pyDAC: Open-Source two-dimensional modeling of s-TVSA for direct air capture: experimental assessment and validity of one-dimensional approximations.
Year: 2026
Type: preprint
Abstract: Open and experimentally validated process models remain scarce in direct air capture (DAC), limiting reproducibility and making it difficult to separate physical conclusions from implementation choices. We present pyDAC, an open-source Python framework for packed-bed DAC simulation with one dimensional (1D) and two dimensional (2D) axisymmetric formulations built on the same governing framework. The 2D model is validated against independent zeolite 13X breakthrough experiments, including temperature measurements at multiple radial positions. We then use pyDAC to assess the validity of the 1D approximation for a steam-assisted temperature–vacuum swing adsorption cycle using Lewatit® VP OC 1065. At fixed operating conditions and bed volume, aspect ratios τ = L/D (bed-lenght/bed- diameter) from 1/8 to 8 are examined. The 1D model systematically predicts more favorable performance, overestimating productivity by up to 14 % and underestimating equivalent work by up to 13 %; discrepancies remain around 4 % even at τ = 8. The residual difference originates from the early, conduction-dominated wall-heating transient, which a constant wall heat-transfer coefficient cannot reproduce. During multi-objective optimization, however, the bias reverses because the optimizer selects shorter pre-heating steps within this transient. At τ = 1/2, the 1D Pareto front requires a median 30 % more equivalent work at equal productivity, and the sampled 2D-optimal designs become infeasible when re-evaluated in 1D. At τ = 8, the fronts nearly coincide, but the corresponding operating decisions differ. These results show that 1D models remain useful for screening, but selected designs should be verified in 2D when radial heat transfer controls cycle performance or constraint margins.
Source: ChemRxiv
Institution: Eindhoven University of Technology
Language: English
Subfield: Filtration and Separation
Field: Chemical Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7221704625