Wistron Boosts Manufacturing Energy Efficiency with AI and NVIDIA Omniverse

Wistron Boosts Manufacturing Energy Efficiency with AI and NVIDIA Omniverse







With the rising emphasis on environmental, social, and governance (ESG) investments and projects, producers are increasingly more in quest of cutting edge techniques to strengthen power potency and sustainability of their operations, in line with the NVIDIA Technical Blog.

Bettering Operational Potency with OpenUSD and AI

Wistron, a well-known world provider of knowledge and communications merchandise, has made important strides on this section. The corporate has advanced a virtual dual platform and AI-enabled simulation gear to optimize the design, efficiency, and effort potency in their run-in take a look at rooms in unused NVIDIA DGX and NVIDIA HGX factories. This development has the prospective to loose power intake by way of as much as 10%.

To reach this, Wistron’s builders hired the NVIDIA Omniverse platform, leveraging Common Scene Description (OpenUSD) to form their Virtual Dual platform. This platform connects to their construction control machine and IoT hub, offering real-time knowledge from hundreds of bodily sensors throughout their amenities. Those sensors track important parameters comparable to core temperature and air con go back temperature.

The combination of OpenUSD has facilitated real-time collaboration for faraway groups, streamlining the overview of facility layouts and accelerating decision-making processes in facility making plans and operations.

Moreover, Wistron’s builders have built-in physics-informed AI fashions into their Virtual Dual platform the use of the open-source NVIDIA Modulus framework. Those fashions backup boost up simulation paintings, fortify thermal dynamics, and loose operational dangers, making sure that cooling techniques carry out optimally even below hard situations.

Construction Virtual Twins of Check Rooms

Wistron has applied OpenUSD to unify their knowledge pipeline and streamline workflows, enabling the origination of 3-d fashions by way of their professionals. This way standardizes asset origination and guarantees compatibility throughout diverse tool and simulation gear. John Lu, plant supervisor at Wistron, highlighted that OpenUSD do business in versatile knowledge modeling, permitting the mix of various knowledge and results from diverse 3-d modeling and simulation gear.

For instance, OpenUSD has facilitated the relationship between Wistron’s Virtual Dual platform and the Autodesk FlexSim simulation tool, improving their groups’ skill to simulate, analyze, and experiment with important production processes. Builders have constructed a customized extension to import knowledge and parameters from FlexSim, the use of an OpenUSD-native Omniverse Connector to combine this knowledge into their Virtual Dual platform.

Accelerating Simulation and Predicting Possibility with Physics-Knowledgeable AI

Wistron’s simulation professionals depend on computational fluid dynamics (CFD) simulations to help the design and control in their run-in take a look at rooms. Then again, the normal CFD approaches, working on general-purpose computing architectures, ceaselessly proved rigid and resource-intensive.

To handle those demanding situations, Wistron built-in spare functions into their Virtual Dual platform, together with:

  • Physics-informed neural networks (PINNs) the use of NVIDIA Modulus, which considerably accelerates airflow simulations from 15 hours to only 3.6 seconds—a fifteen,000X growth.
  • An AI-based extension for high-fidelity visualization and research of CFD simulations, serving to to reduce cooling machine rather a lot and loose running prices.
  • A advice machine for his or her computerized storehouse and retrieval machine (ASRS) to spot optimum checking out places and autonomously playground unused supercomputing baseboards in boxes with the least thermal chance.

Those improvements permit Wistron to approximate the underlying physics in their thermal techniques, offering rapid and correct predictions of temperature distributions and thermal behaviors inside their run-in take a look at rooms. Groups can now determine facility hotspots and forecast core temperatures as much as half-hour in travel.

Learn the untouched NVIDIA announcement at COMPUTEX to be informed how Wistron is adopting NVIDIA era to form and function virtual twins.

Symbol supply: Shutterstock

. . .

Tags


Leave a Reply

Your email address will not be published. Required fields are marked *