The evaluate function in SDF.NDTable was blocking generation of analytical ODE Jacobian for me. I found that the c-function to extract the table derivative is already available, so an easy fix is to add this annotation to evaluate:
annotation(derivative=evaluate_der);
and declare evaluate_der as
function evaluate_der
input SDF.Types.ExternalNDTable table;
input Real[:] params;
input SDF.Types.InterpolationMethod interpMethod;
input SDF.Types.ExtrapolationMethod extrapMethod;
input Real[:] params_der;
output Real value;
external "C" value = ModelicaNDTable_evaluate_derivative(table, size(params, 1), params, interpMethod, extrapMethod, params_der) annotation (
Include="#include <ModelicaNDTable.c>",
IncludeDirectory="modelica://SDF/Resources/C-Sources");
end evaluate_der;The following test will verify that derivatives are correctly evaluated
model InterpolationMethodsDer "Tests table output differentiation"
extends SDF.Examples.InterpolationMethods;
Real der_hold = der(hold.y);
Real der_nearest = der(nearest.y);
Real der_linear = der(linear.y);
Real der_akima = der(akima.y);
Real der_fritsch_butland = der(fritsch_butland.y);
Real der_steffen = der(steffen.y);
end InterpolationMethodsDer;
The evaluate function in SDF.NDTable was blocking generation of analytical ODE Jacobian for me. I found that the c-function to extract the table derivative is already available, so an easy fix is to add this annotation to evaluate:
annotation(derivative=evaluate_der);and declare evaluate_der as
The following test will verify that derivatives are correctly evaluated