The Autonomous Production System is based on the Intelligent Integrated Production System of Daicel Production Innovation, which was completed in 2000. In chemical plants that manufacture chemical products, raw materials are fed into equipment and undergo various processes, including reaction, extraction, distillation, and drying processes, to create products. In many of these processes, it is impossible to directly see inside the equipment. This is a major difference between assembly- and machining-based types of manufacturing, which allow you to directly observe products during production and assess the situation, and process-based manufacturing, such as in chemical plants.

For this reason, in chemical plants, operations are managed by monitoring data from a variety of sensors installed in each unit of the production equipment, and the know-how required for stable equipment operations while maintaining quality is accumulated within the minds of the operators responsible for plant operation. The Intelligent Integrated Production System created through Daicel Production Innovation is a framework that transforms the tacit knowledge of seasoned operators into explicit knowledge, enabling anyone to make appropriate operational decisions.

• 

  A type of production in which materials are processed into components, which are then combined to complete one product
  Automobiles / home appliances / electrical equipment / precision instruments / machine tools

• 

  A type of production in which products are manufactured by changing the properties of substances through chemical reactions and other processes
  Petroleum refining / steel / chemicals / energy

Daicel Production Innovation enabled stable operations and reduced significant quality issues, but it did not enable the full use of all the know-how distilled from seasoned operators. It is common for quality and cost decisions to be a trade-off, such as when improving quality leads to higher costs. Optimizing both characteristics requires the consideration of a vast number of elements. This made it difficult to systematize and implement. This is because it was necessary to process complex and massive calculations in a timely manner to aim for optimal operations that achieve further reductions in energy consumption, resource use, and costs while pursuing high quality based on existing know-how and skills. With the computers available in 2000, the processing power for this task was limited.

However, Daicel possessed data accumulated over more than 20 years.
All the know-how of seasoned operators was visualized through a Comprehensive Operability Study (Comprehensive OBS), which systematically organizes plant operations in terms of safety, stability, quality, and cost, and at the Daicel Aboshi Plant alone, this amounted to 8.4 million bits of know-how.
The Autonomous Production System was born from the idea of making full use of the many years of organized and accumulated know-how leveraging state-of-the-art AI.
A team centered on young workers fostered at manufacturing sites collaborated with the University of Tokyo to develop an original AI application, and in August 2020, they completed the Autonomous Production System, an evolution of the Intelligent Integrated Production System of Daicel Production Innovation. Based on the 8.4 million bits of know-how distilled from seasoned operators, this system visualizes cause-and-effect relationships in manufacturing and uses proprietary AI to predict and calculate optimal solutions for them.

Daicel Production Innovation advances through four steps: preliminary studies to ascertain the necessity of innovation, infrastructure improvement and stabilization to strengthen production infrastructure by reducing the on-site workloads identified in the studies, standardization of operations to universalize the stabilized state, and systematization to prevent regression from standardized practices.

Among them, step 2, standardization, is the step where the know-how of seasoned operators in plant operations is revealed, enabling anyone to perform the tasks. Key points for operation management, such as sensor information, are summarized for each operating pattern of connected plant groups from the perspective of safety, stability, quality, and cost and the viewpoint of manufacturing types and operating loads, and the decision-making flows executed by operators on the production floor are distilled and revealed. At this time, a Comprehensive Operability Study (Comprehensive OBS) is employed to systematically organize operations.

Furthermore, in step 3, systematization, the Intelligent Integrated Production System was built to prevent regression from the standardized operating methods. It is a system that supports the decision-making of operators. Based on the concept of a "system that makes the information needed visible to those who need it when they need it," it is a system that standardizes past know-how into manufacturing technology.

Within the Intelligent Integrated Production System, the importance of monitoring points is divided into three levels.
Monitoring points related to safe and stable operations are considered Priority I, those related to quality and cost are considered Priority II, and those related to optimal operation are considered Priority III.

However, based on the processing power of computers in 2000, the scope of support provided by the system was limited to judgements and operations for Priority I and Priority II.
This is because each of the decision-making elements that lead to optimization of quality, energy intensity, and volume—the monitoring points for Priority III—involved trade-off relationships. For example, the time taken to pursue higher quality could reduce the production volume. In addition, the computing power available at the time was insufficient to fully utilize all the available know-how and skills.