Simulation Data in Context#

Simulation results are most meaningful when viewed alongside the geometry they describe. This example overlays a steady-state thermal simulation onto a full-fidelity NVIDIA GB300 NVL72 server rack, showing hot-aisle air temperature during operation. The thermal data (CGNS) and the rack geometry (USD) come from different sources; Kit-CAE composes them into a single scene via OpenUSD without converting either.

GB300 server rack with thermal simulation volume showing hot-aisle temperature

This example uses simplified thermal data inspired by the type of analysis performed in the NVIDIA Omniverse Blueprint for AI Factory Digital Twins. The thermal field was extracted from a larger computed domain, so the influence of surrounding racks and other infrastructure is present in the results. Here we view only the region near a single GB300.

Note

This example requires sample data. If you have not already downloaded it, see Examples for the download link.

Dataset#

The sample data contains a steady-state thermal simulation result and a detailed rack model. Both are subsets of the larger AI factory digital twin; the thermal field covers the volume around a single rack.

File

Format

Field

Description

compute_thermal.cgns

CGNS

Temperature

Steady-state air temperature near the rack (cell-centered, C)

GB300.usd

USD

(geometry)

Full-fidelity GB300 NVL72 server rack model

  • Mesh: ~71,500 nodes, ~57,100 hexahedral elements

  • Time: Steady-state (single snapshot)

Import the Data#

Note

This walkthrough assumes you have already built Kit-CAE. If not, see Get Started for setup instructions.

  1. Launch Kit-CAE.

    On Linux:

    ./repo.sh launch -n omni.cae.kit
    

    On Windows:

    repo.bat launch -n omni.cae.kit
    
  2. Click File > Open and navigate to:

    {path to}/kit_cae_user_guide_data/examples/02_simulation-data-in-context/GB300.usd

    The GB300 rack model loads and appears in the viewport.

    GB300 rack visible in the viewport after opening the USD file
  3. Click File > Import and navigate to:

    {path to}/kit_cae_user_guide_data/examples/02_simulation-data-in-context/compute_thermal.cgns

    Check Import to Stage and click Import. A new compute_thermal prim appears in the Stage panel. Nothing new is visible in the viewport yet.

    Stage panel showing GB300 and compute_thermal prims

Create Bounding Boxes#

Next, create two bounding boxes. The first covers the full extent of the thermal data. The second is scaled to focus on the region directly behind the rack; this narrower box will serve as a Region of Interest (ROI) for the volume operator.

Remember that sources are simple USD prims. Once created, they are not linked to the imported data. You can duplicate, rename, scale, and reposition them freely.

  1. In the Stage panel, expand the thermal dataset to /World/compute_thermal/compute_thermal_cgns/Base/Zone/Elements. Right-click Elements and select Create > CAE Sources > Bounding Box.

    Full bounding box around the thermal data and rack
  2. Select the bounding box you just created. Press Ctrl+D to duplicate it.

  3. With the duplicate selected, press R to switch to the scale gizmo. Scale it to more closely fit the rack, focusing on the hot-aisle region directly behind the GB300. Press W to return to the translate gizmo when done.

    Narrowed bounding box from perspective view Narrowed bounding box from front view showing rack alignment

Create a Volume Operator#

  1. Right-click Elements and select Create > CAE Operators > Volume.

  2. When prompted, choose NanoVDB.

    Volume type selection dialog showing NanoVDB and irregular options

Note

The Irregular option works directly on the unstructured mesh without voxelization. It preserves exact cell geometry and is useful when accuracy matters more than speed (for example, validation or detailed analysis near boundaries). NanoVDB first resamples the data onto a sparse voxel grid, then renders on the GPU; this is faster for interactive exploration but trades sub-voxel detail for performance.

This dataset uses only triangular and quadrilateral faces, which means Irregular mode is available here. Many of the datasets in the main guide use polyhedral elements that require NanoVDB. If you want to try Irregular mode on this data, select it instead; note that it will take longer to compute. This walkthrough continues with NanoVDB.

Visualize the Thermal Field#

  1. Select the Volume operator prim in the Stage panel. In the Property panel, scroll to Region-of-interest (ROI) and click Add Target. Select the narrower bounding box you created.

    Volume operator with ROI set to the narrowed bounding box
  2. Scroll to Colors [Field Selection] and click Add Target. Navigate to /World/compute_thermal/compute_thermal_cgns/Base/Zone/SolutionCellCenter/Temperature and click Select.

    The thermal field appears in the viewport, overlaid on the rack geometry.

    Initial thermal volume visible around the rack with default coloring
  3. Expand the Volume operator in the Stage panel and navigate to Material > Colormap. Adjust the color range and transfer function to better reveal the temperature gradients near the rack.

    Colormap settings with domain 31 to 37 and clampToEdge Thermal volume with adjusted colormap showing temperature gradients around the GB300

Tip

Toggle the rack visibility (eye icon on the GB300 prim) to see the thermal field alone, then re-enable it to view the results in context with the geometry.