Turbulence Modeling Overview

Principles

Definition

Turbulence modeling is a methodological approach within computational fluid dynamics focused on approximating the effects of turbulent fluid motion. It involves the development and application of mathematical models to predict the statistical behavior of turbulent flows, enabling the simulation of complex systems where direct resolution of all turbulent scales is computationally prohibitive.

Ontological type

Core Models

Engineering Applications

Model Limitations

Key developments

Key Figures

Statistical and Experimental Foundations era

B. E. Launder[1] is associated with the Massachusetts Institute of Technology[3] and Imperial College London[4] in this era. His key contribution is the numerical computation of turbulent flows[7], a foundational method that validated statistical models and spurred development of closure approaches during this era. W. C. Reynolds[2] is associated with Stanford University[5] and Indiana University[6] in this era. His work on the structure of turbulent boundary layers[8] and on the production of turbulence near a smooth wall in a turbulent boundary layer[9] provided essential empirical grounding for near-wall closures.

Computational Modeling and Closures era

Florian Menter [1] was a key figure in turbulence modeling during the Computational Modeling and Closures era, with institutions such as Stanford University [3] and University of Minnesota [4] contributing to the field. His landmark 1994 paper, Two-equation eddy-viscosity turbulence models for engineering applications [5], introduced a practical RANS closure that bridged simple eddy-viscosity models and more rigorous approaches, enabling reliable engineering predictions in this era. Parviz Moin [2] was a central figure in the era, with affiliations at Stanford University [3] and the University of Minnesota [4], advancing subgrid-scale modeling and turbulence analysis. His papers — A dynamic subgrid-scale eddy viscosity model [6], Turbulence statistics in fully developed channel flow at low Reynolds number [7], and A dynamic subgrid-scale model for compressible turbulence and scalar transport [8] — collectively defined dynamic SGS closures and validated them against DNS and experiments, which was crucial for LES adoption in engineering turbulence modeling.

Data-Driven Physics-Informed Closures era

Peter J. Schmid [1] is a leading figure in turbulence modeling within this era, with affiliations at King Abdullah University of Science and Technology [3] and Imperial College London [4]. Schmid [1] pioneered Dynamic mode decomposition of numerical and experimental data [6] as a data-driven modal analysis tool that extracts coherent structures from flows, aligning with the era’s data-centered closure development. Robert Moser [2] is a leading contributor based at The University of Texas at Austin [5] during this era. Moser's direct numerical simulation of turbulent channel flow up to high Reynolds numbers [7] and one-point statistics for turbulent wall-bounded flows at Reynolds numbers up to δ+ ≈ 2000 [8] provide essential benchmarks and validation targets for data-informed closures across Reynolds numbers and geometries.