You run a topology optimization on it ( based on the constrains defined in simulation env )
You get the resulting part ( optimized )
Now you would like to run the same simulations you run on the original model ( to verify performance ).
In order to do so you need to define the constrains from the beginning ( which may lead to mistakes ).
Should you had the ability to "carryover" the constrain from the original model ( as the constrain area haven't changed ) to the resulting ( optimized ) model – It would save considerable amount of time and reduce risk of mistakes.
this shouldn't be hard to implement as the constrained geometry remains "as is" and doesn't changed due to geometry Reconstruction.