Turbulence modeling benchmark experiments

Collaborators: Profs. Chris Roy, William Devenport, Aurelien Borgoltz, and Eric Paterson

Aerodynamic modeling of high Reynolds number flows, while mature in many ways, suffers dramatic inaccuracies in complex cases such as separating flows. The origins of these deficiencies—turbulence and transition models—are themselves incomplete representations of the full physics and require quantitative insights from theory, fully-resolved simulations, and experiments. While experiments are often considered the standard against which to judge models, significant challenges exist in measuring the true boundary conditions present in any given study. These challenges have led many to suspect that unknown and uncertain experimental boundary conditions play a substantial role in the discrepancies between simulation and experimental results.

Our team has focused on conducting combined experimental and computational studies with an emphasis on the stringent demands of computational model validation. For instance, a study already completed on unsteady transitional flow over an airfoil supported the development of a new implementation of transition modeling for delayed detached eddy simulation (see Zhang et al. 2019 reference below). A study of low Mach number, high Reynolds number flow of a plane turbulent boundary layer over bumps, the BEnchmark Validation Experiment for RANS and LES Investigations (BEVERLI) hill case, is also underway with the goal to produce detailed experimental data sets of smooth wall flows with various levels of separation that, for the first time, meet the most exacting requirements of CFD validation as per Oberkampf and Smith (2017, ASME Journal of VV & UQ). To attain this goal, an intensive research program brings together experts in experiments, computations and CFD validation. This effort involves detailed measurements using advanced instrumentation which are complemented by detailed simulations for high fidelity sensitivity and uncertainty quantification. 

Relevant publications:

  1. Lowe KT, Borgoltz A, Devenport WJ, Fritsch DJ, Gargiulo A*, Duetsch-Patel JE*, Roy CJ, Szoke M, and Vishwanathan V*, 2020 "Status of the NASA/Virginia Tech Benchmark Experiments for CFD Validation," AIAA SciTech 2020 Forum paper AIAA-2020-1584.
  2. Szoke M, Vishwanathan V*, Loeschen T, Gargiulo A*, Fritsch DJ, Duetsch-Patel JE*, Borgoltz A, Roy CJ, Lowe KT, and Devenport WJ, 2020 "Developing a Numerical Model of the Virginia Tech Stability Wind Tunnel for Uncertainty Quantification Based On Real-World Geometry," AIAA Scitech 2020 Forum, AIAA-2020-0343).
  3. Gargiulo A*, Beardsley C, Vishwanathan V*, Fritsch DJ, Duetsch-Patel JE*, Szoke M, Borgoltz A, Devenport WJ, Roy CJ, and Lowe KT, 2020 "Examination of Flow Sensitivities in Turbulence Model Validation Experiments," AIAA SciTech 2020 Forum, paper AIAA-2020-1583.
  4. Cadel DR*, Zhang D, Lowe KT, and Paterson EG, 2018 "Unsteady boundary layer development on a wind turbine blade: an experimental study of a surrogate problem," Experiments in Fluids, 59:72, https://doi.org/10.1007/s00348-018-2526-zView-only access: http://rdcu.be/JQGw
  5. Zhang D, Cadel DR*, Paterson EG, and Lowe KT 2019 "Hybrid RANS/LES Turbulence Model Applied to a Transitional Unsteady Boundary Layer on Wind Turbine Airfoil," Fluids 4:3, 128, https://doi.org/10.3390/fluids4030128