The renowned auto parts maker puts the technology into manufacturing and supply operations to improve product quality and efficiency
_¹ÚÁ¤±¹.jpg)
President Park Chung-kook of Hyundai Mobis.
Hyundai Mobis said March 5 that it has started incorporating artificial-intelligence (AI) technology into its manufacturing and supply operations to improve product quality and efficiency.
The auto parts maker said it developed an algorithm to detect product defects and is implementing the system into its production lines for motor-driven power-steering electronic control units in Jincheon County, North Chungcheong Province.
The algorithm looks for defects using its image-based data system and is expected to improve efficiency from the traditional method of physical checks.
The company said the new system has a detection rate of 98 percent, which will improve as it accumulates more data.
Currently applied to one production line, the company expects to expand its use to five by the end of this year. As for the AI application to its supply operations, the parts maker said it is currently using AI to analyze and predict demand for repair parts.
¡°The inefficiencies that existed due to the limits of technology in each of the operations are now largely solved through the development of AI technology.¡± said Lee Sang-hwa, head of Hyundai Mobis IT planning team.
Hyundai Mobis¡¯ announcement follows its adoption of AI for its research and development operations last year.
According to the Hyundai Motor Group unit, it developed an algorithm recently in which an AI can detect faulty steering parts, and began applying the algorithm at its plant in Jincheon, North Chungcheong Province.
Under the system, the AI computer can detect faulty products by looking into the image data of the part - an electronic control unit (ECU) of motor-driven power steering products.
ECU is a chip on which many micro elements are sitting. To check whether it is working properly, Hyundai Mobis has been using a camera-based system to check whether those elements are in the right place.
However, many faulty cases turned out to be functioning properly in actual tests, thus Hyundai Mobis engineers personally had to inspect those parts found to be faulty with their eyes again.
The company said such a process was inefficient so it had the AI computer learn various types of functioning cases, so it can reduce the workload of engineers.
¡°Currently, the algorithm logs a 98 percent proving rate and the rate will be improved when the AI learns more correct samples.¡± a Hyundai Mobis official said.
Since the algorithm is applicable to manufacturing lines for other chips, the company will expand the use of the AI system to four other manufacturing lines. Also, it will apply the system at its plant in Tianjin, China.
According to the company, it also plans to use another AI system that can predict the demand for after-sales service in the first half of this year.
The forecasting system will analyze a vast amount of data on the weather, seasons, drivers¡¯ habits, number of cars on the road and each model¡¯s maintenance period and assist the company to prepare parts to cope with after-sales demand.
As Hyundai Motor Group¡¯s parts company, Hyundai Mobis is supplying 2.7 million parts for 244 vehicle models and struggling to manage its inventories. The company is using previous demand data to predict a year¡¯s demand, and the system will improve accuracy.
¡°The advance of AI technology is helping the company to address various limitations in its businesses.¡± the official said. ¡°We will continue innovating our business practices by embracing more AI technologies.¡±
_Çö´ë¸ðºñ½º ÁøÃµ°øÀå »ý»ê¹°·ù¿¡µµ AI Á¢¸ñºÒ·®À»Àâ¾Æ³½´Ù.jpg)
Researchers at work at a Hyundai Mobis Lab. (Photos: Hyundai Mobis)