High Power Electronics Design

Zed (Zhangjun) Tang

ANSYS, Inc
June 3, 2014
1 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential
 Outline

Part I: EMI/EMC
Part II: Multiphysics Thermal/Structural

2 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Part I: EMI/EMC
EMI/EMC
Magnetic Field
Electric Field
Conducted Fields
Radiated Fields
Testing and solutions
Simulation is solving EMI/EMC problems both quantitatively
and qualitatively

Electronics is outpacing the basic construction of these

boards . . . were going to have to change how we do
things. - Paul Huray, Ph.D.
Professor of Electrical Engineering, University of South Carolina

3 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Agenda

High Power Electronics/Motor Drives EMI/EMC

* Power Device (IGBT, MOSFET, Power Diode) model extraction

* Busbar/Cable Parasitics extraction (frequency dependent LCR

model)

* System Integration PWM frequency/control algorithm

* Conducted/Radiated EMI/EMC

4 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Introduction

6.5kV IGBT Module Characteristics

Collector
Emitter

Packaging

6.5kV-600A Module
Dielectric
24 IGBT and
Gel
Diode 12 Diode Chips
Chip IGBT Baseplate
Ceramic
Chips
Substrate
5 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential
 Introduction
6.5kV IGBT Module Analysis
Include package in IGBT
performance

Find DC current distribution

Find switching currents for

power dissipation

Use power dissipation to

determine environmental
electromagnetic fields

6 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Introduction

Model design developed at Alstom/Pearl

IGBT Module Pack 3D Parameters Design and Couplings IGBT Model Far Field Study
accurate model Extraction Model

Tridimensional IGBT pack model and EM study

Parasitic model extraction
IGBT circuit model
Far Field Study for Electric Field EM
Electromagnetic
(EM) study

7 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 ElectroMagnetic Study

Module layout verification

The module contains 8 IGBTs in parallel: does each

IGBT receive the same amount of current?
If the current flows un-evenly, this will cause
mechanical stress and reliability issues.
Electromagnetic simulation is required. We use
Maxwell3D.

8 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 ElectroMagnetic Study
The layout in imported from the CAD tool
The DC solver is used
The input current (600 A) is defined
The sink (return current path) is defined

600 A Sink

Outputs: conduction path and current distribution

9 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 ElectroMagnetic Study

Current Distribution
IGBTs on, Diodes off

The structure is meshed

using automatic and
adaptive meshing

10 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 ElectroMagnetic Study
The end IGBTs see less current than the
center ones.
This can cause reliability issues as the
center IGBTs will be overloaded
An optimization of the copper tracks can
be made in order to equalize the
currents.

Igbt1a and Igbt4a have the

highest quantity of current
11 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential
 Parasitics Extraction

Extracting parameters is straightforward as the nets are

automatically assigned.

Emitter net

Gate net

Collector net

12 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Parasitics Extraction

The simulation outputs consist of the RLC matrices for different frequencies

13 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 System Integration
How do we import the results from Q3D?: Q3D dynamic link

2 Types of links: Single Frequency or

Frequency dependent
No need to manually import output file
Simplorer incorporates directly the Q3D
project
If some results are not available, Simplorer
dynamically launches Q3D
Parameters and variables can be passed
between S8 and Q3D

14 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 System Simulation

Wattmeter Power Module from Q3D

for board parasitics

IGBT

Vc
Vg

15 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 IGBT Characterization

Accurate models of the semiconductors

are needed to achieve a good circuit
simulation
Simplorer offers a parameterization tool
for IGBTs
The user needs to import the data from
the datasheet

16 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 IGBT Characterization

SheetScan

17 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 IGBT Characterization
Once all the curves and data are entered, start extraction
The tool fits the data to the internal Simplorer model using Genetic Algorithm

Characterization tool
Component dialog

18 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 IGBT Characterization
Switch on
Ansoft Corporation switch_on Simplorer1
50.00 3000.00 15.00

Test Circuit 40.00 2500.00

Vce 10.00

30.00 2000.00 Ic 5.00

VM2.V [V]
U1.VCE
R2.I [A]
20.00 1500.00 0.00

10.00 1000.00 -5.00

Curve Info
U1.VCE
TR
VM2.V
0.00 500.00 TR -10.00
R2.I
TR

-10.00 0.00 -15.00

499.90 499.95 500.00 500.05 500.10 500.15 500.20 500.25 500.30
Time [us]

Ansoft Corporation
Switch off
switch_off Simplorer1
50.00 2500.00 15.00
Curve Info
U1.VCE
TR

40.00
2000.00
Vce TR

TR
VM2.V
10.00
R2.I

30.00 5.00

1500.00

VM2.V [V]
rise time= 40 s
U1.VCE
R2.I [A]

20.00 0.00

1000.00

fall time = 50 s
10.00 -5.00

0.00
500.00 Ic -10.00

-10.00 0.00 -15.00

999.00 999.50 1000.00 1000.50 1001.00 1001.50 1002.00 1002.50 1003.00
Time [us]

19 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 System Simulation

Vce Ic

Power
Vg

Vge

The power pulse duration is much smaller than the rise/fall time of Ic and Vce

20 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 System Simulation

FTT of the power through Igbt1a

Most of the power level is below 110 MHz

21 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Emitted Fields

There is very high power going through the IGBTs (almost 60 000 W in this
study) during a very short period of time (60 ns). This switching can cause
EMI issues in the inverter, but also in the surrounding equipment

To be answered using the finite element method in HFSS:

Will the module radiate?

Are the field levels surrounding the module within mandated levels?

22 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Emitted Fields

Regulators impose
maximum levels of
electric fields close to
electric equipment.
In the 10-110 MHz
range:
Emax=61V/m

Exposure limits defined by European Community

23 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Emitted Fields
The structure is discretized with adaptive meshing. The meshing
frequency is 100 MHz
The frequency sweep ranges from 15MHz to 120 MHz

24 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Emitted Fields

For each frequency, the power amplitude is entered

Power E field at 1m for 1000w E field at 1m

Spectrum (MHz) (W) (V/m) (V/m)
16.52892562 21439.97604 2.6312 56.41286497
33.05785124 8635.09049 2.7994 24.17307232
49.58677686 5579.619715 2.8731 16.0308054
66.11570248 4131.16773 3.063 12.65376676
82.6446281 3276.823585 3.4045 11.15594589
99.17355372 2712.888158 3.8924 10.55964586
115.7024793 2308.359536 4.4861 10.35553171
132.231405 2022.75744 4.905 9.921625241

Spectrum from Simplorer

Outputs from Simplorer Outputs From HFSS Fields Levels

Inputs for HFSS (normalized results)
25 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential
 Emitted Fields
The E field is very localized
close to the module even at
100 MHz
However, the very high
power can lead to large
Power Power
E field at 1m for 1000w E field
E field at at
1m1m f
Spectrum (MHz) (W)
values
Spectrum (MHz)
of
(V/m)
E field
(W)
even far
(V/m) (V/m
16.52892562 21439.97604 from2.6312
16.52892562the module
21439.97604 56.41286497 2.631
33.05785124 8635.09049 33.05785124
This2.7994
design8635.09049
is fine at 24.17307232 2.799
49.58677686 5579.619715 110MHz.
49.58677686
2.8731 5579.619715 16.0308054 2.873
66.11570248 4131.16773 66.11570248
3.063 4131.16773 3.063
12.65376676
82.6446281 3276.823585
Power E 82.6446281
field at3.4045 3276.823585
1m for 1000w
Power 11.15594589
E field
E field 3.404
at at
1m 1m f
99.17355372
Spectrum (MHz) 2712.888158
(W) 99.17355372
Spectrum (V/m) 2712.888158
3.8924
(MHz) (W) (V/m) 3.892
10.55964586 (V/m
115.7024793
16.52892562 2308.359536
21439.97604 4.4861
115.7024793
2.6312 2308.359536
16.52892562 21439.97604 10.35553171
56.41286497 4.486
2.631
mag E @ 100 MHz, Power = 10 000W2022.75744
132.231405
33.05785124 8635.09049 4.905
132.231405
33.05785124
2.7994 2022.75744
8635.09049 9.921625241
4.905
24.17307232
2.799
49.58677686 5579.619715 49.58677686
2.8731 5579.619715 16.0308054
2.873
66.11570248 4131.16773 66.11570248
3.063 4131.16773 12.65376676
3.063
82.6446281 3276.823585 82.6446281
3.4045 3276.823585 11.15594589
3.404
26 2013 ANSYS, Inc. June 3, 2014 99.17355372 2712.888158
ANSYS Confidential 99.17355372
3.8924 2712.888158 10.55964586
3.892
 Part II: Multiphysics -- Outline

Context / Introduction
Component level Modeling
System level Modeling Electrical Domain
Multiphysics Modeling - Workbench
Conclusions

27 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Context/ Introduction
A (Power) Component is the combination of semi-conductors
(IGBT, diodes, ) and associated BusBar or PCB. This
presentation details how this component is designed:
1 First stage consists in optimizing the geometry layout of the
BusBars, Connectors, PCBs and finding the best
Semiconductor types for a given specification.
This design is established with a pre-defined test circuit
Its a mix of Electrical Modeling and 3D Modeling.
2- Second stage consists in using this component in its real
environment (loads, sources, actual waveforms) to determine
the exact current waveform that are seen by the Component.

3 -Finally, The current waveforms are used as a source for a

Electro-thermal-Mechanical analysis

28 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Context/Introduction

Multi-physics Modeling is used at the component level and a the system level
Each physics is simulated with High Fidelity Solvers
Workbench is used to couple the different physics
Simplorer is used as a Power Electronics simulator and System integrator

Components level System Level

Simplorer

ANSYS Workbench
29 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential
 Component Level Modeling

30 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Component Level Modeling
Example of EMC/EMI Oriented Model design developed at Alstom *

IGBT Module Pack 3D State Space Design and Couplings IGBT Model Far Field Study
accurate model Extraction Model

Tridimensional IGBT pack model and EM study

State space extraction
IGBT circuit model
Far Field Study for Electric Field EM
* Emmanuel Batista, Vincent Delafosse, Simulating
Electromagnetic
(EM) study EMC/EMI Effects for High Power Inverter Systems ,
Ansoft Inspiring Engineering, 2008
31 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential
 Component Level Modeling
Double-pulse tests and measurements are standard for
validating components in the Train industry. The Busbar, the
IGBT and Diode are characterized on a pre-defined test
procedure.

Questions to be answered:
For an imposed IGBT type, what is the best Busbar layout ?
We are able to perform virtual testing of these components
combining :
Circuit simulation for the Electrical Characteristics of semiconductors
3D Modelisation of the RLC
32 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential
 Component Level Modeling
The switching of these power modules are operated at low frequency
(1kHz-20kHz). However, the highest frequency that we need to capture is
related to the rising time of gate voltage. Its about 10ns, therefore Q3D
is solved up until 100MHz ( f 2 1t
max )
sw

Several busbars geometries

have been tested and measured

33 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Component Level Modeling
We use ROM to capture electrical characteristics of the BusBar layout

Test Procedure
Q3D

Power Module

Load

RLC Parameters

Chopping
Filter Capacitor
Capacitor
StateSpace
Model

34 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Component Level Modeling

This virtual prototype of the power component enables Geometry

validation of the busbar based on electrical criteria. Its a mix of 3D
and circuit simulations.

Obtained Results:
Validation of the component
(IGBT/ Busbar layout)

Double-pulse Virtual Test Results

35 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Component Level Modeling
Moreover, in train applications, power components are used in
parallel as it is necessary to achieve high power requirements ad
high reliability.
It is very easy to validate the power components in parallel in their
system environment looking at: Overshoot

Overshoot, Oscillations, Losses, Static Derating

Derating
Dynamic Derating

Voltages /currents at the switch off of the IGBT

BUSBAR COMPARISON Icom=1000A Vbus=500V

GEOMETRY dI/dt [kA/s] dV/dt [kV/s] DV [V] Lloop[nH] Static Derating [%]
1 4 16 238 59.5 11
2 4 15 186 46.5 11
3 3.6 15 223 62 22
36 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential
 System level Modeling Electrical Domain

37 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 System Level Modeling
Power Cables Characteristics are simulated

Specifics of the Motor cables:

- High Currents/ Low Frequency with
almost sinusoidal at the Stator
- Very high dV/dt, Voltage Switching
- Forts dV/dt, Voltage-based PWM (MLI)

3D layout of the Cables

ROM models contain:

- Partial Inductances of Cables
38 2013 ANSYS, Inc. June 3, 2014 -
ANSYS Confidential Capacitance effects are also included
 System Level Modeling

Driving Logic

Ordre de pilotage

Filter Capacitor
Capacit filtre LOAD
Charge
Choppingde
Capacit
Capacitor
dcouplage
ROM Models of Busbar
Power
Composants de and Cables
Components
puissance

39 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 System Level Modeling
Voltage waveforms of the PWM

Switching events are defined with Asynchroneous PWM with Fcom>>10xFstat

State Graphs
Average IGBT is used because:
Very long simulation: t > 1 s
We dont need switching details
Switching losses are fitted to the
datasheet values and measured
values.
Rfrences
1. Pierre Solomalala, Emmanuel Batista, Diffrents
aspects de la modlisation multi-physique au sein du
Command signal applied at the IGBT gate laboratoire PEARL/PRIMES , UGM ANSYS, 2009.
2. Emmanuel Batista, Vincent Delafosse, Simulating
EMC/EMI Effects for High Power Inverter Systems ,
40 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential Ansoft Inspiring Engineering, 2008.
 System Level Modeling
Some Results

41 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Multiphysics Modeling
in Workbench

42 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Multiphysics Modeling
ANSYS Workbench

Electromagnetic Thermal Mechanical

TC

Electric +Q3D
Semiconductor

Simplorer
43 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential
Drive
 Multiphysics Modeling

Workbench enables a seamless 3D coupling flow.

W/m3 TC

h [m]

A parameter Analysis is used to determine what insulation material is needed to

achieve a maximum displacement level

44 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Multiphysics Modeling
Maxwell 3D simulation: Current Calculation
Input: average current that is computed in the system simulation in Simplorer
Output: Ohmic losses are calculated based on the volumic current calculations.

45 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Multiphysics Modeling
Static Thermal Module of ANSYS: Temperature distributions

Heat Flux on Copper parts Maximum temperature on Copper parts

46 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Multiphysics Modeling
Static Thermal Module of ANSYS: Temperature distributions

Insulation part Temperatures Phase Conductors Temperatures

47 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Multiphysics Modeling

Static Stress Analysis

Y axis Displacements X axis displacements

48 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Multiphysics Modeling

Static Stress Analysis

Total Displacements

49 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

 Conclusion

A Multi-physics Multi level modelisation method has been presented

Dynamic IGBTS have been used at the component level
Average IGBT have been used at the system level
State Graphs enable easy development of the command algorithm used to drive each
inverter
The Stress values are accurate as we have loaded the structure with real waveforms
from system analysis.

50 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential

51 2013 ANSYS, Inc. June 3, 2014 ANSYS Confidential