Municipal wastewater UV Disinfection modelling based on artificial neural networks for real-time process control

Accurate modelling of UV disinfection of wastewater to predict efficiency is fundamental to control the process and balance inactivation goal and energy consumption. In this study, a “black-box” model of UV disinfection is proposed, based on artificial neural networks (ANNs), which can predict the initial and residual E. coli concentrations, using real-time monitored wastewater matrix characteristics and UV dose as predictors. The model was calibrated on data coming from a full-scale municipal wastewater treatment plant, delivering treated wastewater to an agricultural district in the peri-urban area of Milan (Italy). The model can be used to optimize UV dose according to wastewater flow rate and quality, guaranteeing reliability of the system compliance with agricultural reuse limits and minimizing energy consumption.