This summary concerns one of American Electric Power (AEP) 765-kV transmission systems (“the System”) in Southern Ohio, located in the vicinity of AEP’s 2600-MW Gavin Plant. On or about December 1, 1979, the subject system experienced a major disturbance as a result of several extraordinary events. The sequence of events began with the failure of a 765-kV surge arrester at Gavin, to which the system responded as normal. Contemporaneously however, there were also relay and communications malfunctions which resulted in two islanding conditions for a 1300-MV coal-fired generating unit at Gavin, which was eventually reset to synchronize with the rest of the interconnected system.
Islanding conditions refer to a condition in which a distributed generator continues to power locations in spite of the fact that the grid power for the largest electric utility is not functioning. In the instant scenario, the malfunctions at the Gavin plant resulted in islanding conditions, whereby the distributed generators at Gavin were still able to power the three circuits, despite lacking the electrical grid power actually required to do so.
By design, the Gavin Plant was AEP’s second largest generating plant on the AEP system, with two 1300-MV coal fired units situated therein. The Gavin Plant was integrated into the overall system by way of three 765-kV circuits: the Gavin-Amos circuit, the Gavin-Marysville circuit second, and the Gavin-Mountaineer circuit. The Amos plant was 40 linear miles southeast of the Gavin Plant, which was the largest plant on the AEP system, generating 2900 MW of power.
On the day of the subject incident, where a 765-kV surge arrester failed at the Gavin Plant, a Phase B arrestor on the Gavin-Marysville line sparked and exploded, leading to the failure. This explosion resulted in a complete shunt reactor transfer bus toppling from the supports and also caused a permanent fault on the existing system. At the time of the incident, no inclement weather was occurring. Nonetheless, Gavin Unit Number 1 went out of service and Gavin Unit Number 2 was operating at an output level of 1150 MV. With the AEP internal load at 8000 MW, these totals only equaled 60% of the projected winter peak load for that lead. At the same time, AEP was delivering an additional 3000 MW from internal sources to systems around the Southeastern and Eastern areas of Ohio, transferring 1000 MW from nearby neighboring systems.
Inspections of the relevant data in the aftermath of the event showed that two distinct events took place sequentially, and that each contributed to the disturbance. The first, was the failure of the surge arrester at the Gavin-Marysville 765-kV circuit, which led to the permanent phase-to-ground fault. And, the high speed re-closure of that circuit was unsuccessful because of the permanent nature of the fault. The second event was the Gavin-Mountaineer circuit tripping because of an error resulting from the carrier blocking signal due to the original Gavin 1 failure. The resulting system condition thereafter, led to the Gavin Unit Number Two being adversely affected.
In analyzing the sequence of events and looking at the dynamics of such an event, involving multiple units and multiple events, it should be noted that the transmission outlets associated with the power plants on this system were typically operating with double contingency criteria. This would tend to suggest the potential for the system to remain at least somewhat stable following such an event sequence, as this double contingency criteria requires that all of the generating units have to remain stable after a disturbance, even where one transmission element was actually out of service prior to the subject disturbance. In off peak conditions, like that of the Gavin incident of 1979, an analysis of the system’s dynamic performance would tend to suggest that it was in a steady state of operation, and that the Unit was actually quite stable. Conversely however, had the disturbance actually taken place during peak-load conditions, the Unit would most likely have been left in an unstable state.
As such, one might anticipate anomalies like overvoltage situations in the midst of sequential events occurring within the system during peak load times. As gleaned from computer programs used by AEP in their load planning processes, the transient stability program indicated that there was a definitive need to better account for overall load representation, and of course without such accounting, overvoltage situations may be expected.
The expected electromagnetic response would appear to be tied to interactions such as torque accelerations and decelerations, and speed deviations, during events such as islanding and resynchronization. This is a real unknown however, given the lack of monitored data in some areas at the time of the incident, such as speed and rotor angles, which would make it more difficult to really understand and predict the net impact of such disturbances on the generating units and their electromechanical functions.
Generally speaking however, a power system’s ability to recover from such unexpected events is all contingent upon the functionality of its infrastructure. A sophisticated, functional, and fully operational monitoring system and protocol must be in place for all transmission, distribution, and generation facilities, including plants like Gavin.
Equally critical is the speed of the response to an event (fast versus slow), as well as a thorough analysis of each and every “major” disturbance within a given system, so that the net effects of such an event can understood and predicted with greater accuracy. This would give AEP a better idea of how system stability is impacted.