+ All Categories
Home > Documents > Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software...

Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software...

Date post: 19-Oct-2020
Category:
Upload: others
View: 0 times
Download: 0 times
Share this document with a friend
20
Embedded control of an Adaptive Magnetic Flux Array challenges of distributed software in Power electronics system Sylvain Laversanne - August 14 2014
Transcript
Page 1: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Embedded control of an Adaptive Magnetic Flux Array challenges of distributed software in Power electronics system

Sylvain Laversanne - August 14 2014

Page 2: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Summary

1. AMFA technology explained

2. Architecture and challenges

3. Software development

4. Conclusions

Page 3: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

About AXIFLUX Axiflux® was incorporated in Melbourne, 2011 to research, design, manufacture and distribute a revolutionary range of electric motors and generators. The founding group are highly entrepreneurial focused on pushing the innovation envelope to its limits. Initial areas of focus encompass: • High-efficiency, high power density electric vehicle motors • Commercial pumps and ventilation fans • Conveyor belt motors • Direct-drive wind turbines

Page 4: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

AMFA technology explained

Concept of the Adaptive Magnetic Flux Array : - Permanent Magnet Motor with an Axial Flux configuration:

- Compact & Efficient - High Torque low speed

- Independently controlled coils:

- Low ripple, low cogging torque - High Reliability - Freedom of design

- Applications:

- Direct drive - High Torque electronically speed/torque controlled

motors - In built redundancy

Page 5: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

AMFA technology explained

Electronically controlled integrated motor : lab version, AMFA1

Stator

Rotor

Power electronics

Page 6: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

AMFA technology explained

Electronically controlled integrated motor : lab version, AMFA2

Stator

Rotor

Power electronics

Page 7: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

AMFA technology explained

Concept of an axial flux motor

Magnets Coils

Radial type

Page 8: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

AMFA technology explained

High torque and high efficiency motor

Page 9: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

AMFA technology explained

Independently driven coil motor

1 2 3 n-1 n

pos

neg

pos

neg

Each coil is driven by its own power electronics module

Page 10: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Architecture and challenges

Sensors and peripherals

0 1 2 3 4 5 6 7 8 9

x 10-3

-6

-4

-2

0

2

4

6

I2C low speed peripherals

High speed CANbus and Isolated Serial bus

Analog signals (I,V,etc.)

Piccolo DSP (1 of n slaves)

Delfino DSP Master/Interface DSP

Complex network & architecture ( analog, CANbus, I2C, serial, SPI, etc.)

Page 11: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Architecture and challenges

Fast Control loop

0 1 2 3 4 5 6 7 8 9

x 10-3

-6

-4

-2

0

2

4

6

Analog signals (I,V,etc.)

Piccolo DSP

Each Module is an arbitrary high power module: - Current waveform up to 1200Hz and more - Phase is critically locked - Feedback regulation in a single phase non stationary frame - Software fault detection is critical - Sampling Time is <35µs

Real Time, synchronised, Multitasking software

Page 12: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Software development

Classic approach, requiring : - Model in the Loop simulation - Software in the loop simulation - Rapid control prototyping

- Model based design

- Data logging and analysis

Page 13: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Software development

INPUTS DATA PROCESSING &

SAFETY

CALCULATION &

CONTROL

OUPUT

ADCs

CANbus

Serial

I2C

CANbus

ePWM

GPIOs

FAULT

Handling

Conditioning &

Error check

State

calculation

Observers

Control

Modulation &

conditioning

Delayed loop DATA Bus

Software architecture

Page 14: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Software development

SIMULATION

DATA PROCESSING &

SAFETY

CALCULATION &

CONTROL

FAULT

Handling

Conditioning &

Error check

State

calculation

Observers

Control

Modulation &

conditioning

Delayed loop DATA Bus

Simulation architecture

Page 15: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Software development

Tools used: MATLAB R2014a Simulink Fixed Point Designer Embedded Coder (C2000 DSP family) Vehicle Network Toolbox TI Code Composer Studio v5.5

Page 16: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Software development

Example of time saver: Full DSP configuration with GUI

Page 17: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Software development

Example of time saver: Direction implementation of standard CANbus databases

Page 18: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Software development

Example of time saver: GUI aided fixed-point design Access to optimised fixed-point mathematical functions

Page 19: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Software development

Example of time saver: Multitasking OS and easy use with minimum overhead

3 µs overhead for context switching between main interrupt and task start

Page 20: Sylvain Laversanne - August 14 2014 - MathWorks€¦ · Architecture and challenges 3. Software development 4. Conclusions . About AXIFLUX Axiflux® was incorporated in Melbourne,

Conclusions

- Fast tasking 25khz and more multitasking software is possible without

compromising with easy integration of network interfaces & peripherals

- Advantages of a model-based software (understandable, sharable, etc)

- MATLAB tools are time saving: for the same software design time has been divided by 5

- Rapid prototyping with embedded coder doesn’t require a specific hardware

- Versatile tool allowing rapid prototyping of any ECUs

- Versioning tools already integrated in the package

- Generated code is compliant with classic standard


Recommended