En.Light and KoreaElectricPower Conduct Pilot Test of Physical AI-Based Digital Twin for Power Facilities
Conducted an 8-month pilot project with KEPCO
Building Digital Twins for Power Facilities
Automatically Generating 3D Assets with ‘Trinics’
Efforts to Expand the Practical Application of Robots and Drones
[Edaily Reporter Shin Yeong-bin ] 3D artificial intelligence (AI) company N.DotLight is launching a pilot project with KoreaElectricPower (KEPCO) to demonstrate the application of digital twins and robotic AI to power facilities. The goal is to recreate high-risk power facility environments in a virtual space and verify the feasibility of deploying robots in the field.
EnDotLight announced on the 2nd that it will jointly carry out the project titled “Securing Technology for the Demonstration of Physical AI Robots Utilizing Power Facility Data” with KoreaElectricPower(015760)through an open innovation initiative. The core of this project is to precisely construct digital twins of power facilities and verify the reliability of robots in real-world applications by linking them with KoreaElectricPower’s actual operational data. The demonstration period will last a total of 8 months.
Power facility sites are highly hazardous and feature unpredictable environments, making it costly and risky to train robots directly on-site. Robotic reinforcement learning requires large-scale simulations, and the process of mass-producing “Sim-Ready” 3D assets that can be immediately utilized in simulations has long been cited as a bottleneck in technology adoption.
EnDotLight plans to address this challenge by leveraging its proprietary physical AI technology, “TRINIX.” TRINIX is a solution based on a neuro-symbolic architecture that combines a 3D CAD engine with generative AI. It automatically generates 3D assets—including physical properties, collision meshes, and joint structures—from text, images, and existing CAD data.
The generated assets can be exported in standard formats such as USD, URDF, and MJCF, enabling integration with simulation environments like NVIDIA Omniverse and Isaac Sim. In this project, N.DotLight will build an automated pipeline for creating SIM-ready assets for power facilities and support verification by linking simulations with real-world environments using Omniverse-based digital twins and Isaac SIM.
KoreaElectricPower (KEPCO) is providing power facility specifications and operational data. It will also conduct proof-of-concept (PoC) and technical validation of autonomous operation at power facility sites using its own robots.
Starting with this demonstration, the two companies plan to broaden the scope of application and expand the platform to support various robotic applications, such as surveillance drones, in the future.
Park Jin-young, CEO of N.DotLight, stated, “The combination of our AI-based generation and 3D data automation technologies with KoreaElectricPower’s infrastructure and operational data has laid the groundwork for enhancing safety and efficiency in the energy industry.” He added, “By successfully completing this demonstration, we will lead the physical AI ecosystem in the domestic energy sector and drive digital twin innovation in high-risk industrial sites.”
The biggest bottleneck for AI servers is memory. As the KV cache—where large language models (LLMs) store past computations—accumulates, the required memory capacity increases exponentially. This prob…
Concerns about “tax risks” have been raised in some quarters of the financial investment, pharmaceutical, and biotech markets regarding Genosco, a subsidiary of OSCOTEC Inc.(039200)specializing in new…
As July began (June 29–July 3), the pharmaceutical and biotech industries turned their attention to Celltrion Pharm Inc.’s large-scale investment in production facilities and AriBio’s successful fundr…