From a single monitor to fleet visibility
An online dissolved gas analysis (DGA) monitor’s value is not the parts-per-million reading on its display — it is whether that reading reaches the people and systems that act on it, in time, in the right context. This is where remote monitoring and the industrial internet of things (IIoT) come in: moving DGA data from the transformer plinth to the SCADA screen, the asset-performance platform, and the cloud, without a human visiting the site.
The three-layer architecture
The whitepaper structures online DGA systems in three layers:
| Layer | Location | Function |
|---|---|---|
| Sensing | Beside the transformer | Degassing, detection, local computation, alarms |
| Station control | Station control room | Aggregation of devices, station-level monitoring, uplink transmission |
| Management | Group / cloud | Cross-station aggregation, AI diagnostics, asset management |
The layers connect directly or through an edge gateway that performs protocol conversion and data pre-processing, accommodating both traditional and digital substations.
Integration: SCADA, APM, and cloud
Integration is not wiring; it is data semantics. The whitepaper is explicit that the decisive factor is master-data consistency — equipment ledgers, channel point tables, range and unit conversion, and time synchronization (NTP/SNTP) must be managed uniformly, or you get “protocols connected but data mismatched.”
- SCADA / station monitoring — gas concentrations, moisture, and device status map to telemetry and telesignal points via MODBUS, IEC 60870-5-104, or DNP3.0 point tables. 4-20 mA analog upload carries few channels and no diagnostics, and is progressively being replaced by digital protocols.
- Asset performance platforms (APM) — platforms such as Rugged Monitoring’s RM EYE accept MODBUS, MQTT, and IEC 61850 at the platform layer, with the edge gateway supporting DNP3.0, IEC 60870-5-104, and IEC 61850. DGA data is aggregated with load, temperature, and partial discharge to support health-index and maintenance decisions.
- Cloud / IIoT — data uplinks to a group cloud via MQTT or HTTPS, enabling centralized monitoring of many plants and mobile push.
In a digital substation, the monitor connects as an IEC 61850 IED at the station-control level, so its data is uniformly readable by station monitoring and fault-recording systems.
Edge vs. cloud: where the intelligence lives
Intelligence is split deliberately between layers. The edge layer — the on-site unit and the station — handles real-time tasks: built-in diagnostic libraries (Duval triangle and ratio methods), multi-level alarms, trend storage, and store-and-forward of interrupted data. It remains fully autonomous during a network outage.
The cloud layer handles batch tasks: cross-site aggregation, fleet-wide comparison, heavy-model inference such as machine-learning fusion and remaining useful life (RUL), plus visualization and reporting. The design principle: the higher the real-time requirement, the closer to the edge; the larger the data volume, the more it moves to the cloud.
Remote O&M for unmanned sites
For unmanned and remote sites — wind and solar step-up substations, distribution vaults, industrial plants without a resident electrical team — remote operation and maintenance (O&M) is the point. The recommended cadence is a remote inspection every 1–3 months covering data completeness, alarm records, and self-diagnostic results, achieving “unattended operation with remote inspection.” A laser photoacoustic spectroscopy (L-PAS) system supports this because it has no carrier gas and no routine consumables: the maintenance event that forces a site visit simply does not exist.
Outdoor-rated hardware is a prerequisite: an IP55 enclosure, a −40 to +55 °C operating range, and modular spares (laser source, photoacoustic cell, microphone) that can be swapped as a whole unit on site when a visit does happen.
The MaaS option
Where the user does not want to run the platform or the O&M at all, monitoring-as-a-service (MaaS) packages the analyzer, maintenance, and diagnostics into a subscription. The service provider handles calibration, spare parts, and diagnostic support; the user pays per device. Forecasts project MaaS penetration in transformer monitoring to rise from about 8–10% in 2026 to about 20–25% in 2035, and it is especially attractive to users with weak maintenance capability or only a few monitors at a site.
Data governance and time synchronization
Remote monitoring multiplies the data flow, which makes governance matter. The whitepaper’s guidance: measurement data needs integrity validation, unified timestamps, and traceable calibration records; abnormal data — flushing-transition segments, device self-test anomalies — is flagged rather than deleted, so the trend database is not polluted. Cybersecurity follows the IEC 62443 defense-in-depth approach: zone isolation, least privilege, TLS on remote access, firmware signature verification, and logging and auditing.
PAS DGA for remote and unmanned monitoring
PAS DGA’s L-PAS monitors are built for remote operation: no carrier gas, no consumables, low routine maintenance, and protocol support for MODBUS, IEC 60870-5-104, IEC 61850, DNP3.0, and MQTT. The DGA-900 delivers 9 gases plus moisture; the DGA-500 provides field-standard hydrogen monitoring where screening suffices. For the physics behind the low-maintenance design, see how L-PAS works and the fourth-generation overview in laser photoacoustic spectroscopy.
Whether you integrate with your own SCADA and APM or take a MaaS subscription, the data model and time sync come ready. Contact PAS DGA to plan a remote-monitoring rollout for your fleet.