[bcade]

Hello,

I have been working on a simulation of water flowing through a microchannel heat sink that I anticipate boiling will occur in.

I have had success getting converged simulations using the RPI boiling model that seem physically realistic, however the convergence is extremely messy and temperamental and tends to break if I change the parameters I want to design to (heat flux in, or flow rate).

I have discovered that I get beautiful and reliable convergence if I simply disable the RPI boiling model yet leave Thermal Phase Change on. I also see physically reasonable vapour production (albeit less than when I happen to get RPI to converge).

What is the physical interpretation of running Thermal Phase Change *without* the RPI wall boiling? Should I be able to do this, or are these fictitious results? I haven't found the ANSYS documentation specifically says anything about what Thermal Phase Change does by itself.

Thanks so much!

[ghorrocks]

Good question. I am not an expert on modelling phase change, but I have a few comments:

The RPI wall boiling model is developed for modelling of large structures, things like power station heat exchangers and industrial processes. It assumes that there are lots of nucleation sites and that the bubbles are small relative to the surface size. In microchannels you have few (or even zero) nucleation sites and the bubble size is not small compared to the surface size. So I would not think the RPI boiling model is a suitable physical model for boiling in a microchannel.

So I would think that I would probably just use the simple phase change model without RPI if I was you - but I am no expert on the area and have not checked this in detail.

[bcade]

Thanks! Just knowing the history of RPI gives me some confidence that it might be inappropriate and causing my solution to break. I'll perform some more simulations using Phase Change and then find some data to verify.