To promote labor-saving and manpower reduction in the operation of drinking water treatment plants, a control system was developed that combines machine learning-based prediction of coagulant dosing rate with coagulation status evaluation. 　Long-term field verification was conducted at the Inagawa Water Treatment Plant. The results confirmed that the system functioned effectively, demonstrating its potential for reliable practical operation. Furthermore, an influent turbidity prediction method was developed using meteorological information of the water source area. Verification at the same plant showed that the timing of turbidity rise could be predicted without delay, indicating its potential for operational management support.

View full abstract

[in Japanese]

View full abstract