Zig Zag Overvoltages

[mls1]

A post a few years ago discussed the calculations associated with zig-zag grounding transformers. In it one user posted the following:

"You might need to reconsider an almost entirely reactive impedance for the grounding. It is generally considered a bit of a no-no to have reactance grounded systems, due to the problems with power frequency overvoltages under arcing earth fault conditions. I don't recall all the details, but the mining industry in South Africa had some problems with it in the 1970s. There were a few papers in the SAIEE journal - people like a Dr Pretorius at the CSIR were involved with it back then, I think. Eastern Natal Region of Eskom used an NEC / NER combination at 11kV to keep earth fault levels to 300A, and the R componenet was to ensure that the system was not reactively earthed. "

We are involved in a project in which a 22.9kV system with a zig-zag grounding transformer has experienced some surge voltage problems. The zig-zag is directly grounded without a resistance so that minimum ground resistance will be leakage inductance only. I have been searching for the above papers describing possible power frequency overvoltages during arcing earth fault conditions. This seems like a plausible cause if there is phase angle displacement on each arc fault. I did search SAIEE (I am in the USA) but could not find anything that old. Can anyone point me to where I might find the above references or any other references that may cover arcing fault transient voltages on zig-zag grounded systems? Much appreciated!

 

[waross]

One of my instructors many years ago had experience working in mills with ungrounded 440 Volt delta systems. They experienced a couple of overvoltage events with motors failing all over the plant. As I remember the cause was an arcing ground fault part way into a winding in a transformer. It acted like a spark gap transmitter and developed a high frequency. The transformer acted as an auto transformer and boosted the high frequency voltage up to several thousand volts. The transformers and motors had a very high impedance at the fault frequency so the high frequency high voltage floated on the 440 Volt lines and equipment started experiencing failures at the winding ends.
This points to two effects.
1. An arcing fault may generate high frequencies.
2. The high frequencies will see motors and transformers as very high impedances.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[prc]

The story waross mentioned is well described in Donald Beeman's 1955 classic" Industrial Power System Handbook".
The condition to limit transient over voltages from arcing faults is Ro/X0> 2.Please check this. The grounding resistance shall be more than 2 times (reactance of grounding transformer +zero sequence reactance of line transformer) Please see above hand book and GE Industrial Power System data book SECTION 0.3-

http://digital.lib.uiowa.edu/cdm/ref/collection/uidb/id/10896

The overvoltages are created when system capacitance to ground reach the grounding reactance value,setting in resonnance.
I doubt whether your problem is due to transients. Recently I came across some grounding transformer failures.Root cause was found to be poor quality of core laminations used in the transformer. Please share failure mode details.

 

[waross]

Thanks for supplying a good reference for my anecdote prc.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[mls1]

The plot thickens a little bit here. The issues do not appear to be related to transient overvoltages from arcing ground faults. What is really at issue is the relay setup for ground faults on the 22.9kV system that are causing fairly long sag voltages (over 30 cycles) in the plant. There is another issue in this system that I'd appreciate some feedback on. The zig zag is solidly grounded at the substation and has a fairly low impedance. However, a neutral was not run with the overhead conductors. For the utilities we work with it is standard practice to run a neutral and ground it ~4 times for each mile. With this system not having a neutral, all ground faults must travel entirely through earth. The operators have indicated a wide range of phenomenon occurring during ground faults such as 480V HRG systems alarming, etc. This system has been in service like this for a very long time but it would seem to me that a neutral should be run out with at least the main backbone lines and grounded along the main run to provide a reduced and more stable return impedance. Would I be mistaken in that logic?