24 Hours of Water: AI-based Guidance System of Operating Parameters for Wastewater Treatment

Описание: Providing a sample of the global content available on WEFTEC Connect, Hiroki Itokawa (Japan Sewage Works Agency) presents on AI-based guidance systems of operating parameters for municipal wastewater treatment. Operating biological wastewater treatment processes still needs expertise, since we cannot fully control both the influent (constituents and their fluctuations) and inherent biological reactions with highly nonlinear nature. This is particularly prominent in large-scale municipal wastewater treatment plants (WWTPs), where operators have to control multiple treatment trains with mostly ununiform characteristics. However, the number of operating engineers/staffs has been likely to be reduced due to cost constraint and other local circumstances. This is of great concern in some nations like Japan, where shrinking population has become evident and the shortage of skilled engineers is getting a big social issue.

In this situation, operators' support system for municipal WWTPs is of great help. Widely used on-line control technologies of operating parameters (e.g. air flow (AF) control for biological treatment) are one option, although they are basically dependent on limited number of input online data. In manual operation, plant operators determine operating parameters by integrating large number of data items with their own knowledge and experience. Other option would be a statistical prediction approach, which can deal with a large number of data and make an output with pre-defined or self-trained model(s).
In 2018, the authors launched a research project, where a guidance system of operating parameters for municipal WWTPs using AI-based predictive model was developed and evaluated under 'B-DASH Project', government-commissioned R&D projects conducted by MLIT, Japan. This presentation describes the outline of the guidance system and preliminary demonstration results, in which prediction by the system was compared with actual WWTP operation data in order to verify whether its predicting power would be practically acceptable.

For more information about WEFTEC visit: https://www.weftec.org/

#WEFTEC, #24HoursOfWater, #WaterEnvironmentFederation,