💡

Daily Synthesis

AI is Revolutionizing Aerodynamics 🚀💨

NAVASTO is breaking new ground in engineering with AI-driven tools that redefine aerodynamic design and slash simulation times from hours to milliseconds. Their flagship software, Navpack, streamlines workflows, delivers real-time insights, and enables engineers to tackle complex aerodynamic challenges—no wind tunnel required.

⚙️ What’s New?
1️⃣ Real-Time Feedback: Engineers receive instant predictions for drag coefficients, flow fields, and design sensitivities, empowering faster, smarter decisions.
2️⃣ Blender Integration: NAVASTO’s navDesign plugin pairs AI models with Blender’s powerful 3D geometry tools, enabling seamless design, optimization, and real-time visualizations.
3️⃣ EV Aerodynamics: With EV efficiency heavily dependent on aerodynamic performance, NAVASTO’s AI tools optimize vehicle profiles to reduce drag and maximize range.

🌟 Why It Matters:
Traditional workflows rely on manual tests, lengthy simulations, and physical prototypes. NAVASTO’s AI models accelerate design cycles, save costs, and unlock performance gains that manual CAD work can’t achieve.

💡 From fluid flow predictions to real-time geometry refinements, NAVASTO is leading a new era of AI-powered aerodynamics.

🔗 Learn more about their tools and approach: NAVASTO.

This is engineering, accelerated by AI.

Jousef Murad | Deep Dive: Altair’s PhysicsAI Uses Geometric Deep Learning to Accelerate Engineering ⚙️🌍

On the latest episode of Jousef Murad | Deep Dive, Jonathan Ersson and Eamon Whalen of Altair unveil how PhysicsAI is revolutionizing engineering workflows with Geometric Deep Learning.

🔍 What’s PhysicsAI?
PhysicsAI delivers lightning-fast physics predictions by training deep learning models on historical simulations. Instead of waiting hours for simulations like FEA or CFD to finish, engineers can now get results in seconds.

🚀 Why It’s a Game-Changer:

• Eliminates the need for explicit parameterization. The shape itself is the input.

• Supports transfer learning: update models with minimal new data.

• Provides a similarity score to ensure reliability and trust in predictions.

🎙️ In the podcast, Jonathan and Eamon break down the differences between traditional machine learning and geometric deep learning, showcasing its transformative power across structural analysis, aerodynamics, and manufacturing workflows.

🌟 Host: Jousef Murad
📢 Guests:

• Jonathan Ersson: Development Manager for Geometric Deep Learning at Altair

• Eamon Whalen: Engineering Data Science Manager at Altair

🔗 Full Episode & Resources:
Watch here: YouTube Link
Learn more: Altair PhysicsAI

AI and simulation are shaping the future of faster, smarter engineering.

⚒️

Tools of the Day

• ⚙️ NAVASTO Navpack API - A powerful Python toolkit that supercharges digital engineering workflows by harnessing machine learning and deep learning capabilities

• ⏱️ Altair PhysicsAI - Leveraging Geometric Deep Learning, it delivers lightning-fast physics predictions by learning from historical simulation data, freeing engineers from the constraints of traditional parametric studies

🎓

Paper of the Day

• Physics-informed generative neural networks for RF propagation prediction with application to indoor body perception (arXiv) - Kindly shared and explained by Yan Barros’s post on LinkedIn.

🏆

Top Tweet