Postprocessing Sensitivities
When a multiphysics problem is solved using sensitivity analysis, the generated solution contains stored sensitivity data. You can access this data in postprocessing using the fsens and sens operators:
fsens(<control_variable>) evaluates the sensitivity (derivative) of the objective function with respect to the specified control variable. This result is available for all sensitivity methods. The result of fsens can be evaluated on the geometric entities where the control variable is defined. For a global control variable, fsens is available everywhere. In the same way, fsensimag(<control_variable>) evaluates the sensitivity (derivative) of the objective function with respect to the imaginary part of the specified control variable.
sens(<dependent_variable>,<control_variable>) or sens(<dependent_variable>,<control_DOF>) evaluates the sensitivity (derivative) of the specified dependent variable with respect to the specified control variable degree of freedom. This is only possible when forward sensitivity has been used, which computes and stores derivatives of the entire solution vector with respect to each control variable degree of freedom.
Global control variables can be identified by name. Otherwise, control variable degrees of freedom are identified by their index (starting from 1) among all control variables in the solution vector. The result of sens has the same geometric scope as the dependent variable argument; it can be plotted or evaluated wherever the dependent variable itself is available.