No abstract

No abstract

This study focuses on the non-linear effect of gas hourly space velocity (GHSV), oxygen (O2) concentration in the feed, the reaction temperature, and the CH4/CO2 ratio on hydrogen production by catalytic methane dry reforming using artificial neural networks (ANN). Ten different ANN models were configured by varying the hidden neurons from 1 to 10. The various ANN model architecture was tested using 30 datasets. The ANN model with the topology of 4-9-2 resulted in the best performance with the sum of square error (SSE) of 0.076 and coefficient of determination (R2) greater than 0.9. The predicted hydrogen yield and the CH4 conversions by the optimized ANN model were in close agreement with the observed values obtained from the experimental runs. The level of importance analysis revealed that all the parameters significantly influenced the hydrogen yield and the CH4 conversion. However, the reaction temperature with the highest level of importance was adjudged the parameter with the highest level of influence on the methane dry reforming. The study demonstrated that ANN is a robust tool that can be employed to investigate predictive modeling and determine the level of importance of parameters on methane dry reforming.

Conventional treatment methods such as chlorination and ozonation have been proven not to be effective in eliminating and degrading contaminants such as Bisphenol A (BPA) from wastewater. Hence, the degradation of BPA using a photocatalytic reactor has received a lot of attention recently. In this study, a model-based approach using a multilayer perceptron neural network (MLPNN) coupled with back-propagation, as well as support vector machine regression coupled with cubic kernel function (CSVMR) and Gaussian process regression (EQGPR) coupled with exponential quadratic kernel function, were employed to model the relationship between the textural properties such as pore volume (Vp), pore diameter (Vd), crystallite size, and specific surface area (SBET) of erbium- and iron-modified TiO2 photocatalysts in degrading BPA. Parametric analysis revealed that effective degradation of the Bisphenol up to 90% could be achieved using photocatalysts having textural properties of 150 m2/g, 8 nm, 7 nm, and 0.36 cm3/g for SBET, crystallite size, particle diameter, and pore volume, respectively. Fifteen architectures of the MPLNN models were tested to determine the best in terms of predictability of BPA degradation. The performance of each of the MLPNN models was measured using the coefficient of determination (R2) and root mean squared errors (RMSE). The MLPNN architecture comprised of 4 input layers, 14 hidden neurons, and 3 output layers displayed the best performance with R2 of 0.902 and 0.996 for training and testing. The 4-14-3 MLPNN robustly predicted the BPA degradation with an R2 of 0.921 and RMSE of 4.02, which is an indication that a nonlinear relationship exists between the textural properties of the modified TiO2 and the degradation of the BPA. The CSVRM did not show impressive performance as indicated by the R2 of 0.397. Therefore, appropriately modifying the textural properties of the TiO2 will significantly influence the BPA degradability.

No abstract