Published In June 2001
I recently received a telephone call from a power quality (PQ) engineer with an electric utility company to help settle a debate on the proper classification for a zero- voltage condition that lasted 45 cycles. To me, this underscored the fact that the Institute of Electrical and Electronics Engineers, Inc. (IEEE) Std. 1159 definitions are still not well known, despite the fact that the document was published over six years ago. The most common power-quality phenomena at most locations is what we refer to as a Root Mean Squared (RMS) variation, or a change in the magnitude of the voltage, which is calculated using the RMS method. The RMS calculation gives a DC equivalent value of sinusoidal or other periodic signal by calculating (for each cycle) the square root of the mean value of the squared value of the digitized samples of the waveform. Based on the magnitude of the voltage and how long this change from the nominal value lasts, there are different categories or labels used to describe the disturbance. Several IEEE and International Electrotechnical Commission (IEC) committees are working on how to actually carry out these measurements in a repeatable fashion over a plethora of different types of events, but that is beyond the scope of this discussion. As shown in Table 1 below, the first characterization is whether the disturbance lasts more or less than one minute, called long or short duration, respectively. RMS variations beyond one minute fall into what some utility people refer to as reliability issues, instead of power quality, since most automatic fault protection schemes would have cleared whatever faults they could by that time. After that, it often requires manual intervention due to a permanent fault or system malfunction. Within the short-duration variations, there are three further divisions of duration, which also have origins with the source of the disturbance and the system reaction to it. Less than 1/2 second increments in duration are called instantaneous, 1/2 to 3 seconds are considered momentary, and 3 seconds to 1 minute are temporary. For long-duration events, “sustained” becomes the adjective. The magnitude is determined by the remaining voltage, not by the difference from what was before the event as compared to what is during the event. The IEEE terminology uses “per unit (pu),” which is an engineering term. It can be thought of as “percent of nominal.” If the system is a 480V circuit, the nominal value is 1.0 pu, and the 432V would be 0.9 pu or percent of nominal. The 48V would be 0.1 pu. If the disturbance has a magnitude of less than 90 percent of nominal but more than 10 percent, we call it a sag (or a dip in Europe). Below 10 percent is an interruption, and above 110 percent of nominal is a swell. A swell used to be called a surge, but people use surge suppressors to clamp transients from lightning and such, so a new word was chosen. Hence, a typical sag that was caused by a fault on parallel feeder on the distribution system would be an instantaneous sag. An example of such is shown in Figure 1a (waveshape) and Figure 1b (RMS timeplot). The system protection breaker operated after 28 cycles, so that is how long the sag lasted. Since the monitoring point wasn’t on the same feeder as the breaker and the fault, when the breaker opened, it isolated the monitoring point from the fault, and nominal voltage was restored. If the monitoring point had been on the downstream side of the faulted feeder, the result when the breaker operated would have been an interruption until it reclosed, as shown in Figure 2, where there is a sag followed by an interruption. The breaker didn’t cause the fault, but it did have an effect on how long it lasted. My friend at the utility company experienced a “momentary interruption” on the transmission system. While the words alone without any numerical details may not seem significant, the two-word phrase is enough to convene an image of what it was, possible causes, and that there were probably a lot of companies where processes were interrupted, unless they had a UPS or other form of backup. Few processes can ride through near-zero-volt conditions. The good news is that it was cleared automatically, so things could get back to normal as soon as they restarted their processes. Next month’s column is the final part on distortion and the remaining categories of PQ phenomena. BINGHAM, manager of products and technology for Dranetz-BMI in Edison, N.J., can be reached at (732) 287-3680.