The recently developed mathematical model for simulation of a PSA system (Raghavan et al., 1985) has been extended to a nonlinear Langmuir equilibrium system. The pressure swing adsorption separation of a small concentration of ethylene from helium has been studied experimentally on 4A and 5A molecular sieves. In the 4A system mass transfer is controlled by intracrystalline diffusion and is slow, whereas in the 5A system mass transfer is very much faster and is controlled by macropore diffusion. With rate and equilibrium parameters derived from experimental breakthrough curves, the theoretical model provides a good representation of the experimentally observed behavior of both PSA systems. A mathematical model has been developed for a n isotherma1 trace PSA system with finite mass transfer resistance and a nonlinear (Langmuir) equilibrium isotherm. The model equations are solved by the method of orthogonal collocation. Using a small laboratory system, the PSA separation of ethylene from helium has been studied experimentally with two different adsorbents. In the 4A sieve mass transfer is controlled by intracrystalline diffusion and is relatively slow, whereas in the 5A sieve the uptake is macropore-controlled and very much faster. The model parameters were derived by matching experimental adsorption and desorption breakthrough curves to the theoretical curves calculated from the mathematical model for a single adsorbent bed subjected to a step change in feed concentration.
CONCLUSIONS AND SIGNIFICANCEThe numerical simulation provides a good representation of the experimentally observed behavior of both the 4A and the 5A systems, suggesting that the model does indeed represent the essential features of the real systems. In the case of 4A sieve, mass transfer resistance is high and a relatively long cycle time is therefore necessary to achieve efficient separation. The model assumes isothermal behavior. It is in principle possible to extend the model to allow for heat effects but only at the expense of a considerable increase in complexity. The results suggest that the present model is probably adequate for most practical purposes and could usefully be extended to the simulation of the more complex multiple-bed PSA cycles used in commercial hydrogen purification processes. The main practical value of such a simulation is that it makes possible a detailed optimization of the PSA cycle.