High speed – The enabler for cost reduced and power dense sustainable electric drive units

Mathias Deiml, AVL Software and Functions GmbH
Katharina Berberich, AVL Software and Functions GmbH
Wilhelm Vallant, AVL List GmbH

The e-mobility market is facing the challenge of an increasing pressure to reduce unit costs and increase power density and performance at the same time. AVL’s second-generation high-speed e-axle increases power density to reduce material usage, focusing on the e-motor and gearbox, with CO2-equivalent footprint as a key comparison criterion.

With the updated design in Generation 2 it achieves a power density of over 4.3 kW/kg, features a 30,000rpm motor, silicon carbide double inverter, and loss optimized gearbox. It minimizes magnet and copper use, reducing unit costs, CO2 emissions and improving efficiency.

Speed and cost – Why 30.000 RPM

Electric motors, such as PMSM based machines widely used for automotive applications, consist of high masses of expensive materials: dynamo sheet metal, magnets, copper. Reducing the size of the motor reduces these costs. Downside of this characteristic is that scaling down motor size means to reduce its torque.

Pmech = ω * M; M ~ d 2 * l;
d: diameter; l: lamination length

To maintain the power rating the rotational speed needs to be increased when reducing motor dimensions. An increase of factor 2 in motor speed results in a reduction of active motor material by a factor of 2. Provided that special or expensive technologies in motor and transmission can be avoided there is a substantial cost saving in the motor. Weight and size of the motor can be reduced accordingly and are additionally beneficial for vehicle features and again costs.

Fig. 1: Speed vs Cost

These relations between speed and weight/volume/cost are shown in Fig. 1. AVL’s engineers were looking
for the optimal speed, at which weight and cost are both low [1] and found it at 30.000 RPM.

High speed e-machine development

To develop this e-motor for the required 30.000 rpm without using expensive technology like cobalt steel lamination, the following challenges had to be overcome:

• Mechanical rotor strength due to high centrifugal forces
• Bearing technology at high rotational speed in partnership with SKF
• High frequency losses in copper windings
• Increased iron losses due to higher fundamental frequency of 1500Hz

The developed Solutions included new innovative designs for

• Hairpin Windings
• Cooling concept

Selecting hairpin winding technology helps to deal with the high frequency losses: A six-layer hair pin winding with bended copper on one side and copper forming and welding on the other side was chosen. A copper fill factor of 60% was achieved. This factor includes the oil channels for cooling, described in the cooling chapter. A slot-liner with only 0.1 mm thickness was used. The slot was optimized for coolant flow in combination with copper fill factor.

There is no need for resin in the slot. Thus, additional micro-oil channels along the copper improve coolant flow. Recycling of copper is facilitated; the winding can be removed in one piece without shredding. [2]

The prototypes were manufactured using SLM selective laser melting, a 3D copper printing process. This has further advantages:

• Smaller winding heads possible
• Low cost for tooling
• Smaller resistance compared to welding

Copper printing, may be only the beginning, further optimizations like free shaping of super short winding heads can be realized in later steps. It is suitable for small series production. For large scale volumes welding shall be used. AVL’s current design enables both manufacturing processes. [2]

The sophisticated direct cooling concept shown in more detail:
Direct oil cooling means that the coolant is directly in contact with the copper winding in the slots and winding heads. The stator is fully filled with oil, which is pumped along the stator slots. Thus, the copper losses are very efficiently cooled. An air gap tube keeps the oil away from the rotor space, so no friction or splashing losses are present. [3]

The transmission uses a dedicated lubrication fluid with an own oil pump and filter. In addition – compared to Gen 1 – there is now a transmission fluid cooler integrated. This device can be installed optionally, for high performance application, where it is needed. [2]

The magnets are cost-effective quality with low heavy rare earth content.

Achieved eMotor performance

The efficiency map shows a maximum of over 97%, in a wide range. Due to NO20, segmented magnets and high copper fill factor the losses at high speeds are well balanced. This is shown in Fig. 2.

The map is generated for 180°C copper temperature. Thanks to direct oil cooling, the copper temperature is typically 90°C, resulting in ever better efficiency.

Fig. 2: High Speed E- Machine Performance and Efficiency at 180° C Machine Temperature

Overall EDU system implementation

The EDU design with the new e-Motor continued to focus on overall cost reduction, reduction of material usage and increased power density without losing efficiency. Scalability and high-volume rollout were also key considerations. With these requirements the resulting concept is described:

1. EDU:
a. dual motor, dual torque path to enable torque vectoring for premium segment
b. central gearbox for optimal length of drive shafts in
various vehicles
c. low profile design to support flat chassis

2. High speed motor:
a. standard materials for sheet metal and magnets
b. PSM technology for best power density and system efficiency
c. high frequency winding for low factor AC_loss / DC_loss
d. cooling with best-in-class effectiveness, combined with the inverter
e. mechanical rotor stability up to 1,2 * n max

3. SIC inverter:
a. high power density in complete package
b. best efficiency, fast switching slopes, variable switching frequency
c. combined DC link, interleaved switching, low EMC emissions

4. Gearbox:
a. 2 stage transmission, single speed
b. lay shaft concept
c. high efficient tooth geometry, optimal NVH tradeoff
d. injection lubrication, separate fluid than motor

CO2e footprint and comparison to other systems for AVLs e-axle technology

Besides the already discussed technological and cost targets, a future proven EDU concept must also comply to a sustainable design to reduce CO 2 e footprint and increase efficiency in the life cycle.

„Design-to-CO 2 e“ as a holistic approach in the course of the systems engineering meant a transition from an original focus on „design to function“ and „design to cost“, to the additional dimension of CO 2 equivalent over the life cycle. [4] The total life cycle includes Production, Use and End of Life with. recycling or 2 nd life.

Cost, weight & CO 2 e comparison

To compare our EDU system with different solutions, three systems with a peak power rating of 160 kW were evaluated: a baseline PMSM system, AVLs new high-speed PMSM system, and an externally excited synchronous motor (EESM) system.

For the complete EDU system, the evaluation included weight, cost, and CO 2 e of active motor parts, power inverter, and transmission, considering additional components like rotor shaft, bearings, and motor housing. Rotor excitation for the EESM system is considering cabling, sensors, and rotor interface as part of the power inverter.

Collector ring and copper bars/wires to the rotor windings are already considered with the E-Motor active parts. A potential for saving in terms of weight and cost was identified in the magnitude of 17 kg and 8% respectively with the use of high-speed motor, where EESM shows the potential of 5% cost saving but a disadvantage of up to 13kg in terms of system weight (see Fig. 3).

Fig. 3: Weight, cost & C0 2 e evaluation – EDU System

Evaluating the sustainability of EDU systems by comparing calculated CO 2 e emissions shows an advantage
of 10% with the High Speed PMSM System and a potential increase in terms of CO 2 e emissions of up to 6%
with the EESM system. [3]

Conclusion and outlook

AVL’s high speed EDU is in the second generation with an efficiency and performance above state of the art and improved sustainability due to high power density of 4.3kW/kg.

The described technology is tested continuously, e.g. in 2 AVL demo cars, which are running since 2020. Furthermore, the AVL torque vectoring function has been developed for and demonstrated in the vehicle.

Planning of our Gen 3 includes further improvement of inverter power density by 50%, a single motor derivate EDU for compact cars with only 25kg at 120kW/3400Nm, and intensive use of AI based machine control, such as rotor temperature determination.

References
[1] M. Deiml, A. Simonin, P. Jafarian: High Speed Electric Drive by AVL, CTI Berlin, 2018
[2] M. Deiml, M. Schneck: Increasing the Sustainability of Electric Drives by High Speed Motor Technology, CTI Berlin, 2022
[3] A. Angermaier; M. Deiml; W. Vallant; G. Fuckar:, Electric drive units with high power density and sustainability through high speed and maximum efficiency, Wiener Motoren Symposium, 2024
[4] Sams, C.; von Falck G.; Sorger H.: Cost Engineering as an Essential Part of Systems Engineering. In Systems Engineering for Automotive Powertrain Development. Schweiz: Springer, 2021

We should make a clear commitment to electromobility

Dr Norbert Alt, COO of FEV

There is uncertainty about which drive solutions will be best in the near and longer term. In our interview, Dr Norbert Alt, COO of FEV, an innovation driver in the mobility sector, advocates a technology-open approach but with a clear commitment to electric vehicles. Hybrid BEVs are included, he says – but e-fuels will not be realistic for passenger cars in a market-relevant quantity by 2035.

Dr Alt, what do you think about the current reluctance to buy battery-electric cars?

Looking at BEV sales figures in China, the USA, and the EU, you see continuous growth and an all-time high in absolute terms. If you count BEVs, PHEVs, and REEVs together, the overall market share has grown from 22 in the first half of 2023 to 25 percent in the first half of 2024. In China, for example, New Energy Vehicles – the NEVs – have a market share of 45 percent. In Germany, we did see a decline of 16 percent for BEVs and an increase of 13 percent for PHEVs in the first half of the year. In contrast, 94 percent of new cars registered in August were electric in Norway. These are just examples. So, we should look more at international developments, not so much at ourselves in Europe. When I talk to industry stakeholders, everyone agrees that in the long term, we‘ll mainly drive electric, meaning BEVs with some PHEVs or REEVs and continuously decreasing shares of ICE in passenger cars.

How do you define REEV as being distinct from PHEV?

A while ago, we had REX, which was a Range Extender with a low-power combustion engine. Today’s REEVs – range-extended electric vehicles – have a powerful combustion engine that keeps the electric drive motor operating at full power even when the battery is low. At FEV, we actually call them Hybrid BEVs. So a Hybrid BEV / REEV should be‚ battery-electric born’, meaning it’s an electric car that we then hybridize. You can only have one vehicle platform, a BEV platform because that will be the mainstream. Hybrid BEV / REEV gives you the driving dynamics of a BEV plus additional range from the combustion engine. In China, Hybrid BEV / REEVs have a significant market share. In the spirit of technological openness, we in Europe could learn something from this. Hybrid BEVs would also help people make the transition to pure electric cars.

Interest in REEVs seems to be growing in North America, too …

Under the ACC-II Regulation of the California Air Resources Board, 35 percent of all vehicles must be ZEVs, or zero-emission vehicles, from 2026 – and 100 percent by 2035. Almost one-third of the US states are adopting this legislation. Here‘s the surprising part: Of those ZEVs, 20 percent can be hybrids, provided they have a minimum electric range of 70 miles. So basically, this opens a little window where Europe can say to the authorities: ‚Hey, look at the rest of the world; these concepts are successful in China, and they’re enshrined in law in the USA’. As engineering service providers, we always have our ears to the ground so we know what’s coming down the line. And right now, the Hybrid BEV or REEV topic is in a strongly increasing development focus. We believe that taking a similar approach as China or the USA would make more sense than unproductive debates about banning ICEs.

And what if e-fuels were to become established?

The future will be significantly electric. In Europe, we discuss using e-fuels after 2035, which, in theory, are great. But with an existing fleet of a good 1.3 billion cars worldwide, we’d need them very quickly. In the years leading up to 2035, they cannot help. It takes eight to ten years to approve and build the production
facilities. Even if we built 100 percent of the plants being discussed worldwide, the total amount of e-fuel would cover just 10 percent of German requirements – in areas outside individual mobility! Even the aviation industry is concerned about how to hit its targets. By 2032, the goal is to replace 1,2 percent of aviation fuel with sustainable e-fuels known as RFNBOs, or renewable fuels of non-biological origin’. So, we don’t see a realistic way of getting meaningful quantities of e-fuels on the road in 2035. One option not yet on the table would be similar as in US and China allowing Hybrid BEVs that burn fossil fuels and can drive 150 km electrically for example, to stay on the road longer than planned. If we were to regulate for 70 percent BEV and 30 percent Hybrid BEV, for instance, we would still be driving around 90 percent electric overall.

Moving to vehicle technology – how would you design a Hybrid BEV drive architecture?

The question is, do you want just serial, or do you want series-parallel? Obviously, engineers want the latter because you have high efficiency on the highway, too. But if one has a battery range of 150 or 200 km, maybe one can live with a purely serial drive that consumes a little more fuel at 140 or 150 km/h, which you can only do in Germany anyway. And in the city, the serial hybrid drive is 5 to 7 percent better than a conventional parallel hybrid. Also, we’re continuing our efforts to make combustion engines more efficient. We already have engines with 43 percent efficiency in production, and we’ll soon be heading for 45 percent and more.

How will electric drives and charging technology develop in the future?

There’s a clear trend towards 800V. But market penetration will be top-down for cost reasons, with 400V systems widespread in the lower sectors. In the truck sector, we’re talking about megawatt charging with even higher voltages. Another exciting development path involves optimizing the charging curve for 400V systems. The aim is to minimize losses even when fast charging to 80 percent. We are currently developing 400V systems, which you can charge from 20 to 80 percent in 19 minutes, and which are on a similar level as an 800V system. Several factors make this possible: a special cell chemistry that is well understood, an accurate battery management system, and improved heat dissipation. So, it’s not just about high charging power – you also need the smoothest possible charging curve. That’s something that offers excellent customer benefits.

What about trends in electric motors?

There is a somewhat unexpected trend here. In the past few years, motor speeds have increased – up to 30,000 rpm in the Far East. Higher speeds mean smaller motors, which is good, but also more friction, which is not good for efficiency. Some German manufacturers have now reduced EDU rpm to 13,000 and are seeing excellent consumption at 120 or 130 km/h. You only need higher rpm for sporty applications or when the package is a high priority. Another trend – one we are working on with a German company – is dual-rotor motors, which enable high efficiency at part load, high power and torque density and lower costs. And a topic the whole industry is working on is externally excited electric motors. These require no heavy rare earths and can reduce our dependencies on raw materials.

Which way is battery chemistry heading?

Nickel-manganese-cobalt (NMC) and lithium-iron-phosphate – LFP – batteries are used mainly in passenger cars, especially LFP in the cost-sensitive segment. We’re also developing sodium-ion batteries. Sodium Chloride is available in large quantities in nature, so that’s more favorable concerning raw materials. Also, sodium-ion is already not far behind LFP regarding energy density. And since Hybrid BEVs batteries require less space in a standard BEV platform, they can also cut production costs for Hybrid BEVs. Then, in the high-performance sector, we’re talking about solid state, but that will take a while. Semi-solid-state batteries will come sooner. We’re currently working with manufacturers on joint developments here.

What drive technologies do you foresee in trucks in the next few years?

Surprising as it may seem, there is also a trend towards full electrification in Europe. At the last IAA, almost all major OEMs shared that view. The trend also applies to long-haul trucks because this is an industry where TCO rules. We are talking about 30, possibly 40 percent BEV share in 2035. Hydrogen will also play a role. FEV is working with manufacturers to develop both fuel cells and hydrogen-powered combustion engines. But the majority will be battery-electric trucks. Some OEMs have allied to install 1,700 charging stations along main routes in Europe by 2027. The electric trucks and infrastructure operators are there, but we need support and commitment at the political level.

Some people say China is well ahead of Europe in BEV technology. What’s your view?

The BEV vehicles that German and European manufacturers offer today are on a high level of technology. They have good efficiency, high charging power, system design, drag coefficient, and more. On the other hand, the Chinese are ahead in battery cell technology and production know-how; they have invested strongly in the development and have better access to raw materials. But in the future, more and more processes will be sustainable, through to a circular economy. This will reduce the need to add further raw materials. That will put the issue of „who has the raw materials?“ back into perspective. What’s more, German manufacturers are spending billions on setting up battery cell production. At FEV, we have our own cell chemistry department, too – and as a development service provider, we work with the Chinese. Another topic is automation, which means ADAS systems, autonomous driving, etc. Here, some other countries are nowhere near as advanced as German manufacturers. So, from a technology perspective, the outlook is a lot brighter than some people say, and we will see continuous innovations in all areas. Additionally, one has to mention that low-cost BEVs are very important for the expected market penetration of BEVs, and the capability of the Chinese to produce on low cost is very high which is the major challenge for the European.

How could decision-makers in politics and the industry advance electric mobility?

We would like the EU to look at the regulatory approaches of the USA and China, and to use them as a model to some extent. At the moment, we are letting car owners hope they can have e-fuels in 2035. That’s not going to happen, only for a very small share of the market. We should also make a clear commitment as a tech community, saying we will drive more and more BEVs, some of them Hybrid BEVs. We should allow a defined percentage of Hybrid BEVs and stop the unhelpful discussion about the ICE ban. By 2045, we may reach the point where filling stations no longer stock fossil fuels, just bio and e-fuels. Another critical point we really need to tackle is electricity prices. You can generate electricity with a photovoltaic system for 10 cents per kWh, or even less. If people could charge their vehicle for 10 cents, electric cars would sell like hot cakes. Hybrid BEVs would benefit, too, because people would prefer electric driving whenever possible.

Interview: Gernot Goppelt

e-NVH Design of Electrical Drives with Collaborative, Multidisciplinary Modeling & Simulation

Tackling noise issues after manufacturing can be expensive and may degrade electric drive performance for efficiency, cooling, and weight. Therefore, an efficient virtual prototyping workflow of e-NVH is critical for accelerating development times, reducing prototyping, manufacturing & testing costs, and improving engineering productivity.

The prediction of acoustic noise and vibration in electric drives requires modeling and simulation of all excitation mechanisms, including mechanical forces (e.g. tire/road, gear friction forces), aerodynamic forces (e.g. windshield, fans), and electromagnetic forces. These three sources of acoustic noise must be considered together due to masking effects. For example, the unpleasant tonalities of magnetic noise produced by inverter switching harmonics might be covered up by windshield noise at high speed.

In electrical drives, magnetic excitations come from the converter and the associated traction motor. The frequencies of magnetic forces depend on slot/pole/phase combination, while their magnitude depends on magnetic circuit geometry and control. This makes electromagnetic noise, vibration, and harshness (e-NVH) a unique discipline that requires dedicated troubleshooting and mitigation tools.

Manatee software, now a part of the SIMULIA brand of Dassault Systèmes, is a specialized computer-aided engineering (CAE) software for the assessment and control of magnetic noise at all design stages of electrical machines and drives. It helps designers identify the best tradeoff between electromagnetic & NVH performance. Manatee expands the SIMULIA multiphysics simulation portfolio, which includes Abaqus, CST Studio Suite, Opera, Simpack, and Wave6. Leading vehicle and powertrain developers use these robust and proven simulation technologies for the virtual prototyping of vehicle components, subsystems, and systems, including electric drives.

e-NVH simulation requires electrical, magnetic, structural dynamics, and acoustic models. The traditional method for evaluating e-NVH has been for a CAE expert to use general-purpose FEA software and manually combine all the physics. This process requires setting-up each solver and their interfaces, such as mesh-to-mesh projections, time vs. frequency domain and discretization. It also requires scripting to call this customized model combination at variable speed and to obtain relevant output quantities such as A-weighting for human ear sensitivity. This complex workflow, which is generally maintained and run by a single CAE expert, results in a bottleneck in the numerical simulation process, preventing the use of NVH metrics in design iterations.

To overcome this issue, Manatee provides a unique algorithm to accelerate variable speed e-NVH calculations without loss of accuracy. It contains its own multiphysics models with predefined couplings, but can also be easily interfaced with other CAE software.

In addition, Manatee comes with predefined workflows adapted to each engineering role. The multiphysics simulation can be set up with the click of a button thanks to standardized interfaces that enable collaboration between electromagnetic and mechanical departments. Electrical engineers can set up the electrical machine and run electromagnetic calculations at variable operating points; the result is a Magnetic Look-

Up Table (MLUT). In parallel, mechanical engineers can import the electric powertrain modal basis. The MLUT can be provided to NVH engineers and the modal basis to electrical engineers. This way, noise mitigation techniques based on electromagnetic and mechanical design can be investigated in parallel. NVH engineers can track if NVH targets are fulfilled at all stages of new product development.

An efficient collaboration of control, electrical, mechanical & NVH engineers also requires a user-friendly interface, where all physical quantities can be easily visualized and post-processed. Manatee comes with insightful plots to help quickly identify which magnetic force excites which structural mode. New specialized e-NVH solutions are regularly developed, supporting all engineers finding solutions to reduce noise & vibration levels (skewing, notching, Harmonic Current Injection, etc).

Manatee also provides parameter sweeps, global optimization with predefined e-NVH metrics and design exploration tools to find the best tradeoffs between electromagnetic & NVH performances.

The robust and user-friendly features of Manatee provide a unique, collaborative CAE environment specialized in the assessment and control of electromagnetic noise & vibrations. By using Manatee during the design stages, designers and engineers can significantly accelerate the virtual prototyping of electric drives, resulting in shorter development cycles, reduced prototyping costs, and better NVH risk management.

For more information visit: https://www.3ds.com/products/simulia/manatee