Transformer capacitive leakage current

[hhhansen]

Hi all,

I have made some records on a single phase (132/27,5 kV,23 MVA) railway transformer.
Normally there are 2 earthing connection point situated on the tank of the transformer. We have removed one earthing connection, see phote and then one should expect to measure the sum of the capacitive leakage currents, see attached current path diagram.

From this diagram I would expect the capacitive leakge current to be constant. However the enclosed records (0,5 sec. loggings) indicate that the earthing current fluctuates from app. 5 - 250 mA. The fluctuations follows the load current on the transformer.

I am not able to explain these fluctuations.

Can anyone assist?

I would be especially happy to know how to represent this i my model file. By now I have conected two capacitances from phase R resp. T yielding a constant capacitive current for each.

Best regards
Hans-Henrik

 

[LiteYear]

There's a few things going on here. Fundamentally yes, capacitive earth leakage current should not be a function of load current but predominantly a function of load voltage. However there's a few confounding factors here.

1. There could be some inductive coupling going on. I'm not sure where, but I suggest it because inductive coupling would be load current sensitive.
2. Since the tx secondary appears to use the earth as a conductor, any earth resistive will cause an earth potential rise. Since the tx primary is star-point earthed, variations in the earth potential will appear across the parasitic capacitances on the primary side.
3. You appear to have missed the parasitic capacitance from tx secondary active to earth.
4. Leakage current from the tx secondary will not tend to find the primary as a return path - secondary leakage current must return to the secondary coil.
5. If there are any non-linear loads, they can generate higher frequency voltages that are load current sensitive. Those higher frequencies could be contributing to your leakage current.
6. It's going to be very difficult to separate the current paths and sources. Lots more measurements may help.

 

[Skogsgurra]

This looks like one of the problems we have had with several applications where traction is by means of a VFD. There, you have a lot of funny things going on and it is usually not possible to measure with normal current clamps. You need clamps that can handle up to around 1 MHz to be able to see what is actually happening.

The load dependency that you have, is that a load that you can vary yourself? Or is it the effect of a train running on the track? I guess the latter. In that case, it is surely the effect of HF components in the lok's traction system.

I attach a short description of one HF transformer that I use in cases like this. You are free to make or roll your own. No patents or anything! PDF here:

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.