Modeling and optimization of photocatalytic treatment of landfill leachate using tungsten-doped TiO2 nano-photocatalysts: Application of artificial neural network and genetic algorithm

• Published in: Process safety and environmental protection Vol. 117; pp. 267 - 277
• Main Authors: Azadi, Sama, Karimi-Jashni, Ayoub, Javadpour, Sirus
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2018
• Subjects: Artificial neural network; Genetic algorithm; Landfill leachate treatment; Optimization; W-doped TiO2; Landfill leachate treatment; Artificial neural network; Genetic algorithm; W-doped TiO2; Optimization
• ISSN: 0957-5820; 1744-3598
• DOI: 10.1016/j.psep.2018.03.038

One of the most important practices in each water and wastewater treatment process is the accurate modeling, optimization, and finding the best condition which leads to achieve maximum efficiency. Recently, artificial neural network and genetic algorithm have been accepted as efficient tools for empirical modeling and optimization, especially for non-linear phenomena. In the present study, Artificial Neural Network (ANN) was applied to model the temporal variations of landfill leachate COD in the photocatalytic treatment process using tungsten-doped TiO2 (W-doped TiO2) nano-photocatalysts. Four influential parameters on the process efficiency, pH, tungsten content (wt.%), calcination temperature (Temp), and exposure time (T) of leachate were considered to predict temporal variations of the leachate COD concentration. Different ANN structures were developed, trained, validated and tested using the data from 150 experiments. Optimal ANN structure was determined based on three performance measures, MAPE, NRMSE, and R. Prediction process inside the optimal ANN was extracted in the form of simple and user-friendly mathematical formulas. Genetic Algorithm (GA) was used to find the most efficient W-doped TiO2 nano-photocatalysts in the COD removal of landfill leachate. The process optimization was conducted at a fixed exposure time using a GA whose objective function was the mathematical formulas obtained from the optimal ANN model. Based on the modeling results, the ANN model, as a non-linear model, has a high predictive accuracy (4% mean error and 0.98 correlation coefficient) when it comes to prediction of temporal variations of the leachate COD in the photocatalytic treatment process using W-doped TiO2 nano-photocatalysts. Based on the optimization results, the most efficient W-doped TiO2 nano-photocatalysts were provided when tungsten content, calcination temperature, and leachate pH were 2.2 percent by weight, 529°C, and 6.3, respectively.