In this lecture, we continue our turbulence modeling journey by moving from RANS (k-ω SST) to Large Eddy Simulation (LES). Before running the LES case in OpenFOAM, we first understand the core concepts behind LES — including filtering, subgrid scale stresses, the closure problem, and the role of Kolmogorov scales. What you will learn in this lecture:

• What is Large Eddy Simulation (LES)?

• Deriving LES equations from the Navier–Stokes equations

• Filtering operation, properties, and physical interpretation

• Subgrid-scale (SGS) stresses and why closure is needed

• Kolmogorov scale, DNS vs LES vs RANS

• How LES filter width (Δ) is chosen from grid spacing

• Cube-root volume method vs max(Δx, Δy, Δz) method

• Smagorinsky model and SGS viscosity formulation

• Setting up a complete LES case in OpenFOAM

• Differences between LES and RANS results for flow past a square cylinder

• Visualization and comparison of vortical structures and flow physics

We convert an existing k-ω SST RANS case to LES using the Smagorinsky model, edit the turbulence properties, remove unused fields like k and omega, run the case, and finally compare the resolved flow structures. The lecture shows how LES captures large-scale vortices and near-wall dynamics much better than RANS on the same mesh. In the next lecture, we will explore additional LES models in OpenFOAM and compare their accuracy, stability, and suitability for different flow problems.

In this lecture, we refine our previous LES simulation of flow past a square cylinder using the Smagorinsky model. We diagnose errors such as non-uniform pressure, incorrect velocity distribution, excess cells in the depth direction, and unstable Y+ values. The video covers step-by-step mesh improvement using blockMesh and snappyHexMesh, adding layers, fixing feature angles, adjusting resolution, resolving divergence issues, correcting boundary conditions, and validating mesh quality. Finally, we run a stable LES case and prepare for switching turbulence models in the next lecture.

In this lecture, we continue our Large Eddy Simulation (LES) study in OpenFOAM by comparing old and improved mesh configurations and analyzing their impact on flow physics. We start by visually comparing two simulations—one with a coarser mesh and another with refined resolution near the body. Using velocity fields, slices, and LIC plots, we demonstrate how mesh resolution and prism (boundary) layers significantly influence vortex formation, diffusion, and shedding behavior. Key focus areas include:

• Effect of near-wall mesh refinement on velocity gradients

• Importance of prism layers for LES accuracy

• Comparison of vortex diffusion in coarse vs fine meshes

• Role of Smagorinsky subgrid-scale model

• Practical snappyHexMesh troubleshooting for layer generation

We further refine the mesh by adding 20 prism layers with controlled expansion, rerun the simulation, and show how finer near-wall resolution helps capture multiple small-scale vortices that are completely missed with coarse grids.

In this lecture, we compare RANS and LES simulations of flow past a square cylinder using the same mesh to ensure a fair comparison. Results from the k-ω SST RANS model are compared against multiple LES turbulence models, including Smagorinsky, WALE, k-equation, and Deardorff DiffStress models.

We analyze velocity fields, vortex structures, near-wall behavior, and subgrid viscosity differences, and explain why LES behaves differently from RANS despite identical mesh resolution. The lecture also includes a walkthrough of OpenFOAM LES source code, turbulence model selection, delta definitions, and practical issues such as initialization, boundary conditions, and solver stability.

In this lecture, we explore hybrid RANS–LES turbulence modeling, focusing on Detached Eddy Simulation (DES), Delayed Detached Eddy Simulation (DDES), and Improved Delayed Detached Eddy Simulation (IDDES). We begin with the motivation behind hybrid approaches, highlighting the strengths and limitations of RANS and LES when used independently. You’ll learn why RANS performs well near walls, why LES becomes expensive at high Reynolds numbers, and how hybrid models combine the best of both worlds. The session then covers:

• Fundamental principles of DES, DDES, and IDDES

• Grid dependence and shielding issues in DES

• How DDES delays grid-induced separation

• Why IDDES is the most robust and widely used hybrid model

• Length-scale blending and near-wall shielding concepts

• Practical reasons why k-ε DES models are not used

• Evolution of hybrid RANS–LES models from 1997 to present

In the second half, we move to a hands-on OpenFOAM walkthrough, where we:

• Locate DES/DDES/IDDES models in the OpenFOAM source code

• Set up Spalart–Allmaras DDES and IDDES simulations

• Configure turbulence models, delta functions, and wall distance requirements

• Compare results against LES (WALE model) for flow past a square cylinder

• Visualize and analyze differences in vortical structures and flow separation using ParaView