The protection of high-voltage shunt capacitor banks

Bibliography: pages 86-88.The use of shunt capacitor bank equipment is essential if a utility wishes to control the flow of reactive power effectively. The most significant results stemming from this will be lower losses on the system and an increased power transfer capability. Thus it is important that the methods used to protect a shunt capacitor bank will ensure that the bank is available when required. While the more common shunt capacitor problems are related to capacitor unit failures, conditions such as bank over currents, surge voltages and harmonics can cause extended undesired conditions. Today's protection methods are able to remove a shunt capacitor bank from service before extensive damage is done, although the location of the faulty capacitor units will not be known (if this was in fact the reason for the protection tripping the bank). This thesis explores the subject of improving the protection of high-voltage shunt capacitor banks, specifically with respect to the detection of unhealthy fuseless capacitor units. An extensive literature search was carried out on the theory pertaining to the protection of shunt capacitor banks, and a model of a fuse less shunt capacitor bank was built in the laboratory to better understand the failing process of an element within a capacitor unit. The changes in the capacitor unit's current and voltage profiles, as well as those of the remaining healthy capacitor units, were monitored as an element failure was simulated (whereby the element forms a solid weld, or short circuit).Stemming from these experiments, it was found that where a bank consists of strings of units with no interconnection between the units of different strings, an element failure in a capacitor unit would cause a significant decrease in voltage across the affected unit. This voltage change could be used to identify when elements are failing in capacitor units, and the location of the unhealthy unit could also be determined. One potential method would be to have capacitor units with built-in voltage transformers attached across each element section in the unit. As element failures occur either send this information to ground level, where it can be read by a microprocessor relay device, or have a display on the outside of the capacitor unit. In the case of the change in unit current, it was found to be very small and thus had no function for detecting unhealthy capacitor units