Téléchargez notre catalogue de produits 2025 pour obtenir les schémas détaillés et les paramètres techniques de tous les composants des appareillages de commutation.
Obtenir le catalogue Téléchargez notre catalogue de produits 2025 pour obtenir les schémas détaillés et les paramètres techniques de tous les composants des appareillages de commutation.
Obtenir le catalogue
Transformer overheating shortens insulation life faster than almost any other operating stress. IEEE C57.91 establishes that every 6 °C rise above rated temperature roughly halves insulation life, so identifying the root cause early is an economic necessity, not a maintenance preference. This guide sequences the diagnostic process from quick field observations through quantitative testing to procurement decisions, covering the four root causes responsible for the majority of overheating failures: overloading, cooling system failure, harmonic distortion, and connection defects.
Before investing in outage time or specialized testing, use this table to identify the most probable root cause from the first observable evidence.
| Symptôme | Premier test | Cause première probable | Action suivante |
|---|---|---|---|
| WTI alarm active; load appears high | Current clamp on all three phases; compare to nameplate kVA | Surcharge | Log load profile for 7 days; review demand peaks |
| Temperature rises faster than load increases | Confirm fan operation and oil level | Cooling system failure | Inspect radiators, fans, pumps; clean or repair |
| Elevated temperature at moderate apparent load; audible hum | Power quality analyzer; measure THD-I and K-factor | Harmonic distortion | Calculate Factor-K; derate or filter |
| Localized hot spot at one terminal; metering shows no load anomaly | IR thermography on all external connections | Connection defect | DLRO test on flagged joint; re-torque or replace |
| Overheating only during seasonal peaks | Check ambient temperature against nameplate cooling class rating | Ambient derating exceedance | Reduce load or add supplemental cooling |
| WTI and TOT inconsistent with each other | Compare instrument readings against a calibrated reference | Instrument fault | Calibrate or replace temperature indicators |
| Instrument | Application in This Guide | Source d'acceptation |
|---|---|---|
| True-RMS clamp meter or current transformer logger | Load current measurement; phase balance check | IEEE C57.91; nameplate kVA |
| Power quality analyzer (to 50th harmonic) | THD-I, individual harmonic orders, K-factor input | IEEE 519; IEEE C57.110; IEC 61378 |
| Infrared camera (<= 0.1 °C NETD; >= 320×240) | Connection defect location; radiator uniformity check | NETA MTS-2019 (ΔT criteria) |
| Low-resistance ohmmeter / DLRO (test current >= 100 A DC) | Contact resistance at terminals, tap changer, cable lugs | IEEE C57.152; IEC 60076-1 |
| Insulation resistance tester (500–5000 V DC) | Winding insulation check following thermal event | IEEE C57.12.90; IEC 60076-1 |
| Oil sampling kit and laboratory (DGA, moisture) | Detect dissolved combustion gases; moisture in oil | IEC 60599; IEC 60422 |
| Calibrated torque wrench | Connection re-torque verification | Connector manufacturer specification |
| Anemometer | Fan airflow measurement at fan outlets | OEM cooling design specification |
| Ultrasonic clamp-on flow meter | Oil pump flow measurement (OFAF/ODAF units) | OEM pump rating |
| Winding temperature indicator (WTI) / oil temperature indicator (OTI) | Continuous thermal monitoring | IEC 60076-2; IEEE C57.91 |
| OEM installation and maintenance manual | Setpoints, torque values, contact resistance baselines | OEM documentation |
| Project specification and one-line diagram | Rated cooling class, load assumptions, harmonic requirements | Project engineering package |
Overloading is frequently misread in the field because demand peaks are intermittent; a single spot measurement taken during a low-demand period will miss the thermal event entirely.
Step 1 – Check nameplate kVA against connected load. Calculate apparent load from metered voltage and current. A load factor above 100% is an immediate flag; a load factor between 80% and 100% is not automatically safe because ambient temperature, cooling condition, and load shape all affect the available thermal margin.
Step 2 – Review thermal indicator history. Pull the maximum-demand pointer reading from the WTI or OTI. A WTI reading that has reached or exceeded the alarm setpoint—typically 120 °C for ONAN-rated units per IEC 60076-2—confirms thermal stress has occurred even if current load appears normal.
| Metric | Acceptable | Enquêter | Action requise |
|---|---|---|---|
| Peak kVA as % of nameplate | <= 100% | 100?120% | > 120% |
| Duration of peaks above 100% | < 15 min/day | 15?60 min/day | > 60 min/day or recurring daily |
| Load factor (avg kVA / nameplate kVA) | <= 75% | 75?90% | > 90% |
| Peak-to-average ratio | < 1.5 | 1.5?2.0 | > 2.0 |
A transformer can be correctly loaded and free of harmonic distortion yet still overheat if its cooling system cannot dissipate heat fast enough. Cooling failure is among the most actionable root causes because the fault is usually visible, measurable, and correctable before winding insulation degrades.
| Classe de refroidissement | Winding Medium | Circulation | Performs Well When | Becomes Risky When |
|---|---|---|---|---|
| ONAN / ON | Mineral oil | Convection naturelle | Low-maintenance sites, stable load | Ambient > 40 °C or steep load cycles |
| ONAF / OF | Mineral oil | Forced air (fans) | Moderate overload capacity needed | Fans fail silently or filters clog |
| OFAF | Mineral oil | Forced oil + forced air | High continuous load, compact footprint | Oil pump seals age or flow sensors are absent |
| ODAF / OD | Directed oil | Forced directed oil + air | Large power transformers, tight thermal margins | Pump cavitation or blocked oil ducts go undetected |
| ANAN / AN | Dry-type, air | Convection naturelle | Indoor, fire-sensitive locations | Enclosure ventilation is restricted or ambient rises |
| ANAF / AF | Dry-type, air | Forced air | Indoor with variable load | Fan failure or duct blockage causes rapid hot-spot rise |
Radiators and cooling fins (ONAN/ONAF): Blocked fins from dust, paint overspray, or biological growth reduce effective surface area. Pass criterion: fin passages visually clear; IR scan shows uniform temperature gradient from top to bottom of each radiator bank. Fail indicator: one or more radiator sections significantly cooler than adjacent sections on IR scan, indicating blocked oil flow.
Cooling fans (ONAF/ANAF): Fan motor failure, reversed rotation after maintenance, or seized bearings reduce airflow without triggering an alarm if current monitoring is absent. Measure airflow at the fan outlet with an anemometer; a reading below 80% of rated CFM warrants investigation.
Harmonics increase losses without increasing the fundamental-frequency load current that most protection relays monitor. A transformer running at 70% of nameplate kVA can still overheat if the load is rich in harmonics, and a standard ammeter will not reveal the problem.
| Paramètre | K-Factor (IEEE C57.110) | Factor-K (IEC 61378 / BS 7821) |
|---|---|---|
| Origine | Amérique du Nord | Europe / IEC regions |
| Objectif | Rate a new transformer for a known harmonic load | Derate an existing transformer |
| Eddy loss exponent | 2.0 (conservative) | 1.7 (empirically derived) |
| Sortie | Dimensionless multiplier; transformer K-rating >= calculated K | Applied to nameplate kVA to get derated capacity |
| Where it wins | Specifying new transformers for VFD or UPS loads | Evaluating whether an existing standard transformer is adequate |
| Where it becomes risky | Applying to a transformer not built to IEEE C57.110 | Using without measured harmonic data |
Practical derating using Factor-K: Derated kVA = Nameplate kVA / Factor-K. A Factor-K of 1.15 means the transformer should be treated as having 87% of its nameplate capacity.
Clamp current probes on all phase conductors at the transformer secondary; measure all three phases simultaneously.
| Mesures | Acceptable | Enquêter | Action requise |
|---|---|---|---|
| THD-I | < 8% | 8?15% | > 15% |
| Individual harmonic (any order) | < 5% of I1 | 5?10% | > 10% |
| Neutral / phase current ratio | < 0.5 | 0.5?1.0 | > 1.0 |
| Factor-K | < 1.05 | 1.05?1.20 | > 1.20 |
Loose or corroded connections are among the most underdiagnosed overheating root causes. A bolted lug that has relaxed by even a few milli-ohms can dissipate enough heat to carbonize surrounding insulation while nameplate load stays within rating and the cooling system shows no fault.
| Pre-Scan Condition | Exigence |
|---|---|
| Load at time of scan | >= 40% of rated current; document actual load |
| Minimum soak time at load | 30 minutes before scanning |
| Wind speed | < 3 m/s |
| Emissivity setting | 0.90–0.95 for oxidized copper or aluminum; 0.85 for painted steel |
| Camera sensitivity | <= 0.1 °C NETD; minimum 320×240 detector |
| ΔT Above Reference Phase or Ambient | Sévérité | Action |
|---|---|---|
| 1–3 °C | Possible defect | Re-scan at next opportunity; monitor trend |
| 4–15 °C | Defect confirmed | Schedule repair within 30 days |
| > 15 °C | Serious defect | De-energize or reduce load; repair before returning to full load |
| Type de connexion | Acceptable | Enquêter | Reject / Remediate Immediately |
|---|---|---|---|
| Bushing terminal pad (HV, >= 15 kV) | < 10 µΩ | 10–50 µΩ | > 50 µΩ |
| Bushing terminal pad (LV, < 1 kV) | < 15 µΩ | 15–60 µΩ | > 60 µΩ |
| Cable lug to busbar, bolted | < 20 µΩ | 20–100 µΩ | > 100 µΩ |
| Ground strap connection | < 25 µΩ | 25–100 µΩ | > 100 µΩ |
| OLTC contact finger set (per phase) | Per manufacturer spec ±20% | > 20% above spec | > 50% above spec |
The following four-stage workflow sequences decisions by evidence cost and probability, starting with observations that require no outage and progressing to tests that require one.
Maintenance history: Date of last oil sample and DGA results; date of last cooling system inspection; any recent load increases or added nonlinear loads on the bus.
Conditions environnementales : Ambient temperature relative to the transformer’s rated cooling class ceiling; altitude above 1,000 m; dust accumulation on radiator fins; recent high humidity or flooding history.
| Priorité | Cause profonde | Immediate Action (within 24 h) | Short-Term (within 30 days) | Long-Term |
|---|---|---|---|---|
| 1 – Critical | Contact defect with ΔT > 40 °C at bushing or tap changer | De-energize; repair before re-energizing | Full contact resistance survey; DGA for arcing by-products | Establish thermography and contact resistance baseline; revise inspection interval |
| 2 – High | Overload > 120% continuous | Shed load; enable all available cooling stages | Install metering to track load growth | Load forecast review; upgrade or parallel transformer |
| 2 – High | All fans inoperative | Manual load reduction to 60–70% of nameplate; emergency fan repair | Replace failed components; inspect control circuit | Implement cooling system health monitoring with remote alarm |
| 3 – Elevated | K-factor exceedance | Derate transformer to safe K-factor limit | Measure harmonic spectrum at all major loads | Replace with appropriately rated unit or install harmonic mitigation |
| 4 – Moderate | Single blocked radiator or partial fan failure | Clean or restore affected cooling section | Full radiator inspection and cleaning schedule | Environment-specific maintenance intervals based on contamination rate |
| 5 – Routine | Ambient temperature exceedance during seasonal peak | Confirm load is within corrected nameplate rating; monitor continuously | Evaluate supplemental cooling | Include ambient derating in annual capacity planning |
Site context: A 1,000 kVA, 13.2 kV / 480 V, ONAF-cooled transformer serving a manufacturing plant’s VFD-heavy production floor received a WTI alarm at 118 °C during a mid-afternoon production peak.
Initial observations: Load current measured at the secondary showed approximately 880 kVA—88% of nameplate rating. Ambient temperature was 36 °C, within the transformer’s 40 °C cooling class ceiling.
Cooling check: Two of four cooling fans were rotating. A third fan’s contactor had tripped on thermal overload. The fourth fan had been reconnected in reverse after a recent motor replacement, reducing its effective airflow contribution. Total measured airflow was 62% of combined rated CFM for all four units.
When a transformer has a documented overheating history, a like-for-like replacement rarely solves the problem. The procurement process must address the root causes identified during troubleshooting before a purchase order is issued.
| Critère | Minimum Acceptable | Red Flag |
|---|---|---|
| Temperature rise rating | 80 °C or 115 °C rise for dry-type; class confirmed | 150 °C rise for a K-rated unit without thermal justification |
| K-factor documentation | Factory test report included | K-factor on nameplate only, no test data |
| Cooling class documentation | ONAN/ONAF/OFAF clearly stated with rated capacity at each stage | “Self-cooled” with no thermal model |
| Loss data | No-load and load losses at rated current provided | Efficiency percentage only |
| Thermal model basis | IEEE C57.91 or IEC 60076-7 stated | No thermal model provided |
| Warranty scope | Covers winding insulation failure, not only manufacturing defects | Excludes overloading without defining the threshold |
| Spare parts commitment | Replacement windings or cooling components available within stated lead time | Custom design with no spare parts commitment |
Utilisez ces références XBRELE pour relier la décision sur le terrain au produit correct, au test et au flux de travail de l'approvisionnement : Page produit XBRELE, Gamme de disjoncteurs à vide XBRELE, Guide de notation de la VCB, Liste de contrôle pour l'acceptation du TFA/TSA par le VCB, XBRELE power distribution transformer range.
Pour le contexte de la méthode externe, comparez la procédure du site avec la procédure publique. Page des normes IEEE C37.09 et appliquer le manuel de l'équipementier et les spécifications du projet pour l'équipement fourni.
Exemple de terrain : lors d'une inspection de service, une phase a été mesurée en dehors de sa ligne de base de mise en service, alors que les deux autres phases sont restées stables. L'équipe a répété la mesure avec des fils vérifiés, a contrôlé la synchronisation et la course du contact, et a utilisé la divergence mesurée pour distinguer un problème de pression de contact d'un problème générique de nettoyage de surface.
In commercial buildings, harmonic distortion from VFDs, UPS systems, and switched-mode power supplies is the most frequently overlooked cause. Overloading is often suspected first, but a power quality measurement frequently reveals that a transformer operating at 70–80% of nameplate kVA is still overheating because its eddy current losses are elevated by high THD-I.
For distribution transformers in standard industrial or commercial service, a contact resistance survey at all external terminals every 3 years is a reasonable baseline. Transformers in high-vibration environments, coastal or humid locations, or applications with frequent load cycling should be surveyed annually.
Not indefinitely. At 110% load in a standard 40 °C ambient, IEEE C57.91 indicates insulation life consumption roughly doubles compared to rated load.
A K-factor rating indicates that a transformer has been designed with reinforced windings and a reduced eddy current loss coefficient to handle harmonic-rich loads. Standard distribution transformers are K-1 rated; units rated K-4, K-13, and K-20 are progressively more tolerant of harmonic currents.
Restore full fan operation and observe whether the WTI reading returns to normal at the same load level within 2–4 hours. If temperature drops significantly, cooling failure is confirmed as a primary contributor.
Methane (CH4) and ethylene (C2H4) are the primary markers of thermal decomposition of oil at moderate temperatures (150–500 °C). Acetylene (C2H2) appears at temperatures above 700 °C and is associated with arcing or very intense localized heating.
Replacement becomes the more economical decision when two or more of the following are true: the unit has experienced multiple thermal alarms within a 3-year period despite corrective action; DGA results show sustained or growing concentrations of thermal decomposition gases; contact resistance on internal components cannot be restored to specification without a full rewind; the load environment has changed to the point where the existing unit cannot be adequately derated; or the unit is beyond the manufacturer’s recommended service life with spare parts no longer available. A like-for-like replacement should always be preceded by the procurement checklist above to avoid repeating the same failure mode.
