Automotive Manufacturing

Paving the Future of Factories with NVIDIA Omniverse Enterprise

Objective

NVIDIA and BMW collaborate to showcase how digital twins and virtual worlds will drive the future of manufacturing.

Customer

BMW

Partner

Bentley Systems, Autodesk

Use Case

Industrial Virtualization

Technology

NVIDIA Fleet Command, NVIDIA Isaac Sim, NVIDIA Omniverse, NVIDIA Quadro RTX 8000, NVIDIA RTX-powered server on EGX Platform

Driving Efficiencies–Worldwide

BMW produces 2.5 million cars every year, and customers tailor 99 percent of those cars before purchase. The challenges for manufacturing are complex and, to stay competitive, BMW must continuously find solutions to meet the high demands of their customers. Using NVIDIA Omniverse™ Enterprise, BMW aims to simulate many aspects of an entire factory in real time and produce a digital twin or virtual model designed to accurately reflect a physical object. By integrating NVIDIA Omniverse into their workflows, the team at BMW can enhance their construction and management processes, achieve their production goals, and exceed customer expectations

Customizing on a Massive Scale

To deploy industrial digital twins effectively for designing a factory’s workflow, BMW must collect the most accurate data from its physical counterpart. The company traditionally used various software platforms to collect operational data from global teams, but with over 30 factories worldwide, it became challenging and time-consuming to track the extensive amount of data. This left the data prone to inaccuracies and discrepancies.

Image courtesy of BMW.

NVIDIA Omniverse Enterprise provides an open, virtual space that allows the BMW teams to collaborate on the same models in real time from any location

Through Omniverse Connectors, different users can easily access applications, such as Bentley Microstation and Autodesk Revit, and work on the same projects together.

With NVIDIA Isaac Sim, BMW can deploy a fleet of intelligent robots for logistics to improve the material flow in their production.

The team can also leverage NVIDIA Isaac for synthetic data generation and domain randomization.

With NVIDIA Fleet Command software, BMW workers can also securely and safely orchestrate robots and machines in the factory.

img1
img2
img3
img4
img5

Images courtesy of BMW

Additionally, planning a factory reconfiguration typically occurs when launching a new vehicle, and BMW must take into account worker safety and ergonomics, as well as retraining a robot’s interactions to the new environment. Executing the most efficient plan is an arduous task, given BMW’s 57,000 employees working alongside thousands of robots. In addition to humans, programming a robot’s system for multiple environmental scenarios can take weeks or months to complete. Even using AI to train the robots requires hundreds of thousands of virtual scenarios and increases in complexity with the number of robots in use.

Plus, when it comes to optimizing a factory’s reconfiguration to find the most efficient and safest path for workers, the process can take months to implement and cost hundreds of thousands of dollars. The inability to synchronize data between the physical factory and its digital twin can lead to faulty decisions, potentially resulting in higher costs.

To overcome all these challenges and continue making high-quality vehicles, production planners at BMW sought to simulate a true-to-reality planning and deployment lifecycle for every factory within their global production network.

Streamlining Factory Collaboration

NVIDIA Omniverse Enterprise allows BMW to take data captured in real time or stored in separate databases and use AI to simulate a photorealistic, accurate living digital twin of the physical factory. The platform, which is built for 3D design teams to work together simultaneously across multiple software suites in a shared virtual space, enables BMW designers and operators to see every detail and collaborate on one model.

BMW used NVIDIA-Certified Systems HPE Apollo 6500 servers equipped with NVIDIA RTX™ 8000 GPUs to host Omniverse Kit, a platform for creating apps and microservices used for editing and rendering, alongside workstations from Altair. A single lightweight server hosts Omniverse Nucleus, the central database hub for digital content creation (DCC) that users employ to connect and collaborate in a shared virtual space. This is possible because the Nucleus server reformats files from other DCC apps into Universal Scene Description (USD) format.

BMW uses applications such as Bentley Microstation, Autodesk Revit, and Dassault Systèmes CATIA export to Omniverse so that users working in any of the apps can connect to a project on the server and manipulate the same asset as others.

BMW employs NVIDIA Isaac Sim to train their delivery robots. With Isaac Sim, BMW can generate synthetic data with domain randomization for machine learning. This means using millions of generated images with photo realistic objects of infinite variations—including textures, orientations, and lighting conditions—in different simulated environments, enabling virtual robots to learn the appropriate behavior under any condition. Once the desired sequence is found, the real-world materials and robots can be built to reflect this information.

“Omniverse greatly enhances the precision, speed, and consequently the efficiency of our planning processes.”

Milan Nedeljkovic
BMW AG Board Member for Production

Connecting Virtual Worlds

With Omniverse Enterprise, BMW’s production designers and planners can use the same virtual environment as the robotic teams using Isaac Sim. This eliminates the need to rebuild the environment in every alternate simulation software used by other teams.

Deploying software and monitoring system statuses across so many edge AI locations worldwide can be a costly challenge to execute efficiently. But by using NVIDIA Fleet Command, not only can robots and machines be controlled and updated remotely, but data captured from sensors is uploaded back into Omniverse. This continuous data loop between physical and virtual environments synchronizes the most current information and makes a true- to reality digital twin possible, allowing BMW to make the most informed decisions at any given moment.

Image courtesy of BMW.

Together, we’re about to make a huge leap forward and open up completely new perspectives in the field of virtual, digital planning,” said Milan Nedeljkovic, BMW AG Board Member for Production. In the future, a virtual representation of our production network will allow us to realize an innovative, integrated approach to our planning processes. Omniverse greatly enhances the precision, speed, and consequently the efficiency of our planning processes.”

BMW is paving the way for tens of millions of other factories in the world to augment their production workflows. This increased efficiency and productivity will reduce consumption, improve worker conditions, and ultimately produce better products.