modelling of multi-domain systems€¦ · analysis –specification- design model research...

1© 2015 The MathWorks, Inc.

Modelling of Multi-Domain Systems

Dr. Michael Kitz

MathWorks

June 9, 2015

2

What Is This?

V+

V-

ω

dt

diLRiKV m

mmmbin

dt

dJbiKT mmt

3

Three Key Takeaways

Build physical models to test for integration issues early

Improve your design efficiency using domain-specific modelling

Choose the modelling strategy suited to your task

5

Product Innovation

‘Going Smart’

Model-Based Design:

a Smart Design Process

for Smart Products

Smart Product

The Challenge

Smart Design

Process

6

Smart Product

Model-Based Design

TE

ST

&

VE

RIF

ICA

TIO

N

INTEGRATION

IMPLEMENTATION

ANALYSIS – SPECIFICATION- DESIGN

MODEL

RESEARCH REQUIREMENTS

MCU DSP FPGA ASIC

Structured

TextVHDL, VerilogC, C++

Architecture

Algorithms

Schematics

TEST

CASES

Environment

Constraints

Physical Domains

TEST

CASES

PLC PAC

7

Smart Product

Model-Based Design

TE

ST

&

VE

RIF

ICA

TIO

N

INTEGRATION

IMPLEMENTATION

ANALYSIS – SPECIFICATION- DESIGN

MODEL

RESEARCH REQUIREMENTS

MCU DSP FPGA ASIC

Structured

TextVHDL, VerilogC, C++

Architecture

Algorithms

Schematics

TEST

CASES

Environment

Constraints

Physical Domains

TEST

CASES

PLC PAC

8

Programming/modelling style

Time and frequency

Continuous and discrete

Multiple physical domains

What does Multi-Domain mean?

Physical Modelling (Schematic)

Data Flow (Block diagram) Event-Driven Systems Programming Language (Textual)

10

Relative Effort of Plant Model Representations

Plant

Specification

Mathematical

Code

Programming EffortDomain Expertise

MATLAB and Simulink

Programming Effort

C, C++, FORTRAN

Domain Expertise

Programming EffortDomain Expertise

Physical Modelling Tools

Physical

Structure

Data Flow

Representation

Duration

11

Data-Driven ModellingFirst Principles Modelling

Neural Networks

Physical Networks System Identification

Parameter Tuning

Programming

Block Diagram

Modelling Language

Symbolic Methods

Modelling Approaches

Statistical Methods

13

p1 p2

qp1 p2

p3

p4

Physical Network ConceptThrough & Across Variables

All nodes have the same pressure (across variable)

Sum of flows (through variables) at a node is zero

Each component must specify an equation involving the through and/or

across variables at its boundary

14

Physical Modelling with Simscape

Mechanical(translational)

Hydraulic

Thermal

Liquid

Custom Domains via Simscape Language

3-D Mechanical Systems (multi-body)

Electric Drives, Power Networks

Pneumatic Magnetic

N S

Electrical

Thermal

Mechanical(rotational)

Domain-specific modelling:

Models that reflect structure of

physical system

Common basic concept:

Bidirectional flow of power

between components

Custom domains and components

Libraries and special technology

add-ons

Co-simulate with FEM-tools

15

Model Diagram Examples

Hydraulic-mechanical modelPneumatic-mechanical model

Electrical modelThermal model (detail)

17

Grid

Pitch

Yaw

Rotor

Speed

Blades

Tower

Geartrain GeneratorHub

Lift

Wind

Multi-Domain Model: Wind Turbine

18

Multi-Domain Model: Wind Turbine

19

Wind Turbine – Architectural Base Model

20

Wind Turbine – Basic Functional Models

21

Wind Turbine – Elaboration, Exploration, Integration

22

Three Key Takeaways

Build physical models to test for integration issues early

Improve your design efficiency using domain-specific modelling

Choose the modelling strategy suited to your task