Chemical addition significantly affects the efficiency of sludge dewatering at wastewater treatment plants (WWTP). While necessary for effective separation, chemical represents a substantial operational cost. An optimal approach to minimizing chemical dosage while ensuring optimal sludge dryness is critical.The objective of this project is to demonstrate, in the sludge dewatering process, the feasibility of using machine learning (ML) models to predict: 1) how machinery variables effect the centrifuge performance, and
2) if machinery variables could inversely predict the optimum chemical dosage for maximum cake dryness.
This proof-of-concept aims to establish a scalable ML-based model for potential implementation in full-scale wastewater treatment operations, to minimize chemical consumption while maximizing sludge dryness. (Xiangyuan)
