[juanltm]

I want to simulate the Heat Transfer in easy geometries using Liquid Metals. Thats very interesting at Nuclear Fusion Research because of its great Heat Transfer capabilities.

Did somebody work with all this stuff before? Do you have any idea of which solver will be easiest to try to modify, for this type of problems?

Thank you very much

[mike_jaworski]

Hi Juan,
I work in this field at the University of Illinois at Urbana-Champaign (USA). We do work with liquid lithium PFCs and I've done some simulations with OpenFOAM on it though not many owing to some other time constraints. My long-term intention is to add some solvers for the general MHD flow problems.

That problem, as you state it, is very vague. I've looked primarily at liquid metal (LM) use in the divertor and with lithium, there's a free-surface associated with it, but the flows are slow so a laminar solver could be applicable. In blanket modules, you don't worry about the free surface (most designs I've seen), and you should *still* be laminar due to MHD reactions.

If you just want to calculate heat transfer from the fluid to the solid, the conjugateHeatFoam solver (in the development version currently, maybe in 1.5 beta?) works well for it. This is currently based on the icoFoam solver but that could be modified to include MHD effects. Modifying solvers is not hard (check the OF wiki).

Regards,
Mike J.

P.S. if you're attending PSI-18, I have a poster presentation there if you'd like to meet.

[mkraposhin]

Hi, Michael Jaworski, i'm also interested in liquid metals simulations.

is it normal to use icoFoam (which simulates Newtonian liquids) with such task?

p.s.
sorry for bad english

[mike_jaworski]

Hi Kraposhin,
The liquid metals I work with all behave as normal Newtonian fluids. In situations of zero magnetic field, any regular hydrodynamic solver for an incompressible is fine. It's just a matter of setting appropriate material constants.

For use in a magnetic field, icoFoam would have to be modified to include these effects. There is an mhdFoam solver packaged with openFOAM right now, however this doesn't presently solve for electric fields and currents so you can't specify all the boundary conditions necessary for general MHD simulations. Current/charge tracking is needed for that (conducting non-conducting boundary conditions, etc).

Regards,
Mike

[mkraposhin]

Michael Jaworski, thank you

[alexandrepereira]

Hi Michael

I am studying at Universidade de Coimbra, my base Education is Mechanical Engineering.
I have this putative design of a MSBR ( Molten Salt Breeder Reactor ) for submarine propulsion application in which the working fluid ( and fuel ) is a solution of fissile material which generates heat in a reflected/moderated cavity, expands as a gas to a low pressure chamber, releasing its excess kinetic energy in an MHD generator, before condensing end releasing its heat enthalpy to a heat sink before completing the rankine heat cycle.

Part of the electromotive force produced in the MHD generator could be used in a MHD thruster using sea water as a working fluid.

Controlling carefully the flow regimes in the MHD generator and MHD thruster, very stealthy operation could be acheived.

I am "endeavouring" to create an application for MHD flow of liquid metal ( some sort of mhdIcoFoam, solving the conservation of moment equation plus the source term due to EM coupling, kJxB , the continuity eq in incompressible regime, the, the heat transport eq with the ohmic source term, plus the Maxwell eqs.

Since my base education is in Mechanical Engineering and i do not know much of Computational Electromagnetics, I ask you:

Which kind of simplifications should i make to the Maxwell Equations so as to solve the Viscous Incompressible MHD flow, besides the strategy of calculation that I outlined above ( finding J and B, calculating vector product, and patching in on the source term momentum eq of Navier Stokes equations before solving for U and p, and using a function of J density as a source term in the heat diffusion process)...?

Thanks in advance... :-)

Alexandre

[mike_jaworski]

Hi Alexandre,

The "standard" simplification in MHD with regard to Maxwell's equations is to neglect the displacement current term. The challenge in doing these flows is "simultaneously" solving all the fields required. I'd suggest looking first at mhdFoam to see how that one was implemented since it illustrates a method for solving the B field equation and enforcing zero divergence. If you want to brute-force the equations, I once imagined a "simple" extension of mhdFoam that retained a current continuity equation (and field, obviously), so that other boundary conditions could be applied besides purely conducting.

An alternative method that I was leaning toward was to implement a vector-potential equation and reducing the EM equations that way. Similar to the stream function/vor

One of the people I collaborate with uses an MHD code and they have described the equations they solve in some publications. check for

N. B. Morley et al., Fusion Engineering and Design, v.72 (2004) 3-34.

to see what equations they solve.

Regards,
Mike J.

[PetSul]

Originally Posted by mike_jaworski

Hi Juan,
I work in this field at the University of Illinois at Urbana-Champaign (USA). We do work with liquid lithium PFCs and I've done some simulations with OpenFOAM on it though not many owing to some other time constraints. My long-term intention is to add some solvers for the general MHD flow problems.

That problem, as you state it, is very vague. I've looked primarily at liquid metal (LM) use in the divertor and with lithium, there's a free-surface associated with it, but the flows are slow so a laminar solver could be applicable. In blanket modules, you don't worry about the free surface (most designs I've seen), and you should *still* be laminar due to MHD reactions.

If you just want to calculate heat transfer from the fluid to the solid, the conjugateHeatFoam solver (in the development version currently, maybe in 1.5 beta?) works well for it. This is currently based on the icoFoam solver but that could be modified to include MHD effects. Modifying solvers is not hard (check the OF wiki).

Regards,
Mike J.

P.S. if you're attending PSI-18, I have a poster presentation there if you'd like to meet.

Does conjugateHeatFoam apply for solid-liquid cases? I'm trying to simulate heat transfer in a case where there is a solid cheramic particle and liquid metal surrounding it. It such a simple case, the velocity of liquid phase is zero. So it's basicly just the heat transfer between the phases. But I'm not getting anywhere with it. I'm using OF 1.6 and there is chtMultiRegionFoam solver but it doesn't apply for my cases as it is.

Can I solve this by modifying some other solver, icoFoam perhaps?

Regards,

Petri Sulasalmi