Application of Artificial Neural Networks for Prediction of Photocatalytic Reactor

Abstract

In this paper, forecasting of kinetic constant and efficiency of photocatalytic process of T i02 nano powder immobilized on light expanded clay aggregates (LECA) was investigated. Synthetic phenolic wastewater, which is toxic and not easily biodegradable, was selected as the pollutant. The efficiency of the process in various operation conditions, including initial phenol concentration, pH, Ti02 concentration, retention time, and UV lamp intensity, was then measured. The T i02 nano powder was immobilized on LECA using slurry and sol-gel methods. Kinetics of photocatalytic reactions has been proposed to follow the Langmuir-Hinshelwood model in different initial phenol concentration and pH. Several steps of training and testing of the models were used to determine the appropriate architecture of the artificial neural network models (ANNs). The ANN-based models were found to provide an efficient and robust tool in predicting photocatalytic reactor efficiency and kinetic constant for treating phenolic compounds. Water Environ. Res., 87, 113 (2015).

Application of multilayer perceptron network and random forest models for modelling the adsorption of chlorobenzene on a modified bentonite by intercalation with hexadecyltrimethyl ammonium (HDTMA)

“…A large number of simple processing elements (neurons) are arranged in three layers: the input layer (independent variables), output layer (dependent variables), and one or more hidden layer(s) based on the complexity of the problem at hand. However, theoretical studies have shown that the single hidden layer is sufficient for an ANN to approximate any complex ( 1) nonlinear function [27,47]. The general expression of MLP network can be given as:…”

“…ANNs, inspired by biological nervous processing, are a parallel distribution processing method combined of neurons and weights and are based on the principle that highly connected system of simple processing elements. They can train complex interrelationships between inputs and outputs variables [27,28]. Neurons are the basic element in ANN, which is linked to neurons in the next layer and there forming different types of ANNs.…”

Modeling, Instrumentation, Automation, and Optimization of Water Resource Recovery Facilities

“…Cheng and Zhi (2015) ANN-based models were found to provide an efficient and robust tool in predicting photocatalytic reactor efficiency and kinetic constant for treating phenolic compounds (Delnavaz, 2015).…”

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

“…Traditionally, ANN models have been applied to photocatalytic degradation processes in order to study the effect of a variety of reaction conditions on the photocatalytic performance by means of percentage of degradation or removal efficiency. It is only recently that the apparent reaction rate constant of a first-order photocatalytic degradation curve was considered as the simulated response of the developed ANN model (Behnajady & Eskandarloo, 2015;Delnavaz, 2015). The characteristics of a series of similar recent studies are summarized in Table 1. As can be seen, ANN models can be used to assess the effect of numerous important factors of the process, such as light intensity, organic/inorganic ions concentration and oxygen dose, which influence the photocatalytic performance but are rarely considered in kinetic models.…”