{ "cells": [ { "cell_type": "markdown", "id": "cell-00", "metadata": {}, "source": [ "# Propane parameter estimation\n", "\n", "\n", "**Evidence: calibration.** Fit three PC-SAFT pure-fluid parameters to six\n", "observations, saturation pressure and saturated liquid mass density at\n", "210, 270 and 330 K. Start from `gross-2001-propane/3`, Gross and Sadowski (2001),\n", "Table 2 ([paper](https://doi.org/10.1021/ie0003887)).\n", "\n", "The observations are Glos, Kleinrahm and Wagner (2004), Table 2\n", "([paper](https://doi.org/10.1016/j.jct.2004.07.017)), retained as\n", "`engine/tests/cases/pure-saturation-regression/inputs/source/propane-baseline.csv`.\n", "Zotero `NF5HJCV7` is metadata only; no local PDF attachment is claimed.\n", "The [official NIST/TRC ThermoML record](https://trc.nist.gov/ThermoML/10.1016/j.jct.2004.07.017.html)\n", "provides the machine-readable measurements. The retained source record\n", "`engine/tests/cases/2004-glos/inputs/source/glos-2004-propane-saturation.yaml`\n", "documents the compact factual extraction and source-specific use basis.\n", "Source pressures in kPa were multiplied by exactly 1000 to obtain the Pa\n", "values read from the teaching CSV; mass densities remain in kg/m³.\n" ] }, { "cell_type": "markdown", "id": "cell-01", "metadata": {}, "source": [ "## Read the inputs\n", "\n", "Use a repository checkout with the environment described in the installation\n", "guide. The wheel alone does not contain these data. All density, association,\n", "equilibrium and parameter-fit iterations below run in the native Engine.\n" ] }, { "cell_type": "code", "execution_count": 1, "id": "cell-02", "metadata": { "execution": { "iopub.execute_input": "2026-10-20T20:17:35.608951Z", "iopub.status.busy": "2026-10-20T20:17:35.608776Z", "iopub.status.idle": "2026-10-20T20:17:35.878529Z", "shell.execute_reply": "2026-10-20T20:17:35.877949Z" } }, "outputs": [], "source": [ "from pathlib import Path\n", "import csv\n", "import pandas as pd\n", "from IPython.display import display\n", "from epcsaft import (\n", " Parameters,\n", " Pure,\n", ")\n", "from epcsaft import equilibrium as eq, regression as rg\n", "\n", "pd.set_option(\"display.max_columns\", None)\n", "pd.set_option(\"display.max_colwidth\", None)\n", "\n", "ROOT = next(\n", " folder\n", " for folder in (Path.cwd(), *Path.cwd().parents)\n", " if (folder / \"ARCHITECTURE.yaml\").is_file()\n", ")\n", "PACKETS = ROOT / \"data\" / \"packets\"\n" ] }, { "cell_type": "markdown", "id": "cell-03", "metadata": {}, "source": [ "## Declare observations, coordinates and one native fit\n", "\n", "Each observation uses a zero-vapor-amount saturation declaration. Log-ratio\n", "residuals have unit weight, a calibration choice rather than source measurement\n", "uncertainty. Coordinates use origin/scale/bounds:\n", "\n", "| Parameter | Origin | Scale | Bounds |\n", "|---|---:|---:|---:|\n", "| Dispersion energy/k / K | 221 | 10 | 150–260 |\n", "| Segment count / 1 | 1.9489 | 0.1 | 1.5–2.5 |\n", "| Segment diameter / Å | 3.7209 | 0.1 | 3–4.2 |\n", "\n", "The native budget is 100 iterations and 60 s, with one published-packet start.\n", "Require usable parameters, lower cost, and independent native re-evaluation\n", "of all fitted predictions within 10⁻¹⁰ relative.\n" ] }, { "cell_type": "code", "execution_count": 2, "id": "cell-04", "metadata": { "execution": { "iopub.execute_input": "2026-10-20T20:17:35.880188Z", "iopub.status.busy": "2026-10-20T20:17:35.880024Z", "iopub.status.idle": "2026-10-20T20:17:35.954531Z", "shell.execute_reply": "2026-10-20T20:17:35.954000Z" } }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
ObservationTemperature / KObservedFitted predictionLog-ratio residual / 1
0Pressure / Pa21036032.0003.602740e+04-0.000128
1Liquid density / kg m^-3210604.5946.033307e+02-0.002092
2Pressure / Pa270430425.0004.305303e+050.000245
3Liquid density / kg m^-3270532.8535.337605e+020.001702
4Pressure / Pa3301983000.0001.983084e+060.000042
5Liquid density / kg m^-3330434.8694.349694e+020.000231
\n", "
" ], "text/plain": [ " Observation Temperature / K Observed Fitted prediction \\\n", "0 Pressure / Pa 210 36032.000 3.602740e+04 \n", "1 Liquid density / kg m^-3 210 604.594 6.033307e+02 \n", "2 Pressure / Pa 270 430425.000 4.305303e+05 \n", "3 Liquid density / kg m^-3 270 532.853 5.337605e+02 \n", "4 Pressure / Pa 330 1983000.000 1.983084e+06 \n", "5 Liquid density / kg m^-3 330 434.869 4.349694e+02 \n", "\n", " Log-ratio residual / 1 \n", "0 -0.000128 \n", "1 -0.002092 \n", "2 0.000245 \n", "3 0.001702 \n", "4 0.000042 \n", "5 0.000231 " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "start = Parameters.from_json(\n", " PACKETS / \"gross-2001-propane/3/parameters/parameters.json\", components=[\"propane\"]\n", ")\n", "rows = list(\n", " csv.DictReader(\n", " (\n", " ROOT\n", " / \"engine/tests/cases/pure-saturation-regression/inputs/source/propane-baseline.csv\"\n", " ).open()\n", " )\n", ")\n", "observations = []\n", "for row in rows:\n", " problem = eq.Problem(\n", " T=float(row[\"temperature_k\"]),\n", " P=eq.Free(float(row[\"saturation_pressure\"])),\n", " feed=eq.Amounts({\"propane\": 1}),\n", " phases=[\n", " eq.Phase(\"L\", kind=\"liquid\"),\n", " eq.Phase(\"V\", kind=\"vapor\", amount=eq.Pinned(0)),\n", " ],\n", " row_tolerance=1e-10,\n", " )\n", " observations += [\n", " rg.saturation_pressure_observation(\n", " start,\n", " component=\"propane\",\n", " T=float(row[\"temperature_k\"]),\n", " observed=float(row[\"saturation_pressure\"]),\n", " form=\"log_ratio\",\n", " ),\n", " rg.observation(\n", " start,\n", " \"phase_density\",\n", " problem,\n", " observed=float(row[\"saturated_liquid_mass_density\"]),\n", " phase=\"L\",\n", " form=\"log_ratio\",\n", " mass_basis=True,\n", " ),\n", " ]\n", "coordinates = [\n", " rg.coordinate(\n", " start, [(name, \"propane\", 1)], origin=origin, scale=scale, bounds=bounds\n", " )\n", " for name, origin, scale, bounds in [\n", " (\"dispersion_energy_over_k\", 221, 10, (150, 260)),\n", " (\"segment_count\", 1.9489, 0.1, (1.5, 2.5)),\n", " (\"segment_diameter\", 3.7209, 0.1, (3, 4.2)),\n", " ]\n", "]\n", "fit_result = rg.fit(\n", " start,\n", " coordinates,\n", " observations,\n", " controls=rg.FitControls(maximum_iterations=100, maximum_elapsed_time_seconds=60),\n", ")\n", "assert fit_result.usable, fit_result.message\n", "assert fit_result.final_cost < fit_result.initial_cost\n", "fitted_parameters = rg.fitted_parameters(start, fit_result)\n", "predictions = [\n", " rg.predict(Pure(fitted_parameters), observation).value\n", " for observation in observations\n", "]\n", "assert all(\n", " (\n", " abs(a - b) < 1e-10 * max(abs(a), 1)\n", " for a, b in zip(predictions, fit_result.predictions)\n", " )\n", ")\n", "values = dict(\n", " physical=fit_result.physical,\n", " initial_cost=fit_result.initial_cost,\n", " final_cost=fit_result.final_cost,\n", " iterations=fit_result.iterations,\n", " predictions=fit_result.predictions,\n", " residuals=fit_result.residuals,\n", " covariance_rank=fit_result.covariance.rank,\n", " covariance_assumption=fit_result.covariance.assumption,\n", " observed=[o.observed for o in observations],\n", ")\n", "values[\"prediction_round_trip_relative_error\"] = max(\n", " (abs(a - b) / abs(a) for a, b in zip(predictions, fit_result.predictions))\n", ")\n", "assert values[\"prediction_round_trip_relative_error\"] <= 1e-10\n", "display(\n", " pd.DataFrame(\n", " {\n", " \"Observation\": [\"Pressure / Pa\", \"Liquid density / kg m^-3\"] * 3,\n", " \"Temperature / K\": [210, 210, 270, 270, 330, 330],\n", " \"Observed\": values[\"observed\"],\n", " \"Fitted prediction\": values[\"predictions\"],\n", " \"Log-ratio residual / 1\": values[\"residuals\"],\n", " }\n", " )\n", ")\n" ] }, { "cell_type": "markdown", "id": "cell-05", "metadata": {}, "source": [ "## Inspect calibration residuals\n", "\n", "Rows alternate saturation pressure and liquid density at the three temperatures.\n", "These are residuals against the same measurements used in the fit.\n" ] }, { "cell_type": "code", "execution_count": 3, "id": "cell-06", "metadata": { "execution": { "iopub.execute_input": "2026-10-20T20:17:35.955783Z", "iopub.status.busy": "2026-10-20T20:17:35.955659Z", "iopub.status.idle": "2026-10-20T20:17:36.276114Z", "shell.execute_reply": "2026-10-20T20:17:36.275778Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from io import BytesIO\n", "from IPython.display import Image\n", "import matplotlib.pyplot as plt\n", "\n", "plt.plot(range(6), values[\"residuals\"], marker=\"o\")\n", "plt.axhline(0, color=\"black\")\n", "plt.xlabel(\"Declared observation row\")\n", "plt.ylabel(\"Calibration log-ratio residual / 1\")\n", "buffer = BytesIO()\n", "plt.tight_layout()\n", "plt.savefig(buffer, format=\"png\")\n", "plt.close()\n", "display(Image(data=buffer.getvalue()))\n" ] }, { "cell_type": "code", "execution_count": 4, "id": "cell-08", "metadata": { "execution": { "iopub.execute_input": "2026-10-20T20:17:36.277239Z", "iopub.status.busy": "2026-10-20T20:17:36.277092Z", "iopub.status.idle": "2026-10-20T20:17:36.279413Z", "shell.execute_reply": "2026-10-20T20:17:36.278988Z" } }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Dispersion energy/k / KSegment count / 1Segment diameter / Å
0208.5112491.9964563.619429
\n", "
" ], "text/plain": [ " Dispersion energy/k / K Segment count / 1 Segment diameter / Å\n", "0 208.511249 1.996456 3.619429" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Initial cost / 1Final cost / 1Iterations / 1Covariance rank / 1Covariance assumptionPrediction round-trip relative error / 1
00.0000140.00000443unit-variance weighted residuals, independent errors and a regular local linearization at an interior point; multiply by variance_factor for an unknown common variance; no claim under bounds, rank deficiency or model discrepancy2.625431e-15
\n", "
" ], "text/plain": [ " Initial cost / 1 Final cost / 1 Iterations / 1 Covariance rank / 1 \\\n", "0 0.000014 0.000004 4 3 \n", "\n", " Covariance assumption \\\n", "0 unit-variance weighted residuals, independent errors and a regular local linearization at an interior point; multiply by variance_factor for an unknown common variance; no claim under bounds, rank deficiency or model discrepancy \n", "\n", " Prediction round-trip relative error / 1 \n", "0 2.625431e-15 " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "display(pd.DataFrame({\n", " \"Dispersion energy/k / K\": [values[\"physical\"][0]],\n", " \"Segment count / 1\": [values[\"physical\"][1]],\n", " \"Segment diameter / Å\": [values[\"physical\"][2]],\n", "}))\n", "display(pd.DataFrame([{\n", " \"Initial cost / 1\": values[\"initial_cost\"],\n", " \"Final cost / 1\": values[\"final_cost\"],\n", " \"Iterations / 1\": values[\"iterations\"],\n", " \"Covariance rank / 1\": values[\"covariance_rank\"],\n", " \"Covariance assumption\": values[\"covariance_assumption\"],\n", " \"Prediction round-trip relative error / 1\": values[\"prediction_round_trip_relative_error\"],\n", "}]))\n" ] }, { "cell_type": "markdown", "id": "cell-07", "metadata": {}, "source": [ "## Meaning and limits\n", "\n", "\n", "This is calibration against direct measurements. It supplies no independent\n", "validation set or new predictive accuracy floor. The printed covariance rank\n", "and assumption describe a local linearized calculation conditional on the\n", "declared weights and residual model. They do not establish global\n", "identifiability or predictive uncertainty. The fitted values are a tutorial\n", "result, not a newly adopted canonical parameter set.\n" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.13.14" } }, "nbformat": 4, "nbformat_minor": 5 }