This waveform shows the highly chaotic nature of distribution transformer ferroresonance.  The shape and magnitude of the voltage wave change rapidly and spontaneously.  The waveform is from an open phase of a 300 kVA 24.94 kV silicon-steel wye-wye distribution transformer, with 300' of  #1/0 cable connected to the open phase.

Click here to download a clip containing the sound of a distribution transformer in chaotic ferroresonance (.wav file; 334 kv)

Ferroresonance can create rather severe overvoltages, even when switching ordinary wye-wye transformers with everyday lengths of cable between the switching location and the transformer.  This waveform  shows  overvoltages produced by switching  a conventional 225 kVA 24.94 kV silicon-steel wye-wye distribution transformer, with 220' of  #1/0 cable connected to the open phase, during DSTAR field testing. Initially, this transformer was operating with one phase, only, closed to the source.  At 80 ms on this plot, the the second phase was closed, producing voltage crests of 2.35 p.u. of rated crest voltage.  

Click here to download the sound of a distribution transformer at the closing of the second phase during energization. (.wav file, 227 kb)

A common assumption is that, because ferroresonance can create a high overvoltage, failure of an MOV arrester subjected to these overvoltages is certain.  DSTAR research has shown that the open-phase of the transformer, where the overvoltages occur, provides a weak, high-impedance source.  The clipping action of the arrester sometimes tends to "take the wind out" of the ferroresonance, and overvoltages sometimes subside without the arrester needing to conduct again for some period.  Shown below are open-phase voltage and arrester current for such a case.  There is a burst of ferroresonant overvoltage, which causes the arrester to conduct a very small current for a very few impulses which are only a few hundred microseconds wide.  The clipping "tames" the ferroresonance as seen at the right side of the waveform.   After a period of time, which can be on the order of many seconds, another burst may occur.  The research has concluded that arresters may fail under certain circumstances, due to protracted heating, or can survive extended periods in other situations.