Ported 2p2c and co2 model and all test. Unfortunately most tests still fail. In the co2ni tests the enthalpy is very different from the reference. This might point to something why the 2p2cni tests also fail. The co2_tpfa test probably fails because it is an unstructured grid and we do something different for tpfa there. We might check by computing old style transmissibilities.
Introduces boundary flags for box/tpfa subcontrol volume faces. Not sure if this is the best solution though.