End-to-end Autonomous and Resilient Operability Strategy of Full-Scale PN-SBR System: from Influent Augmentation to AI-aided Optimal Control and Scheduling

This study developed an artificial intelligence (AI)-driven autonomous resilient operation of partial nitrification (PN) process to enhance the performance of a two-stage PN-anammox process by ensuring a specific NO2/NH4 ratio under varying influent conditions before the wastewater enters the subsequent anammox process. First, a mathematical model representing PN process was developed in a full-scale sequencing batch reactor (PN-SBR). That model was then calibrated using AI-driven multi-objective optimization targeting NH4, NO2, and NO3 using a comprehensive rank-based global sensitivity analysis framework to identify the biological kinetic and stoi-chiometric parameters that has high affection on the PN process. After validating the PN-SBR model with real-time measurements of the PN process, an AI-driven optimal aeration strategy (AI-OpAS) was developed for the targeted PN-SBR. The influent characteristics were augmented using the oversampling method and then clus-tered using varying influent conditions. The AI-OpAS used iterative approximate dynamic programming (iADP) to determine the optimal aeration length scheduling and optimal dissolved-oxygen control policy to maintain autonomously the NO2/NH4 ratio while reducing aeration energy under diverse influent conditions. The results show that the proposed AI-OpAS can operate the PN-SBR process by achieving the proposed NO2/NH4 ratio of 1.1 and reducing overall aeration energy consumption by up to 31.38 % under varying influent conditions; furthermore, this AI-aided strategy of resilient operation can be extended to the other variables such as pH for autonomous and sustainable PN-SBR system simultaneously.