Digitale Transformatorüberwachung: Sensoren, Konnektivität, Alarme und Fernabfrage von Daten

Veröffentlichungsdatum:4. Oktober 2026, 09:55:14 Uhr

  • Definition: digital transformer monitoring turns physical measurements such as temperature, gas, partial discharge and load into time-stamped data, and sends it to operators over standard industrial protocols.
  • Sensors: winding and oil temperature come first, then dissolved gas, partial discharge, bushing and tap changer sensors where the asset is critical.
  • Connectivity: local devices talk RS485/Modbus RTU to an edge gateway. The gateway forwards data by Modbus TCP, IEC 60870-5-104, DNP3 or IEC 61850 to SCADA or a cloud platform.
  • Alarms: use staged alarms (fan start, pre-alarm, alarm, trip). Trip and fan relays must work locally even when the network is down.
  • Remote data: trends, rate-of-change alerts and event logs are more useful than single readings. Good data quality (time sync, tag naming, sensor-fault flags) matters as much as the sensors.

Inhaltsverzeichnis

1. Temperature Sensors, Pt100 Probes and 4–20 mA Transmitters

Why temperature is the first digital signal

Winding temperature is the main driver of insulation ageing, so it is the first parameter to digitize. Every 6–8 °C above the design hot spot roughly halves the expected life of paper insulation, which is why accurate readings support safe loading decisions. For background on how heat builds up, see transformer overheating causes and prevention.

Common temperature sensing options

  • Pt100 probes: low cost and easy to read by a controller or RTU. Use three-wire connection to cancel lead resistance.
  • Oil temperature indicator (OTI): a pocket-mounted probe that reads top oil temperature with local dial and alarm contacts.
  • Winding temperature indicator (WTI): estimates hot spot from top oil plus a load-current heating offset. The difference between the two gauges is explained in oil temperature gauge vs winding temperature gauge.
  • 4–20 mA or Pt100 transmitter: converts a dial gauge signal into a digital-friendly output for the gateway.

Sensor output reference

Sensor Typical Output Typical Range Digital Path
Pt100 probe Resistance (100 Ω at 0 °C) -50 to 250 °C Controller or RTD input module
Oil temperature indicator Contacts, Pt100 or 4–20 mA 0–120 °C Transmitter to gateway
Winding temperature indicator Contacts, Pt100 or 4–20 mA 0–150 °C Transmitter to gateway
Fiber optic probe Optical signal to demodulator -40 to 200 °C RS485 Modbus or relay from demodulator
Current transformer (load) 1 A or 5 A secondary 0–120% rated Meter or transducer

2. Fiber Optic Temperature Sensor and Demodulator for Winding Hot Spots

What makes fiber optic sensing suitable for digital systems

  • Direkte Hotspot-Messung: the probe sits against the winding, so the value does not depend on a thermal model or CT ratio.
  • Dielectric and EMI-immune: no metal at the tip, so readings stay clean in strong electric and magnetic fields.
  • Clean digital output: a 6-channel fluorescent fiber optic demodulator converts the optical signal to RS485 Modbus registers and relay outputs.

Typical fiber optic system parameters

Artikel Typischer Wert
Messbereich -40 °C bis 200 °C
Genauigkeit ±1 °C
Reaktionszeit ≤ 1 s
Channels per unit 1, 3, 6 oder bis zu 64
Digital output RS485 Modbus RTU, 4–20 mA, Relais
Länge der Messleitung Maßgefertigt, mit Verlängerungskabel erweiterbar

Auswahlkriterien

  • Channel count: three channels for one three-phase unit, six or more when a transformer needs several hot spot points, and a high-channel system when one panel serves a whole substation.
  • Aufbau der Sonde: armored probes for oil-immersed windings, polyimide-coated probes for resin and dry-type windings.
  • Embedded or retrofit: embedded winding probes are installed during manufacture or rewinding. In-service units normally use top-oil or surface probes instead. The measurement principle is covered in Glasfaser-Temperaturmessung in Transformatoren.

3. Dry-Type Transformer Temperature Controller with RS485 Output

Functions a digital controller must provide

  • Three-phase sensing: one Pt100 per phase plus an optional core channel.
  • Fan control with hysteresis: prevents fans from chattering around the set point.
  • Independent alarm and trip relays: hard-wired to the protection circuit.
  • Sensor-fault detection: an open or shorted probe should raise a fault flag and not read as a normal temperature.
  • RS485 Modbus RTU: temperatures, relay states and fault flags available to the gateway. A device such as the IB-S201 dry-type temperature monitor and controller combines these functions in one panel unit.

Insulation class reference for set points

Isolationsklasse Maximale Wicklungstemperatur Average Winding Rise
Klasse B 130 °C 80 K
Klasse F 155 °C 100 K
Klasse H 180 °C 125 K

4. Online DGA Monitor for Dissolved Gas Data

Where online DGA adds value

  • Critical assets: transformers feeding hospitals, data centers, rail or process lines where outage cost is high.
  • Remote sites: unmanned substations where manual sampling is slow and expensive.
  • Developing faults: a rising gas trend between lab samples can be caught within hours rather than months.

Gas indicators

Gas Typical Fault Indication
Wasserstoff (H₂) Partial discharge, low-energy fault
Methane (CH₄), Ethane (C₂H₆) Low to medium temperature overheating
Ethylen (C₂H₄) High temperature overheating
Acetylen (C₂H₂) Arcing or very high temperature
CO, CO₂ Paper insulation overheating

Auswahlkriterien

  • Single-gas vs multi-gas: a hydrogen or composite-gas monitor suits screening, while multi-gas units support Duval triangle and IEC 60599 interpretation.
  • Trend over absolute values: use IEEE C57.104 or IEC 60599 limits and watch the rate of gas increase.
  • Valve mounting: install on the drain or sampling valve through an isolation valve to allow service without draining the tank. A practical comparison with lab testing is in online DGA vs oil sampling.

5. Partial Discharge, Bushing Adapter and OLTC Motor Current Sensors

Partial discharge sensors

  • UHF sensors on the oil valve give good noise rejection for internal discharge.
  • HFCT sensors on bushing taps or the ground strap suit retrofit work.
  • Acoustic sensors on the tank help estimate the source location.

Bushing monitoring adapters

  • Measured values: capacitance change and tan δ, compared across phases.
  • Why it matters: bushing insulation faults can escalate to tank rupture, so trending these values is worthwhile on high-value units.

OLTC monitoring sensors

  • Measured values: tap position, drive motor current, operation count and compartment temperature difference.
  • Typical fault clues: rising motor current and slower tap change time, as outlined in OLTC transformer monitoring.

Sampling and data reference

Data Type Typical Sampling Typical Retention Alarmgrundlage
Winding / oil temperature 1–60 s Years (aggregated) Absolute and rate of rise
Load current / voltage 1 s Years (aggregated) Overload percentage
Dissolved gas 1–24 h Life of asset Level and gas generation rate
Teilentladung Continuous, stored as events Life of asset Magnitude and pulse count trend
Bushing tan δ / capacitance Continuous, stored hourly Life of asset Change from baseline

6. Edge Gateway, Modbus and IEC 61850 Connectivity

Typical data path

  • Field layer: sensors, temperature controllers and demodulators on shielded twisted pair RS485.
  • Edge layer: a gateway polls devices, buffers data during network loss and converts protocols.
  • Control layer: SCADA, EMS or a cloud platform receives the converted data.

Protocol reference

Interface / Protocol Typische Anwendung Mittel Anmerkungen
RS485 / Modbus RTU Devices to gateway Shielded twisted pair Up to about 1200 m at 9600 bps
Modbus TCP Gateway to PLC or SCADA Ethernet Simple and widely supported
IEC 60870-5-104 / DNP3 Utility telecontrol Ethernet / IP Common in grid operations
IEC 61850 Digital substations Ethernet Requires ICD/SCD file mapping
MQTT / OPC UA Cloud and IIoT platforms Ethernet / IP Use TLS and certificates
4G / 5G / fiber backhaul Remote unmanned sites Wireless / fiber Add VPN and firewall

Connectivity rules

  • Local protection stays local: alarm and trip relays must never depend on the network.
  • Buffer on the gateway: store readings during outages and back-fill them when the link returns.
  • Match the control system: choose the protocol your SCADA already speaks rather than adding a new one.

7. Alarm Logic, Relay Outputs and SMS/Email Notification

Four-stage alarm scheme

Bühne Relay / Notification Action Setting Basis
Lüfterstart Start cooling fans Below alarm, per manufacturer
Pre-alarm Notify operator, review load Insulation class and loading guide
Alarm Dispatch inspection, reduce load Nameplate temperature rise
Reise Disconnect the transformer Maximum permissible temperature

Example starting points

Transformatortyp Lüfter starten Alarm Reise
Dry-type Class F winding 100–110 °C 130 °C 150 °C
Dry-type Class H winding 120–130 °C 155 °C 170–180 °C
Oil-immersed top oil 65–75 °C 80–85 °C 95–105 °C

Digital alarm improvements

  • Rate-of-rise alarms: a fast temperature climb can matter more than the absolute value, especially under sudden overload.
  • Alarm delay and hysteresis: a short delay and a 2–5 °C hysteresis prevent nuisance alarms and relay chatter.
  • Priority levels: separate informational, warning and critical messages so operators respond in the right order.
  • Notification routing: send critical alarms to SCADA and by SMS or email to the on-call engineer.
  • Set point governance: treat all values as starting points, confirm against the nameplate and transformer temperature rise limits, and log every change with date, reason and approver.

8. Remote Data Platform, SCADA Dashboard and Cloud Analytics

What a useful dashboard shows

  • Asset overview: one health indicator per transformer, color-coded by alarm state.
  • Trend charts: winding temperature against load, so abnormal heating stands out.
  • Event log: alarms, relay actions and sensor faults with time stamps.
  • Fleet comparison: similar transformers compared side by side to spot outliers.

Platform options

Option Best For Typical Protocol Consideration
Existing SCADA / DCS Utilities and large plants IEC 104, DNP3, Modbus TCP Needs tag mapping and HMI pages
Substation monitoring server Digital substations IEC 61850 Aligns with substation automation
Cloud platform Distributed or small assets MQTT, HTTPS Needs secure cellular or fiber link
Local touch-screen HMI Single transformer or room Modbus RTU No remote access unless added

Making the data actionable

  • Standardize tag names: one naming and register convention across all transformers keeps the platform maintainable, as discussed in transformer monitoring dashboard and SCADA integration.
  • Link data to maintenance: trigger work orders from sustained warnings, not only from trips.
  • Keep baselines: store readings at known loads so later drift is easy to see.

9. Cybersecurity and Data Quality for Remote Monitoring Devices

Cybersecurity checklist

  • Network separation: keep monitoring devices on a segmented network behind a firewall.
  • Encrypted remote links: VPN or TLS for any connection leaving the substation.
  • Access control: role-based accounts, no shared passwords and unused ports disabled.
  • Firmware management: track versions and apply updates through a controlled process, referencing IEC 62443 practices.

Data quality checklist

  • Time synchronization: NTP or IRIG-B so events from different devices can be correlated.
  • Quality flags: pass sensor-fault and communication-loss flags to the platform so bad data is not trended as real.
  • Calibration records: keep dated records for every sensor and transmitter.
  • Range checks: reject values outside physical limits before they reach analytics.

10. Installation and Commissioning Checklist for Monitoring Hardware

Before installation

  • Collect rating, cooling type, insulation class, temperature rise and test report data.
  • Decide whether sensors can be embedded (new unit) or only retrofitted (in-service unit).
  • Define the sensor list, controller, gateway and auxiliary power supply.
  • Agree the register map and protocol with the control team.
  • Plan the outage window and safety clearances.

During installation

  • Route fiber and signal cables away from live HV parts and keep the minimum bend radius.
  • Ground cable shields at one end only.
  • Seal cable glands against moisture on outdoor units.
  • Label each channel, cable and terminal to match the drawings.

Commissioning tests

Test Verfahren Bestehenskriterium
Temperature channel check Compare with a reference thermometer Im Rahmen der Gerätegenauigkeit
Relay output test Simulate set points and check contacts Correct fan, alarm and trip action
Sensor fault test Disconnect a probe Fault flag raised
Communication test Read all registers from SCADA Die Werte entsprechen der lokalen Anzeige
Network loss test Unplug the uplink Local alarms work and data is back-filled

Nach der Inbetriebnahme

  • Write a response procedure for every alarm level and train the operators.
  • Store baseline readings at known loads.
  • Schedule an annual check of sensors, wiring, contacts and communication.

11. FAQ: Digital Transformer Monitoring Questions

What is digital transformer monitoring?

It is the use of electronic sensors, controllers and communication devices to measure transformer condition, such as temperature, load, gas and partial discharge, and to send that data to operators in real time over standard protocols.

Which sensors should I install first?

Start with winding temperature, top oil temperature and load current. Add online DGA, partial discharge, bushing and OLTC sensors when the transformer is critical, high-value or hard to access.

What is the difference between a temperature controller and a full monitoring system?

A temperature controller handles local fan, alarm and trip functions for one transformer. A full monitoring system adds more sensor types, a gateway, a data platform and trend analysis, as described in Online-Transformatorüberwachungssysteme.

Which communication protocol is best?

Modbus RTU and Modbus TCP suit most industrial and commercial projects. Utilities usually use IEC 60870-5-104 or DNP3, and digital substations use IEC 61850. Pick the protocol your control system already supports.

Will monitoring still protect the transformer if the network fails?

It should. Alarm, fan and trip relays must operate locally in the controller or indicator, and the gateway should buffer data and back-fill it once the link returns.

Can digital monitoring be added to an existing transformer?

Yes. Oil temperature probes, winding temperature indicators with transmitters, current-based load monitoring, online DGA and external PD or bushing sensors can be retrofitted. Embedded winding fiber probes are normally fitted during manufacture or major overhaul.

How accurate are fiber optic temperature sensors compared with Pt100?

Fiber optic systems typically reach ±1 °C and measure the hot spot directly with strong EMI immunity. Pt100 probes are accurate at their location but usually measure oil, air or surface temperature, not the winding itself.

How often should sensors and alarms be checked?

Follow the manufacturer's interval. A common practice is an annual functional check of sensors, alarm contacts and communication, with recalibration when drift is found.

How do I get a recommended configuration for my transformers?

Prepare the transformer type, rating, cooling method, insulation class and your SCADA protocol, then Leiten Sie diese an das Entwicklerteam weiter for a matched monitoring scheme.