Sensitivity
Handle sensitivity solver parameters.
Syntax
model.sol(sname).feature(solv).create(fname,"Sensitivity")
model.sol(sname).feature(solv).feature(fname).set(pname,value)
Description
Attribute feature. This feature can be used to make analytic forward or backward (adjoint) sensitivity analysis. This analysis is done after the main problem has converged. The solution approach (coupled or segregated, Jacobians, and so on) for the main problem is reused.
Table 6-64:
Valid Sensitivity Property/value Pairs
Property
Value
Default
Description
sensfunc
String
Sensitivity functional variable name
sensmethod
none
|
adjoint
|
forward
|
numeric
none
Sensitivity analysis method
The forward numeric method (
numeric
) is a variant of the forward sensitivity method where the right-hand side of the sensitivity problem is computed by numerical differentiation.