Skip to content

0024. Adsorber performance basis: an equilibrium TSA cycle and external results, outside the simulation

  • Status: Accepted
  • Date: 2026-10-27

Context

An adsorber (PSA, TSA) in process/simulation.yaml is a component_separator: the engineer states its recoveries and DWSIM closes the balances around them. A converged flowsheet then says nothing about whether the split is attainable, and the model basis reports it as component-split (specified). Process-library work (the ASU front-end purification, hydrogen PSA, gas dehydration) needs a basis that is calculated or at least checked, without making such a model a prerequisite for a concept, a BFD or a sourced inventory.

Options considered: a dynamic PSA/TSA bed model (PDEs for breakthrough and cycles), which needs kinetics and heat-transfer data that a design rarely has and a solver core does not ship; a DWSIM unit operation, which DWSIM does not provide; coupling an external model as a unit inside the DWSIM solve, as the fixed bed is coupled, which needs an outer loop; or a calculation beside the simulation whose recoveries the component_separator carries.

Decision

  • A plant repository may hold process/adsorption.yaml (schema cheme/adsorption/v1, chemeascode.domains.process.separations.schema, cheme schema process-adsorption). It is optional; without it a sourced component_separator split stays the representation, labelled as specified.
  • An entry with technique: tsa, performance: calculated is evaluated by core's equilibrium TSA cycle (chemeascode.domains.process.separations.adsorption, cheme process adsorption): local equilibrium from sourced isotherms (Langmuir, dual- and multi-site Langmuir, Sips, Toth, extended Langmuir or Sips within a layer), breakthrough at the stoichiometric time less half a measured mass transfer zone, the cyclic-steady-state balance for recoveries and product impurity, regeneration heat carried by the hot purge with an ideal thermal front, the purge needed to carry the desorbed components, and adsorbent replacement. Every isotherm states the temperature and partial-pressure range it was measured over; core evaluates it nowhere else and refuses the calculation instead (cheme.adsorption-isotherm-range).
  • An entry with performance: external holds an external model's recoveries, utilities and operating range. Core checks the range, the balance and the specification; the performance stays the external model's. A calculated PSA is refused (cheme.adsorption-unsupported): core has no PSA cycle model.
  • The simulation keeps the component_separator. When an entry qualifies (every case inside the ranges, the specification met, the regeneration feasible, every input sourced) and the simulator's recoveries agree with it within split_tolerance, the model basis reports adsorption-equilibrium (predictive, not simulated) or external-performance (specified); otherwise the basis stays as it was and the reason is a limitation. predictive.model-basis therefore still fails for an adsorber ("not in the simulation"): a calculated split is not a simulated one.

Consequences

  • Positive: a TSA's recoveries, regeneration duty and replacement rate follow from data with provenance; the report keeps prescribed and calculated performance apart; nothing new is required for conceptual work.
  • Negative: a second place (the adsorption record) has to agree with the simulation spec; core reports a difference but does not write the spec.
  • Known limitations: no thermal front, breakthrough dynamics, heat losses or PSA cycle; a component without an isotherm is treated as not adsorbed. The reference case (CO2 on 13X, NASA/TM-2017-219847 isotherm, US 6,506,236 mass transfer zone) is qualified for its equilibrium inputs and the cycle relations, not against a measured TSA cycle. Coupling the calculation into the DWSIM solve, as for the fixed bed, is a possible follow-up.

Evidence

  • src/chemeascode/domains/process/separations/{schema,adsorption}.py, src/chemeascode/domains/process/model_basis.py (apply_adsorption), src/chemeascode/project/maturity.py, tests/test_adsorption.py, tests/test_adsorption_reference.py, docs/framework/maturity.md.