Engineering at AI Speed: Near-CFD Accuracy in 0.1 Seconds


October 27, 2026
11:00 AM ET / 10:00 AM CT / 8:00 AM PT / 4:00 PM GMT
Duration: 1 hour
Already registered? Click here to log in.
Summary
Centrifugal compressors are critical assets underpinning energy and industrial systems, from LNG and gas pipelines to refining, hydrogen, and data center power. Yet optimizing these complex machines depends on Computational Fluid Dynamics (CFD) simulations that can take hours or days to complete, limiting the speed and breadth of design exploration. Baker Hughes set out to change that.
- Learn how energy technology company Baker Hughes developed a physics-informed AI surrogate that generalizes across compressor families — a critical piece of equipment at the heart of key energy operations — eliminating the need to build and train a separate AI model for every centrifugal-compressor stage while preserving near-Computational Fluid Dynamics (CFD) accuracy.
- Examine blind-test results demonstrating 99.5% median accuracy in predicting 3D flow fields and compressor performance for geometries the model had never seen before.
- Understand how inference times were reduced from hours to fractions of a second, transforming engineering workflows from batch simulation to interactive design.
- See how NVIDIA PhysicsNeMo and the Dell Pro Max with GB300 accelerated computing enabled full-resolution model training, overcame memory constraints, and unlocked design-space explorations that were previously impractical at industrial scale.
Join Baker Hughes, Dell Technologies, and NVIDIA to discover how a physics-informed AI surrogate, trained on advanced deskside accelerated computing infrastructure, is collapsing simulation timelines and unlocking a new era of interactive turbomachinery design.
Speaker
Sponsored by:







