The Problem arises when defining a Simulink Model with a 1D, 2D, 3D, nD-Lookuptable; The Table is in Workspace as Simulink.LookupTable with dimension 1,2,3, n initialized; Compiling with the Catia FMIKit-Simulink v2.9 works fine, since the modelDescription.xml shows correctly defined BP1 and Table Values;
For Example, for the first point of the 1D Table:
<ScalarVariable name="Parameters.TEST_1D.BP1[0]" valueReference="66" causality="parameter" variability="tunable">
<Real/>
</ScalarVariable>
...
<ScalarVariable name="Parameters.TEST_1D.Table[0]" valueReference="71" causality="parameter" variability="tunable">
<Real/>
</ScalarVariable>
but when compiling with the FMI Kit Version 3.0 or higher, there is a problem in the definition of the Points; first you inverted since Version 3.0 the Breakpoint Definition and the Table Values (that could be a Problem!); but the Definition of the Table Point looks wrong, since you define two dimensions in the table; here for example, the same first Breakpoint with the first value:
<ScalarVariable name="Parameters.TEST_1D.Table[0,0]" valueReference="66" causality="parameter" variability="tunable">
<Real start="0"/>
</ScalarVariable>
....
<ScalarVariable name="Parameters.TEST_1D.BP1[0]" valueReference="71" causality="parameter" variability="tunable">
<Real start="0"/>
</ScalarVariable>
I have noticed, that the FMU compiled with FMIKit 2.9 simulates correctly (although version 2.9 had a Problem with 3D -> nD Lookuptables, so I could test it only with 1D and 2D LookupTables), but the one generated with 3.0 or 3.1 still interpolates the 1D Lookuptable surprisingly correctly, but the 2D-3D-4D Lookuptables are completely wrong interpolated.
Hier is an Example, with a small Simulink Model with a 0D, 1D and 2D Lookuptable and a .mat File with the Workspace Variables of type Simulink.LookupTable TEST_0D, TEST_1D and TEST_2D; the three FMUs were compiled with the FMIKit v2.9, v3.0 and v3.1-rc respectively
TEST_0D_1D_2D_Lookuptable.zip
And a Second Example with also 3D and 4D Lookuptables compiled just with the FMIKit 3.1rc:
TEST_0D_bis_4D_Lookuptable.zip
The Problem arises when defining a Simulink Model with a 1D, 2D, 3D, nD-Lookuptable; The Table is in Workspace as Simulink.LookupTable with dimension 1,2,3, n initialized; Compiling with the Catia FMIKit-Simulink v2.9 works fine, since the modelDescription.xml shows correctly defined BP1 and Table Values;
For Example, for the first point of the 1D Table:
but when compiling with the FMI Kit Version 3.0 or higher, there is a problem in the definition of the Points; first you inverted since Version 3.0 the Breakpoint Definition and the Table Values (that could be a Problem!); but the Definition of the Table Point looks wrong, since you define two dimensions in the table; here for example, the same first Breakpoint with the first value:
I have noticed, that the FMU compiled with FMIKit 2.9 simulates correctly (although version 2.9 had a Problem with 3D -> nD Lookuptables, so I could test it only with 1D and 2D LookupTables), but the one generated with 3.0 or 3.1 still interpolates the 1D Lookuptable surprisingly correctly, but the 2D-3D-4D Lookuptables are completely wrong interpolated.
Hier is an Example, with a small Simulink Model with a 0D, 1D and 2D Lookuptable and a .mat File with the Workspace Variables of type Simulink.LookupTable TEST_0D, TEST_1D and TEST_2D; the three FMUs were compiled with the FMIKit v2.9, v3.0 and v3.1-rc respectively
TEST_0D_1D_2D_Lookuptable.zip
And a Second Example with also 3D and 4D Lookuptables compiled just with the FMIKit 3.1rc:
TEST_0D_bis_4D_Lookuptable.zip