Bring your own binary VLE data

Your measurements.
Inspect, test, fit.

Paste a spreadsheet table or load a CSV. Inspect inferred activities, compare applicable consistency diagnostics, and fit NRTL or Margules to the same observations.

Start here: read one row before fitting

Week 2 clarifications · Printable Lab 08 guide · Excel-to-Lab demonstration · Lab 08 video walkthrough · Example Excel workbook

Load the Session 2 example. Select Load synthetic example. It uses 350 K = 76.85 °C, measured P in kPa, and Antoine equations that also use K. Enter 350 directly in these equations. The worksheet Section C exercise at 50 °C uses different constants; do not copy them here.
Read the inputs and units. Each row is one equilibrium observation: T is temperature, P is total pressure, x1 is the liquid mole fraction of component 1, and y1 is its vapor mole fraction. Check the first-row preview below. An isothermal dataset keeps T fixed while P and compositions vary.
Validate and check one row. Select Validate & inspect data. Find the same row in Imported observations. Compare your hand-calculated γ₁ with exp(table ln γ₁), not directly with ln γ₁. Save the baseline study JSON.
Fit and interpret. Fit Margules and NRTL to the same observations and properties. Record both error metrics, a residual pattern and one limitation. Save each study and residuals before switching models.

Describe the experiment

Component 1 · x₁ and y₁ refer to this substanceComponent 2Measurement and vapor-pressure sources Experimental pathMeasured T column unit (input table)Measured P column unit (input table)

The unit selectors describe the numbers in your input table. Changing a selector reinterprets those numbers; it does not rewrite them. The results table always displays K and kPa.

Saturation pressures are calculated separately using the Antoine constants below, at each row’s measured T. HE, when supplied, must be independent excess-enthalpy data in J/mol.

Pure-component vapor pressures

Choose the exact equation printed by your source for each component. Constants are not automatically converted when changing the form. Extended Antoine and Wagner correlations are not supported. The correlation units can differ from the measured-data units. Valid-range limits below are always in kelvin.

Component 1 · AntoineEquation form

ABCMinimum valid T / KMaximum valid T / KComponent 1: T unit required by its Antoine constantsPsat unit from the Antoine equation

Component 2 · AntoineEquation form

ABCMinimum valid T / KMaximum valid T / KComponent 2: T unit required by its Antoine constantsPsat unit from the Antoine equation

Load CSV or tab-separated text

This template contains T = 350 K and P in kPa; its headers remain T,P,x1,y1. Loading a CSV does not set units, component names or Antoine constants. For this unchanged template, first use Load synthetic example to set its matching properties. For your own data, supply the units and property constants from your sources.

Or paste the table, including headers

Required headers: T,P,x1,y1. Do not supply Psat columns. Optional: HE. Comma, semicolon or tab delimiters; dot decimals. At least 5 distinct interior points; up to 500 rows / 1 MB. Templates and examples are synthetic, not measurements.

Ready for your data.

First-row unit check

Load the example or enter a table to see how its first row will be interpreted.

This preview checks how units are interpreted. Validate & inspect data still checks all rows and both property correlations.

Start with the data

Enter sources, select units, then paste or load a table. Pure endpoints remain visible but cannot specify an absent component's activity coefficient.

Imported observations Observed inputs and calculated quantities; normalized to K and kPa

Observed inputs: T, P, x₁ and y₁ come from the input table. Calculated quantities: Psat1 and Psat2 come from the Antoine correlations; ln γ₁ and ln γ₂ are inferred using the stated vapor assumptions. HE is optional independent input. The original table and its units remain in the study JSON.

For the synthetic example, the first data row (input row 2) means T = 350 K, P ≈ 95.1231 kPa, x₁ = 0.05 and y₁ ≈ 0.397542. It is a constructed equilibrium observation, not a real measurement.

What these tests can establish Scope and interpretation

Area and Herington calculations use the same teaching implementations as Lab 03. Without an explicit endpoint assumption, only the measured interval is integrated and no full-area criterion is assessed. For isothermal data, pressure/excess-volume terms are neglected. Herington's temperature-range term is empirical and is not a measured HE correction.

The differential diagnostic uses three-point derivatives on the actual composition spacing. It is sensitive to noise. For isobaric data, a verdict is withheld unless independent HE is supplied at every usable point. Supplying HE does not resolve excess-volume or vapor-fugacity approximations. No HE is inferred from the fitted model to certify its training data.

Model reproduction reports mean absolute pressure error (%) and mean absolute vapor-composition error in percentage points. The illustrative thresholds are each less than 1. Fits at parameter bounds are flagged. A fit may be poor because of model choice, property errors or measurement errors; it does not identify the cause. Fitted models are not automatically validated for extrapolation, stable flash prediction, or real-fluid process design.

Replicates at the same x₁ must be reviewed outside this importer. No silent averaging, endpoint filling, row deletion, outlier removal or unit guessing is performed. Rows outside either Antoine range block import; correlations are never extrapolated. Reimport after any data, unit or source change.

Calculation source · Regenerate study figure · NIST area/Herington guidance · NIST modeling-test guidance.