Free module. This page is free to read. It does not ask you to sign in. It does not offer a class or a call. Use the edition of each standard named in the contract. Questions: hello@transformerpath.com.
Volume 5 is how a failed unit is investigated. This module is the discipline of that work. A cause you cannot tie to a record is a guess. A guess offered early closes off the measurements that would have tested it. The rest of the course, on the paid plan, walks mechanisms. This page walks the file.
Freeze the record before you explain it
Write down what is known on the day of the event, and keep it separate from what is inferred later. Protection traces say how large the current was and how long it lasted. The nameplate and the routine-test report say what this serial number was built and tested as. Oil samples taken before the event, if they exist, say whether the unit was already abnormal. Photographs, the position of valves, the state of the breather, and the impact-recorder record from the last move are evidence. A theory about lightning, through-fault, or moisture is not evidence until one of those records points at it.
Do not start by replacing parts to see if the problem goes away. A replaced bushing, a changed tap-changer contact, or a reprocessed oil charge destroys the state you needed to inspect. If the unit is unsafe to approach, make it safe first, and record what you had to disturb. Safety of people on site outranks a neat investigation. It does not require you to invent a cause in the same hour.
A cause has to exclude the neighbours
Several mechanisms leave a similar external mark. A tank rupture can follow an internal arc. It can also follow a pressure event whose electrical source is still unknown. A high combustible gas can follow arcing, a thermal fault, or oil that was already contaminated. The investigation earns a cause when the other plausible causes disagree with the records you hold. “Consistent with” is not “shown by”. The paid modules treat individual mechanisms. They still depend on this rule.
- Separate observation from inference. Date both.
- Keep the pre-event samples and the protection record. They are the history the wreckage no longer shows.
- Do not replace the suspected part before it has been inspected and photographed.
- A cause that fits every failure fits none. Say which records would have contradicted it.
The volume page lists outcomes, not a case file. Worked investigations stay on the paid plan.
Upgrade
Module 1 is free. The rest of this course stays on the paid plan. The plan is a one-time payment for 12 months of access, with no auto-renewal. This page does not offer a class or a call.