Why This Transformer Failed: A Structured Fault Investigation
A systematic route from initial symptoms to root cause: gas-in-oil, electrical tests and physical evidence, and how each narrows the possible failure modes.
A transformer alarm rarely points at a single cause. A structured investigation separates a developing fault from benign operating behaviour and drives an evidence-based repair-or-replace decision.
Start with the symptom. Gas-in-oil, a winding-resistance drift, an SFRA signature shift or a temperature anomaly each narrow the candidate set. Differential protection and dissolved-gas analysis (DGA) are the two highest-value leading indicators.
Interpret DGA in context: acetylene above a few ppm points to arcing (a turn-to-turn, tap-changer diverter or core-joint arc); a strong hydrogen and methane signature with no acetylene suggests a low-energy discharge or a hot spot; ethane and ethylene climb with sustained overheating. The ratio matters as much as the absolute value.
Corroborate with electrical tests. A winding-resistance change on one phase indicates a connection or conductor problem; a ratio change points at a turn fault; a capacitance/power-factor change implicates insulation or bushing condition.
Physical evidence closes the loop. Where possible, inspect the evidence a test points to: the diverter contacts, the winding clamps, the insulation at the suspect location. This is what turns a theory into a confirmed cause.
The decision then becomes an engineering question, not a diagnosis: can the fault be repaired in place or by an off-site rebuild, and does the remaining life, reliability and transportation cost justify it at all? Record the reasoning and the evidence.
Related principle
Transformer failure analysis, DGA interpretation and the repair-or-replace decision (see the Masterclass chapters on losses, testing and protection).