模式l Description

Open the modelsldemo_applyVarStruct.

This model uses a number of numeric parameters from the base workspace.

Example Requirements

This example changes the original model and creates files in the current working directory. If you plan to save the example models, save a copy in the current directory.

Typical Workflow

Migrating a model to use structure parameters follows these steps:

1. Collect all numeric variables used by the model into a MATLAB structure.

2. Modify the MATLAB structure (if necessary).

3. Apply the MATLAB structure to the model.

4. Configure the appearance of the parameter structure in the generated code (optional).

Collecting Numeric Variables into a MATLAB Structure

You can useSimulink.findVarsto find out what variables are used by the Simulink model.

model = 'sldemo_applyVarStruct'; varList = Simulink.findVars(model, 'WorkspaceType', 'base')

varList = 20x1 VariableUsage array with properties: Name Source SourceType Users

var = varList(1)

var = VariableUsage with properties: Name: 'Ka' Source: 'base workspace' SourceType: 'base workspace' Users: {'sldemo_applyVarStruct/Controller/Gain3'}

You can useSimulink.BlockDiagram.createVarStructto generate a MATLAB structure containing all numeric variables from the base workspace or data dictionary used by the model.

Simulink.BlockDiagram.createVarStruct(model, 'ModelParam')

Created variable 'ModelParam' in base workspace.

Variable模式lParamis a MATLAB structure that contains the numeric variables as fields (sorted by name).

NOTE:

Modifying the MATLAB Structure

The original structure contains all the variables in one level of hierarchy. You can change the structure before applying it to the model. For example, reorder parameter fields, add extra levels to group model parameters, or add extra fields.

NOTE:

1. One way to do this is to useSimulink.saveVarsto write the structure to a MATLAB script and make changes in the file.

Simulink.saveVars('sldemo_applyVarStruct_data.m','ModelParam') edit sldemo_applyVarStruct_data.m

2. Edit the MATLAB script as needed. For example, split the structure into two structures.

ControlParam = struct; ControlParam.Ka = 0.677; ControlParam.Kf = -1.746; ControlParam.Ki = -3.864; ControlParam.Kq = 0.8156; ControlParam.Tal = [0.3959 1]; ControlParam.Ts = [0.1 1]; ControlParam.W1 = [1 2.971]; ControlParam.W2 = [1 4.144];

ModelParam =结构;ModelParam。Md = -6.8847;ModelParam。Mq = -0.6571; ModelParam.Mw = -0.00592; ModelParam.Swg = 3; ModelParam.Ta = 0.05; ModelParam.Uo = 689.4; ModelParam.Vto = 690.4; ModelParam.Zd = -63.9979; ModelParam.Zw = -0.6385; ModelParam.a = 2.5348; ModelParam.b = 64.13; ModelParam.g = 32.2;

Save the modified data file. You can use it as a preload function.

3. Run the modified MATLAB script to create the MATLAB structures.

sldemo_applyVarStruct_data;

Applying the MATLAB Structure to the Model

You can useSimulink.BlockDiagram.applyVarStructto automatically replace the variable references in the block diagram by the equivalent structure references.

1. Apply the MATLAB structure to the model. The model will be changed. You can then save it to a writable directory.

[applied,unapplied]=Simulink.BlockDiagram.applyVarStruct(model,'ControlParam')

[applied,unapplied]=Simulink.BlockDiagram.applyVarStruct(model,'ModelParam')

NOTE:

Simulink.BlockDiagram.applyVarStruct(model,'ModelParam', 'ApplyMode', 'preview')

Simulink.BlockDiagram.applyVarStruct(model,'ModelParam', 'SearchMethod', 'cached')

2. Test the MATLAB structure in the modified model.

Clear unnecessary variables and run the simulation to check results.

clear; model = 'sldemo_applyVarStruct'; sldemo_applyVarStruct_data;

sim('sldemo_applyVarStruct');

Configuring the Structure Parameter to Appear in Generated Code

By default, the MATLAB structure does not appear in generated code because numeric values are inlined. Code generation requires Simulink® Coder™.

1. To make the structure variable tunable, createSimulink.Parameterwith the structure value and 'ExportedGlobal' storage class.

p=Simulink.Parameter; p.Value=ControlParam; p.CoderInfo.StorageClass='ExportedGlobal'; ControlParam=p; clear p;

rtwbuild('sldemo_applyVarStruct');

The generated code now contains a structure type definition for the tunable structure in themodel_types.hheader file. By default, Simulink Coder generates an automatic name for the type of parameter structures. This name is unique, but not easy to recognize.

typedefstruct{real_TKa; real_TKf; real_TKi; real_TKq; real_TTal[2]; real_TTs[2]; real_TW1[2]; real_TW2[2]; } struct_mn4cJ7zsH8aqc8bBlbIkqC;

2. You can control this type name by usingSimulink.Busto specify the data type of the Simulink.Parameter object.

You can useSimulink.Bus.createObjectto create a bus object with the same shape as the MATLAB structure.

busInfo 金宝app= Simulink.Bus.createObject (ControlParam.Value)

Assign the bus object name to the Simulink.Parameter DataType attribute.

ParamType=eval(busInfo.busName); ControlParam.DataType='Bus: ParamType'; clear(busInfo.busName); clear busInfo;

NOTE:OnlySimulink.Parametercan accept the bus object name as a data type. For MATLAB structures, Simulink Coder generates an automatic name for the type.

3. Resave the data file to include the modified workspace variables.

Simulink.saveVars('sldemo_applyVarStruct_data.m', '-append')

4. Regenerate code for the modified model (requires Simulink Coder).

rtwbuild(model);

The generated type definition now appears as:

typedef struct { real_T Ka; real_T Kf; real_T Ki; real_T Kq; real_T Tal[2]; real_T Ts[2]; real_T W1[2]; real_T W2[2]; } ParamType;

The parameter definition is:

ParamType ControlParam = { 0.677, -1.746, -3.864, 0.8156, {0.3959, 1.0}, {0.1, 1.0}, {1.0, 2.971}, {1.0, 4.144} }