EVALUATION OF KINETIC MODELS VIA COMPUTATIONAL OPTIMIZATION TECHNIQUES FOR DIRECT SYNGAS-TO-OLEFINS PROCESS

In this study, the kinetics of Fischer-Tropsch (FT)-to-olefins was thoroughly investigated on activated carbon supported Fe-Cu-K catalyst. Among the evaluated methods, it was observed that the Runge-Kutta method gave better results. In general, the model that pertains to the study of Fernandes introduced fewer errors than Wang et al.'s model. In the simulations based on the Fernandes model, the H2 values were estimated within error limits between 2% and 8% validated for 7 distinct experiments. CO, on the other hand, was estimated more accurately with the error limits in the ranges of 4-7% for high gas hourly space velocity (GHSV) values. However, error band of the model in the latter case was increased up to 30% for low GHSV values. Regarding hydrocarbon products, an average error value of 30% was obtained for ethylene while the best estimations were obtained for propylene with an error of 0.46% and 13.27%. However, Fernandes model neglects olefin re-adsorption, unlike the Wang's model, so Wang model seems more appropriated to estimate paraffin products. All in all, as per calculations on experimental analysis, both models need improvement in terms of product distribution.