit based on the technology platform always depends on the boundary parameters required by the overall powertrain and – in part – the vehicle concept as well. This can be illustrated by a wheel-hub drive design in which disturbing acoustic phenomena occurred during operation. A systematic analysis of all of the components, the software, and the transfer paths that was conducted together with research partners KIT, FAST, and ETI revealed the cause: Magnetic field fluctuations generated longitudinal and transverse forces in the electric motor’s stator, which were transferred to the bodywork via the chassis.

The initial countermeasure that presented itself was a change in the electric motor’s design in order to prevent excitations in critical frequency ranges or orders. When this method reached its limitations, a reduced physical model was able to be used to adjust the engine control system. In the remaining critical frequency ranges, the electric motor worked as a damper by making a slight, targeted change in the motor torque. Figure 13 illustrates the reduction in vibration amplitudes achieved in relation to the vehicle velocity.

Figure 13 Vibration reduction in a wheel-hub drive through targeted torque variation in the e-machine

Another example of the high relevance of a system-oriented development approach is torque accuracy, which is part of the specifications for every e-machine. However, each physical component in the torque generation chain is subject to certain tolerances that can cause the output torque to fluctuate. If the voltage in an analog signal generator for the rotor position varies by only 1 %, this can cause the angle during activation to be incorrect, resulting in a torque deviation of 0.5 Nm. Since the complete chain has been modeled – Figure 14 – it is possible to determine the effect of each individual component on the torque accuracy. It is necessary to estimate the anticipated deviation in relation to the operating point, which will make it possible to initiate any targeted countermeasures that are necessary.

Figure 14 Propagation of torque generation inaccuracies in a typical e-machine

VI. Summary and Outlook

Due to a growing variety of electrified powertrains in ever greater quantities, new solutions are needed for the electric drive. Schaeffler has found an answer to the balancing act between variety and standardization with its scalable technology platform for electric drives.

This platform includes both the electric motor as well as the power electronics and the hardware/software for drive control. On the basis of different applications (P2 hybrid modules with high-voltage and low-voltage technology, P4 axle drives), it has been shown that a targeted approach is able to cover a very broad application spectrum.

The technologies used in the platform correspond to the very highest demands for efficiency, power density, and scalability. It has been developed in a modular fashion so that hardware and software from third-party providers can be seamlessly integrated.

Future generations of electric drives will be characterized by increasing integration of electrical, electronic, and mechanical elements. If the opportunities offered by the Schaeffler technology platform are consistently put to use, it will be possible to obtain considerable improvements at the system level. For example, the current development stage of the axle drive mentioned above at the beginning, with a 15-kg reduction in the overall weight, has doubled the maximum torque, which is now available for 60 seconds instead of only 10 – Figure 15 – while the nominal power has more than doubled, going from 60 to 145 kW.

Figure 15 Comparison of electric axle drives from 2011 and 2017

Literature

[1] Englisch, A.; Pfund, Th.: Schaeffler E-Mobility – With Creativity and System Competence in the Field of Endless Opportunities. 11. Schaeffler Kolloquium, Baden-Baden, 2018

[2] Reitz, D.: One Idea, Many Applications – Further Development of the Schaeffler Hybrid Module. 10. Schaeffler Kolloquium, Baden-Baden, 2014