Remove components after setup in OpenMDAO

I would like remove (or redefine) a component after setup. Is this possible with the reconfigurable model execution?

Solution

that is technically possible with reconfigurability. However, as of openmdao 2.2.1 that is not a fully supported feature. You'll need to experiment a bit on your own.

There are some tests that can show you how to get started.

Final boundary constraint with states from two tandem phases in Dymos
How to handle large sparse inputs/outputs within an OpenMDAO model?
KeyError when instantiating Birkhoff transcription in Dymos
Is there any way to have tandem phases in dymos that don't share a t_initial and t_duration?
'Output states' in Dymos?
Constraints or objectives cannot be impacted by the design variables of the problem
Access values of design variables while solver is running (OpenMDAO with IPOPT)
Dymos: How to parallelize compute across collocation nodes when using a black-box function
OpenMDAO 3.x: relevance graph for ParallelGroup approximated totals
Modelling components in OpenMDAO where the number of residuals does not match the number of state variables
Can a SubmodelComp have its own driver?
How to tie SNOPT jacobian row and column back to OpenMDAO objective / constraint and variable
Dymos: How to specify the values of a time-varying variable
How can you force instances of MATLAB Engine to operate in parallel from Python within a Python function?
OpenMDAO Dymos Simulate Method Calls Setup Multiple Times
Proper setup of an ExternalCodeComp optimization
Runtime Error while using the class:SellarMDAPromoteConnect, as given in the OpenMDAO tutorial
OpenMDAO genetic algorithm not running in parallel?
Can multiple components edit parts of the same output in OpenMDAO?
Iterating a model in OpenMDAO
Change parameters before running simulation with Dymos for monte-carlo analysis
SqliteCaseReader' object has no attribute 'system_options' in openAeroStruct
How to fix values to component variables in OpenMDAO
How does `KSComp` behave when using the options `lower_flag`, `ref`, `ref0` and `units` and `upper` at the same time?
KSComp with variable-size constraint vector input
OpenMDAO outputs scaling in a coupled model
OpenMDAO Metamodel is not respecting training data
Is there a way to "clean up" an OpenMDAO component to be pickled after an optimization has been run?
What is openMDAO's reasoning for returning 'None' for self cancelling units
Loading KrigingSurrogate Trained Model without Storing Training Data