(click here to view the DSTAR gallery of interesting ferroresonance results, including sound clips)
An inherent consequence of the utility industry's shift to higher efficiency distribution transformers is an increased tendency for ferroresonant overvoltages to occur in many everyday situations. DSTAR research has been focused on both defining the problem as well as exploring solutions. To determine the system parameters which may result in excessive ferroresonant overvoltages, extensive full-scale tests were performed on a number of transformers under realistic field conditions. Literally hundreds of switching events were performed on 12.47, 24.94 and 34.5 kV padmount transformers, including both grounded-wye and delta-primary units. This project overwhelmingly proved that the efficient transformers now used by the industry are much more susceptible to ferroresonant overvoltages than those of a decade or more ago. For example, crest overvoltages 2.35 times normal crest were measured in DSTAR field tests where a conventional silicon-steel core 225 kVA, 24.94 kV wye-wye padmounted transformer was switched at the riser pole with only the capacitance provided by 220’ of cable to support the ferroresonance!
As you can hear in the sound clips provided in the DSTAR Ferroresonance Gallery, transformer noise during ferroresonance is very unusual, and can often lead field crews to believe that the unit is failing. Actually the noise is due to changes in the core dimensions as the core goes in and out of saturation.
DSTAR investigations of ferroresonance disproved several widely-held beliefs regarding this complex overvoltage phenomenon. For example, a transformer's core loss and not its exciting current is the best indicator of how much cable can be safely single-phase switched with the transformer. The research results were the subject of a recent IEEE paper which received considerable attention.
Although this effort yielded a significant amount of new information regarding ferroresonance susceptibility, mitigation techniques were also examined. For example, a minimal amount of load is sufficient to eliminate ferroresonance. Another research project was performed by DSTAR to define the duty placed on metal-oxide surge arresters by ferroresonant overvoltages. The research has disproved the widely-held belief that a high ferroresonant overvoltage will immediately fail a gapless arrester. Because of the very high impedance of the ferroresonant circuit, arrester heating is slow. The DSTAR investigation has shown that arresters may be considered as a viable means to control these overvoltages in many circumstances. Extensive guidelines have been prepared which indicate the limiting system conditions for which various types of arresters can be expected to survive several minutes of ferroresonant overvoltages, as during transformer switching, and indefinite exposure, as might occur from fuse operation which open-phases an unloaded transformer and cable. The convenient form of these guidelines is depicted in the figure below.