Noise prediction from Electric

Motors
Vehicle NVH & Acoustic Innovation Area
Electric and hybrid electric vehicles – Challenges

200+ (H)EV New NVH Balancing

models issues performances
Increased engineering High-frequency noise in EV Conflicting performance
complexity and transient noise in HEV characteristics (NVH, fuel
220
impact the driver comfort economy, drivability)
170

120

70

20

-30

Unrestricted © Siemens AG 2016

Page 2 16.05.2016 Siemens PLM Software
Vehicle NVH & Acoustic Innovation Area
Electric and hybrid electric vehicles – Applications

Analysis and Reduce Warning

optimization e-motor noise systems
Analyzing the relationship Simulation-based approach Simulating emitted sound
between the electrical and combining electromagnetics waves using acoustic
mechanical design and acoustic radiation numerical approaches

Unrestricted © Siemens AG 2016

Page 3 16.05.2016 Siemens PLM Software
So what is the difference?

Motor order
Resonance
Motor order
Resonance
3500.00 80.00
Tacho1 (T1)

dB(A)
rpm

Pa
AutoPower mic (A) WF 126 [981.36-3466.8 rpm]

900.00 -20.00
0.00 Hz 2600.00
mic (CH1)

Off-zero harmonics. Origin?

ICE driven Electric driven

(more high frequency content)

A electric driveline is not silent!

Unrestricted © Siemens AG 2016
Page 4 16.05.2016 Siemens PLM Software
Electric Motor Noise – Siemens activities
Global perspective

Design space 1-D Simulation 3-D Simulation Test

Full vehicle diagnostics
Motor Design Functional Motor Model

Magnetic simulation

Electric Design
Acoustic characterization
10000.00 20.00

PSD c urrent1 WF 381 [425.76-9947.7 rpm]

Thermal Model

Tacho_SRM_Filt2 (T3)

dB/Hz
rpm

A2
Phase current 400.00
0.00 Hz 10000.00
-90.00

c urrent1 (CH105)
10000.00 20.00

PSD Mntl:1_06:+Z WF 381 [425.76-9947.7 rpm]

Tacho_SRM_Filt2 (T3)

dB/Hz
rpm

g2
Acceleration
400.00 -80.00
0.00 Hz 10000.00
Mntl:1_06:+Z (CH3)

10000.00 100.00

PSD M1 WF 381 [425.76-9947.7 rpm]

Controls Design

Tacho_SRM_Filt2 (T3)

dB/Hz
rpm

Pa2
Vehicle Energy Model Vibro-Acoustic Model
Sound
400.00 0.00
0.00 Hz 10000.00
M1 (CH233)

Electric system models including From sub-system characterization

Vibro-acoustic models based on
controls and extendable to thermal to full vehicle diagnostics using
electro-magnetic loads calculation
and energy management TPA, ODS, OMA,…
Unrestricted © Siemens AG 2016
Page 5 16.05.2016 Siemens PLM Software
 Fundamentals of NVH for e-drives: A magnetic transmission

Permanent magnets in the rotor guarantee a

magnetic lock (for Synchronous operation)
between the rotating field and the rotor itself Coils in the machine are supplied to create a
rotating magnetic field in the stator

Energy is exchanged in the

air-gap: looking at this region
can provide meaningful insight

Unrestricted © Siemens AG 2016

Page 6 16.05.2016 Siemens PLM Software
 Fundamentals of NVH for e-drives: PWM supply distorts the current signal
even further

PWM is a common way to

provide power to the machine
by discretely injecting energy
to the system from a DC
source (to create an AC)

This small variations on the current will

also affect the magnetic flux density in
the air-gap and consequently on the
reaction forces on the structure

Unrestricted © Siemens AG 2016

Page 7 16.05.2016 Siemens PLM Software
Illustration: Noise optimization of an SR electric motor

Electric motors are key to all electric and hybrid-electric automotive propulsion
solutions.
Main advantage of Switched Reluctance (SR) type electric motors
• Cost-effective technology, simple construction
• No rare-earth materials (permanent magnets)
• No electric or magnetic components in rotor
Main drawback in automotive applications
• Control for maximum efficiency leads to high torque ripple
• High torque ripple creates NVH problems.

Unrestricted © Siemens AG 2016

Page 8 16.05.2016 Siemens PLM Software
Noise optimization of an Switched Reluctance motor
Global flowchart

Electro-magnetic
Current waveforms Stator Forces
FEM

Electro-Magnetic
200
Rpm = 1000
180 Rpm = 2000
Rpm = 3000
160 Rpm = 4000
Rpm = 5000
140 Rpm = 6000
1 120
Rpm = 7000
Rpm = 8000

Current (A)
100

80

60

40

20

0
10 20 30 40 50 60 70 80 90 100
Rotor angle (deg)
Vibro-Acoustic
2

Structural FEM Stator Deformation Radiated Noise

Unrestricted © Siemens AG 2016

Page 9 16.05.2016 Siemens PLM Software
Noise optimization of an Switched Reluctance motor
Integrating current control in the process

Flowchart of the complete calculation process for noise radiation of an SRM

Unrestricted © Siemens AG 2016

Page 10 16.05.2016 Siemens PLM Software
Noise optimization of an Switched Reluctance motor
SRM Motor FE Model for Magnetic Data

Magnetic data is extracted from 2D FEM simulation

• Geometric characteristic, material properties and rated current needed
• Rotor position is changed in fixed steps
• For each range of position, a fixed current is used, and on each iteration the current is incremented
• Torque and Flux data are collected for each step  Lookup tables

Unrestricted © Siemens AG 2016

Page 11 16.05.2016 Siemens PLM Software
 Noise optimization of an Switched Reluctance motor
SRM Motor Functional Model

Nonlinear behavior of SRM: Look-up tables:

• Flux/current/position
• Current/torque/position

Flux[k-1]

Look-up
Table 1 I[k-1]
Voltage I(Flux, Angle) Torque
Look-up
Table 2 Angle[k-1]
Te(I, Angle)

SRM
Unrestricted © Siemens AG 2016
Page 12 16.05.2016 Siemens PLM Software
Noise optimization of an Switched Reluctance motor
SRM Motor Functional Model in LMS Amesim

 The SRM model is implemented in LMS Amesim using look-up tables

 A customized SRM block is created using the LMS Ameset tool

Stator poles: Rotor shafts Thermal port

Input: Voltage Output: Torque Input: Temperature
Output: Current Input: Velocity Output: Heat flow

Unrestricted © Siemens AG 2016

Page 13 16.05.2016 Siemens PLM Software
Noise optimization of an Switched Reluctance motor
Global flowchart

Electro-magnetic
Current waveforms Stator Forces
FEM

Electro-Magnetic
200
Rpm = 1000
180 Rpm = 2000
Rpm = 3000
160 Rpm = 4000
Rpm = 5000
140 Rpm = 6000
1 120
Rpm = 7000
Rpm = 8000

Current (A)
100

80

60

40

20

0
10 20 30 40 50 60 70 80 90 100
Rotor angle (deg)
Vibro-Acoustic
2

Structural FEM Stator Deformation Radiated Noise

Unrestricted © Siemens AG 2016

Page 14 16.05.2016 Siemens PLM Software
Noise optimization of an SR electric motor
Structural Model

TEST CAE
Stator modelling Freq Freq
• Stator FRFs testing Mode Freq
[Hz] [Hz]
Description
[Hz]
• Stator Modal analysis original updated

• Stator modelling with anisotropic 2,½ 1255

material properties 2,0a 1318 1293 1319 1st ovalisation

• Correlation and updating to identify 2,0b 1342 1295 1343 1st ovalisation (45 deg)

properties 2,1 1632

3,0 3527 3426 3526 3rd ovalisation

Highest focus on ovalisation modes 4,0 5745 5541 5701 4th ovalisation
• Good match with forcing pattern
• Efficient radiator
Stator FRF test Stator TEST modes Stator CAE modes

Unrestricted © Siemens AG 2016

Page 15 16.05.2016 Siemens PLM Software
Dealing with non-linear behavior of the Stator
Research at Siemens

Example 1
Linearity and Reciprocity Check on Stator

Unrestricted © Siemens AG 2016

Page 21 16.05.2016 Siemens PLM Software
Assessment of non-linearity based on test
Test Setup

• 6 x 3D-accelerometers (5 roving + 1 reference)

Mini
• 2 shakers for structural excitation
• 1 mini shaker (XY plane, XY = -18°)
• 1 micro shaker (Z direction)

• Measurement point resolution

Y
• Radial: 16 points (1 for every stator pole)
• Axial: 3 Z
X
 48 measurement points

• Points measured in 10 separate runs Micro

Unrestricted © Siemens AG 2016

Page 22 16.05.2016 Siemens PLM Software
Example
Linearity checks

Z
X
X – X / Radial - Radial Z – Z / Axial - Axial

Unrestricted © Siemens AG 2016

Page 23 16.05.2016 Siemens PLM Software
Example
Reciprocity checks - Axial

Z – X / Axial - Radial

Unrestricted © Siemens AG 2016

Page 24 16.05.2016 Siemens PLM Software
 Multi-scale approach

MICRO-SCALE

zoom

Perform multiple static analyses to obtain the equivalent

parameters

with groups
Homogenization

With (anisotropic) equivalent

stiffness matrix

Unrestricted © Siemens AG 2016

Page 28 16.05.2016
MACRO-SCALE
Siemens PLM Software
 Advantages of the multi-scale approach
Er = Eθ Gθz = Gzr
Material parameter Ez (GPa) Grθ (GPa)
(GPa) (GPa)
Isotropic model 185.0 185.0 71.9 71.9

Homogenized model using [3] epoxy thickness 183.2 115.7 30.5 71.1

Optimized parameters model 185.0 185.0 40.5 10.0

100

10

1
Acceleration/force (m.s-2.N-1) Log

1e-1

1e-2

1e-3

1e-4
(3,0) (3,1)
1e-5
EXPERIMENT – Synthetized
1e-6 (2,0) (2,1) FE MODEL 1 – Isotropic properties
FE MODEL 2 – Anisotropic hom. properties
1e-7
FE MODEL 3 – Optimized orth. values
1e-8
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Unrestricted © Siemens AG 2016 Frequency (Hz)
[1] P. Millithaler, E. Sadoulet-Reboul, M. Ouisse, J.-B. Dupont, and N. Bouhaddi, “Structural dynamics of electric machine stators: Modelling guidelines and identification of three- dimensional equivalent material properties for multi-layered
Page 29 16.05.2016
orthotropic laminates,” Journal of Sound and Vibration, Mar. 2015.
Siemens PLM Software
Noise optimization of an SR electric motor
Electromagnetic Simulation

2D Electro-Magnetic FE-model (stator, rotor, windings)

Time domain simulation based on current waveforms for 60º rotor rotation
• For motor and generator mode
• For rpms from 0 – 5.000 rpm
Results
• High tangential forces between rotor and stator (motor torque)
• Lower radial forces, but more likely to cause noise

Magnetic flux lines Stator forces order 6 Force pattern for order 6,18,…

Unrestricted © Siemens AG 2016

Page 31 16.05.2016 Siemens PLM Software
Noise optimization of an SR electric motor
Mapping EM Forces

JMAG: Export to VL functionality: Export of forces on stator surface in time

domain and extrusion from 2D to 3D model
LMS Virtual.Lab Acoustics:
• Conservative mapping from EM mesh onto structural model mesh
• Conversion from time to frequency domain
Magnetic Flux Lines Force Export to Forces on SRM stator in
in JMAG LMS Virtual.Lab LMS Virtual.Lab

Unrestricted © Siemens AG 2016

Page 32 16.05.2016 Siemens PLM Software
Noise optimization of an SR electric motor
Vibro-Acoustic Forced Response

Structural Forced Response  surface vibrations for main motor orders

(6,12,…)
Creation of Acoustic model
• Wrap of the structural model with free-field termination at boundary
• 6 field microphones @ 1 m
Calculation of pressure in field points resulting from surface vibrations

Acoustic model Evaluation microphone positions Noise at field point

High levels when motor

orders strike ovalisation
mode @ 1.400 Hz
Unrestricted © Siemens AG 2016
Page 33 16.05.2016 Siemens PLM Software
 Noise optimization of an SR electric motor
Optimization

Optimization of the motor control strategy allows to reduce the torque ripple, but can also be used to
significantly reduce the noise radiated by the SRM
Structural optimization of the motor and its housing allows to further reduce the noise radiated by the SR
motor.
• Simulation allows to understand the reasons behind high noise levels in certain rpm-ranges. Noise peaks
typically occur when a motor order strikes a structural resonance and the stator force pattern matches well
with the mode shape.
• Design changes can be proposed, typically addressing resonance issues by adding damping to the
structural modes.
Effect of Optimized Control Strategy Effect of Structural Modifications

Unrestricted © Siemens AG 2016

Page 34 16.05.2016 Siemens PLM Software
Electric motor noise optimization
Integration of the motor in its environment

• The Noise and Vibration performance has to be assessed considering the operational
environment of the motor

• In the following example, the noise in the Vehicle is estimated combining the motor
excitation with vehicle FRF characteristics.

• Eg account acoustic transfer to the interior:

.  p
Pt arg et   Q j . 
.
Acoustic transfer Qj
Qj function (measured or
Sound radiation
typical curve)
from motor areas Fj

• Eg structure borne transfer, and if relevant mechanical coupling to the vehicle, through
FRF-based coupling
 p
Pt arg et   F j . 
 F j

Unrestricted © Siemens AG 2016

Page 35 16.05.2016 Siemens PLM Software