Product Documentation

Search MATLAB Documentation
R2012a Documentation → Model-Based Calibration Toolbox
View documentation for other releases	Learn more about Model-Based Calibration Toolbox

Contents

Index

• Getting Started

• Product Description

• What is Model-Based Calibration?

Designs and Modeling in the Model Browser

Calibration Generation in CAGE

• Examples in Model-Based Calibration Toolbox

Examples of Engine Modeling and Calibration

Tutorials for Design and Modeling

Tutorials for Calibration Generation

Training Material

• System Requirements

Operating System Requirements

Required MathWorks Products

Optional MathWorks Products

• Limitations

Waitbar May Have Transparent Background

• Gasoline Case Study Overview

Case Study Introduction

Why Use Design of Experiment and Engine Modeling?

Problem Definition

Introduction to Two-Stage Modeling

• Designing the Experiment

Overview of Design Process

Creating a Test Plan

Specifying the Model Inputs

Creating Designs

• Importing and Filtering Data

Import and View Data

How Is a Two-Stage Model Constructed?

• Building the Models

Completing Model Setup

Specifying the Local Model Type

Selecting Data and Responses to Model

Selecting Local Models

Creating Boundary Models

• Selecting Global and Two-Stage Models

Inspect the Global Models

Create Multiple Models to Compare

Create a Two-Stage Model

Adding New Response Models

Using Validation Data

Exporting the Models

• Optimized Calibration

Problem Definition

Benefits of Automated Calibration

Importing Models into CAGE

• Setting Up Calibration Tables to Fill

Altering Variable Ranges

Setting Up Tables

Setting Up the Optimization

Defining Variable Values

Running the Optimization

• Setting Up the Sum Optimization

Setting Up the Optimization Initial Values and Objective

Creating Table Gradient Constraints

Running the Sum Optimization

Filling Tables with Optimization Results

• MBT Spark Estimator Problem

What Is an Estimator?

View the Feature

Edit and Import Boundary Model

Use the Feature Fill Wizard

Inspect Results

CAGE Import Tool

• Diesel Case Study Overview

Problem Definition

• Design of Experiment

Introducing Design of Experiment

Constraining the Design

Creating Candidate Designs

Data Collection

Overview of Modeling Process

Building Models

Building and Evaluating Alternative Models

Creating Boundary Models

• Optimized Calibration

Problem Definition

Benefits of Automated Calibration

Importing Models of Engine Responses into CAGE

Defining Additional Variables and Models

Setting Up Calibration Tables to Fill

Setting Up the Point Optimization

Setting Up Constraints

Defining Variable Values

Running the Optimization

• Setting Up the Sum Optimization

Setting Up Initial Values and Sum Objective

Creating the Brake Specific NOx Constraint

Setting Weights for the Sum Objective and Constraint

Set Parameters and Run Optimization

Filling Tables with Optimization Results

• Point-by-Point Diesel Case Study Overview

What Is Point-by-Point Modeling?

Engine Calibration Specifications and Problem Description

• Create Designs and Models Programmatically

Overview of Programmatic Design and Modeling

Creating Designs and Models Programmatically

• Verify and Refine Models in the Model Browser

Open the Project and View Designs

Analyze and Refine Local Fits

Point-by-Point Optimization Overview

• Create Point Optimizations in CAGE

Load Models to Optimize

Create Part Load Point Optimization

Create Full-Load Point Optimization

• Create Multiregion Sum Optimization

Create Data Set from Point Optimization Results

Create Sum Optimization

Add Application Point Sets

Set Up Constraints

Define Weights and Fixed Variables and Run Optimization

• Use Optimization Results

Create Tables to Fill

Fill Tables from Optimization Results

Introducing Composite Models and Modal Optimization

• Example Composite Model and Modal Optimization Projects

Gasoline Example with Four Cylinder and Eight Cylinder Modes

Diesel Example with Low and High EGR Modes

Composite Model Example with Separate Tables for Each Mode

Introduction to the Command-Line Interface

• Tutorial: Design and Modeling Scripts

Processes You Can Automate

Scripting Demos

• Understanding Model Structure for Scripting

Projects and Test Plans for Model Scripting

Response Model Scripting

Boundary Model Scripting

How the Model Tree Relates to Command-Line Objects

Two-Stage Models

Starting the Toolbox and Loading Data

• Setting Up the Model

Creating the Test Plan

Setting Up the Local Model

Setting Up the Global Model

Specifying the Response Model

• Verifying the Model

Verifying the Local Model

Verifying the Global Model

Selecting the Two-Stage Model

Comparing the Local Model and the Two-Stage Model

Maximum Likelihood Estimation

Exporting the Model

• Creating Multiple Models to Compare

Methods For Creating More Models

Creating New Local Models

Adding New Response Features

Comparing Models

Creating New Global Models

Creating Multiple Models Using Build Models

• What Is Design of Experiment?

Why Use Design of Experiment?

Structure of This Design Editor Tutorial

• Getting Started with the Design Editor

Setting Up a Model

Starting the Design Editor

Creating a New Design

• Creating Optimal Designs

Introducing Optimal Designs

Start Point Tab

Candidate Set Tab

Viewing Design Displays

• Using the Prediction Error Variance Viewer

Introducing the Prediction Error Variance Viewer

Improving the Design

• Creating Classical Designs

Adding a Classical Design

Classical Design Browser

Setting Up and Viewing a Classical Design

Using the Design Evaluation Tool

Creating Space-Filling Designs

Applying Constraints

Introduction to the Data Editor

• Loading the Data

Entering the Data Editor

Loading a Data File

• Viewing and Editing the Data

Using Notes to Sort Data for Plotting

Removing Outliers and Problem Tests

Reordering and Editing Data

• User-Defined Variables and Filtering

Adding New Variables

Applying a Filter

Sequence of Variables

Deleting and Editing Variables and Filters

• Matching Data to Designs

Introducing Matching Data to Designs

Tolerances and Cluster Information

Understanding Clusters

What Are Feature Calibrations?

• Setting Up Calibrations

Setting Up Variables

Setting Up Models

• Setting Up a Feature Calibration

Setting Up a New Feature

Setting Up the Strategy

Setting Up the Tables

• Calibrating a Feature

Process For Feature Calibration

Calibrating the Normalizers

Calibrating the Tables

Calibrating the Feature

Exporting Calibrations

What Is a Tradeoff Calibration?

• Setting Up a Tradeoff Calibration

Creating a Tradeoff

Adding Tables to a Tradeoff Calibration

Displaying the Models

• Performing the Tradeoff Calibration

Process Overview

Checking the Normalizers

Setting Values for Other Variables

Filling Key Operating Points

Filling the Table by Extrapolation

Exporting Calibrations

• Setting Up the Data Set

Tutorial Overview

Opening an Existing Calibration

Importing Experimental Data into a Data Set

Adding an Item to a Data Set

• Comparing the Items in a Data Set

Viewing the Data Set as a Table

Viewing the Data Set as a Plot

Displaying the Error

Coloring the Display

Reassigning Variables

• Setting Up a Table and Experimental Data

Tutorial Overview

Adding Variables

Adding a New Table

Importing Experimental Data

Filling the Table from the Experimental Data

Selecting Regions of the Data

Exporting the Calibration

• Tutorial Overview

Tutorial Problem Definitions

The Optimization View

Import Models to Optimize

• Single-Objective Optimization

Process Overview

Using the Create Optimization from Model Wizard

Setting Constraints and Operating Points

Running the Optimization

Using Optimization Results to Fill Tables

Using a Custom Fill Routine to Fill Tables

• Multiobjective Optimization

Setting Up and Running the Multiobjective Optimization

Optimization Output View

Selecting Best Solutions

• Sum Optimization

What Is a Sum Optimization?

Automated Tradeoff

• Worked Example Optimization

Example Overview

Using the Worked Example Optimization

• Creating an Optimization from Your Own Algorithm

Process Overview

Step 1: Verify the Algorithm

Step 2: Create a CAGE Optimization Function

Step 3: Define the Optimization Options

Step 4: Add the Algorithm to the Optimization Function

Step 5: Register Your Optimization Function with CAGE

Step 6: Verify Your New Optimization

• User's Guide

• Release Notes

Summary by Version

Version 4.4 (R2012a) Model-Based Calibration Toolbox Software

Version 4.3 (R2011b) Model-Based Calibration Toolbox Software

Version 4.2 (R2011a) Model-Based Calibration Toolbox Software

Version 4.1 (R2010b) Model-Based Calibration Toolbox Software

Version 4.0 (R2010a) Model-Based Calibration Toolbox Software

Version 3.7 (R2009b) Model-Based Calibration Toolbox Software

Version 3.6 (R2009a) Model-Based Calibration Toolbox Software

Version 3.5 (R2008b) Model-Based Calibration Toolbox Software

Version 3.4.1 (R2008a+) Model-Based Calibration Toolbox Software

Version 3.4 (R2008a) Model-Based Calibration Toolbox Software

Version 3.3 (R2007b) Model-Based Calibration Toolbox Software

Version 3.2 (R2007a) Model-Based Calibration Toolbox Software

Version 3.1 (R2006b) Model-Based Calibration Toolbox Software

Version 3.0.1 (R2006a) Model-Based Calibration Toolbox Software

Version 3.0 (R14SP3+) Model-Based Calibration Toolbox Software

Version 2.1.1 (R14+) Model-Based Calibration Toolbox Software

Version 2.1 (R14) Model-Based Calibration Toolbox Software

Version 2.0 (R13+) Model-Based Calibration Toolbox Software

Compatibility Summary for Model-Based Calibration Toolbox Software

pevEvaluate - Evaluate prediction error variance (PEV)

Syntax

Y = pevEvaluate(optimstore, X)

Description

Warning The evaluation of PEV is no longer supported in cgoptimstore and this method will return PEV values of zero (as detailed below) if called.

A method of cgoptimstore.

Y = pevEvaluate(optimstore, X, itemnames)

returns PEV values of zero for objectives/constraints at the free variable values X. X is a (NPoints-by-NFreeVar) matrix where NPoints is the number of points to be evaluated and NFreeVar is the number of free variables in the optimization.

Y = pevevaluate(optimstore, X, objconname, datasetname)

returns PEV values of zero for the objectives/constraints at the operating points in the data set specified by the string datasetname.

Y = pevevaluate(optimstore, X, objconname, datasetname, rowind)

returns PEV values of zero for the specified objectives/constraints at the points of datasetname given by rowind. X must be a (NRows-by-NFreeVar) matrix where NRows is the length of rowind. rowind must be a list of integer indices in the range [1 NumRowsInDataset].Y is a (Nrows-by-NItems) matrix.

How To

gridPevEvaluate

Recommended Products

Get MATLAB trial software

Includes the most popular MATLAB recorded presentations with Q&A sessions led by MATLAB experts.

Get the Interactive Kit

© 1984-2012- The MathWorks, Inc. - Site Help - Patents - Trademarks - Privacy Policy - Preventing Piracy - RSS