Modeling and optimization of a photocatalytic process: Degradation of endocrine disruptor compounds by Ag/ZnO

• Published in: Chemical engineering research & design 2017-12, Vol.128, p.174-191
• Main Authors: Jasso-Salcedo, Alma Berenice, Hoppe, Sandrine, Pla, Fernand, Escobar-Barrios, Vladimir Alonso, Camargo, Mauricio, Meimaroglou, Dimitrios
• Format: Article
• Language: English
• Subjects: Artificial neural networks; Bisphenol A; Chemical and Process Engineering; Chemical engineering; Chemical Sciences; Computer Science; Computer simulation; Contaminants; Endocrine disruptors; Engineering Sciences; Evolutionary algorithms; Modeling and Simulation; Modelling; Neural networks; Optimization; Photocatalysis; Photocatalysts; Photodegradation; Polymers; Studies; Synthesis; Zinc oxide
• ISSN: 0263-8762; 1744-3563
• DOI: 10.1016/j.cherd.2017.10.012

[Display omitted] •Two-stage decoupled ANN modeling approach of the degradation rate of bisphenol-A.•Study of the dependence of the photocatalyst performance on its synthesis conditions.•Assessment of the contaminant degradation in terms of its apparent rate constant.•Two-stage inverse optimization in terms of maximizing the degradation rate. Artificial neural network (ANN) modeling was applied to study the photocatalytic degradation of bisphenol-A. The operating conditions of the Ag/ZnO photocatalyst synthesis and its performance were simultaneously modeled and subsequently optimized to target the highest efficiency in terms of the degradation reaction rate. Two ANN models were developed to simulate the stages of the photocatalyst synthesis and photodegradation performance, respectively. A direct dependence between the two networks was also established, thus making it possible to directly relate the degradation rate of the contaminant, not only to the photodegradation conditions, but also to the photocatalyst synthesis conditions. In this respect, an optimization study was carried out, by means of an evolutionary algorithm, in order to identify the optimal synthesis and photodegradation conditions that would result in the degradation of a maximal amount of the contaminant. Through this integrated approach it was demonstrated that neural network models can be proven valuable tools in the evaluation, simulation and, ultimately, the optimization of different stages of complex photocatalytic processes towards the maximization of the efficiency of the synthesized photocatalyst.