[danbence]

In order to visualise a volume of interest through the domain of a transient simulation, I have tried to use temperature as a marker. I set the thermal conductivity of air to be practically zero and wrote a UDF to initialise the volume of interest with one temperature and the rest of the domain with a second temperature. I also used a very fine mesh, in order to minimise the smear of temperature caused by the discretization.

To my great surprise, after one small timestep, the hard transition between the two temperatures has been spread/smeared out by orders of magnitude greater distance than the thermal conductivity would allow.

Is my methodology flawed ? My guess is that even though it is a transient simulation with a small time step, the solver is trying to solve the temperature from the boundary conditions instead of being faithful to the initialisation temperatures.

I will probably swap to using particle tracking, but it has shaken my faith that fluent will properly represent the physics of the mixing of different temperature liquids.