Appendix: VLE and flash calculations

Session 1 · optional reference

Soorathep Kheawhom

30 September 2026

Optional reference material

Return to the 12-slide classroom deck.

Use these details when a question from your investigation needs them. They are outside the required 45–60 minute practice.

Choose one extension: compare liquid activity models, examine zero versus nonzero virial coefficients in Lab 09, or add the energy balance in Lab 07.

The equilibrium problem

At fixed temperature and pressure, a stable equilibrium minimizes total Gibbs energy subject to material conservation.

\mu_i^L=\mu_i^V\quad\Longleftrightarrow\quad \hat f_i^L=\hat f_i^V

A calculation also needs a phase model, composition constraints and a stability check.

Liquid and vapor descriptions

y_i\hat\phi_i P=x_i\gamma_i P_i^{sat}\phi_i^{sat}\Pi_i

  • Ideal vapor, ideal liquid: all correction factors are one.
  • Modified Raoult: retain liquid activity coefficients.
  • Gamma–phi: also retain vapor and reference-state corrections.

Lab 01 uses ideal vapor. Lab 09 adds second-virial vapor corrections.

Four boundary calculations

Specified Unknown boundary Composition supplied
T Bubble P Liquid x
T Dew P Vapor y
P Bubble T Liquid x
P Dew T Vapor y

A dew calculation generally needs an inner liquid-composition iteration because γ depends on x.

A check independent of the solver

0.5=(1-0.493506)(0.388889)+(0.493506)(0.614035)

The rounded result closes the component-1 balance. Component 2 must close as well.

Also check x_1+x_2=y_1+y_2=1 and equality of component fugacities. A small balance residual alone does not validate the physical model.

Activity models change the liquid

One-parameter Margules, two-parameter Margules and NRTL share the same equilibrium criterion but predict different γ(x).

Hold the synthetic component properties and T fixed. Change only the liquid model or one interaction parameter.

A nonconvex liquid needs an LLE/VLLE stability analysis; it cannot simply be accepted as a stable VLE flash.

Vapor corrections need a reference

\ln\hat\phi_i=\frac{P}{RT}\left(2\sum_j y_j B_{ij}-B_{mix}\right) B_{mix}=\sum_i\sum_j y_i y_j B_{ij}

Lab 09 also evaluates pure saturated-vapor φ consistently. With all Bᵢⱼ=0, the ideal-vapor limit must be recovered.

A model comparison has controlled inputs

Compare the same T, P, feed and liquid model. Change only the vapor description first.

Record differences in boundary pressures, phase compositions and β. Then examine the declared B(T) range and the truncation guard.

PR/SRK, dense vapor and virial adiabatic flash are outside these lab solvers.

Adiabatic flash as an extension

H_{feed}=(1-\beta)H^L(T,x)+\beta H^V(T,y)

At specified pressure and feed enthalpy, T is another unknown. A TP flash at the feed temperature does not impose this energy balance.

Lab 07 uses synthetic caloric data. Independent practice only if the core TP calculation is secure.

Sources and further reading

Module reference deck · Lab assumptions and sources

The worked examples use synthetic inputs. Each lab states its supported models and validity limits.