polarisation index problem

[mottt]

we got an 160 rpm aprox. generator, 18 MW, manufactured two years ago with resin rich insulation, and the machine was stopped in the hydro power plant 1,5 years, after this we measured insulation....it was in the beginning low, less than 100 Megas, after we swicht on heating resistences, the insulation start to rise... till 300 megas in 1 minute, and 600 megas, after 10 minutes... this means one IP of 2..... the temperature inside machine was 30 º C,.. we think the low insulation and IP, is caused by humidity and water in the area.... then we solve leakages and condensation of water, and start with heating in short circuit, with a winding temperature of 105 ºC... after this we stopped, and cheek insulation and IP, values are 600 Megas and 1000 with a lower than 2 IP... this makes me be very confused... is IP dependent on temperature of the winding? I thought it was independent...? how do you see the machine? could a resin rich insulation without cure enought be the cause?

 

[electricpete]

Just to check - assuming these are not temperature corrected readings - does your instrument have the ability to measure beyond 1000 megaohms and the resolution to distinguish between 1000 megaohms and 1200 megaohms? If not, could be ok but just can't measure it.

If your instrument goes to 5000+ megaohms, I would say there still is some reason for concern. With winding temperature 105C (wow!), it is probably not moisture.

In theory PI should not depend on temperautre of the winding since the same correction factor applies to the 1 minute reading as to the 10 minute readings.

I do know that some insulation systems when not fully cured test bad on the partial discharge test. I haven't heard of not fully cured systems testing bad on PI test.


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[mottt]

yes, this is what i thought, PI is not dependent on temperature.... and for me other question to worry about is that in the factory, they write in the text bench, that the PI is 6...¿?.
About the equipment used is a Chauvin Arnoix megger that could go till 1000 Gigas, 5000 V, and the voltage of the generator is 6000 V.

other point, is that when we heating it... there is a very strong smeel to varnish or something like that.. i thought, cause the whole machine didn´t go inside of oven... and the windings aren´t manufactured using VPI system aren´t cure.

thanks for your attention....