« Back

Using Tecplot 360 to Perform Inferencing with Luminary’s SHIFT-Wing Model 

Tecplot 360, Tecplot Blog July 29, 2026

Our team recently had the opportunity to work alongside Luminary to develop a prototype integrating our two products. Luminary develops Physics AI models for several industries, including aerospace, automotive, and defense, to reduce the time required to run CFD simulations and evaluate designs. These surrogate models allow engineers to:

  • Evaluate more design alternatives
  • Explore trade-offs earlier
  • Assess risk before committing to a design
  • Reduce the time spent waiting for simulation results

However, removing the CFD runtime bottleneck introduces a new challenge: engineers now have far more results to analyze. As Physics AI enables more design evaluations in less time, the demand for post-processing increases as well. Tecplot 360 helps address this challenge by providing visualization and data comparison tools that make it easier to evaluate additional design candidates and identify meaningful differences.  

See how Tecplot 360 connects with Luminary’s SHIFT-Wing model to generate and compare model-predicted results.

Running an Inference in Tecplot 360 

In the demonstration above, we integrated Luminary’s SHIFT-Wing surrogate model directly into Tecplot 360. 

The workflow allows the user to: 

  1. Adjust the parameters that define the wing geometry. 
  2. Load the geometry using the Load Geometry button. 
  3. Select the desired Mach number. 
  4. Submit the geometry and parameters to the model by clicking Inference. 
  5. Visualize the predicted surface quantities in Tecplot 360. 

Tecplot sends the geometry and selected parameters to Luminary’s SHIFT-Wing model, which predicts surface quantities such as pressure and temperature. 

This allows users to evaluate multiple design variations in a fraction of the time required for traditional CFD workflows.   

Performing Data Differencing on Similar Geometries  

The final component of this workflow uses a custom Tecplot 360 add-on to visualize differences in scalar values across similar geometries. It leverages a k-d tree to identify the nearest corresponding node on an alternate geometry and calculate differences between the two datasets. This provides a practical way to compare design candidates and quickly identify regions where meaningful changes occur.  

Accelerating Engineering Decisions with Physics AI 

This prototype demonstrates how Physics AI workflows can be integrated directly into Tecplot 360, giving engineers a practical way to evaluate and compare model-predicted results. As Physics AI capabilities continue to evolve, we look forward to exploring additional applications for CFD post-processing and analysis. If you have questions or are interested in similar integrations, contact support@tecplot.com.