Delta Electronics has launched new AI-based digital twin applications for smart manufacturing and building automation. The applications, built on NVIDIA Omniverse libraries, are designed to enhance operational resilience and energy efficiency through real-time simulation.

In manufacturing, Delta has implemented the DIATwin system at its facility in Thailand to support the production of AI server power supplies. The system uses high-fidelity simulations to validate production line designs and optimise robotic paths. This process allows for faster deployment of production lines and supports decentralised manufacturing with centralised management.

“Real-time system behaviour simulations and the ability to model complex ‘what-if’ scenarios even before production ramps up are key advantages we are experiencing,” said Austin Tseng, President of Delta Electronics (Americas). He noted that the technology helps reduce deployment time for new facilities.

The manufacturing solution integrates product design, equipment, and robotics data into a unified model. It uses NVIDIA PhysX for accurate process simulation and NVIDIA Cosmos to generate synthetic defects for training automated optical inspection models. This approach is intended to improve inspection accuracy and accelerate the transition toward autonomous factories.

Delta has also demonstrated digital twin applications for building automation, which merge environmental sensors and HVAC controls with photorealistic simulations. The company stated that these simulations have the potential to unlock up to 20% in energy savings by optimising the interplay between lighting, sunlight, and climate control systems.