In what they report is a world's first accomplishment, Yokogawa Electric Corp. and JSR Corp. reported March 22 that they successfully concluded a field test in which artificial intelligence (AI) was used to autonomously run a chemical plant for 35 days. They add the test confirmed that reinforcement-learning AI can be safely applied in an actual plant, and demonstrated that their technology can control operations that have been beyond the capabilities of existing control methods, such as PID control and APC, which have up to now necessitated manual operation of control valves based on staff judgments.
In the field test, the AI solution successfully dealt with the complex conditions needed to ensure product quality and maintain liquids in the distillation column at an appropriate level, while making maximum possible use of waste heat as a heat source. It also stabilized quality, achieved high yield, and saved energy. The AI used in this control experiment was Factorial Kernel Dynamic Policy Programming (FKDPP) protocol, which was jointly developed by Yokogawa and the Nara Institute of Science and Technology (NAIST) in 2018, and was recognized at an IEEE conference on automation science and engineering as being the first reinforcement-learning AI in the world that could be used for plant management.