1© 2016 The MathWorks, Inc.

Model-Based Design:

Design with Simulation in Simulink

Ruth-Anne Marchant

Application Engineer

MathWorks

2

3

Outline

Model-Based Design Overview

Modelling and Design in Simulink

– Modelling Physical Systems

Control logic

– Simulation System-level optimisation

Verification of design changes

Summary

4

Electrical

Components

EDA

Mechanical

Components

MCAD/

MCAE

Embeddable

Algorithms

Algorithm

Design

Traditional Development Workflow

SPECIFICATIONS

DESIGN AND IMPLEMENTATION

RESEARCH REQUIREMENTS

Embedded

Software

C/C++

INTEGRATION AND TEST

Electrical

Components

EDA

Mechanical

Components

MCAD/

MCAE

Embeddable

Algorithms

Algorithm

Design

DESIGN AND IMPLEMENTATION

Embedded

Software

C/C++

5

Electrical

Components

EDA

Mechanical

Components

MCAD/

MCAE

Embeddable

Algorithms

Algorithm

Design

Problems in Traditional Development Workflow

SPECIFICATIONS

DESIGN AND IMPLEMENTATION

RESEARCH REQUIREMENTS

Embedded

Software

C/C++

Paper Specifications

Physical Prototypes

Manual Coding

Requirement Documents

INTEGRATION AND TEST

Traditional Testing

6

INTEGRATION

IMPLEMENTATION

DESIGN

Environment Models

Physical Components

Algorithms

Structured

TextVHDL, VerilogC, C++

TE

ST

& V

ER

IFIC

AT

ION

RESEARCH REQUIREMENTS

PLC, PACMCU DSP FPGA ASIC

Model-Based Design Workflow

Traditional System

Development Workflow

1. Research

2. Requirements and

Specifications

3. Design

4. Implementation

5. Test and Verification

7

Model-Based Design: Specifications

CAPABILITIES

•Executable specification

•Executable constraints

•Links to requirements

BENEFITS

•Early validation and test

development

•Clear specification

•Simulate whole system,

including environment

•Tight link to requirements

DESIGN

Environment Models

Physical Components

Algorithms

RESEARCH REQUIREMENTS

8

Model-Based Design: Requirements

Trace to requirements in DOORS, Word, Excel, etc.Model system response bounds

Formalize requirements as properties and objectives

DESIGN

Environment Models

Physical Components

Algorithms

RESEARCH REQUIREMENTS

9

CAPABILITIES

•Refine model description

•Add fixed point, timing,

component interface

details

DESIGN

Environment Models

Physical Components

Algorithms

RESEARCH REQUIREMENTS

BENEFITS

•Fast design exploration

•Design optimization

•Find flaws before

implementation

Model-Based Design: Design

10

CAPABILITIES

•Refine model description

•Add fixed point, timing,

component interface

details

Physical Components

Algorithms

IMPLEMENTATION

Structured

TextVHDL, VerilogC, C++

PLC, PACMCU DSP FPGA ASIC

DESIGN

Environment Models

Physical Components

Algorithms

RESEARCH REQUIREMENTS

BENEFITS

•Fast design exploration

•Design optimization

•Find flaws before

implementation

Model-Based Design: Design

11

CAPABILITIES

•Rapid Prototyping

•Automatic Code

Generation:

•C/C++

•HDL

•PLC

BENEFITS

•Eliminate hand-coding

•Eliminate hand-code errors

•Hardware target portability

•Better testability and reuse

•Bridge between domain,

software, and hardware

knowledge experts

IMPLEMENTATION

DESIGN

Environment Models

Physical Components

Algorithms

RESEARCH REQUIREMENTS

Structured

TextVHDL, VerilogC, C++

PLC, PACMCU DSP FPGA ASIC

Model-Based Design: Implementation

12

CAPABILITIES

•Model Verification

•Software Verification

•Hardware-in-Loop

•Test and Measurement

BENEFITS

•Detect errors earlier

•Reduce use of

physical prototypes

• Implementations that

work first time

•Reuse tests throughout

development stages

TE

ST

& V

ER

IFIC

AT

ION

INTEGRATION

IMPLEMENTATION

DESIGN

Environment Models

Physical Components

Algorithms

Structured

TextVHDL, VerilogC, C++

RESEARCH REQUIREMENTS

PLC, PACMCU DSP FPGA ASIC

Model-Based Design: Test and Verification

13

Building a Wave Farm with Model-Based Design

14

Outline

Model-Based Design Overview

Modelling and Design in Simulink

– Modelling Physical Systems

Control logic

– Simulation System-level optimisation

Verification of design changes

Summary

15

Outline

Model-Based Design Overview

Modelling and Design in Simulink

– Modelling Physical Systems

Control logic

– Simulation System-level optimisation

Verification of design changes

Summary

16

Mechanical System Modelling

17

Mechanical System Modelling

18

Mechanical System Modelling

19

Electrical System Modelling

Bridge Rectifier

(AC to DC)

20

Electrical System Modelling

Speed Controller

DC Motor

Motor

Servoamplifier

Hall Effect

Sensor

21

Electrical System Modelling

22

Multi-Domain Modelling of Physical Systems

23

Control System Design

24

Defining System Mode Logic

Define mode logic using

state machine in Stateflow

Generate production code

directly from model

Motion

Normal Mode

Cruise Mode

Charge

No

Charge

Acceleration

Mode

Start Mode

RPM

> 800RPM

< 790

Accel is low

&&

Charge is low

Accel is high &&

Charge is high

Charge

is high

Charge

is low

Brakes

applied

Brakes

released

Gen = 1

Engine = 0

Motor = 1

Gen = 0

Motor = 1

Mot = 1

Gen = 0

Engine = 0

Specification

Brake ModeEngine = 1

25

Outline

Model-Based Design Overview

Modelling and Design in Simulink

– Modelling Physical Systems

Control logic

– Simulation System-level optimisation

Verification of design changes

Summary

26

Optimise Entire System

Use optimisation algorithms to

automatically tune parameter values

– Match response

– Meet requirements

Optimise system performance

(controller and physical system)

HEV

Model

IC Engine On Spd

IC Engine Off SpdBattery Capacity

Battery Max Discharge Power

Fuel Consumed Over

100 Different Drive Cycles

Calculate

Cost

Optimisation

AlgorithmParameter Values

Motor Torque

Motor Speed

27

Distributing Simulations

with Parallel Computing

Simulating in parallel

– Distribute simulations to

multiple cores/processors

– Dramatic speedup for sets

of simulations (parameter

sweeps, flight cycles

optimisations, and more)

for parfor

…

Computer Cluster

Workers

… …

Desktop System

Workers

… …

Simulation 1

Simulation 2

HEV Model

28

Shorten Simulation Times With Parallel Computing

Problem: Minimize simulation time to run

a parameter sweep on the HEV model

Solution: Use Parallel Computing Toolbox

to speed up the sweep

Model:

Parameter Set 1

Parameter Set 2

Parameter Set N

…

Drive Cycles

Weights

Ratios

…

HEV Model

Multi

Core

29

Outline

Model-Based Design Overview

Modelling and Design in Simulink

– Modelling Physical Systems

Control logic

– Simulation System-level optimisation

Verification of design changes

Summary

30

Automatically Run Tests And Document Results

Problem: Evaluate test results

quickly to make design changes

and document the results

Solution: Use Simulink Report

Generator to automatically

document tests and results

Situation:

Design

Change

Test

Evaluate

Results

Document

Report

31

Outline

Model-Based Design Overview

Modelling and Design in Simulink

– Modelling Physical Systems

Control logic

– Simulation System-level optimisation

Verification of design changes

Summary

32

Key Points

Simulink is a multi-domain modelling and simulation environment

facilitating Model-Based Design

Optimise the system-level performance

Accelerate your development– Speed up simulations using Parallel Computing Toolbox

– Speed-up processes using Simulink Report Generator

33

Call to Action

Learn more about Model-Based Design

with Simulink

Explore our website

– au.mathworks.com

Contact me:

– Ruth-Anne Marchant

ruth-anne.marchant@mathworks.com.au

34

Q & A