If I’m not mistaken, the adaptation of time-step is made only in the diffusion problem. Assume that the HeatTransferProblem is also complicated, e.g. heating due to high-flux pulse-periodic loads. Won’t it be useful to adapt the time-step during solving the heat equation, so the solver could have more attempts if it faces difficulties?
You are correct. This was because the original heat transfer problems were very simple.
It would be nice to have adaptive timestepping for heat transfer too.
Ideally, for coupled problems, we would have 2 timesteps for each problem (Heat transfer and H transport) and the timestep of the coupled problem is the minimum of both.