Smart Lightning Protection

Configuring Transformer Online Monitoring: Oil Temperature, DGA, and PD in Tiers

Direct Answer A substation retrofit wants transformer online monitoring — is an online DGA (dissolvedgas analy

2026-09-19 Smart Lightning Protection 微物联 5 min
Configuring Transformer Online Monitoring: Oil Temperature, DGA, and PD in Tiers

Direct Answer

**A substation retrofit wants transformer online monitoring — is an online DGA (dissolved-gas analysis) device worth it? The answer: tier by voltage class and importance — distribution transformers first get oil-temperature, oil-level, and load-current monitoring; important power transformers add bushing and core-ground-current monitoring; HV large-capacity transformers then consider online DGA. Online DGA suits transformers with high failure cost; routine offline testing is usually enough for ordinary distribution transformers.**

At a campus substation with two 110 kV power transformers and eight 10 kV distribution transformers, the retrofit budget was limited and the vendor suggested "online DGA on everything" — but one 110 kV online DGA unit costs about 5–10× the whole online-monitoring package of a distribution transformer, so blanket deployment made no sense. Tiered configuration: the eight 10 kV units got **oil temperature + oil level + load current** (covers common faults, low cost, best value); the two 110 kV units (supply hubs, high fault cost) added **bushing/core-ground-current monitoring** (early insulation defects); online DGA went on only one — the transformer with the worst failure consequences — while the other used offline testing + online oil-temperature linkage. Over two years, oil temperature flagged one distribution transformer overloaded above 90% and ground current found slight bushing-insulation degradation — the tiered approach covered the main risks at reasonable cost without burdening distribution units with expensive DGA.

Because:

- **Monitoring items must match transformer class**: common distribution faults (overload, high oil temperature) are covered by oil temperature/level/load current; insulation defects on power transformers need deeper means; - **Online DGA vs offline testing is an arithmetic problem**: online DGA is expensive; small transformers with low fault cost are usually served well by offline testing (periodic sampling); - **Cost magnitude sets the pace**: oil temperature/level/load current are cheapest and go everywhere; bushing/ground current are medium and go on power transformers; online DGA is the most expensive and goes only on the highest-fault-cost transformers; - **Data upload and alarm linkage share one platform**: all tier data finally enters the Digital Electrical Intelligent Supervision System with unified trends, alerts, and health.

Key insight: **transformer online monitoring is tiered by class, not "the more the better."** Distribution units get basic monitoring first; power transformers go deeper — online DGA only for the highest-fault-cost transformers.

At an airport substation with two 35 kV power transformers, the vendor quoted online DGA and O&M hesitated. The arithmetic: one supply interruption affects flights, so fault cost is high — online DGA was worth it; but only on one unit (the core transformer carrying the station load), the other used offline testing + online oil-temperature linkage. Later, online DGA found acetylene creeping 3 months early (early internal discharge); a planned outage prevented fault escalation. **The value of online DGA = fault cost ÷ price — worthwhile only on high-fault-cost transformers.**

1. Monitoring Items vs Transformer Class

First clarify what each transformer class should get:

**Distribution transformers (10 kV/0.4 kV)**: - Common faults: overload, high oil temperature, low oil level, hot joints; - Monitoring: oil temperature, oil level, load current, low-side joint temperature; - Purpose: cover common faults at the lowest cost.

**Important power transformers (35/110 kV, hub units)**: - Common faults: bushing insulation defects, abnormal core/clamp ground current, internal discharge; - Monitoring: oil temperature + oil level + load current + bushing dielectric/PD + core ground current; - Purpose: cover early insulation defects; high fault cost justifies going deeper.

**HV large-capacity transformers (220 kV and above, critical hubs)**: - Common faults: internal discharge (DGA gases), winding deformation, insulation aging; - Monitoring: all of the above + online DGA + winding temperature (fiber or thermal model); - Purpose: early warning of internal faults; extremely high fault cost justifies the most expensive means.

**Class-matching logic**: - Distribution: basic monitoring (oil temperature/level/load); - Power transformer: + insulation monitoring (bushing/core ground); - Large capacity: + online DGA (internal discharge).

Key insight: **transformer class decides monitoring depth.** A distribution transformer does not need online DGA, just as a small motor does not need the most expensive protection — match the class, do not waste.

2. Online DGA vs Offline Testing

DGA is the "most expensive and most sensitive" transformer tool — but do the arithmetic:

**Online DGA device**: - Pros: real-time continuous monitoring; characteristic-gas trends (H₂, CH₄, C₂H₂, C₂H₄, CO, CO₂) never break; catches sudden discharge; - Cons: high price (one unit equals several distribution-monitoring packages), needs maintenance, possible false alarms requiring re-test; - Applies: high-fault-cost power transformers (hubs, airports, data centers, hospitals).

**Offline testing (periodic oil samples to the lab)**: - Pros: low cost (per sample), high lab accuracy, mature standards; - Cons: periodic sampling — fast-developing discharge can be missed between samples; - Applies: ordinary distribution transformers, low-fault-cost equipment.

**Trade-off criteria**: - High fault cost (one interruption is costly) + high voltage class → online; - Low fault cost + distribution class → offline testing + online oil-temperature linkage; - Compromise: power transformer online oil temperature + periodic offline DGA + denser sampling when oil temperature is abnormal.

**Engineering intuition**: - Online DGA is not a "standard item"; it is a "config for important power transformers"; - Offline testing is not "backward"; it is the "reasonable choice for distribution transformers"; - Compromise: oil temperature online (real time) + DGA offline (periodic) — cost and coverage balanced.

Key insight: **online vs offline is the arithmetic of fault cost ÷ price.** High cost goes online; low cost is served offline — the compromise is the most common.

3. Cost Magnitude of Each Item

Configuration costs money; first see the cost tiers:

**Low-cost tier (distribution full coverage)**: - Oil temperature: sensor + acquisition, low cost; - Oil level: level gauge + alarm, low cost; - Load current: CT + acquisition, low cost; - Verdict: distribution transformers get these three; controllable cost, common faults covered.

**Medium-cost tier (power-transformer insulation)**: - Bushing: dielectric/PD sensors, medium cost; - Core ground current: ground CT + acquisition, medium cost; - Verdict: added on power transformers; covers early insulation defects.

**High-cost tier (large-capacity internal monitoring)**: - Online DGA: gas-sensor array + online analysis, high cost (one unit equals several distribution packages); - Winding temperature (fiber): fiber sensors embedded in windings, high cost; - Verdict: only on the highest-fault-cost transformers.

**Cost and pace**: - Limited budget: low tier everywhere first (oil temperature/level/load), then deepen by importance; - Sufficient budget: power transformers add medium tier (bushing/core), large-capacity add high tier (DGA); - Principle: spend first on the tier that covers the most transformers, then on the deepest/most expensive tier.

Key insight: **the cost tier sets the pace — cheap full coverage first, then expensive and deep.** Online DGA is the most expensive tier; give it only to the most important transformers.

4. Data Upload and Alarm Linkage

Tiered data must close the loop on one platform:

**Upload requirements**: - All tier data integrates into a **supervision system/gateway** (e.g., ESX Edge Computing Gateway aggregating; Digital Electrical Intelligent Supervision System displaying); - Distribution: oil temperature/level/load-current trends; - Power transformer: + bushing/core-ground-current trends; - Large capacity: + DGA gas trends.

**Alarm linkage**: - High oil temperature → check load/cooling; - Low oil level → check leaks; - Core-ground-current over-limit → check core insulation/clamps; - DGA abnormal (fast-rising acetylene/total hydrocarbons) → immediate outage inspection; - Multi-signal linkage: oil temperature + load + DGA abnormal together = high internal-fault probability.

**Health linkage**: - Oil temperature, load, ground current, DGA data become transformer health deduction items; - Worsening trends auto-escalate alerts; maintenance orders generate by deduction item.

Key insight: **without unified platform data, tiering is wasted.** Oil temperature, load, bushing, DGA — all tiers on one platform so trends and alerts can link.

5. Typical Tiered Configuration Examples

Three scenarios show how to configure:

**Scenario 1: campus distribution transformers (eight 10 kV)**: - Configuration: oil temperature + oil level + load current; - Purpose: cover overload/high-oil-temperature common faults; - DGA: no (offline testing suffices).

**Scenario 2: substation power transformers (two 110 kV)**: - Configuration: oil temperature + oil level + load + bushing monitoring + core ground current; - Purpose: cover early insulation defects; - DGA: one unit (highest fault cost); the other offline + linkage.

**Scenario 3: critical-hub power transformers (airport/data-center 220 kV)**: - Configuration: oil temperature + oil level + load + bushing + core ground + online DGA + winding temperature; - Purpose: full-dimensional coverage, early internal-discharge warning; - DGA: must-have (extremely high fault cost).

**Common tier logic**: - Oil temperature/level/load = always the first layer (full coverage); - Bushing/core = power-transformer second layer; - DGA = large-capacity/hub third layer; - Each higher layer requires a higher transformer class and fault cost.

Key insight: **the tier is "full coverage first, then deepen by class."** Oil temperature/level/load for everyone, bushing/core for power transformers, DGA only for the most important.

6. Product Landing: Tiered Transformer Monitoring

FEXLINK's **Digital Electrical Intelligent Supervision System** brings tiered transformer monitoring to practice:

- **EST Multi-Channel Temperature Controller**: oil-temperature and winding-temperature acquisition (low-tier full coverage + high-tier winding sensing); - **Oil-level/load monitoring**: level gauge + CT input (low tier); - **Bushing/core-ground-current monitoring**: sensors + acquisition (medium-tier power-transformer config); - **Online DGA**: characteristic-gas online analysis (high-tier large-capacity config); - **ESX Edge Computing Gateway**: unified aggregation of all tiers, edge computing, alarm linkage; - The platform shows oil temperature/level/load/bushing/core/DGA six-line trends; health scores auto-link deduction items.

Design principle in one sentence: **turn "whether to buy online DGA" into "tier by class."** Distribution basic, power-transformer insulation deeper, large-capacity internal warning — cost matched to risk.

7. Common On-Site Mistakes

1. **Online DGA on everything**: distribution transformers included. Consequence: cost out of control, most functions idle. Correct: tier by class; distribution gets basic monitoring. 2. **Treating offline testing as "backward"**: forcing online on distribution units. Consequence: waste. Correct: distribution = offline testing + online oil-temperature linkage. 3. **Only the cheap items**: power transformers without bushing/core. Consequence: early insulation defects missed. Correct: power transformers add the medium insulation tier. 4. **Installing DGA and ignoring it**: nobody watches gas trends. Consequence: the online unit becomes a decoration. Correct: trend monitoring + abnormal linkage; install means use. 5. **Data islands**: each tier on its own software. Consequence: trends fragmented, alerts missed. Correct: unified platform integration, multi-signal linkage.

Scope and Limitations

This article discusses tiered transformer online-monitoring configuration (oil temperature/oil level/load current/bushing/core-ground-current/online DGA) for distribution transformers and substation power transformers. Cost magnitudes and tiering are engineering experience references; actual choices follow voltage class, fault cost, vendor proposals, and budget. Online DGA should be selected by transformer importance; offline testing is the reasonable choice for distribution transformers. All tier data needs unified platform integration to form trend warnings and alarm linkage. All cases are anonymized (a certain campus / airport) and do not refer to any specific customer.

Want a deeper look at FEXLINK solutions?

Contact the FEXLINK solutions team for customised solutions and technical support.

Frequently Asked Questions

Which transformer monitoring items belong to each voltage class tier?

Distribution units get oil temperature, oil level and load current, important 35/110 kV power transformers add bushing and core-ground-current monitoring, and HV large-capacity transformers then consider online DGA; the tier is judged by class and fault cost, not by blanket deployment.

Is online DGA worth buying for an ordinary distribution transformer?

One 110 kV online DGA unit costs about 5-10x the whole online-monitoring package of a distribution transformer, and periodic oil sampling already serves low-fault-cost units; the purchase is judged as fault cost divided by price and stays reserved for the highest-cost transformers.

Why did an airport substation put online DGA on only one of its two transformers?

Both 35 kV units carried flight-related load, yet only the core unit took online DGA while the other ran offline testing plus online oil-temperature linkage; the split is judged per transformer's failure consequence, so the deep tier went where an interruption hurts most.

How early did online DGA catch an internal discharge at the airport substation?

Online DGA found acetylene creeping 3 months early, an early internal-discharge sign, and a planned outage prevented fault escalation; characteristic gases H₂, CH₄, C₂H₂, C₂H₄, CO and CO₂ keep continuous trends, which is how the warning is judged rather than a single sample.

How should tiered transformer data be uploaded and linked to alarms?

All tiers integrate through the ESX Edge Computing Gateway, and oil temperature, load, bushing, core ground and DGA data on one platform form six-line trends so multi-signal linkage is possible; the anomaly is judged together with load and temperature before a planned outage order.