Monitoring the dynamic process of non-thermal plasma decontaminated water with Raman spectroscopy real-time analysis system

Abstract

This study focuses on the target pollutant, indigo carmine (E132), employing Raman spectroscopy in conjunction with machine learning techniques. An accurate quantitative analysis model was developed to monitor the dynamic plasma decontamination process, utilizing the partial least squares (PLS). The PLS model exhibited excellent predictive performance, achieving adjusted determination coefficients (R2) of 0.9990 and root mean square error (RMSE) values of 0.0058 g/L for the training set, and 0.9730 and 0.0178 g/L for the test set. Real-time monitoring using Raman spectroscopy investigated the impact of discharge voltages, gas flow rates, and solution pH values on E132 degradation. Feedback control based on real-time data led to a 75 % improvement in degradation efficiency by adjusting discharge gas velocity and enhanced energy utilization through discharge time control. Furthermore, the study provided insights into the mechanisms of air plasma-based wastewater treatment. This research presents a convenient and efficient method for online pollutant analysis, optimizing plasma-based water treatment parameters, improving treatment effectiveness, and conserving energy. It lays the groundwork for integrating advanced control algorithms to achieve automated decontamination processes.