Pls-Based Soft-Sensor to Predict Ammonium Concentration Evolution in Hollow Fibre Membrane Contactors for Nitrogen Recovery

Resource recovery from wastewater is a key element for the effective implementation of circular economy in wastewater treatment plants (WWTPs). Usually, biological nitrification-denitrification processes are used in WWTPs to remove nitrogen from wastewater to protect the receiving aquatic ecosystems. However, these biological processes impede the recovery of this valuable nutrient and its removal require both high-energy consumption and high footprint. The hollow fibre membrane contactor (HFMC) has emerged as a promising technology for nitrogen recovery (as ammonium sulphate, which is a marketable fertilizer) from different waste streams. During the operation of a HFMC, the ammonium concentration decreases and real-time measurements would be of great value for process monitoring, optimization and control. Ammonium probes exist in the market, but they are expensive and present noticeably maintenance costs. In contrast, pH is an easy-to-measure process variable that can be routinely acquired. In this paper, a data-driven soft-sensor based on PLS has been developed to model input data related to the pH and return output information associated with the total ammonium nitrogen (TAN) concentration. Feature extraction based on technical knowledge of the process probed to be especially useful, since the pH derivative and the pH increment after each reagent dosing were more relevant for TAN concentration prediction than the pH value itself. The developed soft-sensor was able to accurately reproduce the evolution of TAN concentration during the HFMC process operation, thus making it possible to monitor the process as well as enabling future development of different control and optimization strategies