Data-driven compliance: Optimising MPBR performance for ammonium removal and maximum water reuse for EU standards

This study presents a data-driven approach for the water management sector, utilizing a logistic regression (LR) framework to predict effluent ammonium nitrogen (NH 4 -N eff ) concentrations and assess water reuse suitability. The system under investigation is an outdoor membrane photobioreactor (MPBR) operating as a tertiary treatment step for nutrient removal from the effluent of an anaerobic membrane bioreactor (AnMBR), which is responsible for primary organic matter removal from municipal wastewater. The proposed framework estimates the likelihood of compliance with stringent Italian reuse regulations (NH 4 -Neff < 1.6 mg N/L). Utilizing 1.5 years of continuous outdoor operational data, eight binary LR models were developed. The input variables are explicitly defined and encompass operational parameters, solids retention time (SRT), hydraulic retention time (HRT), aeration flowrate (F air ), and nitrogen loading rate (NLR), as well as environmental and biological factors, including light availability, temperature, dissolved oxygen, and microalgal biomass indicators. An advanced control strategy was simulated to optimize the four operational parameters simultaneously. In-silico projections demonstrated significant practical advantages over fixed operating conditions: enhanced NH 4 -N eff removal efficiency (12–50% reduction), decreased aeration requirements (32–38% reduction in F air ), and increased production of compliant reclaimed water (up to 750 L per m 3 of treated wastewater). The proposed framework could therefore support supervisory decision-making and potential integration with existing SCADA systems while retaining a transparent analytical structure.