A seemingly small geometry decision can compromise an entire transient CFD simulation.
When preparing a CFD model in ANSYS SpaceClaim or DesignModeler, Shared Topology is often used to create conformal meshes between adjacent bodies. In many applications, this is exactly the right approach.
However, when setting up a Sliding Mesh simulation, particularly for rotor–stator interactions, applying Shared Topology at the rotating–stationary interface can create a fundamental problem.
Let's understand why.
1. What Does Shared Topology Actually Do?
Shared Topology connects coincident geometric faces, edges, and vertices between adjacent bodies.
As a result:
Adjacent bodies share common geometric interfaces.
The mesher generates a conformal mesh across these interfaces.
Nodes are aligned and shared between neighboring regions.
The resulting mesh has continuous connectivity across the interface.
This is extremely useful for conjugate heat transfer (CHT), multi-material models, and stationary fluid–solid interactions.
But Sliding Mesh requires something fundamentally different.
2. Why Does Sliding Mesh Require Independent Meshes?
The Sliding Mesh method is designed to simulate relative motion between two or more mesh zones.
Consider a typical centrifugal fan or turbomachinery application:
Rotating zone: Contains the impeller or rotor and its surrounding fluid.
Stationary zone: Contains the casing, volute, or stator.
Sliding interface: Separates the rotating and stationary computational domains.
The key principle is that both mesh zones must be able to move independently.
The rotating mesh must change its angular position while the stationary mesh remains fixed.
ANSYS Fluent achieves this by connecting the two regions through a mesh interface, which allows fluid information to be exchanged even when the nodes on both sides are not aligned.
This is precisely why non-conformal interfaces are so important.
3. What Goes Wrong When You Apply Shared Topology?
When Shared Topology is applied at the interface between rotating and stationary regions, the mesher creates a conformal connection.
This introduces a conflict with the fundamental operating principle of Sliding Mesh.
The problem can be explained in three steps:
1. Shared nodes
Both regions share interface nodes because the topology is conformal.
2. Conflicting motion
The rotating region needs these nodes to move, while the stationary region requires them to remain fixed.
3. Loss of independent mesh motion
The shared topology prevents the two sides from behaving as independent sliding mesh zones.
Depending on the geometry and meshing workflow, this can result in:
Unwanted movement or distortion of stationary mesh elements.
Invalid or missing sliding interfaces.
Incorrect mesh connectivity.
Difficulties defining independent rotating and stationary cell zones.
A simulation that cannot correctly represent relative motion.
In other words, the problem is not simply a meshing inconvenience. It is a fundamental conflict between shared conformal connectivity and independent mesh motion.
4. How Should You Prepare Geometry for Sliding Mesh?
The correct approach is to preserve separate topology between the rotating and stationary computational domains.
A recommended workflow is:
STEP 1 — Geometry preparation
Create separate bodies for the rotating and stationary fluid regions.
Do not apply Shared Topology across the sliding interface.
STEP 2 — Independent meshing
Generate separate meshes for both computational regions.
The mesh density and element distribution on each side do not need to match.
STEP 3 — Import into ANSYS Fluent
Ensure that the rotating and stationary regions are recognized as separate cell zones.
STEP 4 — Create Mesh Interface
Select the corresponding interface boundaries on both sides and create a mesh interface.
STEP 5 — Define mesh motion
Assign the appropriate rotational velocity to the rotating cell zone while keeping the stationary zone fixed.
STEP 6 — Run a transient simulation
Enable transient calculations and define a suitable time step to capture the physical interaction between rotating and stationary components.
5. Shared Topology vs. Sliding Mesh: The Fundamental Difference
| Feature | Shared Topology | Sliding Mesh |
|---|---|---|
| Mesh connectivity | Conformal | Typically non-conformal |
| Interface nodes | Shared/aligned | Independent |
| Relative motion | Not suitable across the shared interface | Supported |
| Mesh zones | Geometrically connected | Independently moving |
| Typical application | CHT, multi-material models | Rotor–stator interaction |
| Interface treatment | Continuous mesh connectivity | Dynamic mesh interface |
6. An Important Engineering Distinction: MRF vs. Sliding Mesh
It is also important not to confuse Sliding Mesh with the Multiple Reference Frame (MRF) method.
MRF is a steady-state approximation in which rotating regions are described using a rotating reference frame. It does not require the mesh to physically slide relative to the stationary mesh at every time step.
Sliding Mesh, on the other hand, explicitly updates the relative position of the rotating and stationary meshes during a transient calculation.
Therefore, the requirements for their interfaces and mesh connectivity are not identical.
Final Takeaway
Shared Topology is not inherently bad. It is an excellent tool when conformal mesh connectivity is required.
However, at the boundary between independently moving and stationary mesh zones, Shared Topology is generally the wrong choice for a conventional Sliding Mesh setup.
The most important rule to remember:
If two mesh regions need to move independently, do not force them to share the same interface nodes.
In CFD, correct geometry preparation is not just about creating a good-looking mesh. It is about preserving the physical and numerical freedom required by the solver to reproduce the actual flow physics.
Have you ever encountered unexpected mesh deformation or difficulties setting up a Sliding Mesh simulation because of Shared Topology?
0 Comments