One Transformer Always Ran Hotter. Why Didn’t We Blame Cooling First?

The problems that do not trip are the ones that can survive for months

Rafael’s two transformers had been operating in parallel for three months. No outage. No major alarm. But Transformer A was consistently a little hotter and carried a little more current than Transformer B.

That kind of problem is easy to dismiss as “normal variation.”

TransformerGrid does not treat the absence of a trip as proof that the pattern is irrelevant. We ask whether the difference existed from day one or appeared later, whether it changes with load, whether taps match, whether %Z matches and what neutral current is doing.

Why we do not pick a favorite cause too early

A hotter unit can point toward cooling, unequal load sharing, circulating current from ratio/tap difference, phase imbalance, a connection problem or even a measurement issue.

If the investigation starts with “it must be cooling,” the team can spend days proving the first guess instead of discovering the real pattern.

TransformerGrid prefers to place both transformers on the same timeline: current, kW/kVA, tap, ambient temperature, oil/winding temperature, neutral current, bus voltage and relay events.

Why delivery documents become future diagnostic assets

Many procurement documents exist only to pass acceptance. FAT gets signed and the file disappears into storage.

We want those records to remain useful. Percent impedance, ratio, tap, connection, cooling configuration and initial baseline can all help explain later behavior.

TransformerGrid should not be valuable only before the purchase. Good project data should still help an operations team answer “what was normal at the beginning?” years later.

We want customers to feel that scattered data becomes clearer when it reaches us

TransformerGrid will not diagnose a complex transformer problem from one temperature photo. We will not invent certainty just to sound impressive.

What we can do is organize the information: which differences are equipment-related, which are system-related, which questions belong to the manufacturer and which observations must be verified on site.

Once a complicated problem is divided into the right categories, the people responsible for solving it know where to start.

The reputation we actually want

Not “TransformerGrid knows everything.”

We want the customer to say: “They do not brush me off. They ask the next question, line up the records and chase the manufacturing-side answers that are missing.”

Do not guess the cause first. Put both transformers’ data together first.

That is the kind of practical memory we want the TransformerGrid name to carry.