Hello, I would like to model the heating of bitumen in one tank. At firt approximation a 2d model composed by one box and three circles (tubes). The tubes are at 460 K, the box walls are adiabatic. The properties of the bitumen ... well thats the problem. I think is not a problem of melting, but the viscosity grows a lot at ambient temperature (probably is a viscoelastic material at 300 k). The density changes too (but an small quantity).

If the bitumen is cold (300 K), and the tubes begins to heat it up, at first stages probably the problem is a solid heat conduction problem. Then the viscosity is reduced and reduced, small zones of buoyancy driven flow begins to occur, and finally there are a big recirculation zone.

Do you think this can be modelled by Fluent ???

Yes, I think you can do this. You will need to have accurate viscosity and density versus temperature data for the bitumen. You may even start by solving only the energy equation until some amount of the bitumen becomes liquid enough to start flowing reasonably. I seriously doubt is there is a steady-state solution here. I think you're going into transient-land.