P2O: AI-Driven Framework for Managing and Securing Wastewater Treatment Plants

Paper Details

• Title: P₂ O: AI-Driven Framework for Managing and Securing Wastewater Treatment Plants
• Authors: Ajay Kulkarni; Mehmet Yardimci, Md Nazmul Kabir Sikder, and Feras A. Batarseh
• Publication/Conference: Journal of Environmental Engineering, American Society of Civil Engineers (ASCE)
• Publication/Presentation Date:  June 2023

Abstract

Wastewater Treatment Plants (WWTPs) are critical infrastructures that process wastewater before it’s discharged as effluent into rivers or used for other purposes. 

WWTPs use tunnels to store wastewater for treatment, using sensors to monitor and maintain wastewater levels to prevent untreated material from overflowing into the environment. This monitoring technology can be vulnerable to cyber threats.

Researchers focused on the role of Artificial Intelligence (AI) in predicting tunnel water levels and detecting security threats, using an AI framework called P₂O (Prediction, Protection, and Optimization). 

• The prediction module forecast the tunnel water level using Deep Learning models based on existing wastewater flow and data from sensors and gauges. 
  
• The protection model uses Recurrent Neural Network models to determine if a situation is an anomaly or an intentional attack. 
  
• The optimization module uses a Genetic Algorithm to recommend actions to pump operators. 
  

Results showed the prediction module can predict tunnel water levels with 85 percent accuracy. The protection module can detect about 97 percent of intentional attacks on WWTPs.