Digital employees make production lines smarter
Persist in placing the focus of economic development on the real economy, and adhere to the direction of intelligence, greenness, and integration.
——Excerpted from the Communiqu é of the Fourth Plenary Session of the 20th Central Committee of the CPC
On the afternoon of October 22nd at 3 o'clock, the electronic screen in the control room of Jiangsu Suzhou Huaxing Optoelectronic Technology Co., Ltd.'s 8.5-generation LCD panel factory was constantly flashing with a defect image of a coating repair equipment. This is a defective product on the production line. In the past, this situation had to be repeatedly identified and confirmed by technical personnel. Now, there is no need for manual intervention. Our 'digital employees' will handle it, "said Zhang Dongliang, the public affairs manager of Suzhou Huaxing.
During the conversation, the 'digital employees' have already taken action. Firstly, the real-time image of the product is compared with the defect sample library to quickly locate a breakpoint on the panel circuit. Then, the operation instruction is issued, and the equipment in the workshop immediately performs "laser surgery" on the product. In a few seconds, the line breakpoint was repaired.
Nowadays, Suzhou Huaxing has "cloned" thousands of equipment and every pipeline in the workshop into the electronic screen in the central control room, which is intelligently managed by the "digital brain".
Producing a LCD panel requires more than 200 processes, making it difficult to tailor a 'digital brain' for such a complex production line, "said You Wei, Director of Intelligent Manufacturing at the company.
In 2021, in order to reduce costs, increase efficiency, and enhance competitiveness, the company decided to digitize this large-scale production line and formed an artificial intelligence project research team consisting of dozens of senior engineers, with You Wei as the leader. We encountered difficulties from the beginning, "You Wei said. Although he has extensive experience in production management, he knows very little about digital algorithms; The talents introduced by the project team are proficient in algorithms, but they are outsiders in panel production technology.
To solve this problem, You Wei delved into digital algorithms. Slowly, the demand for intelligent solutions he proposed became more and more professional, and everyone understood and cooperated seamlessly.
The process of transformation has been climbing and overcoming obstacles along the way. You Wei recalled that when the "digital employees" first went online for testing, their performance was not satisfactory. For example, when black spots appear on defective panels, "digital employees" often cannot distinguish whether it is a foreign object or a broken wire.
In order to make "digital employees" smarter, the project team is feeding a large amount of scene data while continuing to optimize algorithms. After trying thousands of solutions, the system finally debugged to the optimal state, which can accurately determine the causes of various panel micro defects and make repair decisions. Sometimes it's more reliable than a master craftsman, "said You Hao, the production supervisor of the company." Recently, the system analyzed multi-dimensional real-time data such as torque, vibration, pressure, etc. of an important equipment, discovered the risk of equipment failure, and issued warnings, allowing us to intervene in advance and avoid losses
At Suzhou Huaxing, the application of artificial intelligence throughout the entire process of research and development, production, manufacturing, and equipment operation and maintenance is like having a smart brain, sharp eyes, and agile hands. Compared with before the transformation, the company's yield rate has increased from 95% to 98%, operating costs have decreased by 7.74%, and overall labor productivity has increased by 123.91%. Before the production line renovation, the production cycle from a glass substrate to a LCD panel took at least seven to eight days. With the 'digital brain', it can be completed in four days, "said Zhang Dongliang.
Information source: People's Daily
