How do you choose between LES and DNS for turbulent flow simulations?

Choosing Between LES and DNS for Turbulent Flow Simulations: 1. Direct Numerical Simulation (DNS) - Resolution: Resolves all turbulence scales. - Accuracy: High accuracy, ideal for fundamental research. - Computational Cost: Very high; requires significant resources and time. 2. Large Eddy Simulation (LES) - Resolution: Resolves large turbulence scales, models smaller scales. - Accuracy: Good accuracy, suitable for practical applications. - Computational Cost: More feasible than DNS, balanced for complex simulations.

?????????? ???????? ???????????????????? (??????) ???????? ???? ??????: When you want to study big swirls of turbulence and can guess how the small swirls behave. ????????????????: Uses less computer power, works well for complicated flows, and is good for high speeds. Drawbacks: Some guessing is involved for the small swirls, which can lead to small errors. ???????????? ?????????????????? ???????????????????? (??????) ???????? ???? ??????: When you need to see all the details of the turbulence. ????????????????: Very accurate, no guessing needed, great for simple flows at low speeds. ???? ??????????: 1. Use LES for practical problems and complicated flows. 2. Use DNS for detailed studies and simple flows.

Turbulence refers to the irregular, chaotic fluid motion characterized by vortices, eddies, and rapid variations of pressure and flow velocity. It is a common phenomenon in both natural and industrial processes, such as in the atmosphere, oceans, and around vehicles and buildings. The hallmark of turbulence is its unpredictability and the wide range of scales of motion, from large, energy-containing eddies to small, dissipative ones. These varying scales interact with each other, making the flow complex and difficult to model accurately. The study of turbulence is essential in engineering and physics because it significantly impacts the efficiency and effectiveness of various systems.

Turbulence is a randomness in the flow. The erratic behaviour of the fluid with time and space is termed as turbulence. It’s very complex phenomena and a tricky one to model through CFD techniques. The most preferred model to solve for turbulence is the RANS equations.

According to me, turbulence refers to the chaotic, unpredictable fluctuations in fluid flow characterized by vortices, eddies, and rapid changes in velocity and pressure. Understanding turbulence is crucial for accurately predicting and analyzing fluid behavior in engineering and scientific applications.

Large Eddy Simulation (LES) and Direct Numerical Simulation (DNS) are two computational techniques used to simulate turbulent flows. DNS resolves all the scales of turbulence directly by solving the Navier-Stokes equations without any modeling approximations. This approach provides highly accurate results but is computationally expensive and feasible only for flows with relatively low Reynolds numbers and small domains. LES, on the other hand, resolves the larger scales of turbulence explicitly while modeling the smaller scales that are more universal in nature using subgrid-scale models. By filtering the flow field, LES reduces the computational cost compared to DNS while still capturing the essential features of turbulence.

LES and DNS are two computational approaches to simulating turbulent flows: LES: LES models large-scale eddies directly and uses subgrid-scale models for smaller eddies. This approach captures the most energy-carrying structures in turbulence while approximating the smaller, less critical scales. DNS: DNS resolves all scales of turbulence by solving the Navier-Stokes equations directly without any modeling approximations. This method provides highly detailed and accurate representations of turbulent flows but requires immense computational resources.

LES includes modelling assumptions for small-scale turbulence, which can impact the accuracy of the results. In a marine engineering project, we used LES to simulate turbulent flow around a ship hull. The sub-grid scale models in LES adequately captured the large-scale turbulence, providing accurate drag predictions. Determining the impact of model assumptions ensures that you understand the limitations and accuracy of your chosen method, allowing for informed decision-making in your simulations. LES typically requires less computational time compared to DNS, making it suitable for projects with tight deadlines. In an industrial heat exchanger design project, we selected LES to simulate the turbulent flow and heat transfer.

DNS is typically used in academic research and fundamental studies of turbulence because it provides the most detailed and accurate depiction of turbulent flows. Due to its high computational demand, DNS is limited to low Reynolds number flows and small domains, making it impractical for most real-world engineering problems. It is most useful for validating turbulence models and gaining insights into the fundamental nature of turbulence. LES is more commonly used in engineering applications where a balance between accuracy and computational cost is necessary. LES is suitable for simulations where capturing the large-scale structure and dynamics of turbulence is critical, such as in aerospace automotive and environmental engineering.

One of the primary advantages of LES is its ability to provide detailed information about the large-scale structures of turbulence while keeping computational costs manageable. This makes LES a powerful tool for studying flows with complex geometries and high Reynolds numbers that are beyond the reach of DNS. LES also offers better accuracy than traditional Reynolds-Averaged Navier-Stokes (RANS) models, which can oversimplify the turbulent flow structures. Additionally, LES can capture transient and unsteady phenomena in turbulent flows, which are critical in many engineering applications.

Despite its advantages, LES also has some notable disadvantages. The primary drawback is that it is still computationally intensive compared to RANS models, requiring significant computational resources and time, especially for high Reynolds number flows and complex geometries. This limits its use in situations where quick turnaround times are necessary, such as in iterative design processes or real-time simulations. Another disadvantage of LES is the need for appropriate subgrid-scale models to accurately represent the effects of the unresolved small-scale turbulence. The choice of subgrid-scale model can significantly impact the accuracy of the results, and there is no universally best model for all types of turbulent flows.

Improving LES results can be achieved through several strategies. One approach is to use advanced subgrid-scale models that better capture the effects of unresolved turbulence. These models can be tailored to specific types of flows and conditions, enhancing the accuracy of the simulation. Additionally, incorporating dynamic subgrid-scale models that adapt to the local flow conditions can improve the performance of LES in capturing complex turbulence. Another strategy is to optimize the computational grid and increase the resolution where it is most needed, such as near walls and in regions of high turbulence intensity.

To enhance the accuracy and reliability of LES simulations: Refine grid resolution: Increase the grid resolution in regions with high turbulence intensity to capture more detailed flow structures. Optimize subgrid-scale models: Select and tune subgrid-scale models that best represent the specific turbulence characteristics of your flow. Validate with experimental data: Compare LES results with experimental or high-fidelity data to identify discrepancies and adjust modeling parameters accordingly. Use adaptive mesh refinement (AMR): Implement AMR techniques to dynamically adjust the grid resolution based on flow features, improving accuracy without excessive computational costs.