Learn more about Model-Based Calibration

Tutorial Overview

On this page…
Tutorial Problem Definitions The Optimization View Import Models to Optimize

Tutorial Problem Definitions

In this tutorial you will use optimization to find solutions to the following problems:

• A single-objective optimization to find maximum values of torque, subject to a constraint to keep NOX emissions below a specified level. You will export the output and use it to fill a table. See Single-Objective Optimization.

• A multiobjective optimization to maximize torque and minimize NOX emissions. See Multiobjective Optimization.

• A sum optimization to maximize torque while minimizing NOX, weighted to give more importance to idle speed. See Sum Optimization.

• Using any of your optimizations to run an automated tradeoff. Once you have set up an optimization you can apply it to a tradeoff. See Automated Tradeoff.

Back to Top

The Optimization View

You can use the Optimization view to set up, run, view, and export optimizations. You must also set up optimizations here in order to use them for automated tradeoff.

Start the CAGE Browser part of the Model-Based Calibration Toolbox product by typing

cage

at the MATLAB prompt.

To reach the Optimization view, click the Optimization button in the Processes pane.

When you first open the Optimization view both panes are blank until you create an optimization. After you set up your optimizations, the left Optimization pane shows a tree hierarchy of your optimizations, and the right hand panes display details of the optimization selected in the tree, as with other CAGE processes.

For any optimization, you need one or more models. You can run an optimization at a single point, or you can supply a set of points to optimize. The steps required are as follows:

1. Import a model or models.

2. Set up a new optimization.

The following tutorial guides you through this process to evaluate this optimization problem:

MaxTQ (SPK, N, L)

That is, find the maximum of the torque model (TQ) as a function of spark (SPK), engine speed (N), and load (L). You will use the NOXFLOW model to constrain these optimization problems.

Back to Top

Import Models to Optimize

For any optimization, you need one or more models. You can use the CAGE Import tool to import models from Model Browser projects (see CAGE Import Tool in the CAGE documentation). For this tutorial you can load a CAGE project from the mbctraining directory that contains two models for the optimization problems. Load the project as follows:

1. Select File > Open Project (or the toolbar button) to choose the tradeoffInit.cag file, found in the matlab\toolbox\mbc\mbctraining directory, then click OK.

   The tradeoffInit.cag project contains two models and all the variables necessary for this tutorial. For more information about how to set up models and variables, see Variables and Models in the CAGE documentation.

2. CAGE displays the Models view. You can view your models at any time by clicking the Models button in the Data Objects pane.

   

   Observe that the project you have opened contains two models: TQ_Model and NOXFLOW_Model. In this tutorial you use these models to optimize torque values subject to emissions constraints.

3. To view the items in the Variable Dictionary, click the Variable Dictionary button in the Data Objects pane.

   

   The Variable Dictionary view appears, displaying the variables, constants, and formulas in the current project. The project already has the relevant variables defined, so you do not need to import a variable dictionary. Note that the variables have ranges and set points defined.

Back to Top

Tutorial: Optimization and Automated Tradeoff

Single-Objective Optimization

© 1984-2009- The MathWorks, Inc.    -   Site Help   -   Patents   -   Trademarks   -   Privacy Policy   -   Preventing Piracy   -   RSS