Learning Forum

[Jada]

Hi,

I am working on modelling the melting and solidification heat transfer behaviour of a phase change material (PCM) in different shaped containers. The designs are modelled in 2-D. 

For designs with curved surfaces, after some time into the melting simulations, the flow becomes asymmetric when the emerging liquid phase comes into contact with the curved surface (see images below). The flow remains asymmetric, often keeping to one side- what would be the cause of this asymmetry? This is a 2-D transient simulation, where the flow is modelled as laminar, and the magnitude of velocity is very small (max. 10^-3). 

I have tried mesh refinement, and also gradient adaption, but the flow is still asymmetric after coming into contact with the curved surface. The solution achieves convergence within 20 iterations, and at a time step that is still above the minimum that I have set (I am using adaptive time-stepping), so I'm not sure what the problem may be. 

I would like to know/understand what may be the causes of the asymmetry that I am seeing. Any help would be greatly appreciated.

Many thanks,

Jada

 

Velocity contours:

   

 

[Rob]

Welcome to the weird and wonderful world of fluid dynamics. You may want to look up Coanda Effect, Pitchfork Bifurcation and fluidics (not microfluidics) for some background reading. The above looks plausible; does the flow continue to oscillate or is the last image the stable position? 

[Jada]

 

Hi Rob, thank you for getting back to me. I will look into the Coanda effect..etc as you suggested. The last image is the stable position when 100% liquid fraction is achieved – is there a reason why the flow maintains its position on the right-hand side and not the left?

 

 

[Rob]

That's why I gave you a reading list! ;)   In the experiment burrs on geometry, fingerprints on the materials, sunlight etc can cause this. In the CFD model mesh and solver numerics can trigger the initial asymmetry, and once the flow becomes asymmetric the pressure field that's created will amplify the effect: if other flow features are present/absent the flow may return to the symmetric position or oscillate from left to right with time. 

[Jada]

Ok, I will do some reading around this- thank you for your help!