Production & Network Forecasting
From 24-hour gas network demand forecasting to 90-day subsurface production forecasts — MAVEN FORECAST combines weather-driven demand modelling, LP-based network optimisation, and IPR-based reservoir physics into a single browser-based forecasting platform.
Forecast With Confidence — From Reservoir to Customer
A unified forecasting suite that links subsurface physics, plant constraints, network hydraulics, and end-customer demand into a single decision-support tool.
Weather-Driven Demand Forecasting
Hourly 24-hour demand forecasts driven by temperature, wind chill, day-type, and seasonal profiles — calibrated against historical consumption for every customer node and aggregated across the network.
Subsurface IPR Modelling
Physics-based Inflow Performance Relationship curves with downhole pressure depletion modelling — predicting well productivity and reservoir behaviour across a rolling 90-day horizon.
90-Day Production Forecast
Daily production forecasts across multiple wells with plant minimum/maximum capacity constraints, automatic well allocation, and configurable shut-in / ramp-back logic.
Multi-Tier Network Optimisation
Linear programming solver across multi-pressure tier networks (NTS, HP, IP, MP, LP) — optimising injection points, balancing flows, and minimising line-pack drift over the forecast horizon.
Quality & Compliance
Continuous tracking of gas-quality parameters against GS(M)R and equivalent regulations — Wobbe Index, calorific value, hydrogen content, ICF, and trace species — with automatic constraint flagging.
Energy Nominations & Reporting
Automated daily and monthly nomination reports, work-order impact analysis, and exportable forecast vs. actual variance reports for operations, planning, and commercial teams.
Under the Hood
| Architecture | FastAPI backend + browser SPA, REST + SignalR streaming |
| Forecast Horizons | Hourly 24-hour (network) · Daily 90-day (production) |
| Optimisation | LP solver (PuLP/SciPy) with quality, capacity & line-pack constraints |
| Subsurface Model | IPR curves, downhole pressure depletion, well allocation |
| Demand Model | Weather-coupled regression, day-type profiles, customer base-load |
| Charting | Chart.js v4 with synchronised forecast/actual overlays |
| Database | SQLite (WAL mode) historian and configuration store |
| Compliance | GS(M)R UK gas-quality parameters, configurable for equivalent regimes |
| AI Assistant | mAvenIt natural language forecast queries & what-if scenarios |