©20
05 T
he M
athW
orks
, Inc
.
Modeling a Hybrid Electric Vehicle and Controller to Optimize System Performance
Steve Miller
Technical Marketing, Physical Modeling Tools
The MathWorks GmbH, Munich, Germany
Root Locus Bode Plot
Real Axis Frequency
Controller+-Ref.
Speed
Controller+-
Ref. Voltage
2
Key Results
1) Integrating the requirements in model and simulation is critical for effective development
2) Effective use of control design toolsand optimization algorithms improves overall design
3) Simulating plant and controller in one tool allows engineers to understand and optimize performance of the entire system
Root Locus Bode Plot
Real Axis Frequency
3
Agenda
Introduction to HEV Model 5 minImproving Development Process 5 min
– Linking Design and Specification– Comparing Performance and Specification
Optimizing HEV System 15 min– Linearized Model– Nonlinear Model
Questions and Answers 5 min
4
System Optimization Challenge
HEV with two PI controllers– Speed Controller, Voltage Controller
Must meet performance requirementsConflicting goals
– Improving speed response degradesvoltage bus response
Command
Voltage Bus
VoltageDC-DC
Converter
Duty Cycle
+- ControllerCommand
Speed Vehicle
SpeedCommand
Torque
+-Controller
5
HEV Electromechanical System
Battery
SimPowerSystemsSimulink
SimPowerSystemsSimulink
SimDrivelineSimscapeSimPowerSystemsSimulink
BrakingChopper
Three-PhaseBridge
6
Simscape
Extension of Simulink® designed to model multidomain physical systemsEases process of modeling physical systems
– Does not require deriving and programming the equations of motion for the system
Used by system engineers and control engineers to build a model representing the physical structure of the system
V+
V-
MATLAB®, Simulink®
Sim
Pow
erS
yste
ms
Simscape
Sim
Mec
hani
cs
Sim
Driv
elin
e
Sim
Hyd
raul
ics®
P TT
A B
7
Electrical System in SimPowerSystems
Permanent Magnet Synchronous Motor Drive
– Speed controller– Vector controller– Inverter– Permanent magnet
synchronous motor– Braking chopper
Implemented throughSimPowerSystems in the Simulink environment
SimPowerSystems Blocks
8
Mechanical System in SimDriveline
Transmission System– Input Torque– Transmission– Vehicle Inertia
Implemented in SimDriveline within the Simulink environment
TransmissionVehicleInertia
InputTorque
WheelSpeed
SimDriveline Blocks
9
Agenda
Introduction to HEV Model 5 minImproving Development Process 5 min
– Linking Design and Specification– Comparing Performance and Specification
Optimizing HEV System 15 min– Linearized Model– Nonlinear Model
Questions and Answers 5 min
10
Linking Specification and Design
Problem: Matching design tospecification is difficult.
Solution: Use SimulinkVerification and Validation to directlylink the design to the specification.
Situation:
Part Numbers
HEV Requirements1. Drivetrain SystemLayout
11
Comparing Specification and Design
Problem: Measuring design performance relative to specification is difficult.
Solution: Use Simulink Verificationand Validation to automatically compareperformance to specification.
Situation:
HEV Requirements1. Drivetrain System
Wheel Speed5% Setting Timein 0.5 Sec
12
Improved Development Process
Advantages of improved process
1) Enables quick and easy associaton of specification and model
2) Automatically checks design against specification
3) Supports Model-Based Design, helping you discover and solve problems earlier in the design process
13
Agenda
Introduction to HEV Model 5 minImproving Development Process 5 min
– Linking Design and Specification– Comparing Performance and Specification
Optimizing HEV System 15 min– Linearized Model– Nonlinear Model
Questions and Answers 5 min
14
Possibilities for System Optimization
Nonlinear model– Specify response
characteristics– Automatic tuning via
Simulink Response Optimization
Linearized model– Linearize system via Simulink
Control Design– Perform linear control design with
Control System Toolbox andSimulink Response Optimization
– Retest in nonlinear system
Controller+-
Controller+-
A x + B uRoot Locus Bode Plot
Real Axis Frequency
15
System Optimization On Linear Plant
Problem: Design and tune thetwo controllers in this system tomeet system requirements.
Solution: Use Simulink Control Design, Control System Toolbox todesign, tune, and test the controller
Model:
Controller+-Ref.
Speed
Controller+-
Ref. Voltage
16
Steps to Design Controller1) Identify control loops
of interest
2) Identify operating point
3) Linearize model about this point
4) Perform control design
5) Test controller in nonlinear system
A x + B u = 0Comman
dVolta
ge
Vehicl
eSpee
d
+-
+-
ControllerController
Root Locus Bode Plot
Real Axis Frequency
System Optimization On Linear Plant
Command
Speed
BusVolta
ge
17
System Optimization On Linear Plant
Advantages of Simulink Control Design
1) Enable easy application of linear control theoryOperating points from specification or simulationGraphical design with interactive plots
2) Rapid evaluation of designs with interactive analysis plots
3) Optimize performance based on time, frequency, or root locus constraints
18
System Optimization On Nonlinear Plant
Problem: Design and tune thecontroller in this system tomeet system requirements.
Solution: Use Simulink Response Optimization to design, tune, and test the controller
Model:
Kp Ki Kp Ki
3.0 62.80.033
R
0.001 0.08(Kps+Ki)s
Controller+-Ref.
Speed
Controller+-
Ref. Voltage
(Kps+Ki)s
Voltage
Speed
Voltage Speed Motor
Kp Ki Kp Ki
2.2 75.10.023
R
0.003 0.07
Kp Ki Kp Ki
1.9 83.90.014
R
0.005 0.06
Kp Ki Kp Ki
1.5 95.70.001
R
0.006 0.05
Kp Ki Kp Ki
1.3 1200.001
R
0.007 0.05
19
Control Design On Nonlinear Plant
Steps to Optimizing Response
1) Identify parameters to be tunedand their ranges
2) Specify desired response
3) Perform response optimization
(Kps+Ki)s
20
Control Design On Nonlinear Plant
Advantages of Simulink Response Optimization
1) Simulating plant and controller in one tool allows engineers to understand and optimize performance of the entire system
2) Automatic tuning of parameters saves time
3) Graphical interface makes it easy to map specification to tests
21
Summary
1) Integrating the requirements in model and simulation is critical for effective development
2) Effective use of control design toolsand optimization algorithms improves overall design
3) Simulating plant and controller in one tool allows engineers to understand and optimize performance of the entire system
Root Locus Bode Plot
Real Axis Frequency
22
Agenda
Introduction to HEV Model 5 minImproving Development Process 5 min
– Linking Design and Specification– Comparing Performance and Specification
Optimizing HEV System 15 min– Linearized Model– Nonlinear Model
Questions and Answers 5 min
23
MathWorks Physical Modeling Tools Booth(HEV Example, Other Tools)
Series-Parallel architecture usingSimscape, SimDriveline, SimPowerSystems
24
Physical Modeling Master Class(Wednesday, 10:30 – 12:00)
Model and simulate multidomain physical systems all within the Simulink environment
Optimize system performanceby developing the plant and controller in one environment
Use Model-Based Design on your entire system to improveyour development process
s1 s2
s3
ProductionC-Code
SimulinkPlant Model
©20
05 T
he M
athW
orks
, Inc
.
Thank You ForYour Attention!
Steve Miller
Technical Marketing, Physical Modeling Tools
The MathWorks GmbH, Munich, Germany