Modeling a photocatalytic reactor for water purification

Abstract

Complex systems, characterized by their interconnected and dynamic components, can be effectively modeled using computational approaches, particularly neural network models. These models capture intricate patterns and behaviors within complex systems, which makes them powerful tools for modeling and prediction. Building on the strengths of neural networks, deep learning has demonstrated remarkable success across various domains due to its ability to learn different representations and model non-linear relationships. The aim of this work is to leverage deep learning techniques to model photocatalytic reactors, with the goal of optimizing the photocatalytic degradation process. We propose a model based on self attention mechanism for the prediction of the photocatalytic degradation rate and efficiency based on a set of experimental parameters. Base models are combined for data augmentation with a meta-model that incorporates a self-attention mechanism for prediction. The base models achieved excellent fits to the experimental data, and the meta-model attained a mean squared error of 0.0055 through five-fold cross-validation. tworks, deep learning has demonstrated remarkable success across various domains due to its ability to learn different represent