With NVIDIA AI, NVIDIA Omniverse, and the Universal Scene Description (OpenUSD) ecosystem, industrial developers are building virtual factory solutions that accelerate time-to-market, maximize production capacity, and cut costs through optimized processes for both brownfield and greenfield developments.
The journey to building a full-scale digital twin of a factory is challenging. It involves numerous moving parts, requires the integration of traditionally siloed business functions and data, and demands the adoption of new technologies.
To help you get started, NVIDIA has released a Virtual Factory Reference Architecture, available now.
Getting started with the reference architecture
The architecture is based on engagements with developers from global leaders in the semiconductor industry, such as Delta Electronics, Pegatron, and Wistron, and provides an overview of the components and capabilities that developers should consider when developing virtual factory solutions.
IT infrastructure
The design and deployment of the necessary IT infrastructure to support virtual factory solutions is critical. IT infrastructure must remain secure while supporting local and virtual workstations and interconnecting data from various sources.
3D geometry assets
Factory digital twins require the integration of both factory-building models and factory-production models. Consider how to connect the data from leading industry software tools and platforms, such as Autodesk Revit and Flexsim, Bentley Microstation, Emulate 3D by Rockwell Automation, Siemens Teamcenter and NX, and more.
Simulation and synthetic data generation tools
Virtual factory solutions can be used to run robotics simulations for industrial manipulators, autonomous mobile robots, or inspection cameras, as well as the training of computer vision models with synthetic data generated from these simulations.
AI agents and inspection applications
When developing virtual factory solutions, you probably want to easily integrate third-party AI tools and your own AI models into your solutions.
Historical and live synced IoT data
You can harness extensive factory production data encompassing enterprise, live IoT and machine-level, and historical data with OpenUSD to help manufacturing teams construct operational digital twins, facilitating various use cases from remote monitoring to situational awareness. By live-syncing real-time production and 3D data in your solutions, you enable cross-functional stakeholder groups to make decisions with real-time insights.
Platform
The Omniverse development platform provides you with SDKs and APIs to develop virtual factory solutions that bring data interoperability, collaboration, and physically-based visualization to factory planning and operations teams., This enables you to vastly accelerate planning cycles and decision-making, reduce handoffs and costly change orders, and improve overall budget and schedule outcomes.
Delivery
Accessing design applications or AI framework models involves complexities beyond a standalone workstation, requiring consideration of intricate networking and connectivity. Understanding the needs of internal, external, and geographically distributed stakeholders and ensuring their seamless access to factory digital twin solutions is key to successful adoption and value delivery.
Get started developing your own solutions
- Check out the full Virtual Factory Reference Architecture
- Review our brand new lab on How to Build OpenUSD Applications for Industrial Digital Twins
- Download NVIDIA Omniverse today
Sign up for the NVIDIA Developer Program for updates on additional resources and reference architectures to support your development goals.
Get started with NVIDIA Omniverse by downloading the standard license for free, accessing OpenUSD resources, and learning how Omniverse Enterprise can connect your team. Stay up to date on Instagram, Medium, and Twitter. For more information, join the Omniverse community on the forums, Discord server, Twitch, and YouTube channels.
