POME-MAP is a working, local-first methane-accountability prototype for palm oil mill effluent (POME) biogas systems. It compares expected methane from COD removal with measured dry-normal methane, follows persistent deviations through time, and presents transparent checks for biology, gas capture, metering and time alignment.
Project lead: Tan King Yang
POME biogas facilities may already record wastewater flow, COD, biogas volume and methane concentration, but those measurements are often reviewed in separate reports. POME-MAP brings them into one explainable evidence chain:
- a transparent COD-to-methane screening baseline;
- expected versus measured methane with uncertainty;
- a 60-day EWMA monitor and operating-event context;
- independent field-period validation with failed gates kept visible;
- a populated public-evidence workbook and reproducible online-data replay; and
- clear limitations that prevent a balance gap from being called a leak.
POME-MAP is decision support. It is not automatic control, regulatory reporting, carbon-credit verification or gas-safety instrumentation.
Windows PowerShell:
.\start_local.ps1The start script automatically uses either the Windows-native
.venv\Scripts\python.exe layout or the included .venv\bin\python.exe
layout.
Then open http://127.0.0.1:8765.
The application uses Python's standard-library web server and SQLite. It does not require a cloud account. The browser interface and calculations run locally.
The unchanged screening baseline was tested against one independent 42-day operating period from the UNFCCC Lamthap POME Biogas Project. Aggregate methane overprediction was 18.95%, inside the pre-declared +/-20% screening gate, with 88.96% paired-gas coverage.
That is a conditional screening result, not universal validation. The broader transferability check covered only two of four published ranges, and the diagnostic rules have not yet been tested against a sufficiently large set of ground-truth incidents. The numerical extract, source register, source hashes, frozen protocol, result report and populated workbook are included so the current result can be replayed without contacting a mill, requesting private data or obtaining a separate technical sign-off.
app.py- local web server and API routespome_map/- calculation, history, validation and CSV-audit logicstatic/- public browser interfacedata/- demonstration, validation and protocol datatests/- automated checksMETHODOLOGY.md- equations, assumptions and evaluation methodDATA_SOURCES.md- provenance and source boundariesVALIDATION_PROTOCOL.md- pre-declared validation rulesVALIDATION_REPORT.md- independent field-period resultoutputs/POME-MAP-Public-Evidence-Workbook.xlsx- populated, formula-driven evidence replayoutputs/POME-MAP-Public-Evidence-Protocol.md- public-source workflow and claim boundaryoutput/pdf/POME-MAP-2026-Submission.pdf- single upload-ready project documentoutputs/POME-MAP-Application-Form-Copy.md- prepared form textoutputs/POME-MAP-Submission-Checklist.md- final submission checklist
$Python = @(".venv\Scripts\python.exe", ".venv\bin\python.exe") |
Where-Object { Test-Path $_ } |
Select-Object -First 1
& $Python -m unittest discover -s tests -v
node --check static\app.js
node --check static\history.js
node --check static\validation.js
node --check static\project.js
node --check static\pilot.js