Towards the Optimization of Advanced Oxidation Processes Using Machine Learning Modelling: The DIGIT4WATER Project

The use of Machine Learning (ML) based methodologies is becoming an increasingly widespread technique for predicting the efficiency of Advanced Oxidation Processes (AOP) in wastewater treatment and regeneration. However, most studies focus predominantly on one or several organic compounds, without including contaminants of emerging concern. In addition, the outcomes are still far from being applicable to practical scenarios. To fill this gap, we are developing the DIGIT4WATER project, with the main objective of generating ML models to predict degradation levels of various pollutants, many of which are of particular importance concerning current legislation. To this end, a large and comprehensive database is being generated, which includes AOPs and photo-chemical processes based on UV-C or direct solar radiation. The project is aimed at building innovative digital tools for compliance with the new European regulation (EU 2020/741) governing water reuse in agriculture. This paper presents the current results, focusing on the proposed methodological framework, together with the next steps that will be carried out in the final phase of the project.