Expectations for AI utilization to take root and expand applications as smart factories move to the next stage: MONOist 2022 outlook (1/2 page)

The movement toward smart factories has become one of the major themes in the manufacturing industry's promotion of DX (digital transformation), and is steadily expanding. Various solutions have been created for the real-world "data conversion," which has been regarded as a barrier, and the technical barriers are falling.

Advanced companies are beginning to expand various efforts to create new value through analysis, etc., making use of the foundation of data conversion in the physical world. On the other hand, for companies that are reluctant to obtain results, it is necessary to reconsider from the method of organizing the purpose and data acquisition. With the intermittent movement restrictions of people due to the Korona-ka, it is expected that various forms of data utilization will spread in 2022 with the point of "expansion of applications" utilizing AI.

≫ Back number of "MONOist New Year Outlook"

Since the announcement of the manufacturing innovation project "Industry 4.0" in Germany in 2011, efforts to make factories smarter have progressed step by step. Until around 2015, when the "Industry 4.0 Practical Strategy" was announced, the period of searching for concepts and concrete countermeasures continued, but after that, the technologies, solutions, standards, etc. drawn by these concepts were realized one after another. Began to be. Newly constructed factories are beginning to incorporate the functions drawn by smart factories in some way. In addition, there is a growing movement to package and sell the specific solutions used in these products.

Before the 2020 Korona-ka, there were still skeptical views, but due to the situation where people could not go to the site and restrictions on manufacturing that depended on people, IoT (Internet of Things) and AI (artificial intelligence) ) And other data-based new manufacturing values ​​were reaffirmed. Digital investment, including network development in factories, which was previously postponed, is also becoming active.

In 2022, as smart factories are steadily expanding, applications (uses) that utilize the obtained data to produce results are expected to expand.

Regarding "visualization", we have enhanced various sensors, strengthened the information acquisition function for manufacturing machines and control equipment, spread network equipment to collect, and dashboard software that accumulates information and displays it in a meaningful way. It has entered the stage of being easy to use due to expansion, etc., and is actually being introduced in various forms at many manufacturing sites. If the "information you want to see" can be organized on the manufacturing site side, the situation is such that the introduction can be promoted immediately regardless of whether it is developed in-house or procured from outside.

If it is possible to create a form in the factory where the results can be obtained by these "visualizations", it will be possible to naturally accumulate data. The burden of data acquisition was a major obstacle to the shift to smart factories. As a result, an environment was created in the factory where data was not available in a form that could be used for analysis. It can be said that the major point of the last few years is that "data accumulation" has progressed as the environment in which "visualization" can be promoted with less burden gradually expands.

Based on this trend, what I would like to expect in the next phase is to increase the variations that utilize the obtained data. The point here is the penetration of AI. At the concept and demonstration level, applications that utilize various data are already being introduced, but it is expected that the number of situations where data will be actually used in the field and produce results will increase once the foundation for data acquisition has been established.

For example, in the maturity of Industry 4.0 shown by acatech in Germany, it is predicted that the next stage of "visualization" will be "understanding of the cause" and "future prediction", and then "self-optimization". In addition, Yokogawa Electric has narrowed down the three areas of "equipment abnormality prediction," "cause identification," and "product quality prediction" as the uses of AI in the factory through various demonstrations at its own factory. As these show, the use of data and AI is expected to expand in the world of "understanding the cause", "utilizing prediction", and "autonomous control".