A decade of transition
Online transformer monitoring is entering a decade of structural change. The direction is clear from the technology roadmap: fuller gas coverage, less maintenance, faster measurement, smarter diagnosis, deeper integration, and new business models. The dissolved gas analysis (DGA) market is being reshaped by all six of these trends at once.
Technology trends: coverage, maintenance, speed
Trend 1: full-component coverage — 9 gases plus moisture
The market is moving from early single-gas hydrogen detection toward a full-component configuration of 9 gases plus moisture: H2, CH4, C2H6, C2H4, C2H2, CO, CO2, O2, and N2 plus moisture. These nine gases cover every fault-type indicator defined in IEC 60599:2022, while O2 and N2 reveal sealing and degassing status and moisture reflects insulation dampness. A single deployment that captures a relatively complete dissolved-gas profile is a common direction of recent product definitions.
Trend 2: maintenance-free, consumable-free operation
Optical detection requires no carrier gas, no chromatography column, and no periodic calibration-gas replacement, cutting whole-life-cycle operation and maintenance cost and on-site visit frequency. For remote substations with difficult transportation, this is directly tied to whether the system can operate reliably over the long term.
Trend 3: faster cycles, toward near real time
Measurement cycles are advancing from the hour scale to the minute scale. Laser photoacoustic spectroscopy and quartz-enhanced photoacoustic spectroscopy (QEPAS) have compressed cycles to the minute scale in the laboratory (research). A shorter cycle means a smaller response window for sudden faults and greater early-warning value.
Intelligence and business-model trends
Trend 4: AI-based diagnosis and the digital twin
Machine learning that fuses IEC/IEEE traditional criteria with historical operating data improves diagnostic consistency and interpretation efficiency. Digital twin and remaining useful life (RUL) prediction have become key topics, and predictive maintenance is widely regarded as able to significantly reduce unplanned outages. The accuracy improvement of AI diagnosis is a conclusion of laboratory comparison studies (research comparison) — engineering deployment still needs traditional criteria and expert review.
Trend 5: digital substation integration and protocol interoperability
IEC 61850 (MMS/GOOSE), IEC 60870-5-104, DNP3.0, and MODBUS are becoming standard features of online devices. Devices connect to the station level as intelligent electronic devices (IEDs) and interoperate with asset performance management platforms.
Trend 6: monitoring-as-a-service (MaaS)
A model that replaces one-time procurement with a subscription has begun to emerge, with market penetration of about 8–10% in 2026, projected to reach 20–25% by 2035 (market report). MaaS lowers the initial-investment threshold and transfers metrological traceability, maintenance, and calibration responsibility to the service provider — suited to dispersed monitoring points and limited in-house maintenance capability.
The six trends at a glance
| Trend | Key driver | Value to users |
|---|---|---|
| 9 gases + moisture | Full fault-type coverage | Complete condition awareness in one deployment |
| Maintenance-free / consumable-free | Optical detection proliferation | Lower whole-life-cycle cost |
| Fast measurement cycle | Early-warning needs | Earlier detection, faster response |
| AI diagnosis and digital twin | Data accumulation and edge computing | Automation from data to decision |
| Digital integration and interoperability | Digital substation construction | Seamless integration into existing systems |
| MaaS model | Shift from capital to operating expenditure | Lower initial investment threshold |
The 5–10 year outlook
Over the next 5–10 years, the technology direction is toward lower detection limits and smaller form factors. Mid-infrared laser sources (quantum cascade lasers, QCL; interband cascade lasers, ICL) increase absorption cross-section by roughly 18× relative to near-infrared, expected to lower detection limits further; C2H2 detection-limit research has progressed sequentially from 1.4 ppm to 0.5 ppm and 0.2 ppm, narrowing the gap between research and commercial performance. QEPAS is maturing and can be expected to support smaller products with faster cycles (research).
On the intelligence side, AI diagnostics will move from laboratory research to engineering deployment, with explainability, labeled data, and field validation as prerequisites. Digital twin and RUL will push online DGA from “concentration monitoring” toward predictive maintenance, reducing unplanned outages and over-maintenance.
A practical roadmap: pilot, evaluate, scale-up
For asset managers, the recommended path is three stages: pilot → evaluate → scale-up.
| Stage | Timeframe | Key actions |
|---|---|---|
| Pilot | 0–6 months | Select 1–3 critical units; run side-by-side comparison with off-line laboratory GC over 2–3 sampling cycles |
| Evaluate | 6–12 months | Total cost of ownership assessment; quantify consumable savings and avoided outage losses |
| Scale-up | 12 months+ | Expand using the selection framework; connect via IEC 61850 / IEC 60870-5-104; establish calibration and data governance |
Pilot in parallel with lab verification, evaluate on total cost of ownership, then scale what survives the evidence. For selection guidance, see how to select an online DGA monitor. For the diagnostic side, see our multi-method DGA diagnosis workflow.
PAS DGA for the next decade of monitoring
The PAS DGA-900 aligns with the six trends today: 9 gases plus moisture, no consumables, continuous online measurement, and IED-grade protocol integration (MODBUS, IEC 61850, IEC 60870-5-104, DNP3.0). Whether you are piloting on a few critical units or scaling across a fleet, it is designed to support the journey. Contact PAS DGA to discuss a pilot program.