A working oil & gas SCADA demo: simulated gas wellpad → MQTT → Ignition tags → operator dashboard → PostgreSQL historian. Built around the patterns Houston O&G integrators actually deploy.
Most SCADA portfolio projects are generic IoT demos with the word "industrial" sprinkled in. This one targets the domain: a multi-well natural gas pad with realistic decline curves, tubing/casing pressure dynamics, condensate tank fills, and a fault-injection harness so you can demo alarm/recovery in 30 seconds.
┌──────────────────┐ MQTT ┌──────────────────┐
│ Wellpad simulator│ ─────────────▶│ Mosquitto │
│ (Python, 5 wells)│ 1 Hz JSON │ broker │
└──────────────────┘ └────────┬─────────┘
│
▼
┌─────────────────────────────────────────────────────┐
│ Ignition Gateway │
│ ┌────────────┐ ┌────────────┐ ┌──────────────┐ │
│ │ MQTT Engine│─▶│ Tag Provider│─▶│ Perspective │ │
│ │ (Cirrus) │ │ (UDT-based)│ │ Dashboard │ │
│ └────────────┘ └─────┬──────┘ └──────────────┘ │
│ │ │
│ ┌─────▼──────┐ ┌──────────────┐ │
│ │ Jython │ │ Tag Historian│ │
│ │ (totalizer)│ │ │ │
│ └────────────┘ └──────┬───────┘ │
└─────────────────────────────────────────┼──────────┘
▼
┌─────────────┐
│ PostgreSQL │
└─────────────┘
| Path | Purpose |
|---|---|
simulator/ |
Python wellpad telemetry simulator with fault injection |
docker/ |
Docker Compose for Mosquitto MQTT broker + PostgreSQL historian |
ignition/scripts/ |
Jython scripts to paste into Ignition Designer (flow totalizer, anomaly detector) |
ignition/udts/ |
UDT structure spec for the reusable Well type |
docs/ |
Architecture, setup, demo script |
docker compose -f docker/docker-compose.yml up -d— broker + Postgres uppython simulator/wellpad_simulator.py simulator/wellpad_config.yaml— 5 wells start publishing- Open Ignition Designer → tags populate under
[default]Wellpad/SITE01/... - Open the Perspective dashboard — pressures, flows, tank levels live
python simulator/inject_fault.py WELL_03 pressure_spike 30— alarm fires, dashboard turns red, recovers after 30s
- MQTT data ingestion — broker config, topic hierarchy, QoS choices, retained messages
- Ignition platform — UDTs, tag bindings, Perspective views, alarm pipelines, historian config
- Jython scripting in Ignition context — tag change events, gateway timers, history queries, alarm logic
- SCADA standards / reusability — UDT-based well template means scaling from 5 wells to 500 is a parameter change, not a redesign
- Troubleshooting harness — fault injection lets you demonstrate root-cause workflows without waiting for real failures
The simulator implements:
- Decline curves on tubing pressure (slow drift down over time)
- Pressure ↔ flow correlation (drops in pressure reduce flow)
- Diurnal temperature wobble
- Condensate tank fills proportional to flow, with truck-pickup auto-drain at 95%
- Four fault scenarios:
pressure_spike,comm_dropout,tank_overflow,sensor_drift
The numbers are not calibrated against any specific reservoir — they're plausible, not predictive.
- No Sparkplug B encoding — plain JSON over MQTT for setup speed. The topic hierarchy is Sparkplug-friendly, so swapping in a Sparkplug B publisher (
tahulibrary) is mechanical. Seedocs/architecture.md. - No notification routing (email/Slack). Ignition's built-in alarm pipeline is enough for the demo. Real deployments wire to Twilio, PagerDuty, etc.
- No PLC simulation. The MQTT layer represents the "field already publishes via gateway" case, which is how most modern O&G deployments are architected.
See docs/setup.md. TL;DR: install Ignition Maker Edition, install Cirrus Link MQTT Engine module (free, Maker-compatible), docker compose up, run the simulator, follow the Designer steps.
MIT.