I am planning on creating models where a mechanical model (PDE solved in FEniCSx) is coupled to some rigid bodies, so there will be additional degrees of freedom not associated with nodes in the mesh (e.g., a mass connected to an elastic body by a spring). These additional degrees of freedom will be coupled to the nodal degrees of freedom, so if possible, I would like to incorporate them into my energy equations that I then minimize like normal.
Are there ways of introducing (and solving for) non-nodal degrees of freedom into FEniCSx problems, or will this require a very customized solver? I am more than happy to write a custom solver, but I would really like to retain the benefits of things like symbolic differentiation that FEniCSx/UFL provides.
Any guidance would be greatly appreciated!