Automotive parts specialist Korens (KORENS) said on the 13th that it was recently selected as a participating company in the "PINN model manufacturing convergence data collection and demonstration project," a physical manufacturing AI support program led by the Ministry of Science and ICT. Through this, Korens plans to build an autonomous factory that combines artificial intelligence and energy innovation, going beyond conventional factory automation.
Korens plans to transform its factory into an AI autonomous factory that can make decisions and operate without human intervention by introducing Korean-style physical AI technology through the project. Korens forecasts that applying physical AI to existing production equipment could yield effects such as a 60% reduction in process setup time, a 50% decrease in defect rates, and more than a 30% reduction in worker dependency in terms of productivity and quality.
The physical manufacturing AI support project is part of the government's 2025 "industry-specialized physical AI core technology PoC (Proof of Concept)" initiative, and aims to develop the physics-informed neural network (PINN) models needed for physical AI development and to collect and demonstrate manufacturing convergence data.
Whereas conventional AI was limited to text generation or image recognition, physical AI will link robots, sensors and the Internet of Things installed on the production line to act as the brain of an autonomous factory. As a result, it can analyze process data in real time and automatically control equipment, while also improving to the point of detecting and responding to defects in advance.
Korens' plan to establish an autonomous factory goes beyond simply applying AI on site. For AI to function properly, it is essential to be able to collect and analyze vast amounts of data. To that end, Korens built a data lake that integrates various data such as production, quality and energy in one place. The data lake, which stores raw source data such as sensor records, IoT logs and video data without processing, serves as the "fuel" that allows AI to learn and make judgments on its own.
Korens adopted digital twin technology to recreate the actual factory in a virtual space, simulating equipment conditions and process flows to check for potential problems in advance. For example, if a decline in a specific piece of equipment's efficiency is detected, the digital twin predicts the failure point and prepares maintenance measures in advance. Such proactive responses lead to improved equipment uptime and reduced defect rates, minimizing unexpected production disruptions.
Energy efficiency is also an indispensable task in manufacturing sites. Korens raised the completeness of its autonomous factory by introducing a facility energy management system (FEMS) that monitors and automatically controls the energy use of key equipment in the factory in real time. FEMS analyzes energy usage patterns from the entire factory down to production lines and individual equipment to detect anomalies early, and automatically adjusts the rotation speeds of key devices such as cooling water circulation pumps to reduce unnecessary idling. It also links with a solar power generation system to pursue ESG effects such as reduced power expense and lower carbon emissions.
Korens expects that implementing Korean-style physical AI technology in its manufacturing processes will enable manufacturing innovation in which AI is involved not only in controlling processes but also in research and development stages such as new material development and new product design.
A Korens official emphasized that AI autonomous factories and data-driven infrastructure will secure an overwhelming competitive edge and that the company will expand its business into various fields such as electrification, hydrogen and defense to become a leader in global manufacturing innovation, adding, We will do our best to leap forward as a leader in global manufacturing innovation together with our affiliates.
[OSEN]