Manufacturing & Mobility LIVE: Amir Vaziri, PhysicsX keynote

December 18, 2024

undefined mins

Amir Vaziri, Head of Customer Development at PhysicsX

Amir Vaziri, Head of Customer Development at PhysicsX, discusses transforming design and manufacturing with deep physics models

At Manufacturing & Mobility LIVE, Amir Vaziri, Head of Customer Development at PhysicsX, shared insights into how deep physics models are revolutionising the toughest problems in design and manufacturing.

His fireside chat explored the origins of PhysicsX, the applications of deep physics models, and their transformative potential.

To watch Amir's fireside chat, click here.

What are deep physics models?

Amir explains that deep physics models are advanced AI systems that efficiently capture the relationships between geometry and target variables.

Unlike traditional parametric models, these models use advanced architectures to process complex data with remarkable speed and precision. This efficiency significantly reduces the time and computational resources needed for tasks such as optimising designs and simulating performance.

For instance, engineers typically spend hours or days running simulations on designs. PhysicsX’s deep physics models can reduce this process to seconds. Beyond saving time, these models generate practical, manufacturable geometries, enabling companies to accelerate product development cycles and improve performance.

Amir highlights their utility in tasks like optimising automotive aerodynamics or internal cooling systems, where multi-physics simulations – combining thermal, structural and fluid dynamics – are essential. By incorporating these models, manufacturers can design optimised components faster and with greater precision.

Amir Vaziri, Head of Customer Development at PhysicsX

Real-world success stories

Amir showcases several success stories illustrating the potential of deep physics models:

Automotive cooling systems: A leading car manufacturer leveraged deep physics models to optimise the cooling system of its vehicles. This streamlined development and improved the system's efficiency, reducing time-to-market.
Heart impellers: In a medical application, PhysicsX designed a heart pump with 40% improved energy efficiency. The optimised geometry was manufacturable and is currently undergoing trials, demonstrating the model's practical and life-saving applications.
Additive manufacturing: PhysicsX has collaborated with 3D printing companies to optimise tool paths, address thermal effects and reduce material deformation. These improvements enhance the structural integrity of printed components and reduce waste.

Amir stresses the versatility of deep physics models, which can span the entire value chain – from design to manufacturing and operations – offering scalable solutions for complex problems.

Amir Vaziri, Head of Customer Development at PhysicsX

Overcoming challenges and building trust

A key topic of the discussion was trust in AI-driven engineering. Amir explains how PhysicsX builds trust by embedding uncertainty quantification into its models.

When predictions extend beyond the data the model was trained on, the system highlights areas of low confidence and triggers retraining. This iterative process ensures accuracy and reliability, even in novel scenarios.

To further establish trust, PhysicsX starts with pilot projects that address critical customer challenges. These pilots demonstrate tangible value while accounting for manufacturing constraints and operational requirements.

The multidisciplinary PhysicsX team, comprising engineers, data scientists and machine learning experts, collaborates closely with customers to align technology with real-world needs.

The future of deep physics models

Amir outlines the transformative potential of deep physics models in connecting design, manufacturing and operations.

By breaking down silos between these functions, companies can create integrated workflows with a shared ‘ground truth’. He predicts that, as models become more sophisticated, they will enable organisations to explore broader design spaces, improve product performance and reduce costs.

Emerging applications for deep physics models include:

Additive manufacturing: Optimising the manufacturing process by predicting thermal effects and ensuring structural integrity.
Autonomous vehicles: Enhancing simulation accuracy for real-time applications.
Urban mobility: Accelerating the adoption of sustainable infrastructure such as electric vehicle charging stations.

Amir Vaziri, Head of Customer Development at PhysicsX

A new era of engineering

Amir concludes by emphasising that the true value of deep physics models lies in their ability to revolutionise engineering workflows. Whether in optimising design for manufacturability or accelerating time-to-market, these models offer unprecedented possibilities.

PhysicsX envisions a future where AI-infused engineering becomes the norm, empowering companies to innovate faster and tackle challenges with confidence.

Amir notes: “Deep physics models are not just about solving today’s problems; they’re about redefining how we approach engineering for the future.”

To watch Amir's fireside chat, click here.

Essential diary dates for 2025

Discover the essential diary dates for Manufacturing Digital, as well as sister publications Procurement Magazine, Supply Chain Digital and Sustainability Magazine.

To follow Manufacturing LIVE on LinkedIn, click here.

To enter for the Global Procurement & Supply Chain Awards, click here.