[Antanas]

Hello!

I try to calculate flow (laminar) of electoroconductive gas in supersonic channel of MHD generator. The channel consists of supersonic part of Laval nozzle, electrode zone and diffuzer. I use 2D problem definition. At supersonic inlet I define Static Pressure, Normal Speed and Static Temperature. Outlet is supersonic. I define magnetic induction in domain (Magnetic Field Model Option = User Defined) as (0,0,Bz(x)). This's external magnetic field induction, induced magnetic field should be neglected as magnetic Re << 1. As it's impossible to define resistance of external electrical load, I define voltages on electrodes (Electric Field Model Option for domain is Electric Potential). Walls of nozzle and diffuzer is insulators and I define Zero Fluxes there for Electric and Magnetic field. The same condition I define at inlet, outlet and electrodes (Magnetic Field for Electrodes). Electric conductivity of gas is the function of pressure and temperature (Saha equation).

The problem is that I obtain symmetric solution, though it's known that it should be pressure tilt in channel and local pressure growth at downstream end of cathode.

I wonder about defining conditions for Electric and Magnetic fields on boundaries and external megnetic induction (what option to use for Magnetic Field Model) and outlet conditions (as there may be local subsonic regions).

I calculated this flow (Euler statement) using my fortran program. I used stationary Maxwell equations resolved about electric potential Phi (Poisson partial difference equation). On insulators I used Jn=0 condition, at inlet and outlet Ex=0 or dPhi/dx=0 which is equivalent. There was not any condition for magnetic field as it's not required. For gasdynamics at outlet I used zero gradient conditions: du/dx=0, dv/dx=0, dp/dx=0, etc.

There was pressure tilt between walls and local pressure growth as it should be.

But what about CFX? It's lack of info about Magnetohydrodynamics in CFX. Please help! Maybe there's any tutorial (pdf / video)?

[ghorrocks]

The best source of information about the MHD model is likely to be CFX support. I would give them a call. Other than that I can only suggest you read the general FAQ on http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

[Antanas]

Originally Posted by ghorrocks

The best source of information about the MHD model is likely to be CFX support. I would give them a call. Other than that I can only suggest you read the general FAQ on http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

Thanks for advice Glenn, but I think it'll be hard to chat with them with my poor English

[ghorrocks]

I think you will find support is patient with poor English. Good English skills is not required for CFD. And CFX has support in many countries, including many non-English speaking countries.

So I recommend you give them a call anyway. It is likely to be the best source of information on your topic as it is quite specialised.

[fir]

thank you
i have problem in boundarys conditions apply for rectangular microchannal

[ghorrocks]

If you have a question start a new thread.

[Antanas]

I've solved my problem. Thanks to all.