Ebrahim Abouzari-Lotf - Scheduling the blended solution as industrial CO<inf>2</inf> absorber in separation process by back-propagation artificial neural networks

Document created by Ebrahim Abouzari-Lotf on Aug 13, 2015
Version 1Show Document
  • View in full screen mode

  Publication Details (including relevant citation   information):

  Abdollahi, Y., Sairi, N.A., Said, S.B.M., Abouzari-Lotf, E.,   Zakaria, A., Sabri, M.F.B.M., Islam, A., Alias, Y. 150  892-901

  Abstract: Abstract It is believe that 80%   industrial of carbon dioxide can be controlled by separation and   storage technologies which use the blended ionic liquids   absorber. Among the blended absorbers, the mixture of water,   N-methyldiethanolamine (MDEA) and guanidinium trifluoromethane   sulfonate (gua) has presented the superior stripping qualities.   However, the blended solution has illustrated high viscosity that   affects the cost of separation process. In this work, the blended   fabrication was scheduled with is the process arranging,   controlling and optimizing. Therefore, the blend's components and   operating temperature were modeled and optimized as input   effective variables to minimize its viscosity as the final output   by using back-propagation artificial neural network (ANN). The   modeling was carried out by four mathematical algorithms with   individual experimental design to obtain the optimum topology   using root mean squared error (RMSE), R-squared (R2) and absolute   average deviation (AAD). As a result, the final model (QP-4-8-1)   with minimum RMSE and AAD as well as the highest R2 was selected   to navigate the fabrication of the blended solution. Therefore,   the model was applied to obtain the optimum initial level of the   input variables which were included temperature 303-323 K,   x[gua], 0-0.033, x[MDAE], 0.3-0.4, and x[H2O], 0.7-1.0. Moreover,   the model has obtained the relative importance ordered of the   variables which included x[gua] > temperature > x[MDEA]   > x[H2O]. Therefore, none of the variables was negligible in   the fabrication. Furthermore, the model predicted the optimum   points of the variables to minimize the viscosity which was   validated by further experiments. The validated results confirmed   the model schedulability. Accordingly, ANN succeeds to model the   initial components of the blended solutions as absorber of CO2   capture in separation technologies that is able to industries   scale up. © 2015 Elsevier B.V.

  Address (URL):