Calibration and uncertainty analysis of a fixed-bed adsorption model for CO2 separation

Fixed-bed adsorption is widely used in industrial gas separation and is the primary method for atmosphere revitalization in space. This paper analyzes the uncertainty of a one-dimensional, fixed-bed adsorption model due to uncertainty in several model inputs, namely, the linear-driving-force (LDF) mass transfer coefficient, axial dispersion, heat transfer coefficients, and adsorbent properties. The input parameter uncertainties are determined from a comprehensive survey of experimental data in the literature. The model is first calibrated against experimental data from intra-bed centerline concentration measurements to find the LDF coefficient. We then use this LDF coefficient to extract axial dispersion coefficients from mixed, downstream concentration measurements for both a small-diameter bed (dominated by wall-channeling) and a large-diameter bed (dominated by pellet-driven dispersion). The predicted effluent concentration and temperature profiles are most strongly affected by uncertainty in LDF coefficient, adsorbent density, and void fraction. The uncertainty analysis further reveals that ignoring the effect of wall-channeling on apparent axial dispersion can cause significant error in the predicted breakthrough times of small-diameter beds.