Simulink Design Optimization 1 Getting Started Guide

Downloaded from www.Manualslib.com manuals search engine

How to Contact The MathWorks

Web Newsgroup www.mathworks.com/contact_TS.html Technical Support

www.mathworks.com comp.soft-sys.matlab suggest@mathworks.com bugs@mathworks.com doc@mathworks.com service@mathworks.com info@mathworks.com

Product enhancement suggestions Bug reports Documentation error reports Order status, license renewals, passcodes Sales, pricing, and general information

508-647-7000 (Phone) 508-647-7001 (Fax) The MathWorks, Inc. 3 Apple Hill Drive Natick, MA 01760-2098
For contact information about worldwide offices, see the MathWorks Web site. Simulink Design Optimization Getting Started Guide COPYRIGHT 19932010 by The MathWorks, Inc.
The software described in this document is furnished under a license agreement. The software may be used or copied only under the terms of the license agreement. No part of this manual may be photocopied or reproduced in any form without prior written consent from The MathWorks, Inc. FEDERAL ACQUISITION: This provision applies to all acquisitions of the Program and Documentation by, for, or through the federal government of the United States. By accepting delivery of the Program or Documentation, the government hereby agrees that this software or documentation qualifies as commercial computer software or commercial computer software documentation as such terms are used or defined in FAR 12.212, DFARS Part 227.72, and DFARS 252.227-7014. Accordingly, the terms and conditions of this Agreement and only those rights specified in this Agreement, shall pertain to and govern the use, modification, reproduction, release, performance, display, and disclosure of the Program and Documentation by the federal government (or other entity acquiring for or through the federal government) and shall supersede any conflicting contractual terms or conditions. If this License fails to meet the governments needs or is inconsistent in any respect with federal procurement law, the government agrees to return the Program and Documentation, unused, to The MathWorks, Inc.

Trademarks

MATLAB and Simulink are registered trademarks of The MathWorks, Inc. See www.mathworks.com/trademarks for a list of additional trademarks. Other product or brand names may be trademarks or registered trademarks of their respective holders.
Patents

The MathWorks products are protected by one or more U.S. patents. Please see www.mathworks.com/patents for more information.
Revision History

March 2009 Online only September 2009 Online only March 2010 Online only

New for Version 1 (Release 2009a) Revised for Version 1.1 (Release 2009b) Revised for Version 1.1.1 (Release 2010a)

Downloaded from www.Manualslib.com manuals search engine

Contents
Product Overview

1
What You Can Accomplish Using This Product . . . . . . . Learn More . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Required and Related Products . . . . . . . . . . . . . . . . . . . . . Documentation and Demos . . . . . . . . . . . . . . . . . . . . . . . . . Accessing Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . Accessing Demos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2 1-4 1-5 1-6 1-6 1-6

Parameter Estimation

2
Types of Data for Parameter Estimation . . . . . . . . . . . . . Quick Start Estimating Model Parameters . . . . . . . . . Parallel Computing for Parameter Estimation . . . . . . . 2-2 2-3 2-11

Parameter Optimization

3
Types of Time-Domain Design Requirements for Optimizing Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . Quick Start Optimizing Model Parameters 3-2 3-3

........

iii

Downloaded from www.Manualslib.com manuals search engine

Parallel Computing for Parameter Optimization . . . . .

3-11

Optimization-Based Linear Control Design

4
When to Use Optimization-Based Linear Control Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of Time- and Frequency-Domain Design Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quick Start Optimization-Based Linear Control Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2

4-3

4-4

Tutorial Preparing Data for Parameter Estimation Using the GUI

5
About This Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . About the Sample Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Configuring a Project for Parameter Estimation 5-2 5-2 5-2 5-4 5-6 5-6 5-11 5-14 5-16 5-16 5-22

.....

Importing Data into the GUI . . . . . . . . . . . . . . . . . . . . . . . . Importing Input Data and Time Vector . . . . . . . . . . . . . . . . Importing Output Data and Time Vector . . . . . . . . . . . . . . Analyzing Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selecting Data for Estimation . . . . . . . . . . . . . . . . . . . . . . . Selecting Output Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selecting Input Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

iv

Contents

Downloaded from www.Manualslib.com manuals search engine

Removing Outliers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Why Remove Outliers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . How to Remove Outliers . . . . . . . . . . . . . . . . . . . . . . . . . . . . Filtering Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Filtering Output Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Filtering Input Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interpolating Missing Data . . . . . . . . . . . . . . . . . . . . . . . . . Saving the Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5-25 5-25 5-25 5-29 5-29 5-32 5-34 5-37

Tutorial Estimating Parameters from Measured Data Using the GUI

6
About This Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . About the Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2 6-2 6-3

Estimating Model Parameters Using Default Estimation Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-7 Overview of the Estimation Process . . . . . . . . . . . . . . . . . . . 6-7 Specifying Parameters and Estimation Data . . . . . . . . . . . 6-8 Validating Model Parameters . . . . . . . . . . . . . . . . . . . . . . . 6-13 Improving Estimation Results Using Parameter Bounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Strategy for Improving the Estimation Results . . . . . . . . . How to Specify Parameter Bounds . . . . . . . . . . . . . . . . . . . . Validating Estimated Model Parameters . . . . . . . . . . . . .

6-20 6-20 6-20 6-26

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

7
About This Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . About the Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Design Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Configuring a Model for Optimizing Parameters 7-2 7-2 7-2 7-4 7-5

.....

Optimizing Model Parameters to Meet Step Response Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Specify Time-Domain Design Requirements . . . . . . . . . . . . Specifying Parameters to Optimize . . . . . . . . . . . . . . . . . . . Optimizing the Parameters . . . . . . . . . . . . . . . . . . . . . . . . . Refining Model Parameters to Track a Reference Signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Saving the Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7-8 7-8 7-18 7-21

7-25 7-31

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line

8
About This Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . About the Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Design Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Configuring a Model for Optimizing Parameters 8-2 8-2 8-2 8-4 8-5

.....

Optimizing Model Parameters to Meet Step Response Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8-8

vi

Contents

Downloaded from www.Manualslib.com manuals search engine

Refining Model Parameters to Track a Reference Signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8-15

Tutorial Designing a PID Controller Using Optimization-Based Tuning

9
About This Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . About the Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Design Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Configuring a Project for Optimization-Based Control Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements . . . . . . Specifying the Controller Parameters . . . . . . . . . . . . . . . . . Specifying Bode Magnitude and Phase Margin Design Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Designing the Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . Refining the Controller Design to Meet Controller Output Bounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Saving the Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-2 9-2 9-2 9-4

9-5

9-11 9-11 9-15 9-25

9-32 9-48

Examples

A
Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2

vii

Downloaded from www.Manualslib.com manuals search engine

Index

viii

Contents

Downloaded from www.Manualslib.com manuals search engine

1
Product Overview
What You Can Accomplish Using This Product on page 1-2 Learn More on page 1-4 Required and Related Products on page 1-5 Documentation and Demos on page 1-6

Downloaded from www.Manualslib.com manuals search engine

Product Overview

What You Can Accomplish Using This Product

Simulink Design Optimization software is a Simulink-based product that lets you estimate and optimize model parameters using numerical optimization. You can also use this software to estimate initial conditions and lookup table values, and test and optimize designs for robustness. Simulink Design Optimization software also supports parameter optimization for Simulink models that invoke third-party simulation tools or contain legacy simulation code (using S-functions). This software uses methods from the Optimization Toolbox and Global Optimization Toolbox products for design optimization. Using the software with the Parallel Computing Toolbox software can speed up numerical computations during estimation and optimization. The software provides library blocks that let you accomplish the following goals: Model time-varying systems as lookup tables using Adaptive Lookup Table blocks. Optimize model parameters by graphically specifying time-domain design requirements and reference signal using Signal Constraint blocks. Compute and optimize continuous and discrete root mean square values of signals using CRMS and DRMS blocks with Signal Constraint blocks. When the Control System Toolbox software is installed, you can refine controller parameters of linear time-invariant (LTI) models and Simulink models, linearized using the Simulink Control Design software, in the SISO Design Tool. You can work either in the Graphical User Interface (GUI) or at the command line. New users should start by using the GUI to become familiar with the product. The following operations are available only using the GUI: Interactive data preprocessing (see Data Analysis and Processing). Model validation using residual plot (see Comparing Residuals).

1-2

Downloaded from www.Manualslib.com manuals search engine

What You Can Accomplish Using This Product

Optimization-based controller design (see Optimization-Based Linear Control Design).

1-3

Downloaded from www.Manualslib.com manuals search engine

Product Overview

Learn More
The Simulink Design Optimization documentation provides information to use this product. Although this product employs optimization methods for parameter estimation and optimization, using the product does not require you to have a strong background in optimization theory. You may find it helpful to consult the Optimization Toolbox documentation to learn more about optimization theory, and minimizing an objective function.

1-4

Downloaded from www.Manualslib.com manuals search engine

Required and Related Products

Required and Related Products

Simulink Design Optimization software requires MATLAB, Simulink, and Optimization Toolbox software. The following table summarizes MathWorks products that extend and complement the Simulink Design Optimization software. For current information about these and other MathWorks products, visit http://www.mathworks.com/products/product_listing/index.html. Product Control System Toolbox Description Enables you to design controllers for linear time-invariant (LTI) models using optimization methods. Provides genetic algorithms, and direct search methods to estimate and optimize model parameters. Provides Simulink models of neural networks for optimization-based control design. Enables parallel computing on multicore processors and multiprocessor networks to speed up estimation and optimization. Lets you linearize Simulink models. Use Simulink Design Optimization software to design controllers for linearized models using optimization methods. Lets you estimate linear and nonlinear models from measured data. Import the estimated model into Simulink software, and use Simulink Design Optimization software for optimization-based control design.

Global Optimization Toolbox

Neural Network Toolbox

Parallel Computing Toolbox

Simulink Control Design

System Identification Toolbox

1-5

Downloaded from www.Manualslib.com manuals search engine

Product Overview

Documentation and Demos

In this section... Accessing Documentation on page 1-6 Accessing Demos on page 1-6

Accessing Documentation
The Simulink Design Optimization documentation contains the following components: Getting Started Guide Provides information for mapping your problem to the capabilities of the Simulink Design Optimization software. Step-by-step tutorials walk you through the most common tasks for estimating parameters, optimizing parameters, and designing controllers using optimization methods. Users Guide Describes estimation and optimization tasks for using the Simulink Design Optimization software. Reference Describes commands and blocks for design optimization. Release Notes Describes important changes in the current product version and compatibility considerations. If you are new to using this product, the Getting Started Guide helps you begin using this product quickly. You can follow the steps in the tutorials to perform design optimization using the graphical user interface (GUI) or the MATLAB Command Window. You can also search or browse the documentation for information about specific design optimization tasks.

Accessing Demos
The Simulink Design Optimization software provides demo files that show you how to estimate and optimize parameters of Simulink models, design compensators using optimization methods, and model systems using Adaptive Lookup Tables.

1-6

Downloaded from www.Manualslib.com manuals search engine

Documentation and Demos

To access demos in the Help browser, type the following command at the MATLAB prompt:
demo('Simulink', 'Simulink Design Optimization')

1-7

Downloaded from www.Manualslib.com manuals search engine

Product Overview

1-8

Downloaded from www.Manualslib.com manuals search engine

2
Parameter Estimation
Types of Data for Parameter Estimation on page 2-2 Quick Start Estimating Model Parameters on page 2-3 Parallel Computing for Parameter Estimation on page 2-11

Downloaded from www.Manualslib.com manuals search engine

Parameter Estimation

Types of Data for Parameter Estimation

You can estimate model parameters and initial conditions of single or multiple input and output Simulink models from transient data. You measure transient data when the system is not in steady-state to capture the system dynamics expected under normal operating conditions. For example, the response of a system to step or impulse inputs is transient data. Simulink Design Optimization software lets you estimate model parameters from the following types of data: Time-domain data Data with one or more input variables u(t) and one or more output variables y(t), sampled as a function of time. See Importing Data into the GUI. Time-series data Data stored in time-series objects. For more information, see Time Series Objects in the MATLAB documentation. See Importing Time-Series Data into the GUI. Using complex data for parameter estimation is not directly supported. See Importing Complex Data into the GUI. Simulink Design Optimization software estimates model parameters by comparing the transient data with simulation data generated from the Simulink model. Using optimization techniques, the software estimates the parameters and initial conditions of states to minimize a user-selected cost function. The cost function typically calculates a least-square error between the measured and simulated data. To learn more, see Parameter Estimation in the Simulink Design Optimization Users Guide.

2-2

Downloaded from www.Manualslib.com manuals search engine

Quick Start Estimating Model Parameters

Quick Start Estimating Model Parameters

In this quick start, you get an overview of the typical tasks for estimating model parameters using the Control and Estimation Tools Manager GUI:
1 Start a parameter estimation task. 2 Import estimation and validation data sets. 3 Select parameters to estimate. 4 Estimate the parameters from the estimation data. 5 Validate the estimated parameters using the validation data set.

Prerequisites for parameter estimation include: Simulink model that contains inport or outport blocks, or signal logging For more information, see Configuring a Model for Importing Data in the Simulink Design Optimization Users Guide. Transient data in the MATLAB workspace To estimate model parameters:
1 Start a parameter estimation task by selecting Tools > Parameter

Estimation in the Simulink model window.

2-3

Downloaded from www.Manualslib.com manuals search engine

Parameter Estimation

New estimation task

Optional information about the estimation task

2-4

Downloaded from www.Manualslib.com manuals search engine

Quick Start Estimating Model Parameters

2 Import the input and output data for estimating and validating model

parameters.
a Select the Transient Data node, and click New. b Select the New Data node. c In the Input Data tab, select the Data cell corresponding to the model

channel, and click Import. Select the variable to import in the Data Import dialog box.
Model channel for which you import data Select and Select and click Import to click Import to import data import time vector

Select variable to import

Import variable

d Select the Time /Ts cell, and click Import. Select the time vector to

import in the Data Import dialog box.

e In the Output Data tab, repeat steps cd to import the output data

and time vector.

2-5

Downloaded from www.Manualslib.com manuals search engine

Parameter Estimation

Repeat steps ad to import the validation data set.

Validation data node

For more information, see Importing Data into the GUI in the Simulink Design Optimization Users Guide.
3 Specify parameters to estimate by selecting the Variables node, and

clicking Add. Select the parameters in the Select Parameters dialog box.
Select parameters to estimate

Specify parameters

2-6

Downloaded from www.Manualslib.com manuals search engine

Quick Start Estimating Model Parameters

For more information, see Specifying Parameters to Estimate in the Simulink Design Optimization Users Guide.
4 Estimate the parameters. a Select the Estimation node, and click New. b Select the New Estimation node. c In the Data Sets tab, select the estimation data set.
Lists imported data sets Select data set for estimation

2-7

Downloaded from www.Manualslib.com manuals search engine

Parameter Estimation

d In the Parameters tab, select the parameters to estimate.

Keep parameters fixed or estimate (Optional) Specify initial parameter value (Optional) Specify parameter bounds

e In the Estimation tab, begin estimation by clicking Start.

Information about convergence of the estimation algorithm

Start estimation Select to view measured and simulated response plots Estimation report

2-8

Downloaded from www.Manualslib.com manuals search engine

Quick Start Estimating Model Parameters

In the Parameters tab, examine the estimated parameter values. The Simulink model also gets updated with the estimated parameter values.
Estimated parameter values

For more information, see Estimating Parameters in the GUI in the Simulink Design Optimization Users Guide.
5 Validate the estimated parameters. a Select the Validation node, and click New. b Select the New Validation node.

2-9

Downloaded from www.Manualslib.com manuals search engine

Parameter Estimation

c Configure the validation plots and the validation data set.

Select validation Select validation plot type data set

Select to Opens display validation plots validation plot

For more information, see Validating Parameters in the GUI in the Simulink Design Optimization Users Guide. See Also: Chapter 6, Tutorial Estimating Parameters from Measured Data Using the GUI

2-10

Downloaded from www.Manualslib.com manuals search engine

Parallel Computing for Parameter Estimation

Parallel Computing for Parameter Estimation

When you have the Parallel Computing Toolbox software, you can use parallel computing to speed up parameter estimation. When you use parallel computing, the software distributes the independent simulations on multiple MATLAB sessions. Thus, the simulations run in parallel which reduces the estimation time. Using parallel computing may reduce the estimation time in the following cases: The model contains a large number parameters to estimate, and Gradient descent or Nonlinear least squares is selected as the estimation method. Pattern search is selected as the estimation method. The model is complex and takes a long time to simulate. For more information, see Speeding Up Parameter Estimation Using Parallel Computing in the Simulink Design Optimization Users Guide.

2-11

Downloaded from www.Manualslib.com manuals search engine

Parameter Estimation

2-12

Downloaded from www.Manualslib.com manuals search engine

3
Parameter Optimization
Types of Time-Domain Design Requirements for Optimizing Parameters on page 3-2 Quick Start Optimizing Model Parameters on page 3-3 Parallel Computing for Parameter Optimization on page 3-11

Downloaded from www.Manualslib.com manuals search engine

Parameter Optimization

Types of Time-Domain Design Requirements for Optimizing Parameters

You can optimize parameters of Simulink models to meet the following types of time-domain design requirements: Step-response characteristics such as overshoot, and rise time. Lower and upper bounds on signals Reference signal Simulink Design Optimization software optimizes the model parameters by formulating the time-domain requirements into a constrained optimization problem. It then solves the problem using optimization methods. During the optimization, the software performs the following operations: Simulates the Simulink model, Compares the simulation data with the constraint objectives and any specified reference signal Uses gradient methods to modify selected model parameters to meet the objectives To learn more, see Parameter Optimization in the Simulink Design Optimization Users Guide.

3-2

Downloaded from www.Manualslib.com manuals search engine

Quick Start Optimizing Model Parameters

Quick Start Optimizing Model Parameters

In this quick start, you get an overview of the typical tasks for optimizing model parameters to meet time-domain requirements:
1 Specify an input signal in the Simulink system. 2 Specify the design requirements. 3 Specify parameters to optimize. 4 Optimize the parameters. 5 Evaluate the optimization results.

Prerequisites for optimizing model parameters include: Simulink model Time-domain design requirements To optimize model parameters:
1 In the Simulink model, specify an input signal to the system. For example,

add a Step block.

2 Specify the time-domain design requirements:

3-3

Downloaded from www.Manualslib.com manuals search engine

Parameter Optimization

a In the Simulink Library Browser, select Simulink Design

Optimization.
b Drag and drop the Signal Constraint block into the model. c Connect the Signal Constraint block to the signal that should meet the

design requirements.

3-4

Downloaded from www.Manualslib.com manuals search engine

Quick Start Optimizing Model Parameters

d Double-click the Signal Constraint block.

Design requirements appear as line segments in the Block Parameters: Signal Constraint block window. By default, the design requirements are step-response characteristics.

3-5

Downloaded from www.Manualslib.com manuals search engine

Parameter Optimization

e Double-click the lower yellow region on the plot. Specify the design

requirements in the Edit Design Requirement dialog box.

Select design requirement type

Enter design requirement values

For more information, see Specifying Design Requirements in the Simulink Design Optimization Users Guide.

3-6

Downloaded from www.Manualslib.com manuals search engine

Quick Start Optimizing Model Parameters

3 In the Block Parameters window, select Optimization > Tuned

Parameters, and click Add. Select the parameters to optimize in the Add Parameters dialog box.
Lists parameters selected for optimization Select parameters

Specify parameters to optimize

. For more information, see Specifying Parameters to Optimize in the Simulink Design Optimization Users Guide.
4 In the Block Parameters window, start the optimization by selecting

Optimization > Start. The Optimization Progress window opens where you see the optimization progress.

3-7

Downloaded from www.Manualslib.com manuals search engine

Parameter Optimization

For more information, see Running the Optimization in the Simulink Design Optimization Users Guide.

3-8

Downloaded from www.Manualslib.com manuals search engine

Quick Start Optimizing Model Parameters

5 Evaluate the optimization results after the optimization completes. a In the Block Parameters window, compare the response of the system

against the design requirements.

Optimized response shown in black

Initial response shown in blue

3-9

Downloaded from www.Manualslib.com manuals search engine

Parameter Optimization

b In the Optimization Progress window, view the optimized parameter

values.

Optimization report

Optimized parameter values

See Also: Chapter 7, Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI.

3-10

Downloaded from www.Manualslib.com manuals search engine

Parallel Computing for Parameter Optimization

Parallel Computing for Parameter Optimization

When you have the Parallel Computing Toolbox software, you can use parallel computing to speed up optimizing model parameters to meet time-domain design requirements. When you use parallel computing, the software distributes the independent simulations on multiple MATLAB sessions. Thus, the simulations run in parallel which reduces the optimization time. Using parallel computing may reduce the optimization time in the following cases: The model contains a large number of parameters to optimize, and the Gradient descent method is selected for optimization. ThePattern search method is selected for optimization. The model contains a large number of uncertain parameters and uncertain parameter values. The model is complex and takes a long time to simulate. For more information, see Speeding Up Response Optimization Using Parallel Computing in the Simulink Design Optimization Users Guide.

3-11

Downloaded from www.Manualslib.com manuals search engine

Parameter Optimization

3-12

Downloaded from www.Manualslib.com manuals search engine

4
Optimization-Based Linear Control Design
When to Use Optimization-Based Linear Control Design on page 4-2 Types of Time- and Frequency-Domain Design Requirements on page 4-3 Quick Start Optimization-Based Linear Control Design on page 4-4

Downloaded from www.Manualslib.com manuals search engine

Optimization-Based Linear Control Design

When to Use Optimization-Based Linear Control Design

When you have Control System Toolbox software installed, you can design and optimize control systems for LTI models by optimizing controller parameters in the SISO Design Tool. To use optimization methods for linear control design, also known as optimization-based tuning, you must already have an initial controller. You can then use optimization-based tuning to refine the controller design to meet additional design requirements. For more information on designing controllers, see the Control System Toolbox documentation. Note Optimization-based tuning only changes the value of the controller parameters and not the controller structure itself. Optimization-based tuning provides flexibility in terms of specifying additional design requirements for the controller. When you have a large number of design requirements, you can first design an initial controller by selecting a subset of requirements and subsequently select additional requirements to refine the design. Optimization-based tuning also provides flexibility in terms of selecting a subset of controller parameters to optimize, and specifying bounds on the controller parameters. To design linear controllers for Simulink models using optimization-based tuning, you must first linearize the model using the Simulink Control Design software. For more information on linearizing Simulink models, see the Simulink Control Design documentation.

4-2

Downloaded from www.Manualslib.com manuals search engine

Types of Time- and Frequency-Domain Design Requirements

Types of Time- and Frequency-Domain Design Requirements

When you design linear controllers for LTI or Simulink models using the Simulink Design Optimization software, you can specify both time- and frequency-domain requirements on the system response. You can specify design requirements on the following plots: Root Locus plot Open-Loop and Prefilter Bode plots Open-Loop Nichols plot Step/Impulse Response plots For more information, see Supported Time- and Frequency-Domain Requirements in the Simulink Design Optimization Users Guide. Simulink Design Optimization software uses the frequency-domain requirements to compute the frequency response of the system. It then uses optimization methods to reduce the distance between the current response and the requirements by modifying the controller parameters. The software does not change the controller structure when optimizing the controller parameters. To learn more, see Optimization-Based Linear Control Design in the Simulink Design Optimization Users Guide.

4-3

Downloaded from www.Manualslib.com manuals search engine

Optimization-Based Linear Control Design

Quick Start Optimization-Based Linear Control Design

In this quick start, you get an overview of the typical tasks for optimization-based linear control design using the SISO Design Tool:
1 Open a SISO Design Tool session. 2 Configure a project for optimization-based control design. 3 Specify the controller parameters to design. 4 Specify the design requirements. 5 Design the controller. 6 Evaluate the controller design.

Note The same workflow applies to optimization-based control design for LTI models created at the command line using Control System Toolbox software. To learn how to create LTI models, see Linear (LTI) Models in the Control System Toolbox documentation. Prerequisites for optimization-based linear control design include: Simulink Compensator Design Task that contains a linearized version of the Simulink model and, optionally, any response plots you configure. For more information on how to linearize a Simulink model for control design, see Designing Compensators in the Simulink Control Design documentation. Time- and frequency-domain design requirements To design a controller using optimization methods:

4-4

Downloaded from www.Manualslib.com manuals search engine

Quick Start Optimization-Based Linear Control Design

1 Open a SISO Design Tool session by typing the following command at the

MATLAB prompt:
sisotool(projectname.mat)

SISO Design Task for SISO Design Tool

The command also opens a SISO Design for SISO Design Task window by default and any response plots you configured when you linearized the Simulink model using Simulink Control Design software.

4-5

Downloaded from www.Manualslib.com manuals search engine

Optimization-Based Linear Control Design

2 Configure a project for optimization-based control design by clicking

Optimize Compensators in the Automated Tuning tab of the SISO Design Task.

Method for controller design

4-6

Downloaded from www.Manualslib.com manuals search engine

Quick Start Optimization-Based Linear Control Design

This action creates a new Response Optimization node in the Control and Estimation Tools Manager.

4-7

Downloaded from www.Manualslib.com manuals search engine

Optimization-Based Linear Control Design

3 Specify the controller parameters to design in the Compensators tab.

Right-click to change Select parameters controller parameters (Optional) Specify to optimize representation to Simulink initial value (Optional) Specify block mask parameters parameter bounds

4-8

Downloaded from www.Manualslib.com manuals search engine

Quick Start Optimization-Based Linear Control Design

4 Specify the design requirements. a In the Design requirements tab, click Add new design requirement.

Specify the design requirements, for example Bode magnitude lower limit, in the New Design Requirement dialog box.

Requirement type Response type

Requirement values

Add requirement

4-9

Downloaded from www.Manualslib.com manuals search engine

Optimization-Based Linear Control Design

In the SISO Design window, the yellow region with the black line segment represents the design requirement on the response plot.

4-10

Downloaded from www.Manualslib.com manuals search engine

Quick Start Optimization-Based Linear Control Design

The Design Requirements tab also lists the design requirement.

Select the requirement to enforce during optimization Lists all specified design requirements

b Repeat step a to specify additional time- and frequency-domain

requirements.

4-11

Downloaded from www.Manualslib.com manuals search engine

Optimization-Based Linear Control Design

5 Start the optimization to design the controller by clicking Start

Optimization in the Optimization tab.

Optimization progress

Optimization status

4-12

Downloaded from www.Manualslib.com manuals search engine

Quick Start Optimization-Based Linear Control Design

6 Evaluate the controller design. a Examine the systems response in the response plot, for example the

Bode plot, to see if it meets the requirements. The systems response must lie in the white region in order to meet the design requirement.

4-13

Downloaded from www.Manualslib.com manuals search engine

Optimization-Based Linear Control Design

b Examine the controller parameter values in the Compensator tab.

Optimized controller parameter values

7 Write the controller parameter values into the Simulink model. To do so,

click Update Simulink Block Parameters in the SISO Design Task node. See Also: Chapter 9, Tutorial Designing a PID Controller Using Optimization-Based Tuning.

4-14

Downloaded from www.Manualslib.com manuals search engine

5
Tutorial Preparing Data for Parameter Estimation Using the GUI
About This Tutorial on page 5-2 Configuring a Project for Parameter Estimation on page 5-4 Importing Data into the GUI on page 5-6 Analyzing Data on page 5-14 Selecting Data for Estimation on page 5-16 Removing Outliers on page 5-25 Filtering Data on page 5-29 Interpolating Missing Data on page 5-34 Saving the Project on page 5-37

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

About This Tutorial

In this section... Objectives on page 5-2 About the Sample Data on page 5-2

Objectives
In this tutorial, you learn how to import, analyze, and prepare measured input and output (I/O) data for estimating parameters of a Simulink model. Note Simulink Design Optimization software estimates parameters from real, time-domain data only. You learn to perform the following tasks using the GUI: Import data from the MATLAB workspace. Analyze data quality using a time plot. Select a subset of data for estimation. Remove outliers. Filter high-frequency noise. Compute missing data using interpolation.

About the Sample Data

In this tutorial, you use spe_engine_throttle1.mat, which contains I/O data measured from an engine throttle system. The MAT-file includes the following variables: input1 Input data samples position1 Output data samples time1 Time vector

5-2

Downloaded from www.Manualslib.com manuals search engine

About This Tutorial

Note The number of input and output data samples must be equal to the length of the corresponding time vector. The engine throttle system controls the flow of air and fuel mixture to the engine cylinders. The throttle body contains a butterfly valve which opens when a driver presses the accelerator pedal. Opening this valve increases the amount of fuel mixture entering the cylinders, which increases the engine speed. A DC motor controls the opening angle of the butterfly valve in the throttle system. The input to the throttle system is the motor current (in amperes), and the output is the angular position of the butterfly valve (in degrees).
spe_engine_throttle1.mdl contains the Simulink model of the engine throttle system. For more information on building models, see Creating a Simulink Model in the Simulink documentation.

5-3

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

Configuring a Project for Parameter Estimation

To perform parameter estimation, you must first configure a Control and Estimation Tools Manager project.
1 Open the engine throttle system model by typing the following at the

MATLAB prompt:
spe_engine_throttle1

This command opens the Simulink model, and loads the data into the MATLAB workspace.

Simulink Model of Engine Throttle System

5-4

Downloaded from www.Manualslib.com manuals search engine

Configuring a Project for Parameter Estimation

2 In the Simulink model window, select Tools > Parameter Estimation.

This action opens a new project named Project - spe_engine_throttle1 in the Control and Estimation Tools Manager GUI. This project contains the Estimation Task, as shown in the next figure.

Note The Simulink model must remain open to perform parameter estimation tasks.

5-5

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

Importing Data into the GUI

In this section... Importing Input Data and Time Vector on page 5-6 Importing Output Data and Time Vector on page 5-11

Importing Input Data and Time Vector

In this portion of the tutorial, you import measured I/O data into the Control and Estimation Tools Manager GUI. Importing data assigns the data to the corresponding model input and output signals. The model input and output signals are designated with the Inport Input and Outport Position blocks respectively, as shown in the figure Simulink Model of Engine Throttle System on page 5-4. These blocks let you import I/O data into the GUI. To learn more about the blocks, see the Inport and Outport block reference pages in the Simulink documentation. You must have already configured the parameter estimation project, as described in Configuring a Project for Parameter Estimation on page 5-4. To import input data and time vector into the Control and Estimation Tools Manager GUI:

5-6

Downloaded from www.Manualslib.com manuals search engine

Importing Data into the GUI

1 In the Control and Estimation Tools Manger GUI, select Transient Data

under the Estimation Task node, and click New. This action creates a New Data node under the Transient Data node.

5-7

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

2 Select the New Data node.

3 In the Input Data tab, select the Data cell for Channel - 1, and click

Import. This action opens the Data Import dialog box.

5-8

Downloaded from www.Manualslib.com manuals search engine

Importing Data into the GUI

4 In the Data Import dialog box, select input1, and click Import.

This action assigns the input data input1 to the model input signal spe_engine_throttle1/Input.

5-9

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

5 Select the Time / Ts cell for Channel - 1. 6 In the Data Import dialog box, select time1, and click Import.

This action assigns the time vector to the model input signal spe_engine_throttle1/Input.

7 In the Data Import dialog box, click Close.

5-10

Downloaded from www.Manualslib.com manuals search engine

Importing Data into the GUI

Importing Output Data and Time Vector

To import output data and time vector into the Control and Estimation Tools Manager GUI:
1 In the Control and Estimation Tools Manger GUI, select the Output Data

tab of the New Data node.

2 Select the Data cell for Channel - 1, and click Import.

This action opens the Data Import dialog box.

5-11

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

3 In the Data Import dialog box, select position1, and click Import.

This action assigns the output data position1 to the model output signal spe_engine_throttle1/Position.

5-12

Downloaded from www.Manualslib.com manuals search engine

Importing Data into the GUI

4 Select the Time / Ts cell for Channel - 1. 5 In the Data Import dialog box, select time1, and click Import.

This action assigns the time vector to the model output signal spe_engine_throttle1/Position.

6 In Data Import dialog box, click Close.

You have now imported the I/O data into the Control and Estimation Tools Manager GUI, and assigned the data to the corresponding model signals.

5-13

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

Analyzing Data
In this portion of the tutorial, you analyze the output data quality by viewing the data characteristics on a time plot. Based on the analysis, you decide whether to preprocess the data before estimating parameters. For example, if the data contains noise, you might want to filter the noise from the system dynamics before estimating parameters. You must have already imported the data into the Control and Estimation Tools Manager GUI, as described in Importing Data into the GUI on page 5-6. If you have not imported the data, click here. To plot the output data on a time plot, select the position1(:,1) cell in the Output Data tab, and click Plot Data.

5-14

Downloaded from www.Manualslib.com manuals search engine

Analyzing Data

This action plots the measured output data position1(:,1), as shown in the next figure.

The time plot shows the output data in response to a step input, as described in About the Sample Data on page 5-2. The plot shows a rapid decrease in the response after t = 0.5 s because the system is shut down. To focus parameter estimation on the time period when the system is active, you select the data samples between t = 0 s and t = 0.5 s, as described in Selecting Data for Estimation on page 5-16 section of this tutorial. The spikes in the data indicate outliers, defined as data values that deviate from the mean by more than three standard deviations. They may be caused by measurement errors or sensor problems. The response also contains noise. Before estimating model parameters from this data, you remove the outliers and filter the noise, as described in Removing Outliers on page 5-25, and Filtering Data on page 5-29 sections of this tutorial. Tip You can also plot the input data on a time plot by selecting the input1(:,1) cell in the Input Data tab, and clicking Plot Data.

5-15

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

Selecting Data for Estimation

In this section... Selecting Output Data on page 5-16 Selecting Input Data on page 5-22

Selecting Output Data

In this portion of the tutorial, you select a subset of I/O data for estimation. As described in Analyzing Data on page 5-14, the system is shut down at t = 0.5 s. To focus the estimation on the time period before t = 0.5 s, you exclude the data samples beyond t = 0.5 s. This operation selects the data between t = 0 s and t = 0.5 s for estimation. You must have already imported the data into the Control and Estimation Tools Manager GUI, as described in Importing Data into the GUI on page 5-6. If you have not imported the data, click here. To select the portion of data between t = 0 s and t = 0.5 s:
1 In the Control and Estimation Tools Manager window, select the New

Data node under the Transient Data node.

5-16

Downloaded from www.Manualslib.com manuals search engine

Selecting Data for Estimation

2 Select the position1(:,1) cell in the Output Data tab, and click

Pre-process. This action opens the Data Preprocessing Tool GUI.

The Data Editing area of the Data Preprocessing Tool GUI shows the output data and time vector in the position1(:,1) and Time (seconds) columns, respectively. The Data Preprocessing Tool GUI lets you perform the following types of preprocessing operations: Excluding data Detrending and filtering data Handling missing data
3 To exclude the output data beyond t = 0.5 s: a In the Exclusion Rules tab, select the Bounds check box.

5-17

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

b In the Exclude X

field, where X corresponds to the time vector, select > from the drop-down list. Enter 0.5 in the adjacent field to specify the upper limit of the data to select for estimation. The Data Preprocessing Tool GUI resembles the next figure.

5-18

Downloaded from www.Manualslib.com manuals search engine

Selecting Data for Estimation

c Click Add.

This action adds a new Dataset1 node under the Transient Data node in the Control and Estimation Tools Manager GUI. The Dataset1 node contains the modified output data position1(:,1)* in the Output Data tab.

5-19

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

This operation also replaces the output data samples beyond t = 0.5 s in position1(:,1)* with NaNs. You can view the NaNs by selecting the Modified data tab in the Data Preprocessing Tool GUI, as shown in the next figure.

4 To remove the NaNs: a In the Control and Estimation Tools Manager GUI, select the Output

Data tab of the Dataset1 node.

b Select the position1(:,1)* cell, and click Pre-process.

This action updates the Data Preprocessing Tool GUI with the selected data position1(:,1)*.

5-20

Downloaded from www.Manualslib.com manuals search engine

Selecting Data for Estimation

c In the Missing Data Handling area, select the

Remove rows where

data is excluded or missing check box.

Tip You can view the results of this operation in the Modified data tab.

d In the Write results to area, select the existing dataset

option.

e Click Add.

The Update table data dialog box appears. Click Yes to overwrite the position1(:,1)* data set with the modified data.

5-21

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

5 To plot the data, select the position1(:,1)* cell in the Output Data tab

of the Dataset1 node, and click Plot Data. The selected output data from t = 0 s to t = 0.5 s is shown in the next figure.

Selecting Input Data

After you select the output data between t = 0 s and t = 0.5 s, as described in Selecting Output Data on page 5-16, you must also select the corresponding input data samples. This operation makes the number of I/O data samples equal.
1 In the Control and Estimation Tools Manger GUI, select the New Data

node under the Transient Data node.

2 In the Input Data tab, select the input1(:,1) cell, and click Pre-process.

This action updates the Data Preprocessing Tool GUI with the selected data input1(:,1).

5-22

Downloaded from www.Manualslib.com manuals search engine

Selecting Data for Estimation

3 To exclude the data beyond t = 0.5 s: a In the Exclusion Rules tab, select the Bounds check box. b In the Exclude X

field, where X corresponds to the time vector, select > from the drop-down list. Enter 0.5 in the adjacent field to specify the upper limit of the input data to select for estimation. existing dataset option remains selected.

c In the Write results to area, verify that the

d Click Add.

This action adds the modified data input1(:,1)* to the Input Data tab of the Dataset1 node. This operation also replaces the input data samples beyond t = 0.5 s in input1(:,1)* with NaNs.
4 To remove the NaNs: a In the Control and Estimation Tools Manager GUI, select the

input1(:,1)* cell in the Input Data tab of the Dataset1 node, and

click Pre-process. This action updates the Data Preprocessing Tool GUI with the selected data input1(:,1)*.
b In the Missing Data Handling area, select the

Remove rows where existing dataset

d Click Add.

data is excluded or missing check box. option remains selected.

c In the Write results to area, verify that the

The Update table data dialog box appears. Click Yes to overwrite the input1(:,1)* data set with the modified data.

5-23

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

5 To plot the data, select the input1(:,1)* cell in the Input Data tab of the

Dataset1 node, and click Plot Data. The selected input data from t = 0 s to t = 0.5 s is shown in the next figure.

5-24

Downloaded from www.Manualslib.com manuals search engine

Removing Outliers

Removing Outliers
In this section... Why Remove Outliers on page 5-25 How to Remove Outliers on page 5-25

Why Remove Outliers

Outliers are data values that deviate from the mean by more than three standard deviations. When estimating parameters from data containing outliers, the results may not be accurate. Removing outliers replaces the data samples containing outliers with NaNs, which represent missing data. You interpolate the missing data values in a subsequent operation, as described in Interpolating Missing Data on page 5-34.

How to Remove Outliers

In this portion of the tutorial, you remove outliers from the output data. You must have already selected a subset of the data, as described in Selecting Data for Estimation on page 5-16. If you have not done this preparation, click here.
1 In the Control and Estimation Tools Manger, select the position1(:,1)*

cell in the Output Data tab of the Dataset1 node, and click Pre-process. This action updates the Data Preprocessing Tool GUI with the selected data position1(:,1)*.

5-25

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

2 In the Exclusion Rules tab, select the Outliers check box.

By default, the Window length and Confidence limits fields are set to 10 and 95 respectively. The Window length field specifies the number of successive data samples the software uses to compute the mean and standard deviation. The Confidence limits field specifies the threshold number for identifying outliers. In this example, the mean and standard deviation of 10 successive data samples are computed, and data values that exceed 95% of standard deviation are identified as outliers.

Note The data samples containing outliers are replaced with NaNs.

3 In the Write results to area, verify that the existing dataset

option remains selected.

5-26

Downloaded from www.Manualslib.com manuals search engine

Removing Outliers

4 Click Add.

The Update table data dialog box appears. Click Yes to overwrite the position1(:,1)* data set with the modified data.
5 To plot the data, select the position1(:,1)* cell in the Output Data tab

of the Dataset1 node, and click Plot Data. The spikes, which indicate outliers, no longer appear on the time plot, as shown in the next figure.

5-27

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

The missing data samples, represented by NaNs, appear as gaps on the time plot. To see an example, zoom in to the bottom-left corner of the plot. As shown in the next figure, the data values corresponding to t = 0.009 s and t = 0.019 s are missing.

5-28

Downloaded from www.Manualslib.com manuals search engine

Filtering Data

Filtering Data
In this section... Filtering Output Data on page 5-29 Filtering Input Data on page 5-32

Filtering Output Data

In this portion of the tutorial, you filter the noise, and remove any periodic trends from the output data. To avoid relative phase shift between the I/O data, you must also apply the same filter to the input data. You must have already removed outliers from the output data, as described in Removing Outliers on page 5-25. If you have not done this preparation, click here.
1 In the Control and Estimation Tools Manager window, select the

position1(:,1)* cell in the Output Data tab of the Dataset1 node, and click Pre-process.

This action updates the Data Preprocessing Tool GUI with the selected data position1(:,1)*.

5-29

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

2 In the Detrend/Filtering tab: a Select the Filtering check box.

By default, First order is selected as the filter type. To learn more about the filters, see Filtering Data.
b Specify 0.001 in the First order filter with time constant field.

This field specifies the time constant for the first-order filter. Tip For calculating the time constant, you can visually inspect the time plot to determine the frequency components.

3 In the Write results to area, verify that the existing dataset

option remains selected.

5-30

Downloaded from www.Manualslib.com manuals search engine

Filtering Data

4 Click Add.

The Update table data dialog box appears. Click Yes to overwrite the position1(:,1)* data set with the modified data.
5 To plot the data, select the position1(:,1)* cell in the Output Data tab

of the Dataset1 node, and click Plot Data. The noise is filtered and the output data appears smooth, as shown in the next figure.

5-31

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

Filtering Input Data

After you filter the output data, as described in Filtering Output Data on page 5-29, you must also filter the input data with the same filter.
1 In the Control and Estimation Tools Manager window, select the

input1(:,1)* cell in the Input Data tab of the Dataset1 node, and click

Pre-process. This action updates the Data Preprocessing Tool GUI with the selected data input1(:,1)*.
2 In the Detrend/Filtering tab: a Select the Filtering check box.

By default, First order is selected as the filter type.

b Specify 0.001 in the First order filter with time constant field. 3 In the Write results to area, verify that the existing dataset

option remains selected.

4 Click Add.

The Update table data dialog box appears. Click Yes to overwrite the input1(:,1)* data set with the modified data.

5-32

Downloaded from www.Manualslib.com manuals search engine

Filtering Data

5 To plot the data, select the input1(:,1)* cell in the Input Data tab of the

Dataset1 node, and click Plot Data.

5-33

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

Interpolating Missing Data

In this portion of the tutorial, you interpolate the output data to compute the missing values created when removing outliers, as described in Removing Outliers on page 5-25. The interpolation operation uses the known data values to compute the missing data values. You must have already filtered the noise, as described in Filtering Data on page 5-29. If you have not already done this preparation, click here.
1 In the Control and Estimation Tools Manager GUI, select the

position1(:,1)* cell in the Output Data tab of the Dataset1 node, and click Pre-process.

This action updates the Data Preprocessing Tool GUI with the selected data position1(:,1)*.

5-34

Downloaded from www.Manualslib.com manuals search engine

Interpolating Missing Data

2 In the Missing Data Handling area, select the Interpolate missing

values using interpolation method check box. By default, zoh is selected as the interpolation method. This method fills the missing data sample with the data value immediately preceding it.

Tip You can view the interpolated data samples in the Modified data tab.

3 In the Write results to area, verify that the existing dataset

option remains selected.

4 Click Add.

The Update table data dialog box appears. Click Yes to overwrite the position1(:,1)* data set with the modified data.

5-35

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

5 To plot the data, select the position1(:,1)* cell in the Output Data tab

of the Dataset1 node, and click Plot Data. The new estimation data, prepared by removing outliers, filtering noise, and interpolating missing data, is shown the next figure.

5-36

Downloaded from www.Manualslib.com manuals search engine

Saving the Project

Saving the Project

After you prepare the data, you can delete the data in the New Data node, rename the prepared data, and save the session. To skip the data preparation steps, click here.
1 In the Control and Estimation Tools Manger GUI, select the New Data

node under the Transient Data node.

2 Right-click the New Data node, and select Delete. 3 Select the Dataset1 node under the Transient Data node. 4 Right-click the Dataset1 node, and select Rename. Specify

Estim_Data_Prep as the name of the new estimation data set.

The Control and Estimation Tools Manager GUI resembles the next figure.

5-37

Downloaded from www.Manualslib.com manuals search engine

Tutorial Preparing Data for Parameter Estimation Using the GUI

5 Save the Control and Estimation Tools Manager project: a In the Control and Estimation Tools Manager GUI, select File > Save.

This action opens the Save Projects dialog box.

b In the Save Projects dialog box, click OK.

c In the Save Projects window, specify spe_engine_throttle1p.mat in

the File name field, and click Save. The action saves the project as a MAT-file. To learn how to estimate parameters from this data, see Chapter 6, Tutorial Estimating Parameters from Measured Data Using the GUI.

5-38

Downloaded from www.Manualslib.com manuals search engine

6
Tutorial Estimating Parameters from Measured Data Using the GUI
About This Tutorial on page 6-2 Estimating Model Parameters Using Default Estimation Settings on page 6-7 Improving Estimation Results Using Parameter Bounds on page 6-20 Validating Estimated Model Parameters on page 6-26

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

About This Tutorial

In this section... Objectives on page 6-2 About the Model on page 6-3

Objectives
In this tutorial, you learn how to estimate parameters of a single-input single-output (SISO) Simulink model from measured input and output (I/O) data. Note Simulink Design Optimization software estimates parameters from real, time-domain data only. You learn to perform the following tasks using the GUI: Load a saved project containing data. Estimate model parameters using default settings. Validate the model, and refine the estimation results.

6-2

Downloaded from www.Manualslib.com manuals search engine

About This Tutorial

About the Model

In this tutorial, you use the spe_engine_throttle1 Simulink model. This model represents an engine throttle system, as shown in the next figure.

The throttle system controls the flow of air and fuel mixture to the engine cylinders. The throttle body contains a butterfly valve which opens when a driver presses the accelerator pedal. Opening this valve increases the amount of fuel mixture entering the cylinders, which increases the engine speed. A DC motor controls the opening angle of the butterfly valve in the throttle system. The models for these components are described in Motor Subsystem on page 6-4, and Throttle Subsystem on page 6-5. The input to the throttle system is the motor current (in amperes), and the output is the angular position of the butterfly valve (in degrees).

6-3

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

Motor Subsystem
The Motor subsystem contains the DC motor model. To open the model, double-click the corresponding block.

The following table describes the variables, parameters, equation, input, and output of the Motor subsystem. Variables Parameters U is the input current to the motor. T is the torque applied by the motor.
Kt is the torque gain of the motor, represented by Kt in the

model.

td is the input time delay of the motor, represented by input_delay in the model.

Equation

The torque applied by the motor is described in the following equation:

T (t) = K tU (t td )
where t is time. Input Output U T

6-4

Downloaded from www.Manualslib.com manuals search engine

About This Tutorial

Throttle Subsystem
The Throttle subsystem contains the butterfly valve model. To open the model, right-click the corresponding block, and select Look Under Mask.

The Hard Stops block models the valve angular position limit of 15 to 90. The following table describes the variables, parameters, states, differential equations, inputs, and outputs of the Throttle subsystem. Variables T is the torque applied by the DC motor. is the angular position of the valve, represented by x in the model. Thardstop is the torque applied by the hard stop. Parameters
J is the inertia. c is the viscous friction. k is the spring constant.

States

is the angular position. is the angular velocity.

6-5

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

Equations

The mathematical system for the butterfly valve is described in the following equation:

J + c + k = T + Thardstop
where 15 90 , with initial conditions
Hard Stops block is described in the following

0 = 15 , and 0 = 0 .The torque applied by the

equation:

Thardstop

15 90 0, = K (90 ), > 90 K (15 ), < 15

where K is the gain of the Hard Stops block.

Input Output

6-6

Downloaded from www.Manualslib.com manuals search engine

Estimating Model Parameters Using Default Estimation Settings

Estimating Model Parameters Using Default Estimation Settings

In this section... Overview of the Estimation Process on page 6-7 Specifying Parameters and Estimation Data on page 6-8 Validating Model Parameters on page 6-13

Overview of the Estimation Process

Simulink Design Optimization software uses optimization techniques to estimate model parameters. In each optimization iteration, the model is simulated with the current parameter values. The error between the simulated and measured output is computed and minimized. The estimation is complete when the optimization method finds a local minimum. You perform the following tasks to estimate the model parameters: Specifying Parameters and Estimation Data on page 6-8 Validating Model Parameters on page 6-13 Note The Simulink model must remain open to perform parameter estimation tasks.

6-7

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

Specifying Parameters and Estimation Data

To specify parameters and estimation data:
1 Load the preconfigured project spe_engine_throttle1p.mat.

This MAT-file contains the estimation data. To learn more about the data and how to prepare it, see Chapter 5, Tutorial Preparing Data for Parameter Estimation Using the GUI. To load the preconfigured project:
a Open the engine throttle system Simulink model by typing the following

at the MATLAB prompt:

spe_engine_throttle1
b In the Simulink model window, select Tools > Parameter Estimation.

This action opens a new project named Project - spe_engine_throttle1 in the Control and Estimation Tools Manager GUI.
c In the Control and Estimation Tools Manager GUI, select File > Load.

Browse to the matlabroot\toolbox\sldo\sldodemos\estim directory, and select spe_engine_throttle1p.mat. The Confirm Node Replace dialog box appears. Click Yes to load the project into the Control and Estimation Tools Manager GUI.

6-8

Downloaded from www.Manualslib.com manuals search engine

Estimating Model Parameters Using Default Estimation Settings

The Control and Estimation Tools Manager GUI opens as shown in the next figure.

6-9

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

2 Specify parameters for estimation. a Select the Variables node under the Estimation Task node.

b Click Add.

This action opens the Select Parameters dialog box, which shows the model parameters for the Simulink model.
c Select the parameters J, c, input_delay, and k by pressing the Ctrl key

while clicking each name, and then click OK. This action adds the selected parameters to the Estimated Parameters tab.

6-10

Downloaded from www.Manualslib.com manuals search engine

Estimating Model Parameters Using Default Estimation Settings

Tip When estimating a large number of parameters, you can select a subset of parameters to estimate. To learn more, see the Inverted Pendulum Parameter Estimation demo.
d Select the Estimation node, and click New.

This action adds a New Estimation node under the Estimation node.
e Select the New Estimation node.

6-11

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

In the Parameters tab, select the Estimate check box for all the parameters. This action specifies the selected parameters for estimation.

The Parameters tab, as shown in the previous figure, displays the following information for each parameter: Value: Current parameter value. By default, it is equal to the value specified in the Simulink model. During estimation, the optimization method might change a parameter value to minimize the error between the measured and simulated output. Initial Guess: Initial parameter value. By default, it is equal to the value in the Value field and is used at the start of estimation.

6-12

Downloaded from www.Manualslib.com manuals search engine

Estimating Model Parameters Using Default Estimation Settings

Minimum and Maximum: Bounds on the parameter value. By default, they are set to -Inf, and +Inf, respectively. This means that the optimization method searches for a solution in the range [-Inf, +Inf]. Typical Value: Order of magnitude of the parameter value. By default, it is equal to the value in the Initial Guess field.
3 Specify data for estimation by checking the Selected check box for

Estim_Data_Prep in the Data Sets tab.

Validating Model Parameters

After you specify parameters and data for estimation, as described in Specifying Parameters and Estimation Data on page 6-8, estimate the parameters and analyze the results to determine if the model needs to be refined.

6-13

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

1 Create plots to view the estimation results. a Select the Views node under the Estimation node, and click New.

This action creates a New View node under the Views node.
b Select the New View node. c In the View Setup tab, select the following plots from the drop-down

list in the Plot Type column: For Plot 1, select Measured and simulated. For Plot 2, select Cost function.

6-14

Downloaded from www.Manualslib.com manuals search engine

Estimating Model Parameters Using Default Estimation Settings

d In the Options area, select the Plot 1 and Plot 2 check boxes, and

click Show Plots.

Clicking Show Plots opens the following figures: New View - Plot 1 (Measured and simulated): Displays the measured output data Estim_Prep_Data. During estimation, this plot updates to display the simulated response at each iteration. New View - Plot 2 (Cost function): Displays the error between the measured and simulated output, computed at each iteration. By default, the error is computed using sum- of- squared- errors, which is a least-squares method.
2 Estimate the parameters by selecting the Estimation tab of the New

Estimation node, and clicking Start.

6-15

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

The Estimation tab updates at each iteration, and provides information about the estimation progress. When the estimation completes, the Estimation tab resembles the next figure.

Note The results of the optimization may differ slightly because of different numerical precision across platforms. The table displays the following information for each iteration: Cost Function: Error between the simulated and measured output. During estimation, the default optimization method Nonlinear least squares, lsqnonlin, minimizes the cost function by changing the parameter values. As shown in the previous figure, the cost function decreases by only 27% from 2.69e5 to 1.95e5.

6-16

Downloaded from www.Manualslib.com manuals search engine

Estimating Model Parameters Using Default Estimation Settings

Step Size: Displacement of the optimization method in the current search direction when minimizing the cost function. A final value close to zero indicates that the method has found a local minimum. As shown in the previous figure, the final step size in the last iteration is 5.96e-7.
3 View the estimated parameter values in the Value column of the

Parameters tab.

6-17

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

4 Examine the simulated response plot to see how well the simulated output

matches the measured output. The simulated response with the estimated parameters is shown in blue on the New View - Plot 1 (Measured and simulated) plot, and significantly differs from the measured data. This difference indicates that the estimated parameters are not accurate.

6-18

Downloaded from www.Manualslib.com manuals search engine

Estimating Model Parameters Using Default Estimation Settings

5 Examine the cost function plot to see how the cost function changes during

the estimation. The cost function values shown in step 2 are plotted on the New View Plot 2(Cost function) plot. The cost function decreases by only 27% from its initial value, and in conjunction with the measured and simulated output plot, indicates that the estimated parameters are not accurate.

To improve the estimation results, you apply bounds on the parameter values, as described in Improving Estimation Results Using Parameter Bounds on page 6-20.

6-19

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

Improving Estimation Results Using Parameter Bounds

In this section... Strategy for Improving the Estimation Results on page 6-20 How to Specify Parameter Bounds on page 6-20

Strategy for Improving the Estimation Results

Analyzing the Measured and Simulated and Cost function plots, as described in Validating Model Parameters on page 6-13, you see that the model parameters estimated using the default estimation settings are not accurate. You must run additional estimations to improve the accuracy of the model. There are several techniques to improve the accuracy of the estimated parameters. For more information, see How to Specify Estimation Options in the GUI. In this portion of the tutorial, you improve the results by specifying bounds on parameter values. This technique restricts the region in which the optimization method searches for a local minima. Based on physical insight, you know the following characteristics of the engine throttle system: All the parameter values are positive. Maximum time delay of the system, represented by input_delay, is 0.1 s. Therefore, you specify 0 as the minimum value for all parameters, and 0.1 as the maximum value of input_delay. After you estimate the parameters, analyze the results using the Measured and Simulated and Cost function plots.

How to Specify Parameter Bounds

You must have already estimated the parameters using default settings, as described in Estimating Model Parameters Using Default Estimation Settings on page 6-7.

6-20

Downloaded from www.Manualslib.com manuals search engine

Improving Estimation Results Using Parameter Bounds

To improve the estimation results by specifying parameter bounds:

1 Select the Parameters tab of the New Estimation node. 2 Specify the minimum value for each parameter by double-clicking the

corresponding Minimum cell, and replacing -Inf with 0.

3 Specify the maximum value for input_delay by double-clicking the

corresponding Maximum cell and replacing +Inf with 0.1. The Parameters tab resembles the next figure.

6-21

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

4 Select the Estimation tab, and click Start.

The Estimation tab updates at each iteration, and provides information about the estimation progress. When the estimation completes, the Estimation tab resembles the next figure.

Note The results of the optimization may differ slightly because of different numerical precision across platforms. The cost function decreases by 99% from 1.95e5 to 239.94. The final step size 8.75e-7 is close to 0, which indicates that the optimization method has found a local minimum.

6-22

Downloaded from www.Manualslib.com manuals search engine

Improving Estimation Results Using Parameter Bounds

5 View the estimated parameter values in the Value column of the

Parameters tab.

6-23

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

6 Examine the simulated response plot to see how well the simulated output

matches the measured output. The simulated response plot, shown in blue on the New View - Plot 1 (Measured and simulated) plot, is overlaid on the measured output data. The simulated output closely matches the measured data.

6-24

Downloaded from www.Manualslib.com manuals search engine

Improving Estimation Results Using Parameter Bounds

7 Examine the cost function plot to see how the cost function changes during

the estimation. The cost function values shown in step 4 are plotted on the New View - Plot 2(Cost function) plot. The cost function also decreases by 99% from its initial value, and in conjunction with the measured and simulated response plot indicates a good fit of the simulated response with the measured data.

6-25

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

Validating Estimated Model Parameters

After you estimate the model parameters, as described in Improving Estimation Results Using Parameter Bounds on page 6-20, validate the model using another data set (validation data). A good match between the simulated response and the validation data indicates that you have not overfitted the model. To validate the estimated parameters using a validation data set:
1 Load the preconfigured project spe_engine_throttle1_importValData1.mat.

This MAT-file contains a validation data set already imported into the GUI, and the estimated parameters as described in Improving Estimation Results Using Parameter Bounds on page 6-20. Tip To learn how to import data and time vector into the Control and Estimation Tools Manager GUI, see Importing Data into the GUI on page 5-6. To load the preconfigured project:
a In the Control and Estimation Tools Manager GUI, select File > Load. b Browse to the matlabroot\toolbox\sldo\sldodemos\estim directory,

and select spe_engine_throttle1_importValData1.mat. The Confirm Node Replace dialog box appears. Click Yes to load the project into the Control and Estimation Tools Manager GUI. The Control and Estimation Tools Manager GUI now has a node that contains the validation data, as shown in the next figure.

6-26

Downloaded from www.Manualslib.com manuals search engine

Validating Estimated Model Parameters

The validation data contains the input data, output data and time vector in the MATLAB variables input2, position2 and time2 respectively.
2 Plot the measured and simulated response, and residuals. a Select the Validation node, and click New.

This action creates a New Validation node.

b Select the New Validation node. c In the Validation Setup tab, select the following plots in the Plot

Type column. For Plot 1, select Measured and simulated. For Plot 2, select Residuals.
d In the Options area, select the Plot 1, and Plot 2 check boxes.

6-27

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

e Select Validation Data1 from the Validation data set drop-down list.

The Validation Setup tab resembles the next figure.

Click Show Plots. This action opens the following figures: New Validation - Plot 1 (Measured and simulated): Displays the validation data and simulated response. New Validation - Plot 2(Residuals): Displays the difference between the measured data and simulated response.

3 Examine the simulated response plot to see how well the simulated output

matches the validation data.

6-28

Downloaded from www.Manualslib.com manuals search engine

Validating Estimated Model Parameters

The simulated response, as shown in blue on the New Validation - Plot 1 (Measured and simulated) plot, is overlaid on the measured output data, and closely matches the measured data Validation Data 1.

4 Examine the residuals plot to compare the difference between the

simulated response and measured data. The difference between the simulated and measured data, as shown in the New Validation - Plot 2(Residuals) plot, varies between 2 and -2.5. The residuals lie within 6% of the maximum output variation, and do not display any systematic patterns. This indicates a good fit between the simulated output and measured data.

6-29

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

5 Save the Control and Estimation Tools Manager project. a In the Control and Estimation Tools Manager GUI, select File > Save.

This action opens the Save Projects dialog box.

b In the Save Projects dialog box, click OK. c In the Save Projects window, specify the name of the project in the File

name field, and click Save. This action saves the project as a MAT-file.

6-30

Downloaded from www.Manualslib.com manuals search engine

Validating Estimated Model Parameters

Tip You can load a project by selecting File > Load in the Control and Estimation Tools Manager GUI. To view the estimated parameter values, select the Parameters tab of New Estimation node.

6-31

Downloaded from www.Manualslib.com manuals search engine

Tutorial Estimating Parameters from Measured Data Using the GUI

6-32

Downloaded from www.Manualslib.com manuals search engine

7
Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI
About This Tutorial on page 7-2 Configuring a Model for Optimizing Parameters on page 7-5 Optimizing Model Parameters to Meet Step Response Requirements on page 7-8 Refining Model Parameters to Track a Reference Signal on page 7-25 Saving the Project on page 7-31

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

About This Tutorial

In this section... Objectives on page 7-2 About the Model on page 7-2 Design Requirements on page 7-4

Objectives
In this tutorial, you learn how to use the GUI to optimize the parameters of a Simulink model to meet time-domain design requirements. You accomplish the following tasks using the GUI: Specify step response requirements on the models output. Specify a reference signal for the models output to track. Optimize and refine the parameters to meet the design requirements. Tip To learn how to optimize model parameters to meet time-domain design requirements using command line functions, see Chapter 8, Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line.

About the Model

In this tutorial, you use the Simulink model named sldo_model1, as shown in the next figure.

7-2

Downloaded from www.Manualslib.com manuals search engine

About This Tutorial

Tip To open the model, type sldo_model1 at the MATLAB command prompt. The model contains a Controller block, which is a PID controller. This block controls the output of the Plant subsystem. The Unit Step block applies a step input to the system. Note This tutorial uses a step input to produce the models response and optimize the model parameters to meet step response requirements. You can also use other types of inputs, such as ramp, and optimize parameters to meet requirements on the models response to these inputs. Double-click the Plant subsystem to open it. The plant is a second-order system with delay. It contains Transfer Function and Transport Delay blocks, as shown in the next figure.

7-3

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

To learn more about the blocks, see Transfer Fcn and Transport Delay block reference pages.

Design Requirements
The models output must meet the following design requirements: Overshoot less than 10% Rise time less than 10 seconds Settling time less than 30 seconds Track a reference signal specified by 1e(-0.3*t), where t is time

7-4

Downloaded from www.Manualslib.com manuals search engine

Configuring a Model for Optimizing Parameters

Configuring a Model for Optimizing Parameters

In this portion of the tutorial, you configure a Simulink model for optimizing parameters:
1 Open the sldo_model1 model, if it is not already open, by typing the model

name at the MATLAB prompt:

sldo_model1

The Simulink model opens, as shown in the next figure.

To learn more about the model, see About the Model on page 7-2.
2 Add a Signal Constraint block to the model. a In the Simulink model, select View > Library Browser to open the

Simulink Library Browser.

7-5

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

b In the Libraries pane, select Simulink Design Optimization.

c Drag and drop the Signal Constraint block into the Simulink model

window. To learn more about the block, see the Signal Constraint block reference page.

7-6

Downloaded from www.Manualslib.com manuals search engine

Configuring a Model for Optimizing Parameters

d Connect the Signal Constraint block to the models output, as shown in

the next figure.

Note You must connect the Signal Constraint block to the signal to which you want to add design requirements. To learn more, see Constraining Model Signals in the Simulink Design Optimization Users Guide.

7-7

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

Optimizing Model Parameters to Meet Step Response Requirements

In this section... Specify Time-Domain Design Requirements on page 7-8 Specifying Parameters to Optimize on page 7-18 Optimizing the Parameters on page 7-21

Specify Time-Domain Design Requirements

In this portion of the tutorial, you specify the rise time, setting time and overshoot requirements that the models response should satisfy. You must have already configured the Simulink model, as described in Configuring a Model for Optimizing Parameters on page 7-5. To specify the design requirements:

7-8

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

1 Double-click the Signal Constraint block to open the Block Parameters:

Signal Constraint window.

Tip To view the line segment that extends from 0 to 15 seconds, set the upper limit of the Amplitude axis to 1.3. To do so, right-click in the plot, and select Axes Limits.

7-9

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

The Block Parameters window shows an amplitude versus time plot. The line segments map to the step response requirements shown in the next figure.

% Overshoot

% Settling

% Undershoot

Rise time

Settling time

By default, the line segments represent the following step response requirements: Amplitude less than or equal to 0.01 up to the rise time of 5 seconds for 1% undershoot Amplitude between 0.9 and 1.2 up to the settling time of 15 seconds Amplitude equal to 1.2 for 20% overshoot up to the settling time of 15 seconds Amplitude between 0.99 and 1.01 beyond the settling time for 2% settling To specify the design requirements described in Design Requirements on page 7-4, you must modify the time and amplitude values of the line segments, as described in the next steps.

7-10

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

2 Double-click the lower yellow region shown on the plot.

This action opens the Edit Design Requirement dialog box, as shown in the following figure.

The rows in the dialog box define the lower bound on the signal. The time span, displayed in the Design requirement drop-down list, is same as the simulation time specified in the Simulink model. The rows correspond to the following three line segments in the Block Parameters window: First row Lower-line segment that extends from 0 to 5 seconds. Second row Lower-line segment that extends from 5 to 15 seconds. Third row Lower-line segment that extends from 15 to 50 seconds. The default amplitude and time values specified in the rows appear in the Block Parameters window, as shown in the next figure.

7-11

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

3 Specify the rise time requirement: a In the first row of the Edit Design Requirement dialog box, double-click

the Time field in the End column.

b Type 10, and press Enter.

The first row now looks like the following figure.

This action also updates the Block Parameters window to show the rise time requirement. This requirement is represented by the line segment at the bottom left of the window that extends from 0 to 10 seconds, as shown in the next figure.

7-12

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

4 Specify the settling time requirement: a In the second row of the Edit Design Requirement dialog box,

double-click the Time field in the End column.

7-13

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

b Type 30, and press Enter.

The second row looks like the following figure.

This action also updates the Block Parameters window to show the settling time requirement on the lower bound. This requirement is represented by the line segment that extends from 10 to 30 seconds, as shown in the next figure.

7-14

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

c In the Edit Design Requirements dialog box, select Upper time

response bound from 0 to 50 sec in the Design Requirement

drop-down list.

The rows in the dialog box define the upper bound on the signal, and correspond to the following two line segments in the Block Parameters window: First row Upper-line segment that extends from 0 to 15 seconds. Second row Upper-line segment that extends from 15 to 50 seconds. The default amplitude and time values specified in the rows appear in the Block Parameters window, as shown in the next figure.

7-15

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

d In the first row, double-click the Time field in the End column. e Type 30, and press Enter.

The first row looks like the following figure.

This action also updates the Block Parameters window to show the settling time requirement on the upper-bound. This requirement is represented by the line segment that extends from 0 to 30 seconds, as shown in the next figure.

5 Specify the overshoot requirement: a In the Edit Design Requirement dialog box, verify that Upper time

response bound from 0 to 50 sec is selected in the Design

Requirement drop-down list.

b In the first row, double-click the Amplitude field in the Start column. c Type 1.1, and press Enter.

7-16

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

d Double-click the Amplitude field of the End column. e Type 1.1, and press Enter.

The first row now looks like the following figure.

The Block Parameters window updates to show the overshoot requirement. This requirement is represented by the amplitude value of 1.1 for the line segment that extends from 0 to 30 seconds, as shown in the following figure.

When you optimize the model parameters, the models final response must lie in the white region bounded by the line segments in order to meet the design requirements.
6 In Edit Design Requirement dialog box, click Close.

7-17

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

Specifying Parameters to Optimize

In this portion of the tutorial, you specify the model parameters to optimize. When you optimize the model parameters, the software modifies the parameter values to meet the specified design requirements. Before you specify the parameters to optimize, you can specify the design requirements, as described in Specify Time-Domain Design Requirements on page 7-8. To specify the parameters to optimize:
1 In the Block Parameters window, select Optimization > Tuned

Parameters. This action opens the Tuned Parameters dialog box.

7-18

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

2 In the Tuned Parameters window, click Add.

This action opens the Add Parameters dialog box.

7-19

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

3 In the Add Parameters dialog box, select the PID controller parameters

Kd, Ki and Kp, and click OK.

This action adds Kd, Ki and Kp to the Tuned Parameters window.

The Optimization Settings area displays the following parameter settings: Value Current parameter value Initial Guess Initial parameter value Minimum and Maximum Parameter bounds Typical Value Scaling factor for the parameter To learn more about the parameter settings, see Changing Tuned Parameter Specifications in the Simulink Design Optimization Users Guide.

7-20

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

4 To limit the parameters to positive values, enter the minimum value of

each parameter as 0 in the corresponding Minimum field. After you enter these values, the Tuned Parameters window resembles the next figure.

5 Click OK to apply the parameter settings and close the Tuned Parameters

window.

Optimizing the Parameters

In this portion of the tutorial, you optimize the model parameters to meet the specified design requirements. You must have already specified the design requirements and the parameters to optimize, as described in Specify Time-Domain Design Requirements on page 7-8, and Specifying Parameters to Optimize on page 7-18, respectively. To optimize the model parameters:

7-21

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

1 In the Block Parameters window, click Optimization > Start to start

the optimization. This action opens the optimization progress window as shown in the next figure.

During each optimization iteration, the software simulates the model, and the default optimization method Gradient descent modifies the controller parameters to reduce the distance between the simulated response and the design requirement line segments. To learn more about the optimization method, see Selecting Optimization Methods in the Simulink Design Optimization Users Guide.

7-22

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

After the optimization completes, the Optimization Progress window resembles the next figure.

The message Successful termination indicates that the optimization method found a solution that meets the design requirements within the parameter bounds. For more information about the outputs displayed in the optimization progress window, see Displaying Iterative Output in the Optimization Toolbox documentation. The optimized parameter values are displayed in the Optimization Progress window, and also written into the Simulink model.
2 In the Block Parameters window, examine the models response to see if

the response meets the step response requirements. The final response of the system with the optimized parameter values appears in black, as shown in the next figure.

7-23

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

The plot also displays the initial response of the system in blue. The remaining responses show the simulated response at each optimization iteration. The optimized response of the system lies in the white region bounded by the design requirement line segments and thus meets the specified requirements. Next, you refine the model parameters by specifying a reference signal that the system response must track, as described in Refining Model Parameters to Track a Reference Signal on page 7-25.

7-24

Downloaded from www.Manualslib.com manuals search engine

Refining Model Parameters to Track a Reference Signal

Refining Model Parameters to Track a Reference Signal

In this portion of the tutorial, you optimize the model parameters to track a reference signal specified by the equation

y = 1 exp(0.3 t)
You must have already optimized the model parameters to meet the step response requirements, as described in Optimizing Model Parameters to Meet Step Response Requirements on page 7-8. To optimize the parameters to track a reference signal:
1 In the Block Parameters window, select Goals > Desired Response.

This action opens the Desired Response dialog box.

2 Specify the reference signal:

7-25

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

a In the Time vector field of the Desired Response dialog box, enter

linspace(0,50,200).
b In the Amplitude field, enter 1-exp(-0.3*linspace(0,50,200)).

After you make these changes, the Desired Response dialog box resembles the following figure.

7-26

Downloaded from www.Manualslib.com manuals search engine

Refining Model Parameters to Track a Reference Signal

c Click OK to add the reference signal to the block and close the Desired

Response dialog box. The Block Parameters window updates to display the reference signal, shown as the gray dashed curve, in the next figure.

3 In the Block Parameters window, select the Track reference signal

check box.

7-27

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

4 In the Signal Constraint block window, select Optimization > Start to

start the optimization. This action restarts the optimization using the optimized parameters values computed in Optimizing Model Parameters to Meet Step Response Requirements on page 7-8. The Optimization Progress window opens, as shown in the next figure.

7-28

Downloaded from www.Manualslib.com manuals search engine

Refining Model Parameters to Track a Reference Signal

At each iteration, the optimization method modifies the controller parameters to minimize the error between the simulated response and the reference signal. Simultaneously, the method satisfies the step response requirements by reducing the distance between the simulated response and the design requirement line segments. After the optimization completes, the Optimization Progress window resembles the following figure.

The message, Optimization terminated due to slow progress in parameter or objective values appears in the Optimization Progress window. This message indicates that the optimization method cannot minimize the error between the simulated response and the reference signal any further. You must examine the optimized response of the model to check if the response tracks the reference signal, as described in the next step.

7-29

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

5 In the Block Parameters window, examine the models response to see how

well the response tracks the reference signal. The models response before you began the optimization appears as the blue curve. After the optimization completes, the final response, shown as the black curve, tracks the reference signal.

Note To refine the parameters further to track the reference signal more closely, change the parameter and function tolerances, as described in Selecting Optimization Termination Options in the Simulink Design Optimization Users Guide.

7-30

Downloaded from www.Manualslib.com manuals search engine

Saving the Project

Saving the Project

After you optimize the model parameters, you can save the model with the design requirements and optimized parameter values in a response optimization project. To save a response optimization project:
1 In the Signal Constraint block window, select File > Save As.

This action opens the Save Project dialog box.

2 In the Save Project dialog box, enter a variable name in the Model

workspace variable field.

3 Click OK to save the project.

7-31

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI

Tip You can load the response optimization project by loading the Simulink model.

7-32

Downloaded from www.Manualslib.com manuals search engine

8
Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line
About This Tutorial on page 8-2 Configuring a Model for Optimizing Parameters on page 8-5 Optimizing Model Parameters to Meet Step Response Requirements on page 8-8 Refining Model Parameters to Track a Reference Signal on page 8-15

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line

About This Tutorial

In this section... Objectives on page 8-2 About the Model on page 8-2 Design Requirements on page 8-4

Objectives
In this tutorial, you learn how to use Simulink Design Optimization functions to optimize the parameters of a Simulink model to meet time-domain design requirements. You accomplish the following tasks using the command line: Specify step response requirements on the models output. Specify a reference signal for the models output to track. Optimize the parameters to meet the design requirements. Tip To learn how to optimize model parameters to meet time-domain design requirements using the GUI, see Chapter 7, Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI.

About the Model

In this tutorial, you use the Simulink model named sldo_model1, as shown in the next figure.

8-2

Downloaded from www.Manualslib.com manuals search engine

About This Tutorial

Tip To open the model, type sldo_model1 at the MATLAB command prompt. The model contains a Controller block, which is a PID controller. This block controls the output of the Plant subsystem. The Unit Step block applies a step input to the system. Note This tutorial uses a step input to produce the models response and optimize the model parameters to meet step response requirements. You can also use other types of inputs, such as ramp, and optimize parameters to meet requirements on the models response to these inputs. Double-click the Plant subsystem to open it. The plant is modeled as a second-order system with delay. It contains Transfer Function and Transport Delay blocks, as shown in the next figure.

8-3

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line

To learn more about the blocks, see Transfer Fcn, and Transport Delay block reference pages.

Design Requirements
The models output must meet the following design requirements: Overshoot less than 10% Rise time less than 10 seconds Settling time less than 30 seconds Track a reference signal specified by

y = 1 exp(0.3 t)

, where t is time

8-4

Downloaded from www.Manualslib.com manuals search engine

Configuring a Model for Optimizing Parameters

Configuring a Model for Optimizing Parameters

To optimize parameters of a Simulink model, you must first configure the model:
1 Open the sldo_model1 model, if it is not already open, by typing the model

name at the MATLAB prompt:

sldo_model1

The Simulink model opens, as shown in the next figure.

To learn more about the model, see About the Model on page 8-2.
2 Add a Signal Constraint block to the model. a In the Simulink model, select View > Library Browser to open the

Simulink Library Browser.

8-5

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line

b In the Libraries pane, select Simulink Design Optimization.

c Drag and drop the Signal Constraint block into the Simulink model

window. To learn more about the block, see Signal Constraint block reference page.

8-6

Downloaded from www.Manualslib.com manuals search engine

Configuring a Model for Optimizing Parameters

d Connect the Signal Constraint block to the models output, as shown in

the next figure.

Note You must connect the Signal Constraint block to the signal to which you want to add design requirements. To learn more, see Constraining Model Signals in the Simulink Design Optimization Users Guide.

8-7

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line

Optimizing Model Parameters to Meet Step Response Requirements

In this portion of the tutorial, you use Simulink Design Optimization commands to optimize the model parameters to meet step response requirements. You must have already configured your model for optimization, as described in Configuring a Model for Optimizing Parameters on page 8-5. To optimize the model parameters to meet step response requirements:
1 Create a new project using the newsro command.

proj=newsro('sldo_model1',{'Kd','Ki','Kp'})

This command creates a new project object, proj, that contains instructions to optimize the parameters Kp, Ki, and Kd.
Name: Parameters: OptimOptions: Tests: Model: 'sldo_model1' [3x1 ResponseOptimizer.Parameter] [1x1 sroengine.OptimOptions] [1x1 ResponseOptimizer.SimTest] 'sldo_model1'

Response Optimization Project.

2 Retrieve the signal constraint object for the project, proj, using the

findconstr command. constr=findconstr(proj,'sldo_model1/Signal Constraint')

This command retrieves the following signal constraint object:

ConstrEnable: isFeasible: CostEnable: Enable: Name: SignalSize: LowerBoundX: 'on' 1 'off' 'on' 'Signal Constraint' [1 1] [3x2 double]

8-8

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

LowerBoundY: LowerBoundWeight: UpperBoundX: UpperBoundY: UpperBoundWeight: ReferenceX: ReferenceY: ReferenceWeight: Signal Constraint.

[3x2 [3x1 [2x2 [2x2 [2x1 [] [] []

double] double] double] double] double]

The LowerBoundX, LowerBoundY, UpperBoundX, and UpperBoundY properties of the constr object represent the numeric values of design requirements on a signal. Tip You can view the default design requirements by double-clicking the Signal Constraint block in the sldo_model1 Simulink model. The LowerBoundX, and LowerBoundY properties define the lower bound, and the UpperBoundX and UpperBoundY define the upper bound on the signal. These properties are matrices, whose elements represent the values of the bounds as shown in the following table. Property Name
LowerBoundX

Purpose Time values of the lower-bound line segments Amplitude values of the lower-bound line segments Time values of the upper-bound line segments Amplitude values of the upper-bound line segments

LowerBoundY UpperBoundX UpperBoundY

8-9

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line

For example, the elements of the LowerBoundX, and LowerBoundY matrices map to the following values.
Time of all lower design requirement line segments Time of beginning point of first line segment Time of end point of first line segment Time of beginning and end points of second line segment Amplitudes of all lower design requirement line segments Amplitude of beginning point of first line segment Amplitude of end point of first line segment Amplitude of beginning and end points of second line segment

Time of beginning and end points of third line segment

Amplitude of beginning and end points of third line segment

The line segments that correspond to the elements of LowerBoundX, and LowerBoundY matrices appear in the next figure.
(lx (lx ,ly ) b2 b2 b3 ,ly b3 ) (lx ,ly ) e3 e3

(lx ,ly ) e2 e2

(lx

b1

,ly

b1

(lx ,ly ) e1 e1

8-10

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

Similarly, the elements of UpperBoundX and UpperBoundY map to the time and amplitude values of the upper-bound line segments, respectively. To specify the step response requirements, as described in Design Requirements on page 8-4, you must edit these properties as shown in the next step. Note When you add or modify design requirements using functions, the Block Parameters window does not update to display the new requirements. To view the values of the LowerBoundY matrix, type the following command:
constr.LowerBoundY
3 Specify the step response requirements using the following commands:

% Rise time requirement constr.LowerBoundX=[0 10;10 15;15 50]; % Settling time requirement on lower line segments; constr.LowerBoundX=[0 10;10 30;30 50]; % Settling time requirement on upper line segments constr.UpperBoundX=[0 30;30 50]; % Overshoot requirement on upper bound constr.UpperBoundY=[1.1 1.1;1.01 1.01];

Tip You can type constr.PropertyName to view the updated values of the properties.
4 Specify the parameters to optimize: a Retrieve the vector of parameter objects from the project object, proj,

using the findpar command.

param(1)=findpar(proj,'Kd'); param(2)=findpar(proj,'Ki'); param(3)=findpar(proj,'Kp');
b To view the properties of a parameter object, type the object name. For

example:

8-11

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line

param(1)

This command returns the properties of param(1) object.

Name: Value: InitialGuess: Minimum: Maximum: TypicalValue: ReferencedBy: Description: Tuned: Tuned parameter.
c To limit the parameter to positive values, enter the minimum value for

'Kd' 0 0 -Inf Inf 0 {0x1 cell} '' 1

each parameter as 0 using the following commands:

param(1).Minimum=0; param(2).Minimum=0; param(3).Minimum=0;

Tip To view the updated value of the Minimum property for Kd, type param(1).Minimum.
5 Optimize the parameters using the optimize command:

opt=optimize(proj)

This command returns the following results:

max Iter 0 1 2 3 4 S-count 1 14 21 f(x) 0 0 0 constraint 207.1 3.767 0.4641 0.02942 0 1.25 0.541 0.0755 0.0233 0 0 -1.22e-018 3.95e-018 1 6.18 0.000891 0.000114 Hessian modified infeasible Step-size Directional derivative First-order optimality Procedure

28 -9.19412e-020 35 0

8-12

Downloaded from www.Manualslib.com manuals search engine

Optimizing Model Parameters to Meet Step Response Requirements

Successful termination. Found a feasible or optimal solution within the specified tolerances.

Kd =

0.0468

Ki =

0.1455

Kp =

0.1853

ans =

Cost: 0 X: [4x1 double] ExitFlag: 1 Iteration: 4

At each optimization iteration, the software simulates the model and the default optimization method Gradient descent modifies the controller parameters to reduce the distance between the simulated response and the design requirements. To learn more about the optimization method, see Selecting Optimization Methods in the Simulink Design Optimization Users Guide. After the optimization completes, you see the messages Successful termination, and ExitFlag:1 at the MATLAB prompt. These messages indicate that the optimization method found a solution that meets the design requirements within the parameter bounds. For more information about the outputs displayed during the optimization, see Displaying Iterative Output in the Optimization Toolbox documentation. The optimized values of the parameters, Kp, Ki and Kd, are displayed at the command line, and are also written into the Simulink model.

8-13

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line

Next, you refine the parameters by specifying a reference signal, as described in Refining Model Parameters to Track a Reference Signal on page 8-15.

8-14

Downloaded from www.Manualslib.com manuals search engine

Refining Model Parameters to Track a Reference Signal

Refining Model Parameters to Track a Reference Signal

In this portion of the tutorial, you optimize the model parameters to track a reference signal specified by the equation

y = 1 exp(0.3 t)
You must have already optimized the model parameters to meet the step response requirements, as described in Optimizing Model Parameters to Meet Step Response Requirements on page 8-8. To optimize the parameters to track a reference signal:
1 Specify the reference signal to track using the following commands:

% Enable reference tracking constr.CostEnable='on'; % Specify the time values of the reference signal constr.ReferenceX=linspace(0,50,200)'; % Specify the amplitude values of the reference signal constr.ReferenceY=1-exp(-0.3*linspace(0,50,200))';

Tip To verify that the reference signal has been added to the signal constraint object, type constr=findconstr(proj,'sldo_model1/Signal Constraint'). The ReferenceX and ReferenceY properties now contain the reference signal.
2 Initialize the parameters in proj for optimization. a Set the initial guess of the parameters using the initpar command.

% Copy parameter values from Simulink model to the project initpar(proj)

b Update the scaling factor of the parameters with the initial parameter

values using the following commands.

% Update scaling factor with the initial parameter value param(1).TypicalValue = param(1).InitialGuess;

8-15

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line

param(2).TypicalValue = param(2).InitialGuess; param(3).TypicalValue = param(3).InitialGuess;

To learn more about the scaling factor, see Changing Tuned Parameter Specifications in the Simulink Design Optimization Users Guide. When you optimize the model parameters after this initialization, the optimization method uses the updated parameter settings to recompute new parameter values.
3 Start the optimization using the optimize command.

opt=optimize(proj)

This command returns the following results:

max Iter 0 1 2 3 4 5 6 7 8 9 10 S-count 1 20 36 49 60 71 82 91 100 109 116 f(x) 2.1547 0.561528 0.438343 0.360718 0.271571 0.178588 0.1574 0.146542 0.139688 0.135451 0.133726 constraint 0 0.02738 0.03301 0 0 0 0 0 0 0 0 0.622 0.142 0.296 0.173 0.321 0.147 0.0941 0.0443 0.0185 0.0256 -7.04 -1.13 -0.802 -0.64 -0.468 -0.215 -0.168 -0.208 -0.294 -0.0779 74 1.84 3.3 2.69 1.76 1.28 0.824 0.37 0.102 0.109 Step-size Directional derivative First-order optimality Procedure

Optimization terminated due to slow progress in parameter or objective values. To optimize further, go to Optimization Options and decrease the parameter and/or function tolerances.

Kd = 1.5821 Ki =

0.2876

Kp =

8-16

Downloaded from www.Manualslib.com manuals search engine

Refining Model Parameters to Track a Reference Signal

0.6311

ans =

Cost: 0.1337 X: [4x1 double] ExitFlag: 5 Iteration: 10

At each iteration, the optimization method modifies the controller parameters to minimize the error between the simulated response and the reference signal. Simultaneously, the method satisfies the step response requirements by reducing the distance between the simulated response and the design requirement line segments. After the optimization completes, you see the message Optimization
terminated due to slow progress in parameter or objective values. This message indicates that the optimization method cannot

minimize the error between the simulated response and the reference signal any further. If you want to refine the parameters further to track the reference signal more closely, change the parameter and function tolerances, as described in Selecting Optimization Termination Options in the Simulink Design Optimization Users Guide. Tip To view the updated response of the model, connect a Scope block to the model and simulate the model to view the response.
4 Save the project by typing the following command:

save('proj');

8-17

Downloaded from www.Manualslib.com manuals search engine

Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line

8-18

Downloaded from www.Manualslib.com manuals search engine

9
Tutorial Designing a PID Controller Using Optimization-Based Tuning
About This Tutorial on page 9-2 Configuring a Project for Optimization-Based Control Design on page 9-5 Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements on page 9-11 Refining the Controller Design to Meet Controller Output Bounds on page 9-32 Saving the Project on page 9-48

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

About This Tutorial

In this section... Objectives on page 9-2 About the Model on page 9-2 Design Requirements on page 9-4

Objectives
In this tutorial, you learn to use optimization methods to design a PID controller to meet frequency-domain design requirements on a systems response. You accomplish the following tasks using the GUI: Specify frequency-domain Bode magnitude and phase margin requirements. Design an initial controller to meet the frequency-domain requirements. Refine the initial controller design to limit the controllers output signal.

About the Model

In this tutorial, you use the Simulink model named sldo_model2, as shown in the next figure.

9-2

Downloaded from www.Manualslib.com manuals search engine

About This Tutorial

The model contains a Controller block, which is a PID Controller. This block controls the output of the Plant subsystem. Using the Simulink Control Design software has linearized the Simulink model at the operating point specified in the model. The sldo_model2.mat file contains a preconfigured SISO Design Tool session saved after linearizing the model. To learn more about linearizing Simulink models for control design, see Designing Compensators in the Simulink Control Design Users Guide. Double-click the Plant subsystem to open it. The plant is modeled as a second-order system with delay. It contains Transfer Function and Transport Delay blocks, as shown in the next figure.

9-3

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

To learn more about the blocks, see Transfer Fcn and Transport Delay block reference pages.

Design Requirements
The compensator you design in this tutorial must meet the following design requirements: Bode lower magnitude bound of 0 in the frequency range 1e-3 to 1 rad/sec Phase margin greater than 60 degrees Controller output bounds in the range [-250 550]

9-4

Downloaded from www.Manualslib.com manuals search engine

Configuring a Project for Optimization-Based Control Design

Configuring a Project for Optimization-Based Control Design

To design a linear controller for a Simulink model, you must first configure a Control and Estimation Tools Manager project.
1 Open a SISO Design Tool session for the linearized Simulink model by

typing the following command at the MATLAB prompt:

sisotool('sldo_model2.mat')

Note sldo_model2.mat file contains a preconfigured SISO Design Tool session. This session was saved after Simulink Control Design software linearized the sldo_model2 Simulink model. This command opens the following windows: Simulink model

To learn more about the model, see About the Model on page 9-2.

9-5

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

Control and Estimation Tools Manager GUI, which contains a SISO Design Task node under the Simulink Compensator Design Task node.

9-6

Downloaded from www.Manualslib.com manuals search engine

Configuring a Project for Optimization-Based Control Design

LTI Viewer for SISO Design Task window, which contains the following plots: Closed-loop step response of the system in the top plot Output of the Controller block in the bottom plot

9-7

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

A blank SISO Design for SISO Design Task window

9-8

Downloaded from www.Manualslib.com manuals search engine

Configuring a Project for Optimization-Based Control Design

2 In the Control and Estimation Tools Manager GUI, select the Automated

Tuning tab in the SISO Design Task node.

9-9

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

3 Click Optimize Compensators.

This action creates a new Response Optimization node.

9-10

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements
In this section... Specifying the Controller Parameters on page 9-11 Specifying Bode Magnitude and Phase Margin Design Requirements on page 9-15 Designing the Controller on page 9-25

Specifying the Controller Parameters

In this portion of the tutorial, you specify the controller parameters to design. You must have already configured a project for control design, as described in Configuring a Project for Optimization-Based Control Design on page 9-5. To specify the controller parameters to design:

9-11

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

1 In the Control and Estimation Tools Manager GUI, select the

Compensators tab in the Response Optimization node.

The controller parameters appear as poles and zeros in the Compensator elements column and represent the following: Gain Overall gain of the controller Real zeros Zeros resulting from the differentiator and integrator Real pole Pole resulting from the low-pass filter of the differentiator Tip To view the structure of the Controller block, right-click the block and select Look Under Mask.

9-12

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

2 For convenience, change the PID controller parameters to Simulink block

mask parameters format. To do so, right-click the sldo_model2/Controller column, and select Parameterized format.

9-13

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

This action displays the controller parameters as Simulink block mask parameters P, I, and D, as shown in the next figure. To learn more about mask parameters, see Mask Parameters in the Simulink documentation.

The Compensators tab displays the following parameter settings: Value Current controller parameter value Initial Guess Initial controller parameter value Minimum and Maximum Controller parameter bounds Typical Value Scaling factor for the controller parameter To learn more about the parameter settings, see Changing Tuned Parameter Specifications in the Simulink Design Optimization Users Guide.

9-14

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

3 In the Optimize column, select the check boxes for P, I, and D.

Specifying Bode Magnitude and Phase Margin Design Requirements

In this portion of the tutorial, you specify the Bode magnitude and phase margin requirements that the controller must satisfy. Before you specify the design requirements, you can specify the controller parameters to design, as described in Specifying the Controller Parameters on page 9-11. To specify the Bode design requirements:

9-15

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

1 In the Response Optimization node of the Control and Estimation Tools

Manager GUI, select the Design requirements tab.

9-16

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

2 Click Add new design requirement.

This action opens the New Design Requirement dialog box.

9-17

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

3 Specify the Bode magnitude lower limit requirement: a In the New Design Requirement dialog box, select Bode magnitude

lower limit from the Design requirement type drop-down list.

This action updates the New Design Requirement dialog box, as shown in the next figure.

b Select Open Loop at outport 1 of Controller from the

Requirement for response drop-down list.

c In the Frequency field of the Start column, enter 1e-3.

9-18

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

d In the Frequency field of the End column, enter 1.

The New Design Requirement dialog box resembles the following figure.

e Click OK to close the New Design Requirement dialog box.

9-19

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

The Bode lower magnitude limit is added to the Design requirements tab, as shown in the next figure.

9-20

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

The SISO Design for SISO Design task window also updates to show the Bode plot with the design requirement displayed as the black line segment.

9-21

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

4 Add the phase margin requirement: a In the SISO Design window, right-click the white area on the top plot,

and select Design requirement > New.

9-22

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

This action opens the New Design Requirement dialog box, as shown in the next figure.

b In the New Design Requirement dialog box, select Gain & phase

margins from the Design requirement type drop-down list.

The New Design Requirement dialog box updates to display the Gain Margin > and Phase Margin > options, as shown in the next figure.

c Select the Phase margin > check box, and enter 60 in the adjacent field.

Then, click OK to close the dialog box.

9-23

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

The Design requirements tab in the Control and Estimation Tools Manager GUI updates. It now displays the phase margin requirement, as shown in the next figure.

9-24

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

The SISO Design window also updates to display the phase margin requirement.

Designing the Controller

In this portion of the tutorial, you design the controller to meet the Bode magnitude and phase requirements. You must have already specified the controller parameters to design, as described in Specifying the Controller Parameters on page 9-11, and the design requirements, as described in Specifying Bode Magnitude and Phase Margin Design Requirements on page 9-15.

9-25

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

To design the controller:

1 In the Optimization tab, click Start Optimization.

The Optimization tab updates as shown in the next figure.

9-26

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

At every optimization iteration, the default optimization method Gradient descent reduces the distance between the current response and the magnitude requirement line segment by modifying the controller parameters. Simultaneously, the software also computes the phase margin and reduces the distance between the current response and the phase margin. To learn more about the optimization method, see Selecting Optimization Methods in the Simulink Design Optimization Users Guide. After the optimization completes, the Optimization tab displays the optimization iterations and status, as shown in the next figure.

9-27

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

The message Successful termination indicates that the optimization method found a solution that meets the design requirements. For more information about the outputs displayed in the Optimization progress table, see Displaying Iterative Output in the Optimization Toolbox documentation.

9-28

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

2 Examine the controller parameters and the systems response: a View the optimized parameter values in the Value field of the

Compensator tab.

b Examine the systems response on the following plots:

9-29

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

The SISO Design window

The top plot shows that the magnitude of the system, displayed as the blue curve, lies outside the yellow region. This plot indicates that the system has met the Bode magnitude requirement. The bottom plot displays the phase margin (P.M.) value of 86.1 degrees. This indicates that the system has met the phase margin design requirement of >60 degrees.

9-30

Downloaded from www.Manualslib.com manuals search engine

Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements

The LTI Viewer

The top plot shows that the closed-loop response of the system is stable. The system with the designed controller thus meets both the magnitude and phase margin requirements. The bottom plot in the LTI Viewer shows that the peak value of the controllers output is 1000, which is very large and can cause damage to the plant. To limit the controller output, you apply lower and upper bounds on the signal, as described in Refining the Controller Design to Meet Controller Output Bounds on page 9-32.

9-31

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

Refining the Controller Design to Meet Controller Output Bounds

In this portion of the tutorial, you refine the controller to satisfy bounds on the controllers output. You must have already designed an initial controller, as described in Designing an Initial PID Controller to Meet Bode Magnitude and Phase Margins Requirements on page 9-11.

9-32

Downloaded from www.Manualslib.com manuals search engine

Refining the Controller Design to Meet Controller Output Bounds

To tune the compensator parameters to meet the bounds on the controllers signal:
1 Add the upper-bound on the controllers output: a In the LTI Viewer, right-click the white area on the bottom plot, and

select Design requirement > New.

This action opens the New Design Requirement dialog box.

9-33

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

b In the New Design Requirement dialog box, select Upper time response

bound from the Design requirement type drop-down list.

c In the Time field of the End column, enter Inf. d In the Amplitude field of the Start column, enter 550.

9-34

Downloaded from www.Manualslib.com manuals search engine

Refining the Controller Design to Meet Controller Output Bounds

e In the Amplitude field of the End column, enter 550.

The New Design Requirements dialog box resembles the following figure.

Click OK to close the dialog box.

9-35

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

The Design requirements tab in the Control and Estimation Tools Manager GUI updates. It now displays the upper-bound requirement, as shown in the next figure.

9-36

Downloaded from www.Manualslib.com manuals search engine

Refining the Controller Design to Meet Controller Output Bounds

The LTI Viewer also updates to show the design requirement, as shown in the next figure.

2 Add the lower-bound on the controllers output: a Right-click the white area in the bottom plot in the LTI Viewer, and

select Design requirement > New. This action opens the New Design Requirement dialog box.

9-37

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

b Select Lower time response bound from the Design requirement

type drop-down list. The New Design Requirement dialog box updates to show the lower bound, as shown in the next figure.

c In the Time field of the End column, enter Inf. d In the Amplitude field of the Start column, enter -250.

9-38

Downloaded from www.Manualslib.com manuals search engine

Refining the Controller Design to Meet Controller Output Bounds

e In the Amplitude field of the End column, enter -250.

The New Design Requirement dialog box resembles the next figure. Click OK to close the dialog box.

9-39

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

The Design requirements tab in the Control and Estimation Tools Manager GUI updates. It now displays the lower-limit requirement, as shown in the next figure.

9-40

Downloaded from www.Manualslib.com manuals search engine

Refining the Controller Design to Meet Controller Output Bounds

The LTI Viewer also updates to show the lower-bound on the controllers output, as shown in the next figure.

9-41

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

3 Optimize the parameters to meet the design requirements on the controller

output:
a In the Compensators tab, select the rows containing P, I, and D, and

click Use Value as Initial Guess.

9-42

Downloaded from www.Manualslib.com manuals search engine

Refining the Controller Design to Meet Controller Output Bounds

When you run the optimization again, the optimization method uses the updated parameter values as the starting point for refining the values. Clicking Use Value as Initial Guess updates the values in the Initial Guess column, as shown in the next figure.

9-43

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

b In the Optimization tab, click Start Optimization.

At every optimization iteration, the optimization method reduces the distance between the current response and the upper and lower bounds on the signal. After the optimization completes, the Optimization tab resembles the next figure.

9-44

Downloaded from www.Manualslib.com manuals search engine

Refining the Controller Design to Meet Controller Output Bounds

4 Examine the parameter values refined controller and the systems response: a Examine the following response plots:

The LTI viewer

The bottom plot shows that the controller output lies between 550 and -250, and thus meets the design requirement on the controllers output. Note You must also check that the closed-loop response, shown in the top plot, remains stable after refining the controller design.

9-45

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

The SISO Design window

The plots show that after refining the design, the system continues to meet the magnitude and phase margin requirements specified in Design Requirements on page 9-4.

9-46

Downloaded from www.Manualslib.com manuals search engine

Refining the Controller Design to Meet Controller Output Bounds

b View the optimized controller parameter values in the Value field in the

Compensators tab.

5 Select the SISO Design Task node, and click Update Simulink Block

Parameters. This action writes the optimized controller parameter values to the Controller block in the Simulink model.

9-47

Downloaded from www.Manualslib.com manuals search engine

Tutorial Designing a PID Controller Using Optimization-Based Tuning

Saving the Project

To save a project with the optimized controller parameters:
1 In Control and Estimation Tools Manager GUI, select File > Save.

This action opens the Save Projects dialog box.

2 In the Save Projects dialog box, select Project - sldo_model2, and click

OK to open the Save Projects window .

3 In the Save Projects window, enter sldo_model2_optimized.mat in the

File name field, and click Save. The action saves the project as a MAT-file. Tip You can reload this project by typing sisotool('sldo_model2_optimized') at the MATLAB prompt.

9-48

Downloaded from www.Manualslib.com manuals search engine

A
Examples
Use this list to find examples in the documentation.

Downloaded from www.Manualslib.com manuals search engine

Examples

Getting Started
Chapter 5, Tutorial Preparing Data for Parameter Estimation Using the GUI Chapter 6, Tutorial Estimating Parameters from Measured Data Using the GUI Chapter 7, Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the GUI Chapter 8, Tutorial Optimizing Parameters to Meet Time-Domain Requirements Using the Command Line Chapter 9, Tutorial Designing a PID Controller Using Optimization-Based Tuning

A-2

Downloaded from www.Manualslib.com manuals search engine

Index
A
Index

N
newsro function

add Signal Constraint block 7-5

example 8-8

B
block parameters window 7-10

O
optimization-based control design 9-1 optimize function example 8-12

C
configure for parameter optimization 7-5 configuring control design 9-5 Control and Estimation Tools Manager 9-5

P
parameter settings 7-20 plots 9-15

D
design requirement line segments 7-8 design requirements time-domain 7-4 designing optimization-based linear controller 9-25

R
response optimization tutorial using functions 8-1 tutorial using GUI 7-1

E
Edit Design Requirement dialog box rows 7-11

S
signal constraints objects 8-8 Simulink Design Optimization product related products 1-5 required products 1-5 SISO Design Task 9-5 specify reference signal using GUI 7-25 specifying controller parameters 9-11 specifying design requirements 9-15

F
findconstr function

example 8-8 findpar function example 8-11

L
linear control design 9-5

T
time-domain design requirements 8-4 track reference signal 8-15

Index-1

Downloaded from www.Manualslib.com manuals search engine