Require Help for Transformer Flashover Incident

Since you have energised the transformer for two months the effect of the transfo being wet when installed is unlikely, it should have dried out before your load application.

Some questions to ask {I count 8}....

Because you refer to Restricted Earth Fault operation, I guess you had an LV winding fault, is that correct? {1}

Did you do an insulation resistance check on LV windings before energising? What was result? {2}

Since this is a commonly used size/type of transfo, it seems likely it was factory tested off-load and short-circuited but never given a heat run. You should find out what tests were actually done and the results [If there was a heat run, the supplier can claim it was AOK when it left him, otherwise a manufacture defect is more possible]. {3}

A turn to earth fault is far more likely than turn-turn, for same insulation thickness. When you write "bottom" of winding was burnt, is this in the visual way or also in the electrical sense of being at the neutral [earthed] end of winding? {4}

Fault at earthed end is more possible due to mechanical defect/damage than electrical stress for a uniform insulation thickness. Also low end usually runs cooler.

Have you checked if there was any damage to packing of transfo at arrival, suggesting mechanical damage during transport? Or rough handling on site, anything bent or dented or scratched? {5}

Is there any sign of "foreign bodies" near the fault? {6}

Blockage of ventilation passages I would expect to have slow effect, not within minutes of load application. Do the installation instructions require any transit blocks or braces to be removed? {7}

You have not given the transfo losses, but usually the magnetising losses & temperature rise no-load would be much less than full load loss/temperature rise. Full load would cause more expansion of windings than no-load.

Is there any forced-cooling and is it working OK? {8}

I think the advice offered is good and covers the many possible causes of failure.

But...The problem is easily divided into three responsibilities.

First...Has the manufacturer carried out the full amount of tests necessary?..

Load tests

Pressure tests

Most importantly ..as these are epoxy mouldings..the discharge hv tests?

The manufacturer should test for discharge values and then compare them

with historic tests of exactly the same design and performance.

Every design and configuration will give different values, and a product

should only pass if that design with a particular characteristic has proved

with use to be reliable.

Regretably epoxy moulding process is not just simply avoiding creating

voids during the moulding process.

Also we used to first mould conductors in a slightly cushion moulding and once set, complete the final moulding process.

This was to allow for expansion of copper conductors when reaching running temperatures.

So expansion of copper on load may have caused your problem.(sorry I still like the word "problem" and then "answer")

Second... Damage during transport or handling and positioning on site...packing not correct or strong or robust enough.

One would think a solid moulding would be very be strong...it could be

argued that an oil filled transformer is more forgiving.

Ofcourse if moulding was repaired for transport damage..there could be another cause.?

Third...The electrical environment the transformer is connected to?

Type and characteristics of the transformer load....highly inductive or

capacitive..switching devices? etc.

To summarise.... the manufacturer( not agent) should visit site , and they or with others must also investigate the electrical environment.

I personally would not accept any epoxy moulded trasformer that has been repaired,

because all the above mentioned tests and comparisons need to be carried out.

As I was involved for many many years in manufacture and later for many years in investigating power distribution problems and failures...i would consider it a really big favour if and when the event has been sorted out that you tell me what the outcome was.