Skip to main content

Sizing Methods

The correlations behind each sizing basis. All are standard preliminary-design methods (Towler & Sinnott, GPSA, Seider et al.), chosen to match the AACE Class 4–5 accuracy of the economics engine.

Heat exchangers and coolers (area)​

Area from the log-mean temperature difference (LMTD) method:

A = Q / (U · LMTD)

LMTD = (ΔT₁ − ΔT₂) / ln(ΔT₁ / ΔT₂)
  ΔT₁ = T_hot,in − T_cold,out
  ΔT₂ = T_hot,out − T_cold,in
  • Two-sided exchangers (process–process): hottest inlet is assigned to the hot side, coldest to the cold side. U = 300 W/m²·K (liquid–liquid).
  • Coolers (one process side): sized against cooling water at 30 → 40 °C. U = 600 W/m²·K (process–cooling water).
  • A temperature cross (LMTD ≤ 0) is clamped to a 5 K minimum so the size stays finite; the clamp is recorded as an assumption.

Default U values by service (Towler & Sinnott, Table 19.1):

ServiceU (W/m²·K)
Liquid–liquid300
Gas–gas30
Condenser (organic vapor)700
Reboiler / vaporizer800
Process–cooling water600
Process–steam850

Pumps, compressors, expanders (power)​

Sized directly from the simulated mechanical work:

power = |work|   (kW)

No correlation is involved — the simulation engine already computed the thermodynamic work for the unit's actual path.

Heaters (heat duty)​

Sized directly from the simulated duty, heat_duty = |duty| in kW. The economics engine converts to BTU/hr (× 3412.142) where its fired-heater correlation requires it.

Flash drums and separators (volume)​

Two paths, selected by the phase split at the outlets:

Vapor–liquid separators (Souders–Brown)​

Used when the vessel has a vapor outlet with vapor fraction > 0.05:

u_max = K · √((ρ_L − ρ_V) / ρ_V)        K = 0.107 m/s
D     = √(4·Q̇_V / (π·u_max))            vapor disengagement diameter
H     = max(V_holdup / (π/4·D²), 1.5·D)  liquid holdup + disengagement space
V     = π/4 · D² · H

The K value assumes a vertical drum with a mesh mist eliminator (GPSA). Liquid holdup uses a 5-minute residence time at 50% fill.

Liquid-dominated drums (residence time)​

Used when there is no significant vapor outlet (surge and reflux drums):

V = max(Q̇_L · τ / f_fill, 0.1 m³)       τ = 5 min, f_fill = 0.5
D = ∛(4·V / (3π))                        L/D = 3

Columns (volume, diameter, stages)​

Distillation, absorber, and stripper shells are sized with the Fair flooding correlation (Towler & Sinnott eq. 17.34 fit):

u_flood  = (−0.171·lt² + 0.27·lt − 0.047) · √((ρ_L − ρ_V) / ρ_V)
u_design = 0.8 · u_flood                  design at 80% of flooding
A_net    = Q̇_V / u_design
A_column = A_net / 0.88                   ~12% downcomer allowance
D        = √(4·A_column / π)
H        = n_stages · lt · 1.15           lt = 0.6 m tray spacing
V        = π/4 · D² · H
  • Internal vapor traffic: when the reflux ratio is available (Rigorous columns), internal vapor ≈ overhead vapor × (1 + R). Otherwise the overhead flow is used and the approximation is recorded.
  • Stage count: rigorous engines report actual stages. When only a minimum stage count is available, the fallback N_actual ≈ 2 · N_min (Fenske factor) is used and flagged so you can override it.

The shell volume, diameter, and stage count feed the economics engine separately — the shell is costed as a vessel and the internals (trays or packing) as their own item. See Costing Methods.

Reactors (volume)​

Reactors size by residence time with phase-dependent defaults — covered in depth in Reactor Sizing.

Custom units (delegate)​

CustomUnit nodes inherit the sizing of their base type, or carry an agent-defined sizing formula. See Custom Sizing.