Femto Engineering

CFD Tuesday - Taking the meh out of meshing AMR

At our CFD-tuesdays we usually show the results of our simulations. This week we are showcasing one of the physics models that helps us run these simulations in an acceptable timeframe: Adaptive mesh refinement (AMR).

This physics model uses a criterion to select cells in the simulation that could use some attention and remeshes them to either a smaller cell size or a bigger cell size, depending on what the values within your criterion are. Standard uses, such as examples in the movie, are refining the free surface such as for waves or droplets. However, you can also create your own custom AMR criteria and refine (or coarsen) the mesh based on that.

The best part is that it does this automatically, and that it will also work in reverse: if the refined mesh is no longer needed, it can just coarsen it back to the original values.

This means that we no longer need to have conservative, a-priori mesh refinement boxes, which is great for flows where you cannot say where the flow will be!

CFD Tuesday – Overset Mesh

In this weeks CFD Tuesday we are looking at simulating isentropic piston compression again.

During isentropic compression and expansion, no energy loss occurs, meaning the temperature and pressure at the end of a cycle of compression and expansion will be the same as at the start.

In order to simulate this, we move the bottom of our piston up and down, adjusting our volume using either mesh morphing or an overset mesh.

Since the cells themselves change volume in the morphing method, no conservation issues should occur. However, mesh morphing works in a more complicated matter and exact conservation is not guaranteed.

We wanted to see if we could model the exact solution for a cycle using mesh morphing.

To do so we looked at both the mesh conservation options that Simcenter STAR-CCM+ offers and the time-step.

We compressed the results in the visualization below, but we expand on this in an an article about both the simulation and the physics behind it!

You can find the article at: https://www.femto.eu/stories/case-study-overset-mesh-conservation-for-isentropic-piston-compression-in-star-ccm/

CFD Tuesday

Bursting with curiosity

Have you ever wondered why big tanks get their own little lawn, with a nice little fence around them?

These fences serve as a precaution for the worst-case scenario, when a big hole is created and all the liquids burst out, flooding the surroundings. The walls will prevent dangerous situations such as a big wave of liquid hitting the roads, and also prevent contamination of the surrounding area.

We performed the analysis in STAR-CCM+ using the Volume of Fluid (VOF) method to resolve the fluid-to-air interface, and the Adaptive Mesh Refinement to automatically increase the resolution at this interface.

The flyby of the tank and lighting were created using the built-in screenplay director and advanced rendering options in the software.

CFD Tuesday

We've shown the Meshless DEM in Simcenter STAR-CCM+ before, and as time passes we'll likely do so again.

In this case we wanted to simulate an hourglass ticking the seconds away.
We started by filling it with small rubber balls, with a baffle preventing the particles from dropping.

After filling it up, we removed the baffle and let the balls flow down.

CFD Tuesday - Pump it up (again)

This week we are showing another method to simulate the motion of the compression; the Overset mesh.

In this approach, we don't morph the mesh but we have a moving mesh part and a backgroud mesh part instead.
As the moving part goes up, cells in the background region are deactivated, and some cells are used to communicate between the background and overset region.

If you're wondering why would you need two different methods to simulate the same, it is because these methods have different strengths and weaknesses.

The morpher works best for simple geometries that don't change too much, while the overset mesh allows for greater motion and even closing off regions.
The morpher doesn't have the transfer between the overset and background, and is a really stable and robust method.

If you want to know more, don't hesitate to contact us!

FEM of the month –beer crate drop

Today is ‘Domibo’ (also known as Thursday drinks) cheers! And in honor of that, this month's featured Finite Element Method (FEM) analysis focuses on a beer crate dropped from a height of 1 meter. Conducted in Altair Radioss 2022.3, the simulation aims to demonstrate the effectiveness of drop tests using explicit FEM software.

The objective is to show that a drop test simulation can be used to assess behaviour of a structure during impact.

High-Density Polyethylene (HDPE) serves as the crate material. Sixteen empty bottles are positioned inside the crate. A 3D shell mesh is used to model all components.

The graphs in this video showcase the vertical velocities of the empty beer bottles in the bottom left and top right corner and the internal and kinetic energy of the beer crate. However, from this simulation we can get much more information like forces and accelerations, allowing for a thorough analysis.

Fluid Film Transition

Several weeks ago we showed the Fluid film feature for STAR-CCM+

In this week, we're showing that transition again. To better illustrate this this, we modelled a block in the domain shown before.
The blob of water spreads, hits the block (with a very fine mesh), which creates a wave.

In the rest of the domain, there's a much coarser mesh since the thin layer of fluid isn't resolved using mesh cells, but using the fluid film description.

CFD Tuesday

Tank trucks are interesting vehicles: they are big, they are shiny and they don't enjoy cornering or braking.
They also don't like to be the glas half full/ glas half empty kind of tank.

The reason they don't enjoy this is because they contain liquids, which will slosh around independent of the vehicle when braking or turning.
- When nearly full, the liquid isn't going anywhere
- When nearly empty, the mass of the fluid isn't that important.
- When half-full however, the liquid can move, and the mass is big enough to matter.

To mitigate the effect of sloshing, baffles are added.
In this week's CFD tuesday, we've modelled a tank where the water is at the front of the tank initially, and then sloshes back when the braking stops.
Included in the model are three baffles, which are modelled using FSI. Within these baffles, openings are made to allow the fluid to redistribute over the tank. The stresses in these baffles are plotted on the top right. As can be seen from the plot, they go up to nearly 100 MPa at the peak load. The simulation was performed in Simcenter STAR-CCM+.

Stay tuned!

CFD Tuesday

In this weeks CFD Tuesday we are looking at a feature in STAR-CCM+ called Mesh morphing, combined with the isentropic gas relations.
The isentropic relation defines how the volume, density, pressure and temperature are related for a change in volume.

When we shrink the volume adiabatically, by moving the bottom wall up, we compress the gas, increasing the density, pressure and temperature.
Then, when we increase the volume again, the density, pressure and temperature return to their original values.

This shows that mesh morphing is a good method to model this motion.

Although this is a simple example, there are (endless/many) applications for mesh morphing. Think about moving bodies like piston motions, vortex induced vibrations, or complex physics such as growing of ice

CFD Tuesday

For this week's CFD tuesday, we decided to show a simple simulation of the Fluid Film feature in Simenceter STAR-CCM+.

In our demonstration we have a blob of water within a cylindrical domain, modelled using the Volume of Fluid method (VOF). Due to gravity the water spreads out over the horizontal surface, eventually forming a film of water. When the VOF phase in the first prism layer cell falls below the prism layer thickness, the fluid film solver is used to model the fluid film expanding.

Simcenter STAR-CCM+ can model thin layers of liquid using a separate fluid film region. The fluid film region is a shell region, which models the thickness of the fluid as a parameter. When coupled with the Volume of Fluid (VOF) solver the fluid film can be used in two ways:

✅ Cells with a larger amount of liquid than the boundary layer thickness are modelled using VOF

✅ The simulation switches from VOF to fluid film when the thickness of the fluid becomes equivalent or smaller than the boundary layer thickness.

Since the simulation does not need to fully resolve the small thickness fluid film (with very small mesh cells needed as a result), the solution becomes very efficient.

This method can be used for a variety of simulations, think about simulations of rain on a surface (rooftop, car window, etc.), leakage flows on a surface, condensation, and many more!