Abstract:
The adsorptive separation system of water vapor from the natural gas using a milti-layer adsorber was studied. The commercial silica gel and two sizes of 4A molecular sieve were packed in layers in the adsorber. The mathematical model used for predicting the water breakthrough profile was investigated and developed. The experiments were carried out under different humidity levels of the natural gas feed: 7%RH, 30%RH and 50%RH, and different contact times: 17 seconds and 34 seconds, aiming to compare with the theoretical breakthrough curves obtained form the mathematical model. From the sensitivity analysis of parametrical effects on the theoretical breakthrough curves in order to investigate the existing mathematical model, the interstitial velocity (v) and the effective bed voidage (3) were more sensitive to the theoretical breakthrough curves than the effective axial dispersion coefficient (DL,eff). To develop the existing mathematical model, the parameters and the equilibrium adsorption isotherm constructed for each adsorbent were employed specifically in the model. Upon the curve fitting technique for the adsorption isotherm, linear models gave a good corresp[ondence with the experimental data points for the adsorption on the silica gel adsorbent (LCA-94). The Langmuir and exponential models demonstrated a good agreement with the experimental adsorption isotherm of 4ADG 1/8" at the humidity levels of lower than 22%RH and above 22%RH respectively. The equilibrium adsorption isotherm of 4ADG 1/16" was divided into three regions. The Langmuir and the Freundlich models were
successfully accepted to best fit the experimental data at the humidity of lower than 32%RH and above 62%RH, respectively, whereas exponential model revealed a good correspondence with the experimental data at the humidity region between 32% RH and 62%RH. Since the water concentration in the natural gas feed is very low, the decrease in the interstitial velocity due to the water adsorption can be neglected. Therefore, the assumption of constant fluid velocity was applied in the mathematical model. From the sensitivity analysis, in order to achieve the best agreement between the experimental and theoretical breakthrough patterns, the model for all experimental cases. The modified mathematical model for predicting the breakthrough profiles of water adsorption provided an excellent correspondence with the experimenatlly obtained breakthrough curves under various experimental conditions. The differences of the experimental and theoretical breakthrough times were only about 3% to 5%.
