1.2. Behavior of Large and Small Eddies

Another such technique is called Large Eddy simulation. See in the DNS one important thing to note was that we had the best accuracy but lot of computational time is required LES is sort of a tradeoff between the Reynolds average in Reynolds average we do many approximations so the results are not that accurate compare to DNS, but LES is something which is a tradeoff between DNS and Reynolds average Navier-Stokes equation.

Large Eddy Simulation (LES) is a computational technique used to simulate turbulent flows. It balances accuracy and computational cost. Direct Numerical Simulation (DNS) provides the highest accuracy but requires extensive computation time. LES approximates smaller scales of turbulence, focusing on larger eddies, which reduces computation time while still retaining reasonable accuracy. This makes it a practical choice for complex simulations.

Think of LES like watching a live sports event on TV. You can see the big plays and highlights clearly (the large eddies), while the minor actions that happen during the game (the small eddies) aren't captured in detail. You’re getting a good sense of the game without needing to watch every single moment closely.

So there is a big difference in the behaviors of large and small eddies in a turbulent flow field. Large eddies are more anisotropic and their behavior is dictated by the geometry of the problem domain, while small eddies are nearly isotropic and have a universal behavior as demonstrated by Kolmogorov.

Large eddies (more than a few lengths of the flow) react differently based on the shape of the environment they are in (anisotropic). Small eddies, on the other hand, behave more consistently regardless of their environment (isotropic). Kolmogorov’s theories state that small eddies have predictable patterns and behaviors, which simplifies their modeling.

Imagine large waves crashing on a beach (large eddies) differing based on the shore's shape, compared to smaller ripples in a pond (small eddies) that behave similarly no matter where they are. The waves react to the boundaries, while the ripples simply spread out more uniformly.

The important thing to remember is that the large eddies extract energy from the mean flow… whereas small eddies take energy from a little bit larger eddies, which takes more energy from the larger eddies than them.

In turbulent flows, energy is transferred among different scales of motion. Large eddies absorb energy from the overall flow, while smaller eddies derive energy from the larger ones. This creates a cascade effect where the energy flows from large to small scales, with larger eddies playing a critical role in energizing the smaller eddies.

Think of large eddies like a big snowball rolling down a hill, gathering more snow (energy) as it gets bigger. Then, this snowball breaks off smaller pieces which roll into even smaller balls, each one collecting even less snow than the larger one, but still gaining energy from the larger snowballs.

This is actually a major hurdle in the path of research just trying to develop a universal turbulence model, which is the hurdle—the variation of eddies from large scale to small scale.

Creating a universal turbulence model is complicated because the behavior of large eddies is influenced by many factors, while small eddies behave more consistently. The challenge is in accurately modeling the diverse interactions and energy transfers between eddies of different sizes.

Imagine trying to build a single model of a car that works perfectly under every condition—some cars work well in the city (large eddies), others perform better off-road (small eddies). Each situation influences the performance differently, and finding the right balance in a universal design is complex.

In LES, the larger eddies are computed with a time-dependent simulation where the influence of the small eddies are incorporated through a turbulence model.

In LES, simulations focus on capturing large eddies directly while accounting for the effects of smaller eddies indirectly through a turbulence model. This is achieved using spatial filtering, allowing simulations to represent significant flow features without the computational burden of smaller scales.

Consider using a filter to brew coffee. The coffee grounds (small eddies) are captured by the filter, giving you liquid coffee (large eddies) full of rich flavors without the grittiness. The desired outcome is achieved while simplifying the process and maintaining the essence.

The filtering operation that we have talked about is defined by a filter function g of x, x dash, delta as it is a very complex but there is a filtering operation that you must know that in LES we use a filtering operation to cut off the smaller eddies and solve in reality for the larger eddies.

In LES, a filter function is applied to distinguish between the scales of motion. Larger eddies are solved directly, while smaller eddies are handled mathematically. The width of the filter is determined based on the grid size used in simulations, ensuring that the larger eddies are accurately captured.

It's like having a strainer in your kitchen. When rinsing vegetables, the strainer allows water to flow through while keeping the veggies (larger eddies) intact. The water represents smaller eddies that are filtered out consistently through the strainer (the filter function).

The filtered momentum equation using the grid scale variables can be written so... the influence of subgrid scale eddies introduced through SGS stress.

The governing equations in LES take into account both the larger eddies being solved directly and the smaller scales which are modeled using the concept of subgrid scale (SGS) stresses. This formulation ensures that the effects of the small scales are recognized in the overall flow behavior without needing to simulate every detail.

Think of this as editing a video. You focus on the main story and scenes (the larger eddies), while acknowledging cuts, background noise, or transitions (the small eddies) that impact the final storyline without having to show every single detail in real-time.

--