IA2IA – Start the journey of industrial autonomy
From Industrial Automation to Industrial Autonomy (IA2IA), Yokogawa Electric has been advocating the concept for several years.
Unlike automation, which requires human intervention, autonomy means that cognitive systems can perform all operations independently, can control and manage autonomously, plant assets and operations can learn and adapt, can respond with minimal human intervention, and the ultimate ideal is that the facility can operate autonomously.
This is inevitable progress, but it will not be easy to get to the end. “How do you achieve industrial autonomy? This requires more and better technology to be put into businesses, such as oil refineries, chemical plants and even public transportation. However, autonomy is not just about full automation, assets also need to be modernized. If the systems and infrastructure are old, the technology cannot be embedded. Under such circumstances, it is difficult to achieve the goal of autonomy. That’s why we came up with a master plan and developed a maturity model to determine where the company is now and where it’s going. Of course, it’s a long time commitment. It’s a long process.” Takeoka Yihiko admitted.
“But it also depends on what the client is looking for in the process. For some customers, full autonomy may not be the best option for the company’s operations, and some autonomy may be sufficient. So we will come up with a development model and then calculate the cost. That means money has to be invested. But to achieve your goals, you may face a lack of resources or skills. Nowadays, the use of a large number of digital technologies is an inevitable demand, as I mentioned before, including cloud, artificial intelligence, machine learning, VR, AR, XR, etc., all kinds of technologies must be in place. The question is how mature are these technologies? How can we use it easily? What are the chances of failure? These need to be discussed with the client to make the best choice and finally implemented in the master plan. In the long run, we can figure out the direction and methodology from industrial automation to industrial autonomy.”
In March 2022, with the help of advanced artificial intelligence and IIOT technology, Yokogawa Electric achieved autonomous control of a chemical plant in Japan for 35 consecutive days without human intervention, which created the first industrial control, effectively proving that reinforcement learning AI can be safely applied to real factories.
In March 2023, after a year of stable operation, Yokogawa Electric’s autonomous control artificial intelligence was officially used in the chemical plant of ENEOS Materials, also creating a world precedent. This is the world’s first formal example of using reinforcement learning AI to directly control a factory. In the same month, Yokogawa also extended its IA2IA solution to customers in China with a new application running on Alibaba Cloud. In the future, Yokogawa will continue to explore the application of such AI to other types of processes and plants, and will continue to strive to increase productivity and save energy by expanding the scope of autonomy.