[wales]

fluent guide says that:

"you can first solve the flow equations before enabling energy. Once you have a
converged flow-field solution, you can enable energy and solve the flow and energy equations simultaneously to complete the heat transfer simulation."

If I understood correctly, I should calculate without energy equation (say 1000 iterations). Then again I should calculate 1000 iteration more with energy equation ? I cant understand how this helps converge?

And, which mesh type should I use for a 3D rectangular flow field that contains turbulence and natural convection? tetrahedral or hexahedral?

need urgent help thank you.

[CeesH]

Under solution controls > equations you can select which equations you want to solve. This can be quite useful, for example if you have a problem where the flow is steady but heat transfer is not, you can solve only the energy equation without having to solve the others every timestep (thus saving time)

In your particular application, solving first hydrodynamics with a steady temperature, and adding temperature effects later, may contribute to the stability. Starting from 0 with a very complex situation may cause convergence problems, while if you first solve the flow without temperature effects, this may act as a much better 'initial guess' for a flow with temperature effects, thus contributing to stability. Of course, it is necessary that the flow profile without temperature effects, does not differ too much from the profile including temperature. Otherwise, you will still start from a poor situation (perhaps even poorer than a fresh start). Also, there needs to be a driving force already; if the temperature differences are the only driving force, then solving flow without the energy equation will be of little use, of course.

[wales]

ceesh,

Firstly thanks a lot for your interest.I have trouble with converging. I comprehended what you meant quite well. But the question is how I carry this way out. As, I am newly experiencing fluent, I could not understand how first solution acts as a better initial guess as you said. Do you mean that after I have calculated without energy equation, the fluent keeps the solution in its memory then uses it in another calculation that utilizes energy equation ?

So I should first calculate lets say 1000 iterations without energy eq, then reinitialize it and calculate another 1000 with energy eq ? Is it correct?

[LuckyTran]

Originally Posted by wales

So I should first calculate lets say 1000 iterations without energy eq, then reinitialize it and calculate another 1000 with energy eq ? Is it correct?

Calculate 1000 iterations without energy, then enable energy and continue iterating. Do not re-initialize as that overwrites the stored solution to the parameters on the initialization tab.