It was one of those so-called “overvoltage events.” Somehow overhead wires from the top and bottom circuits made contact. The cause was a tree contractor removing a diseased tree at a private residence. One particularly troublesome limb that had been cut had fallen into the line, bridging the top phase with one from the lower circuit. The result was predictable. The isolators on several surge arresters were blown on the lower distribution circuit and some transformers failed, too.

Several thousands of dollars in damage occurred to Dominion Virginia Power equipment. Also, TVs, heat pumps and other appliances of many customers served from the lower circuit were damaged. Appliance repair and replacement costs of US$1,000 to $2,000 or more per customer were typical. Some customers who were home at the time were frightened by the overvoltage event. Others were upset when they realized it could happen again. How can this complex overvoltage problem be solved cost effectively?

Overbuilt Circuits

As the incident illustrated, with two overhead circuits of different voltages on the same poles, contact between the two circuits can and does occur. When this happens, damage and failures occur to both utility equipment and customer equipment. Dominion experiences 13 sustained overvoltage events annually on its distribution system. With more than 600 miles (966 km) of overbuilt construction, mostly 34.5 kV over 13.2 kV, the annual incidence rate works out to be 0.022/mile (0.014/km).

Traditional approaches to minimize or eliminate the risk of these overvoltages have met with limited success largely due to their high costs or operational difficulties. Some of the traditional solutions include:

• Relocating or undergrounding one of the circuits

• Converting the lower-voltage circuit to the same voltage as the upper one

• Insulating the lower circuit.

The combination of high cost and a low incidence rate makes these types of projects undesirable to fund. Although the three measures are quite expensive, utilities often resort to one of them in areas where repeat overvoltage incidents occur.

What can the customer do? Surprisingly, the answer is the customer can do very little. Some customers are often lured to purchase the more expensive home surge protectors on the market because of their warranty, which is normally in the region of tens of thousands of dollars. Unfortunately, the fine print on most of these devices contains a note that the warranty does not cover damage associated with sustained overvoltages.

The Fundamentals

Once a conductive path is established between the two circuits, the higher-voltage circuit dominates. Reclosing usually occurs with both circuits, so the lower-voltage circuit winds up getting energized with the higher-voltage one. Transformers on the lower circuit do what they are supposed to do — transform.

For the two Dominion circuits, the resulting voltage was 2.6 times normal, which would be 300 V at a 120-V outlet. From this level of overvoltage, it is fairly easy to imagine how customer appliances get damaged or destroyed.

The higher voltage impressed on the lower-voltage circuit usually causes some utility equipment on the lower circuit to fail, too. Distribution surge arresters quickly short out and blow their isolators. Several distribution transformers also succumb to the high voltage. Each shot of reclosing results in additional damage to the lower circuit.

The overvoltage problem associated with overbuilt circuits is not unique to Dominion. It is an industrywide problem.

Solution Efforts

The main problem is to be able to detect a sustained overvoltage event involving an overbuilt distribution circuit and to provide protection to both utility equipment on the lower circuit and customer end-use equipment served by the lower circuit.

This problem has been a challenge at Dominion for over a decade. One early idea was to install a normally open recloser on the underbuilt circuit. The load side of the recloser would be connected to ground. In principle, it would operate like a fast-acting grounding switch. Install some voltage sensing and make sure it is not a temporary overvoltage. The last requirement effectively killed the idea. One would not have the luxury to wait very long before equipment was damaged. Besides, it was just too expensive.

Not to be dismayed, the idea of shunting out the lower circuit had some appeal, just not with a recloser. To limit the overvoltage on the lower circuit, why not use an arrester? After all, an arrester operates like a switch, and it does so automatically based on voltage. It would have to be an arrester that did not have an isolator and one capable of absorbing quite a bit of energy.

A station class arrester was a natural fit. To be safer, a high-energy design was selected. For this application, the station class arrester would function as a sacrificial arrester (that is, it would short out for a sustained overvoltage and prevent damage to the utility equipment on the lower circuit). Once it operated, it would have to be replaced.

It was not clear whether this approach would provide sufficient protection for customer end-use equipment and appliances, particularly electronics. If they were not protected, it would be a show-stopper. The Information Technology Industry Council (ITIC) developed a voltage tolerance envelope applicable to 120-V IT equipment. The overvoltage portion of the ITIC curve indicates, for example, that electronic devices are likely to survive if the voltage is limited to 2 per unit for 1 msec or less. From the arrester current-resistance characteristics and the available fault current, it seemed reasonable to expect that a station class arrester could limit the voltage to around 2 per unit. The time duration was unclear.

Surge arresters are normally applied such that their temporary overvoltage (TOV) capability is not exceeded. In this case, the plan was to apply a 10-kV arrester knowing full well that contact with 19.92 kV would put it well beyond its TOV capability. It was expected that the increased duty would quickly (much less than 1 msec) short out the arrester's metal-oxide-varistor blocks. Would the arrester stay intact and remain shorted for a fault current level of around 6000 A in addition to two-shot reclosing? More importantly, would it protect? Then there was the issue about energy ratings, which for station class arresters are based on impulse tests.

Extrapolating these values to something more meaningful that relates to a sustained fundamental frequency overvoltage also was a gray area, and this uncertainty resulted in selecting an energy rating higher than the standard station class arresters used by the Dominion substation group.