The enhancement of Explicit Algebraic Reynolds Stress Models (EARSM) to support machine learning (ML) enables the adjoint solver to support turbulence model optimization of EARSM model coefficients as a beta feature. Explicit Algebraic Reynolds Stress Models (EARSM) represent an extension of the standard two-equation turbulence models and can capture the following flow effects:
Anisotropy of Reynolds stresses
Secondary flows
For more information on EARSM in Ansys Fluent as well as instructions for enabling the model can be found in the Fluent User's and Theory Guides .
After enabling WJ-BSL-EARSM within the Viscous Model, perform the following procedure:
Finish defining the general settings for your case.
Enter the following tui command to make available the Machine Learning EARSM beta coefficients:
(rpsetvar 'turbulence/machine-learning-wj-earsm-beta-coeffs? #t)This command will also make the Machine Learning EARSM beta coefficients available as Design Variables, or tunable coefficients to be optimized as outlined in Using the Turbulence Model Design Tool.
Perform most of the usual steps outlined in Using the Adjoint Solver with the following exceptions:
It is not necessary to start with a converged flow solution; simply initializing the case file is sufficient.
It is not necessary to initialize the adjoint solution and iterate to convergence using the Run Adjoint Calculation dialog box, as these actions will be performed in the Gradient-Based Optimizer dialog box as outlined in Using the Gradient-Based Optimizer.
Additionally,
Adjoint Turbulence must be enabled within the Adjoint Solution Methods dialog box (see Using the Adjoint Solution Methods Dialog Box).
Adjoint Observables specific to turbulence model optimization should be defined (see Defining Observables for Turbulence Model Optimization).
Prior to configuring settings within the Turbulence Model Design Tool group box (see Using the Turbulence Model Design Tool), you can plot sensitivity contours for the EARSM beta coefficients to observe their influence on your RANS simulation. Selecting EARSM coefficients with large sensitivity magnitudes will increase the controllability of tuning during optimization.
Define the general Gradient-Based Optimizer settings for your optimization problem (see Using the Gradient-Based Optimizer).
Set up the Turbulence Model Design Tool (see Using the Turbulence Model Design Tool).
Due to the adjoint equation being harder to solve during WJ-BSL-EARSM model optimization when compared to GEKO model optimization, the optimized WJ-BSL-EARSM model has an increased likelihood of introducing instability into the flow solver compared to the optimized GEKO model.
Note that the procedure for optimizing the EARSM beta coefficient Design Variables is identical to the optimization procedure for GEKO coefficient Design Variables.