Daikin Industries, Ltd. and Hitachi, Ltd. start a joint demonstration to digitalize quality control know-how in the reaction process of chemical manufacturing, especially in fluorochemicals, at Daikin Yodogawa Plant. The companies have introduced a system that instantly generates alarms for workers to ensure appropriate quality control. Hitachi's Lumada image analysis technology visualizes the status of reaction processes such as liquid foaming status and color change, which were previously checked visually by workers at certain designated points. This digitalization technology extracts the operating status data to maintain quality control and improve future efficiency, according to the companies.
In recent years, given the diversification of customers' needs, rapid progress of digitalization and intensifying global competition, there is reportedly a need for high-mix low-volume production that meet customers' individual needs. In small-scale chemical production processes that are difficult to sense using general instrumentation devices, quality is controlled by monitoring the production status visually at designated points. However, as the quality of finished products varies even under the same manufacturing conditions, quality control requires considering factors seen in liquid foaming state, color change and other reaction states while the liquid is stirred, to improve the consistency. Monitoring frequency is also limited by the need to maintain productivity.
Daikin and Hitachi aim to achieve a next-generation production model utilizing advanced IoT. Their first step was to develop the "Brazing Skills Training Support System" that digitalizes brazing skills by using Hitachi's image analysis technology. Brazing skills require advanced skills and expertise within air conditioner manufacturing. The system commenced operation at Daikin's Shiga Plant in October 2017. Daikin is expanding applications to other locations worldwide.
Daikin and Hitachi have also applied Hitachi's image analysis technology to fluorochemicals manufacturing at Daikin's Yodogawa Plant, collecting the time-series numerical values and image data of the reaction statuses of chemicals and equipment operation statuses using cameras and sensors. By converting the image data to numeric data and correlating them with quality, they can establish quantitative criteria that affect the quality of finished products, according to the companies. The joint demonstration offers the prospect of reducing defective rate and improving productivity. In the future, they will add new logic factors through monitoring and analyzing human decision-making.
For more information, visit: www.hitachi.com