Hi,
i want to solve a problem like this example. I understand how the k-omega model is implemented and then resolved in Modulus. However i don’t understand why is important to add the normalization to the variable “x” to “sin(x)” and at the same time why this normalization is used in flow_net, k_net and om_net but not in the p_net.
Running the same problem without this normalization generate a solution far from the solution of the example.
Thanks for your answers.
Hi @tom_02
The sin(x) here is used to implicitly enforce periodic boundary condition in the channel.
Hi, there is a physical reason for this choice?
The problem definition is a periodic channel. We found its better to implicitly force (The model can only output periodically) this as this removed a optimization loss that we would need to worry about.
Normalization is applied to “x” as “sin(x)” to ensure that the range of values of “x” remains within a reasonable and consistent range. This prevents issues like overflow or underflow of values that can lead to numerical instability in the solution.
In the specific case of flow_net, k_net, and om_net, normalization is used because these variables play a significant role in the k-omega model and directly influence the flow behavior and turbulence. Normalizing them helps to achieve a consistent and reliable solution.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.