EFFECT OF IMPREGNATION CONDITIONS ON FINAL METAL DISTRIBUTION INSIDE A SPHERICAL SUPPORT

碩士 === 大同大學 === 化學工程研究所 === 91 === A mathematical model of dry-impregnation and drying for preparing non-uniform catalyst in a spherical support has been developed. The model for dry-impregnation, which was used for co-impregnation of nickel nitrate and citric acid, consists four simultan...

Full description

Bibliographic Details
Main Authors: Song-Jay Chang, 張松節
Other Authors: Jan-Chen Hong
Format: Others
Language:en_US
Published: 2003
Online Access:http://ndltd.ncl.edu.tw/handle/12374399465686239187
Description
Summary:碩士 === 大同大學 === 化學工程研究所 === 91 === A mathematical model of dry-impregnation and drying for preparing non-uniform catalyst in a spherical support has been developed. The model for dry-impregnation, which was used for co-impregnation of nickel nitrate and citric acid, consists four simultaneous PDE’s, while the model for drying consists five PDE’s and one ODE. The dry-impregnation model is a moving-boundary problem which has been transformed to a fixed-boundary one by change of variables. The simultaneous PDE’s in the mathematical model have been transformed to a series of ODE’s by orthogonal collocation. Gear’s method was used to solve the ODE’s. To prepare a catalyst with desired metal distribution, an optimization problem was set up to search for the best impregnation conditions including impregnation time and bulk concentrations of nickel nitrate and citric acid. Generalized reduced gradient method was used to find the optimal. The parameters used in computer simulation are experimental data obtained previously in this research group. Simulation results show that, as impregnation time or Ni concentration increases, the metal distribution moves toward the center of the pellet. In the meantime, the value of the maximum increases. Both the maximum value and the average amount of the nickel nitrate become larger with increasing acid concentration. Results of the optimization problem show that a global maximum can be obtained, and the problem is a multi-optimum one, however, the global optimum is located in a small cuboid.