/tbnnTurbulenceInterFoam

Modelling of turbulence two-phase flow using interFoam solver with tensor based neural network correcting of Reynolds stress tensor

Primary LanguageC++GNU General Public License v3.0GPL-3.0

trnnTurbulenceInterFoam

Calculation of turbulence two-phase flow using interFoam solver & calculation of basis tensors and their invariants for future modelling of turbulent terms arising from Reynolds averaging.

Requirements

  1. OpenFOAM
  2. tensorflow c api ( you can find installation guide here: https://www.tensorflow.org/install/lang_c )

Note

  1. Works correctly only when executed single-threaded (Multi-threaded execution is not supported!)
  2. The pressure field (p_rgh) is only available from the current time step
  3. The trained neural network must be stored in saved_model format
  4. To find out the names of the neural network inputs and outputs, just run the following command:
model_path = "/path/to/your/model/in/saved_model/format/"
loaded_model = tf.keras.models.load_model(model_path)
loaded_model.summary()