Abstract
High-purity oxygen production through ceramic membranes at high temperature is an interesting cost-effective alternative to cryogenic methods for several industrial applications. It was observed that is very difficult to fit a global model including all the processes that take place in the reactor. The reason is that it needs a methodology to fit many variables and to converge to reasonable values for each parameter.
It has been fitted a global model which is able to predict with a high accuracy the oxygen permeation in an axis-symmetrical BSCF membrane module. The model includes the surface resistance phenomena and the diffusion of the oxygen vacancies in the membranes and has 6 parameters to be fitted. For the fitting two series of experiments at different temperatures were developed: i) experiments changing the composition in the inlet of the feed chamber keeping the composition in the sweep chamber; ii) experiments changing the composition in the inlet of the sweep chamber keeping the composition in the feed chamber. To fit the parameters, the model was implemented in COMSOL Multiphysics 4.4. The parameter fitting was performed for each temperature with a genetic algorithm to reach an initial guess followed by a fitting using Nelder-Mead method.
