Hello, I want to ask a few questions as follows:

1)How are T energy imbalance and H energy imbalance? (My system consists of steel block and cooling water which passes the inside of the block. By the way, the domain for water has H-energy imbalance option as well as various options (e.g., P-mass imbalance and etc) while the block domain has the only T-energy imbalance in cfx-solver.

2) How can I interpret the imbalance from monitor in cfx-solver? (As I explained above, my model has two domains of water and block. So, '%imbalance=domain imbalance/maximum over all domain' is needed for my case (i.e., multidomains)? If yes, maximum over all domain means maximum imbalance over all domain? Am I right?)

3) In the block, there are two different sizes of cross-secitonal area diameter (hole diameter) for water to pass. For example, the water passes the inlet hole with big cross-sec diameter first and later the water is divided according to two small holes (this cross section is very small). For a certain water flow rate, the water is therfore turbulent at the inlet hole and simultaneously laminor at the small hole. In this case, how do I have to set simulation options in cfx-pre in order to cover both water's turbulance and laminar flow at different locations?

Thank you so much.

Hi,

1) CFX solves an enthalpy equation in the fluid regions but a temperature equation in the solid regions. The imbalance is error in the conservation of that quantity. Read the documentation for more information.

2) No, the balances are calculated from the entire solution space. It is not a maximum imbalance, but a total imbalance of the entire solution. Again this is described in the documentation.

3) Relaminarisation of flows is not modelled in CFX. However, I have had some success in modelling flows like this using a k-w based turbulence model (such as SST) as these models allow the turbulence intensity to go to zero and predict a laminar-like state with no turbulent viscosity. It's not perfect but good enough for most engineering applications. The turbulence transition model cannot help you as it only predicts laminar to turbulent transitions, not the other way.

Glenn Horrocks

Hi, Glenn. Thank you for your kind answer. If you don't mind, I want to ask additional questions. 1) If I set the default domain-> fluid models->turbulence model option: k-Epsilon (in cfx-pre), then this option can cover both laminar and turbulent cases ( I mean it covers all ranges of reynolds number)? 2) From the energy imbalance graph presented in monitoring window (i.e., plot for energy imbalance in cfx-solver), the unit of variable value (y value) is the percentage? OR unitless?

Could you please let me know answers? Thanks a lot.

Sincerely,

Hi,

1) The k-e model does not handle low turbulence levels well. The k-w family of turbulence models do. Have a look at "Turbulence Modeling for CFD" by Wilcox for details.

2) The imbalances displayed in Solver manager are percent.

Glenn Horrocks

Dear Glenn,

Thank you for your kind answer again. Glenn, I am not sure still about the meaning of y value for the imbalance plot displayed in Solver manager... If y is zero, does it mean energy in = energy out? In my opinion, y is residual value for imbalance or energy equation measurment . So, if y is graudually approaching to 0 (or equal to zero), it means that residual is zero and thus the solution is convering. But this convergence does not gurantee the energy conservation (i.e., energy in = energy out). Because the solution convergence can be also guranteed in case of energy inconservation... If you don't mind, could you please give me an answer? It might be the last question... Thanks again.

Hi,

If the imbalances are zero that means the energy equation has fully balanced. That is, integrated across the entire simulation domain Energy in - energy stored = energy out.

This is another form of convergence criterion. The residuals used by CFX as a default are another. It is quite possible for the residuals to have converged well but the imbalances to not be converged; and also it is quite possible for the residuals to not have converged but the imbalances have converged.

Feel free to ask as many questions as you wish, that is why the forum is here!

Glenn Horrocks

Thank you so much for your kind answer again, Glenn. You are a treasure in this forum! Thanks again. Have a great day~!