Keeping the Lights On

November 1, 2003 By: Paul Stergiou, David Kalokitis GPS World

GPS and Power Grid Intermesh

Sometimes it takes a disaster of the magnitude of the terrorist attacks on the World Trade Center to bring home the need for new systems. In the wake of that event and its elimination of two key substation resources in lower Manhattan, Consolidated Edison (Con Edison) of New York found that the switch to digital communications made it impossible to use conventional, proven means of responding to a desperate need for new power. The traditional intermesh method for bringing substations on line required analog phone lines, and these no longer existed.

Con Edison teamed with Sarnoff Corporation, a Princeton, New Jersey–based research and development company formerly known as RCA Labs, to find a new approach to intermesh, the intermediary period when power load is transferred between area substations. Ultimately they solved the problem with the same digital technology that had made the old method unworkable.

The new methodology has provided a valuable tool for safely and effectively adding resources to the grid, and an energy industry global honors program recognized it as an outside-the-box solution.

Destruction and Resurrection

Con Edison lost two lower-Manhattan substations during the 9/11 World Trade Center disaster. As quickly as possible, Con Edison temporarily transferred load to other nearby substations, so that customers were unaffected. This formed a workable solution for September through the spring because the peak of the summer cooling season had passed. But the area urgently required a new substation to come into service before the following summer, when the cooling season would expose that section of the city to power shortages.

GPS Antenna, controlled clock, modem, and surge arrestor used in the intermesh

Con Edison swung into action, designing and constructing a new substation, Seaport No. 1.

Bringing It Online. Connecting a new substation to an operating, loaded network such as the one on Manhattan is a complicated task. Engineers must carefully establish power flow from the new substation while reducing power flow from the existing substation. Con Edison needed to transfer the power load from the Leonard Street substation to Seaport without interrupting the load and causing a power outage.

To avoid any complications or a power outage during an intermesh, it is necessary to precisely measure the phase displacement and voltage magnitude difference between the two substations. The phase displacement is the difference between the phases of the 60Hz sinusoidal waves at both stations.

The phase displacement between any two substations can vary widely due to a variety of factors such as station loading, alternate transmission source, and varying generation output. When connecting a new substation or transferring load between existing stations, the conditions must be adjusted to avoid a large circulating current flow that could trip breakers or even damage equipment and interrupt service. For Con Edison, because of the low impedance, the phase displacement between the two stations needed to remain almost perfectly matched; the difference could not exceed 3 degrees.

Because the load at Seaport would increase and the load at Leonard Street would decrease, Seaport No. 1's 13.6-kV bus would have to be leading in phase between 0-3 degrees and remain 200 volts higher than Leonard Street No. 1's. This would ensure that the network protectors (NWP) and station breakers would close in and remain closed for the duration of the intermesh.

(Click on image for larger view.) Figure 1: The phase of the AC waveform is continuously sampled and digitized. In the example above, the equipment at substation A reports Phase = 90 degrees at 11:01:13 UTC. At substation B, the report is 180 degrees at 11:01:13, hence the difference is 180-90 = 90 degrees. To keep it simple for the operator, we normalize the output to ±180 degrees. The equipment reports a phase value between zero and 360 degrees. Let’s say substation A reports 270 degrees and B reports 20 degrees. Since we use B as the reference, 20-270 = -250. We’ll add 180 degrees to that and represent the phase difference as -70 degrees.

Simple instrumentation routinely accomplishes relative phase measurements between two points. Connect a phase meter across two phases in a power system and the phase can be read directly. However, the Seaport and Leonard Street substations lie approximately 1.2 miles apart, with no direct path to connect the instrumentation. Traditionally, Con Edison measured phase displacement over copper phone lines. Leased from the phone company, the copper lines connected each substation to a central control center. The phone lines provided the direct path for relative phase measurement. Engineers understood the process well. Using these lines to monitor phase angle required the use of a cumbersome, time-consuming method to calibrate (zero-out) the phase delay caused by the long copper wires. The latency factor involved in this method varies by length of copper wire. In any event, the engineers who developed this method have all since retired.

Time Marched On. During the past decade, telecom companies had replaced their century-old copper lines with fiber optic cables. Fiber optics provided enhanced bandwidth to users, a necessity for companies in New York's financial district that plow through endless amounts of data.

The increased bandwidth is made possible because fiber transmits information as light instead of electricity. As a result of this fundamental difference, power companies like Con Edison cannot measure phase displacement via fiber optics because phase and phase displacement are characteristics of electrical current (electrons) but not of optical transmission (photons) technologies. Photons cannot travel on copper lines - no longer available anyway - and electrons cannot travel on fiber. The process would require digitizing the alternating current (AC) waveform and transmitting it with careful accounting of timing delays and offsets, making it a difficult system to calibrate.

GPS Candidate. New York's daunting infrastructure made it economically prohibitive to lay down copper wire between the two stations, forcing Con Edison to look for a wireless solution. Loran C and GPS appeared to be the only viable candidates.

While Loran-C is based on accurately timed transmissions from a group of base stations, developing an accurate time stamp at a given location in a large city is not easy. Con Edison turned to the Sarnoff Corporation to develop a GPS solution.

Sarnoff staff determined that GPS timing could serve as an independent reference for measuring the absolute phase displacement. In essence, a digital signal processor could sample the 60Hz power waveform and associate an accurate time stamp with the voltage and phase of the waveform. The information would go to a control location to determine the difference in phase angle.

Figure 2. (Click on image for larger view.)

Typically, a GPS clock provides a 1-pulse-per-second (PPS) output, accurate to 1 millisecond. This still did not guarantee the accuracy necessary for a successful intermesh. In the course of a millisecond, the phase of a 60Hz signal advances 21.6 degrees, far in excess of the accuracy requirements for an intermesh. The 1-PPS output was not suitable for triggering a phase measurement.

The Sarnoff team solved this problem by using the higher accuracy timing available at the internal processor of the GPS clock. The GPS-controlled clock can be outfitted with a digital signal processor triggered by the internal timing, resulting in a time stamp accurate to 1 microsecond. The phase of a 60Hz waveform only advances about one fiftieth of a degree in 1 microsecond. The digital signal processor can sample the AC waveform and measure phase accurately to about 0.3 degrees. Add it all up and the system can provide sub-degree accuracy where it is needed.