Using Physics Informed Neural Network (PINN) and Neural Network (NN) to Improve a k-omega Turbulence Model

pyCALC-RANS has now been extended with PINN and NN.

Lars Davidson, "Using Physics Informed Neural Network (PINN) and Neural Network (NN) to Improve a k-omega Turbulence Model", Journal of Turbulence, https://doi.org/10.1080/14685248.2026.2665148, vol. 27, issue 7, pp. 187--208, 2026
view paper
Download the PINN and NN Pythons scripts and the CFD code here (53 MB, 11 September 2026) Read README file.
- 11 Sept 2026: I found two bugs in my PINN-NN simulations. The two bugs lead to incorrect initial values of prand_k_ML, c_k_ML and c_omega_2_ML. This may lead to that the solution proceedure diverges. If it converges, correct results are obtained (provided sufficient number of iteraions are made so that the solution 'forgets' the incorrect initial values). For more info, see README file
The hill flow case is found in pyCALC-LES
Flowchart
pyCALC-RANS report.
Please read the README file