Modeling PVA degradation in a continuous photochemical reactor using experimental step testing and process identification

Abstract

Abstract In advanced oxidation processes (AOPs) operating in continuous mode, the flow rate and the concentration of oxidants such as H2O2 at the mixing point of the influent must be controlled so that the oxidant residuals in the effluent are minimized while the organic degradation is maximized. Designing a controller to regulate the oxidant flow rate and therefore, its concentration at the mixing point of the photoreactor influent, needs the determination of a reliable process dynamic model. In this study, a step testing technique is employed to construct a dynamic model for the degradation of polyvinyl alcohol (PVA) in a UV/H2O2 continuous annular photoreactor. The experimental design consists of preparing three initial PVA concentrations of 60, 280, and 500\xa0mg PVA/L and conducting eight experimental tests for different inlet mass flow rates of H2O2 ranging from 0.336 to 125\xa0mg\xa0H2O2/min. The process responses of pH solution and total organic carbon (TOC) are measured versus time for different step changes of H2O2 inlet flow rates. Process dynamic models represented by transfer functions of the continuous UV/H2O2 photoreactor are determined from the collected input-output data using graphical method, identification algorithm, and ARX (AutoRegressive with eXogenous) in Matlab environment. The dynamic models predict the data trend quite well. Variations of the steady state gain demonstrate that the PVA degradation in UV/H2O2 continuous photoreactor is a nonlinear process. Besides, the ARX algorithm generates reliable dynamic discrete models that could be used to develop digital controllers for computer control of the UV/H2O2 photochemical reactor.