E-Silence

“Silent design of electrical machines; uncertainties and optimization.”

Context & Issues

In all industrial sectors, electrical machines emit electromagnetic noise with high-pitched tonal components that are highly unpleasant in terms of sound perception.

  • In the rail sector, traction motors are the source of exterior and interior noise.
  • In the automotive sector, while traction motors are a plus in terms of exterior noise, this is not the case for interior noise, where they become a nuisance. In a context of increasing power-to-weight ratio, this problem will become even more acute in the years to come.
  • Accessories for the hybridization of internal combustion engines and electric actuators present in a vehicle are also concerned, the acceptance criterion being that they should not be heard.

In addition to high noise levels, a wide dispersion of noise levels – in excess of 5 dB – has been observed on samples leaving production. This dispersion is detrimental to the product’s high-quality image.

Challenges

01

Take acoustics into account right from the start of electrical machine design by developing methods and tools that enable electrical engineers to quickly estimate the impact of electromagnetic design on vibrations and radiated noise at the pre-sizing stage.

02

Develop methods for optimizing electrical machine topology (teeth, rotor, magnets, etc.) with the aim of reducing electromagnetic excitations “at source” while preserving machine performance and taking manufacturing tolerances into account.

03

Develop techniques for reducing electromagnetic and dynamic structure models, in order to reduce calculation times and enable rapid optimization loops and factoring uncertainties into simulations

04

Demonstrate the benefits of adding an “acoustic” function to resins, by optimizing a resin on the machine.

Key development points

01

Take acoustics into account right from the start of the machine design cycle to quickly estimate the impact of electromagnetic design on radiated noise.

02

Develop multi-harmonic optimization methods for the shapes of the machine’s active parts, with the aim of reducing electromagnetic excitations “at source”.

03

Understand the physical phenomena responsible for noise dispersion by means of tests on a batch of machines sampled on leaving the factory.

04

Integrate methods for taking uncertainties into account in simulations

05

Develop techniques for reducing structure electromagnetic and dynamic models, in order to reduce calculation times

Results & Presentation

Implementation and validation of calculation and optimization methods for the design of electrical machines

The project’s work has demonstrated the benefits of coupling electromagnetic and mechanical model reduction techniques to reduce calculation times. In the case of the high-speed motor, calculation time was reduced by >60%. These methods offer significant potential for optimization studies, or for studies of dispersion due to manufacturing uncertainties, all of which require numerous computational iterations.

Arobust multi-criteria optimization technique has been introduced: this method makes it possible to reduce machine noise for several operating points and/or frequencies, while maintaining machine performance (torque, efficiency).

A calculation method has been developed to model and optimize the mechanical properties of resins used in the manufacture of certain electric motors for acoustic purposes.

Finally, some of these methods have been integrated into two industrial software packages (Flux and FluxMotor), enabling designers to analyze the vibro-acoustic performance of machines while reducing the need for prototypes.

Bringing these methods to market represents a very promising step forward for the industry, with very wide-ranging applications:

  • They can be applied at all stages of the design process: from pre-design studies where different design choices need to be quickly discriminated, to detailed design studies where geometries are finely optimized.
  • They are potentially applicable to all types of electric machines (synchronous or asynchronous machines, variable reluctance machines, axial flux machines, etc.), for all industrial sectors.

Demonstrating the feasibility of technical solutions to reduce the noise radiated by electric motors

As part of a partnership with Von Roll, the benefits of optimizing an impregnating resin according to acoustic criteria were demonstrated through experimentation, with very significant gains (in the case of the high-speed machine from 30dB to 40dB), for the entire temperature operating range defined in the specifications.

We were also able to demonstrate the benefits of using stators with linked teeth to increase stator stiffness (gains of the order of 10dB to 20dB), thereby confirming our design choices.

Finally, an approach has been developed to correlate the geometric tolerances of motor stator construction (by X-ray tomography characterization) with radiated noise levels.

Economic impact

The e-Silence project has also enabled all the industrial partners to maintain and create permanent jobs beyond the duration of the project, through increased sales.

  • Between 25% and 30% of MMT sales in 2021 and 2022 came from work related to the design of quieter engines, thus exploiting the work of the e-Silence project.
  • The project will enable Vibratec to maintain its competitive edge. During the life of the project, sales around this theme increased by 50% (representing 4 FTEs). An equivalent increase is expected over the next three years (increase in the number of electrification projects in the automotive and aeronautics industries).
  • Altair has been able to raise the competitive level of its software in relation to its competitors, and thus has major assets to gain market shares: the new functionalities introduced in Flux and FluxMotor are already being used by several customers, both in France and internationally. The number of customers (mainly in the automotive sector) using Flux and FluxMotor has increased by 20% and 50% respectively, between 2020 and 2022. Annual growth of between 20% and 30% is expected over the next five years, given the prospects for growth in the electrification of means of transport.

Key Figures

1.9 M€

funding

42

months of work

3

peer-reviewed scientific publications

9

conferences (including 6 outside France)

9

webinars organisés par Vibratec

alone (5) or with the SPIN company (IT)

2

Altair webinars

1

participation in the CARA Automotive Techdays 2021 virtual show

1

E-silence workshop in December 2022

This project was selected under the FUI-AAP23 call for projects of the Fond Unique Interministériel (FUI). It is subsidized by the Banque Publique d’Investissement (BPI France) and the AURA region via the European Regional Development Fund (ERDF – this project is co-financed by the European Union).