Model-Based

Design for
Engineering &
Technology

Gowtham Rajmohan
Technical Manager-MATLAB,
ARK Infosolutions Pvt ltd, Chennai
[email protected]
Model Base Design Stages:

• Model Creation

• Simulation and Analysis

• Automatic Code Generation

• Research and Development

• Verification and Validation

• Deployment and Maintenance

 So, what is Model-Based Design anyway?

To understand, let’s first look at the conventional method of

building a product…
Conventional Design ProcessProblem Statement / Requirements

1. Define Requirements
• What are the design objectives?
Design and Implement
• What are our constraints?

2. Component-Level Design
Embedded
Mechanical Electrical Control Software
• Mechanical, Electrical, Control
System, Software etc.

3. System Integration Integrate and Test

4. Testing
So what’s the problem?
Requirements & Design Specifications
Problem Statement / Requirements
contained in separate files:

• Difficult to analyze, Easy to misinterpret

• Difficult to integrate with design
• Difficult to manage as they change

Physical Prototypes:
Design and Implement
• Incomplete and expensive
• Prevents rapid iteration
• No system-level testing
Embedded
Manual Coding: Mechanical Electrical Control Software

• Time consuming
• Introduces defects and variance
• Difficult to reuse
Integrate and Test
Traditional Testing

• Design and integration issues found late

• Difficult to feed insights back into design
process, No Traceability
 Requirement errors are most expensive to fix
during testing
Relative cost to fix an error

Errors introduced in:

Requirements phase
Design phase
Coding phase

Requirements phase Design phase Coding phase Testing phase

Project phase where error is fixed

Source: Return on Investment for Independent Verification & Validation, NASA, 2004.
Model-Based Design can help in preventing this
Multi-domain
Save time by developing in a
Simulation Model single simulation environment
Requirements
and
Specifications
Produce better designs by
continuously comparing
Control Electrical design and specification

Lower cost hardware

Embedded
prototypes
Software Mechanical

Continuous Simulation: Early

Error Detection

Easy Integration
Model-Based Design: ‘V-shaped’ Workflow

1. Define 7. Complete
Requirements Integration and
Test

2. System-Level 6. System-Level
Specification Integration and
Test

3. Subsystem 5. Subsystem
Design Integration and
Test

4. Subsystem
Implementation

Design Phase Testing and Integration

Conventional Design ProcessProblem Statement / Requirements

1. Define Requirements
• What are the design objectives?
Design and Implement
• What are our constraints?

2. Component-Level Design
Embedded
Mechanical Electrical Control Software
• Mechanical, Electrical, Control
System, Software etc.

3. System Integration Integrate and Test

4. Testing
So what’s the problem?
Requirements & Design Specifications
Problem Statement / Requirements
contained in separate files:

• Difficult to analyze, Easy to misinterpret

• Difficult to integrate with design
• Difficult to manage as they change

Physical Prototypes:
Design and Implement
• Incomplete and expensive
• Prevents rapid iteration
• No system-level testing
Embedded
Manual Coding: Mechanical Electrical Control Software

• Time consuming
• Introduces defects and variance
• Difficult to reuse
Integrate and Test
Traditional Testing

• Design and integration issues found late

• Difficult to feed insights back into design
process, No Traceability
 Requirement errors are most expensive to fix
during testing
Relative cost to fix an error

Errors introduced in:

Requirements phase
Design phase
Coding phase

Requirements phase Design phase Coding phase Testing phase

Project phase where error is fixed

Source: Return on Investment for Independent Verification & Validation, NASA, 2004.
Model-Based Design can help in preventing this
Multi-domain
Save time by developing in a
Simulation Model single simulation environment
Requirements
and
Specifications
Produce better designs by
continuously comparing
Control Electrical design and specification

Lower cost hardware

Embedded
prototypes
Software Mechanical

Continuous Simulation: Early

Error Detection

Easy Integration
Model-Based Design: ‘V-shaped’ Workflow

1. Define 7. Complete
Requirements Integration and
Test

2. System-Level 6. System-Level
Specification Integration and
Test

3. Subsystem 5. Subsystem
Design Integration and
Test

4. Subsystem
Implementation

Design Phase Testing and Integration

Pilotless Flight
Aurora used Model-Based ofMATLAB
Design with Aurora
® Centaur
and Simulink®to develop Centaur’s flight
control, vehicle, mission, and flight management software.
Ather Energy Develops Electric Two-Wheeled
Scooter and Charging Stations Using Model-
Based Design
Challenge
Design and produce India’s first intelligent electric
scooter

Solution
The Ather 450 intelligent electric scooter.
Use Model-Based Design to simulate and optimize the
scooter design and to generate production code for its
“We had lots of promising ideas, but as a small
control software startup, we did not have the time, money, or
people to build prototypes to test each one.
With Model-Based Design, we identified and
Results
validated the best ideas through simulation,
 Design alternatives evaluated in weeks, not monthsmaking it possible to deliver a full-featured
 Testing time reduced by 50%
scooter in less time.”
 to
Link Field issues rapidly resolved
user story
- Shivaram N.V., Ather Energy
Airbus Develops Fuel Management System for the A380 Using
Model-Based Design

Challenge
Develop a controller for the Airbus A380 fuel
management system

Solution
Use MATLAB, Simulink, and Stateflow for Model-
Airbus A380, the world’s largest commercial
Based Design to model and simulate the control aircraft.
logic, communicate the functional “Model-Based Design gave us advanced visibility
specification, and accelerate the development into the functional design of the system. We also
of simulators completed requirements validation earlier than was
previously possible and simulated multiple
Results simultaneous component failures, so we know what
will happen and have confidence that the control
 Months of development time eliminated
 Models reused throughout development logic will manage it.”
 Additional complexity handled without staff - Christopher Slack, Airbus
Link to user
 Benefits of Continuous Testing and Verification
• Requirements are linked to the model RESEARCH REQUIREMENTS
(reduces chances of errors being caught late)
• Domain-specific detail added to refine the DESIGN

model. Intellectual property and engineering Environment Models

data are reused from existing designs and
Physical Components
CAE tools. (re-use existing work)
Algorithms
• Control algorithms tested against requirements TEST
by simulating the model (robust testing and AND
VERIFICATION
validation) IMPLEMENTATION

• Automatically generate code for testing and C, C++ VHDL, Verilog SPICE

production for: Analog

MCU DSP FPGA ASIC
• Microcontrollers Hardware

• FPGAs
• PLC IDEs INTEGRATION
(less effort than manual coding, fewer errors)
Model-Based Design Example – for a Hybrid
Electric Car

Define Complete
Requirements Integration
and Test

System-Level System-Level
Specification Integration
Battery
and Test
Engine

Energy Management

Subsystem Subsystem
Design Integration
Simulink
Plant Model
and Test
Production
C-Code Subsystem
Implementation
1. Define Requirements
Involve requirements in development
process as early as possible to find
errors in time

Define Complete
Requirements Integration
HEV and Test
Requirements

System-Level System-Level
Specification Integration
Battery
and Test
Engine

Energy Management

Subsystem Subsystem
Design Integration
Simulink
Plant Model
and Test
Production
C-Code Subsystem
Implementation
2. System Level Specification
System level analysis at the start of design
process to set realistic specifications

Define Complete
Requirements HEV Requirements HEV Requirements Integration
1. Control System 1. Drivetrain System and Test
Wheel Speed Gear Ratios
5% Setting Time 1.5:1, 2.1:1, …
in 0.5
System-Level Sec System-Level
Specification Integration
Battery
and Test
Engine

Energy Management

Subsystem Subsystem
Design Integration
Simulink
Plant Model
and Test
Production
C-Code Subsystem
Implementation
3. Subsystem Design
Tradeoff studies on different subsystems to
enable efficient, feasible designs

Define Complete
Requirements Integration
Cruising and Test

Braking
System-Level System-Level
Specification Integration
Battery
and Test
Engine

Energy Management

Subsystem Subsystem
Design Integration
Simulink
Plant Model
and Test
Production
C-Code Subsystem
Implementation
3. Subsystem Design
Tradeoff studies on different subsystems to
enable efficient, feasible designs
Requirements
System-level
Design
Detailed
FE, CFD, Design
single domain
Define tools Complete
Requirements Verify via Integration
co-simulation and Test
Refined
Design
System-Level System-Level
Specification Integration
and Test
Battery
Component
Engine Design
Energy Management

Subsystem Subsystem
Design Integration
Simulink
Plant Model
and Test
Production
C-Code Subsystem
Implementation
4. Convert to C-Code
Enables reuse of work through
end of the development process

Production Simulink
Define C-Code Plant Model Complete
Requirements Integration
and Test

System-Level System-Level
Specification Integration
Battery
and Test
Engine

Energy Management

Subsystem Subsystem
Design Integration
Simulink
Plant Model
and Test
Production
C-Code Subsystem
Implementation
In summary, Model-Based Design:
• Uses a system-level simulation model to
mathematically describe complex behavior
Multi-domain
Subsystem
• Treats the Model as a single source of truth for Modelling
capturing requirements
System- Continuous
level Model Test and
• Provides a solution for designing and integrating Simulation Verification
multi-domain subsystems

• Lets you test and verify continuously throughout Automatic

Code Generation
the development process

• Automatically generates code from the model for

real-time testing and implementation
Different levels of model fidelity for
Ordinary Differential Equations (ODEs)
subsystem design Differential Algebraic Equations (DAEs)

Partial Differential Equations (PDEs)

PDE Toolbox or,

3rd Party tools for FEA/CFD
etc.
Simscape Overview
• Enables physical modeling (acausal)
of multidomain physical systems
• Assemble a schematic
• Equations derived automatically
• Leverage MATLAB and Simulink

• Foundation library of physical elements

• Mechanical, hydraulic, electrical…

• Use with other Simscape add-on products

to model more complex systems
Simscape Add-on Libraries

• Simscape Power Systems

• Three-phase electrical networks

• Simscape Electronics
• Actuators, sensors, and semiconductors

• Simscape Fluids
• Pumps, actuators, pipelines, valves, tanks

• Simscape Multibody
• Multibody systems: joints, bodies, frames

• Simscape Driveline
• Gears, leadscrew, clutches, tires, engines