关于这一集
CDFAM Computational Design Symposium — DC 2026
Mike Park · Flexcompute
Computational design enables wider configuration exploration and reduces time to market, but turning innovation into validated, manufacturable designs is hampered by human-in-the-loop simulation. Manual geometry cleanup, meshing, and data management pace every workflow, while the feedback loops between simulation and the rest of the design organization stay broken by data silos. This talk presents an end-to-end, AI-native approach to closing those loops. GeometryAI applies a novel, topology-aware geometry representation that takes raw CAD to analysis-ready models automatically with mathematical guarantees that facilitate meshing and analysis. Physics-agnostic output feeds CFD, structural, and thermal tools with direct connections to surrogate training and inference. The GPU-native solver Flow360 delivers high-fidelity physics at speeds that tighten feedback loops. Thread ties it together with automatic data lineage, so every run is reproducible, every result is traceable, and institutional knowledge becomes the starting point rather than the bottleneck. Standard web interfaces and Python APIs enable humans and their agents equally. Real-world results are validated with industry-standard AIAA High-Lift Prediction Workshop submissions. Full-aircraft hover analysis is possible in hours. We show how geometry-aware automation, GPU-native simulation, and complete traceability make simulation-driven design genuinely agent-ready, at the scale and speed modern programs demand.
Links
Talk page with full transcript
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.designforam.com
