Cross-Industry Collaboration Delivers Breakthrough in Simulation-Driven Manufacturing

Recent developments announced by Synopsys Inc. and their collaborators at Microsoft Ignite highlight a novel simulation-driven framework that addresses one of manufacturing's persistent challenges: optimizing complex production processes in real time.

This remains a challenging problem due to the computational demands of high-fidelity simulations. Typical computational fluid dynamics (CFD) analyses provide detailed insights that usually require several hours to complete, which limits their use for immediate operational decisions.

READ MORE: From CAD to Co-Design: Mastering AI, Material Science, Digital Twins and MBSE

The collaborative simulation-driven framework aims to improve dynamic manufacturing processes with significantly reduced runtimes, according to the press release. The framework was first implemented by Krones AG, manufacturer specializing in integrated packaging and bottling lines. Krones created a digital model of their assembly line that accounted for manufacturing variables such as bottle geometry, liquid viscosity and fill level. This virtual assembly line enabled the assessment of production scenarios in near real time.

The main technical challenge addressed is the reduction of CFD simulation time from several hours to less than 5 min., which was achieved by using a GPU-accelerated cloud native simulation solver developed by Synopsys. By leveraging parallel processing on GPUs and cloud computing resources, the framework supports rapid scenario analysis and continuous process improvements.

Open USD (universal scene description) plays a central role in enabling data exchange and interoperability between various computer-aided engineering (CAE) tools and platforms within the framework. This facilitates collaboration among engineering operations and research team, allowing faster iteration cycles on simulation data.

READ MORE: Physical AI in Motion: How Machine Learning Drives Next-Gen Industrial Automation

The press release notes that several industry partners contributed to the development and deployment of the framework:

• CADFEM Germany GmbH modified the Ansys Fluent Solver configurations to improve efficiency and maximize GPU utilization specific to Krone’s use case.
• NVIDIA Omniverse and CUDA-X libraries were used to visualize and interact with simulation data in a virtual factory environment.
• Microsoft Azure provided the cloud infrastructure that enabled scalable and high-performance computation.
• SoftServe integrated and deployed the system to operate within the digital twin on the factory floor.

This collaborative approach allowed the creation of a near real-time digital twin model to monitor and optimize the production line.

The framework provides several practical benefits for engineers involved in machine design and process optimization, including real-time comparison of process scenarios for faster decision making; resource use can be optimized to reduce material waste; and feedback loops between engineering teams and operations are shortened, enhancing coordination.

READ MORE: Reverse Mic: NVIDIA, Teradyne Thought Leaders Compare Notes

Nvidia's vice president of simulation technology, Rev Lebaredian, notes that integrating GPU accelerated solvers with cloud-based visualization enables real-time factory floor simulation without impacting physical systems. "For manufacturers, the ability to optimize dynamic processes in real-time has been a massive challenge, as traditional, high-fidelity simulation is often too slow to be practical for immediate factory-floor decisions," he said in the release.

The framework’s use of OpenUSD and Omniverse technologies creates a foundation for other simulation-driven applications beyond manufacturing, potentially extending to medical procedures and other fields that require real-time simulation feedback.

Prith Banerjee, senior vice president of Simulation & Analysis Incubation at Synopsys said, "This framework is a testament to how cross-industry collaboration and open ecosystem partnerships are reshaping the future of industrial innovation. It marks a major step toward scalable, intelligent simulation-driven applications that will be pivotal to ushering in the next phase of digital transformation."