Neqsim Column Internals And Rate Based

Gotchas

• Model-class boundary: PackedColumn subclasses the equilibrium DistillationColumn. Its getTheoreticalStages() is a post-solve hydraulic height/HETP equivalent; getNumberOfTrays() is the stage topology actually solved. Neither is a segmented rate-based mass-transfer result.
• Rate-based is not a general distillation column: RateBasedPackedColumn is counter-current and segmented, with configured transfer species and mass/heat transfer correlations. It requires gas and liquid inlet streams. It is not a general multicomponent crude tower with a condenser, reboiler, side products, tray equilibrium, or an automatic mass-transfer coefficient fitted to vendor data.
• Convergence is a result gate: do not consume outlet composition merely because run() returned. Require solved(), finite getLastConvergenceResidual() <= getConvergenceTolerance(), and component balances. An exhausted fixed-point run throws IllegalStateException; tests verify it does not publish the previous profile as a solved new run.
• Unqualified simultaneous solver path: the source test for simultaneous-segment enthalpy conservation is disabled as tracked issue 4094. Keep the default fixed-point segment solver for the tested path unless the alternative is independently validated; do not claim its energy closure based on convergence alone.
• Hydraulic inputs are SI: standalone calculators take mass flows in kg/s, density kg/m3, viscosity Pa s, surface tension N/m, diffusivity m2/s, geometric dimensions m except tray hole diameter in mm. Their setters are not NeqSim unit-converting setters.
• Do not trust default/zero metrics: run calculate() on the standalone calculators. For rate-based equipment, validateSetup() reports missing gas/liquid feeds, missing fluid systems, nonpositive diameter, negative height, missing segments, or missing packing with specific remediation. For ColumnInternalsDesigner, a missing column logs an error and returns; it does not throw. Check design-result validity and freshness.
• Pressure/temperature conventions: thermodynamic systems use K and bara unless an overload explicitly names a unit. DistillationColumn.addLiquidPumparound(...) takes its temperature drop in K and tray indices are bottom-up; verify tray numbering before mapping a P&ID or reference table.
• Solver and tear interactions: the available DistillationColumn.SolverType values include DIRECT_SUBSTITUTION, DAMPED_SUBSTITUTION, INSIDE_OUT, MATRIX_INSIDE_OUT, WEGSTEIN, SUM_RATES, NEWTON, NAPHTALI_SANDHOLM, MESH_RESIDUAL, and AUTO. NEWTON is a tray-temperature accelerator, not a full MESH Newton solve. Active pumparounds require outer return-stream coupling; the column logs and uses MESH_RESIDUAL instead of NAPHTALI_SANDHOLM when pumparounds are configured.
• Pumparounds versus side products: a pumparound draw is an internal recycle, not an external product. Generic DistillationColumn side-draw fractions and flow specs are supported, but SarirAtmosphericSideStripperContactScreen is only a source-bounded contact screen, not a multistage side stripper.
• No inferential compliance: passing a local flood fraction or pressure-drop check is not a vendor guarantee, API/ASME/NORSOK compliance finding, mechanical rating, or validated mass-transfer design.

Validation / benchmarks

• Use a shortcut Fenske–Underwood–Gilliland result as an independent first estimate of minimum stages/reflux and actual stage demand only for a key-component, approximately constant-relative-volatility split. Then validate with a rigorous equilibrium column; do not compare shortcut results to rate-based packed transfer as though they shared a model basis.
• Preliminary design checks commonly target about 70–85% of flooding for tray operation; packing often uses a lower design fraction such as 65–75%. These are screening bands, not mandatory limits. Check top, bottom, and controlling internal stages rather than only one average condition.
• Inspect tray weeping, entrainment, downcomer backup and pressure drop together. For packing, inspect flooding, wetting, pressure drop per metre, HETP and bed height together; an acceptable flood percentage does not prove adequate separation.
• Use rough HETP ranges only to challenge order of magnitude: random packing often about 0.3–0.6 m, structured packing about 0.2–0.5 m; tray spacing/efficiency depends on service. The calculator tests use broad physical sanity bounds, not a universal performance guarantee. Replace generic values with vendor/test data for final design.
• For rate-based results, check signed component transfer against inlet/outlet material closure, direction of transfer, monotonic trend with packing height under fixed conditions, segment profile finiteness, and energy residual when heat transfer is active. The source tests use relative component-balance tolerances from 1e-5 to 1e-6 for their synthetic cases.
• For PackedColumn, confirm the hydraulic result belongs to the latest run and compare its getTheoreticalStages() to packed height/HETP, not to the stage count solved by the VLE model. For mechanical design, record the chosen internals preset, capacity factor, diameter override, pressure-drop basis, and vendor source.
• ProductBoilingPointDistribution is a discrete mole-basis diagnostic. Use Standard_ASTM_D86 or its qualified D86/TBP conversion where applicable; no D2887 implementation is present in this package.

Related skills

• neqsim-distillation-design for equilibrium-column setup, solver selection, specification combinations, convergence and conventional design heuristics.
• neqsim-process-modeling for flowsheet wiring, stream basis, process assembly and process-level validation.
• neqsim-api-patterns for EOS, mixing rules, property initialization and units.
• neqsim-dynamic-simulation for transient/control behavior; steady-state internals results do not qualify dynamic performance.
• neqsim-benchmark-reference-data for independent benchmarks, tolerance policy and source provenance.
• neqsim-professional-reporting for uncertainty, assumptions and design limitations in deliverables.