Hardcoding the Future of Autonomous Factories

By integrating NVIDIA’s Physical AI into DELMIA’s Virtual Twin technology, Dassault Systèmes is moving from static automation to autonomous software-defined systems.

Dassault Systèmes and NVIDIA are building an industrial AI foundation. DELMIA contributes a virtual twin of production systems, a model of production that connects design intent to real-world execution across engineering, manufacturing and supply chain, the companies report. NVIDIA contributes physical AI and accelerated computing to simulate robotics-grade physics and perception at scale. 

This integration of NVIDIA Omniverse physical AI libraries into the DELMIA Virtual Twin of global production systems establishes a shared industrial AI architecture that grounds artificial intelligence in the laws of physics and validated scientific knowledge. It allows manufacturers to design, simulate and operate complex systems with a system of record for industrial AI, according to the companies.

Virtual Twins

DELMIA enable companies to create virtual twins of their operational ecosystem, including individual machines and robotic workcells to full factory floor layouts and global supply chains. Within this virtual environment, manufacturers can:

• Simulate and validate production processes before equipment is installed.
• Optimize workflows for maximum throughput and efficiency.
• Identify potential bottlenecks and safety hazards without disrupting ongoing operations.
• Train operators and maintenance crews in a risk-free setting.

The virtual twin orchestrates design, engineering, production and supply chain in one environment so decisions can be tested and reused. 

Integrating AI for Autonomous Production

The partnership with NVIDIA brings physical AI into DELMIA virtual twins. NVIDIA Omniverse provides a platform for developing and operating 3D simulations and industrial digitalization applications using OpenUSD-based interoperability. Combined with DELMIA’s production semantics, manufacturers can test autonomous behavior in realistic conditions before deployment.

AI models accelerated by NVIDIA computing can evaluate scenarios across production constraints, resources and variability. The result is the emergence of software-defined production systems. These are factories and operations where decisions remain human-led, but are continuously supported by AI that recommends, tests and validates options in the virtual twin before changes are deployed, the companies explain.

Sources: Press materials received from the company and additional information gleaned from the company’s website.