At Hsinpu Chemical Fiber Plant, the existing wastewater treatment model relies heavily on manual operation for water level control, equipment operation and meter check. In addition, there are no reference parameters such as the electrical current and voltage for over 90% of the operating equipment. The reference has mainly been from sampling, which lacks the precision needed for system calibration. Hence, the plant modified the meters and pipelines. By adding AI algorithms developed by Industrial Technology Research Institute (ITRI), the plant is able to utilize the data to forecast the efficiency of aeration ponds and treatment performance, which optimizes operational performance and ensures the effluent quality.

Project implementation is conducted in 4 phases: 

1. Digitizing – Install electric meters and construct a databank for operational equipment. The phase is projected to conclude in the 4th quarter of 2021.
2. Monitoring – Establish a regional surveillance system and system control platform. The phase is projected to conclude in the 1st quarter of 2022.
3. Automating – Add production monitoring equipment and automatic control logic. The phase is projected to conclude in the 3rd quarter of 2022.
4. Intellectualizing – Build an AI algorithm and effluent COD prediction model. The phase is projected to conclude in the 4th quarter of 2022.