FAN Simulation, NFLOW LBM

Hello. This is E8IGHT.

We brought a news regarding NFLOW LBM!

With NFLOW LBM, we ran an analysis of Axial Fan Case which is a 3-dimensional flow around an axial fan-like structure.

On this analysis, we applied a halfway bounce-back which has a higher accuracy than the standard bounce-back used in the existing 3D LBM solver and analyzed to see the speed of the moving structure additionally.

In order to reflect the structure of arbitrary shape more accurately, we implemented interpolated bounce-back which reflects the position of the structure boundary between grid points and reflects the speed of the structure.

So, to test the two boundary conditions mentioned above, we have verified the correct implementation of the boundary condition by analyzing the flow of an axial fan case caused by the rotation of an arbitrary shape (ellipse) around the Z axis.

The algorithm is as follows.

LBM Algorithm: Static Structure (Left), Dynamic Structure (Right)

The halfway bounce-back we analyzed reflects the speed of the wall.

The second term on the right side of the equation reflects the momentum due to the wall velocity to the existing probability distribution function.

In the case of interpolated bounce-back, as shown in the figure below, it is divided into the case where the boundary of the structure is closer to the boundary node than the midpoint between the grid points (q<1/2) and the case where the boundary of the structure is closer to the solid node than the midpoint between the grid points.

The interpolated bounce-back reflects the wall speed. You can check that the same momentum term is used as Halfway bounce-back!

Back to the case analysis, to describe z-axis rotation of a fan-like, the technique reapplies every iteration only to the boundary conditions of moving structure. As shown in a figure below, the x and y coordinates in the domain are rotated clockwise to rotate counterclockwise in the frame of reference of the domain. A fan-like case has been implemented.

Frame of reference : fan

Frame of reference : domain

There are some cases where an elliptical solid node rotates inside the analysis area and the grid point that is solid at the current iteration becomes the fluid grid point at the next iteration.

These grid points are called fresh nodes, which it doesn’t have density, pressure, and probability distribution values that are the fluid properties required for LBM calculation.

Therefore, the equilibrium probability distribution function is applied to the fluid grid points as the probability distribution function of the fresh node and by applying a technique that gives average macroscopic properties (density and speed) of the surrounding grid points, there becomes no discontinuity in the analysis results.

Rotating an elliptical-like fan blade around the z-axis, you can see the three-dimensional flow of static fluid as shown in the video below.

Reynolds number is 32,435, and domain uses a single orthogonal magnetic field of even size (single block), and Multi-relaxation time (MRT) and Large Eddy Simluation (LES) are applied.

If you plot an isosurface with a positive Q-criterion, you can see the structure with the rotation of the flow. In the video, you can see that there is a rotation of a flow, and below is a video of a flow with rotation.

On the other hand, LBM is inherently an unsteady flow solution method that solves the Boltzmann transport equation. With LBM, remeshing to reflect changes in the flow domain geometry can be done much more quickly than with traditional CFD methods.

Furthermore, the unsteady flow field can be directly solved without the need for subiterations that have to converge to the solution at each time step.

So far, we looked at the analysis of similar Axial Fan Case using LBM.

Isn’t it amazing that LBM can perform analysis that are hard to do with existing CFDs?

We will be back with other news regarding LBM.

It was E8IGHT so far. Thank you.

New Flow Leads Our World!

www.e8-global.com

Tel. 1811-2707

Email. info@e8-korea.com

Reference: (Krüger, Timm, et al. The Lattice Boltzmann Method: Principles and Practice. SPRINGER, 2018.)