Artificial neural network modeling of phenol adsorption onto barley husks activated carbon in an airlift reactor

Loredo-Cancino, Margarita (2013) Artificial neural network modeling of phenol adsorption onto barley husks activated carbon in an airlift reactor. Doctoral thesis, Universidad Autónoma de Nuevo León.

Preview

Text
1080256819.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.
Download (1MB) | Preview

Abstract

Purpose and method of the study: The production of activated carbon from barley husks (BH) by chemical activation with zinc chloride was optimized by using a 23 factorial design with replicates at the central point, followed by a central composite design with two responses (the yield and iodine number) and three factors (the activation temperature, activation time, and impregnation ratio). Both responses were simultaneously optimized by using the desirability functions approach to determine the optimal conditions of this process. The experimental data from the batch phenol adsorption onto barley husks activated carbon (BHAC) was represented by adsorption isotherms (Langmuir and Freundlich) and kinetic models (pseudo-first and pseudosecond order, and intraparticle diffusion models), besides the regeneration of phenolloaded BHAC with different solvents was evaluated. Experimental data confirmed that the breakthrough curves were dependent on BHAC dosage, phenol initial concentration, air flow rate, and influent flow rate. Adaline and feed-forward backpropagation Artificial Neural Networks (ANNs) were developed to predict the breakthrough curves for the adsorption of phenol onto barley husks activated carbon (BHAC) in an airlift reactor. Feed-forward back-propagation networks were tested with different quantity of neurons at the hidden layer to determine the optimal number of neurons in the ANN architecture to represent the breakthrough curves performed at different operational conditions for the airlift reactor. Contributions and conclusions: After the simultaneous dual optimization of BHAC production, the maximal response values were obtained at an activation temperature of 436 °C, an activation time of 20 min, and an impregnation ratio of 1.1 g ZnCl2 g BH-1 , although the results after the single optimization of each response were quite different. At these conditions, the predicted values for the iodine number and yield were 829.58 ± 78.30 mg g-1 and 46.82 ± 2.64%, respectively, whereas experimental tests produced values of 901.86 mg g-1 and 48.48%, respectively. Moreover, activated carbons from BH obtained at the optimal conditions mainly developed a porous vstructure (mesopores > 71% and micropores > 28%), achieving a high surface area (811.44 m2 g-1 ) that is similar to commercial activated carbons and lignocellulosic-based activated carbons. These results imply that the pore width and surface area are large enough to allow the diffusion and adsorption of pollutants inside the adsorbent particles. Freundlich isotherm model satisfactorily predicted the equilibrium data at 25 and 35 °C, whereas the Langmuir isotherm model well represented the equilibrium data at 45 °C. The maximum phenol adsorption capacity onto BHAC was 98.83 mg g-1 at 25 °C and pH 7, similar to phenol adsorption onto commercial activated carbons. The kinetic data were adequately predicted by both the pseudo-first order and intraparticle diffusion models. The external mass transfer was minimized at stirring speeds greater than 400 min-1 , and the adsorption kinetics are affected by both initial phenol concentration and temperature. Adsorption equilibrium was reached within 40 and 200 min at initial phenol concentration of 1000 mg L-1 at 35 °C and 30 °C, respectively. Ethanol/water solutions at 10% V/V were the most effective regenerating agent, with desorption capacity of 47.79 mg g-1 after five adsorption-desorption cycles. The breakthrough curves of phenol adsorption onto BHAC in an airlift reactor in continuous operation were adequately predicted with feed-forward back-propagation ANN architecture with 2 neurons in the hidden layer for the single-input single-output problem. Correlation coefficients higher than 0.95 were observed between the breakthrough curves predicted by the developed Adaline network and those obtained experimentally for the multiple-input single-output problem. Further improvements and generalization of the developed predictive Adaline network are discussed.

Item Type:

doctoralThesis (Doctoral)

Additional Information:

Doctor en ciencias con orientación en procesos sustentables

Creators: