Today’s power quality (PQ) monitors are capable of capturing more data in a second than could be humanly processed in a day. Techniques employed, such as “report by exception,” can help to minimize data, but it still doesn’t answer the question, “Is this meaningful information?” This doesn’t prevent the user from applying the “test of reasonableness” to the data collected.
Remember that it is primarily the firmware or software that makes the PQ monitors determine what data to capture. These programs, though they may involve thousands and thousands of lines of code, still cannot match the human brain’s ability to decipher what is real and what is not. To try to encode all of this expertise and every possible combination of events that occurs in the real world into an instrument would make the development project last seemingly forever. Depending on the monitor, those things deemed most likely to occur are accounted for.
In past articles, we have used Kirchoff’s and Ohm’s Laws to understand what causes sags and swells, as well as harmonics and transients. These same rules can be used to determine if what you are seeing in the data recorded makes sense and passes the test of reasonableness or if some other explanation is needed. For example, the monitor is capable of recording the maximum harmonics distortion for voltage and current (Vthd and Ithd), along with the time and date of occurrence. On the display, it shows that last Friday at 10:02 p.m., the voltage distortion hit a peak of 8.9 percent, whereas the Vthd was typically between 3.2 and 3.8 percent. This sounds like a potential problem for your facility, as triple the distortion can lead to problems with additional power dissipation in motors and transformers.
However, a closer examination of the data shows that there was also another PQ event during the same time. A series of sags and transients were occurring from 9:50 p.m. until around 10:20 p.m. The waveforms of these disturbances show considerable distortion that is typical of arcing, such as when a distribution wire shorts to a tree branch. Then, you remember that there was a storm that night around the same time. Whereas it would still be wise to notify the electric utility company to remedy the problem, it is not a harmonic concern for your facility. The maximum distortion was during a PQ disturbance, not a steady-state condition, for which harmonics are of concern.
Figures 1 and 2 show another example of the need to apply the test of reasonableness. This data came from a chemical plant in western Pennsylvania, where they were monitoring a three-phase delta circuit at a second-floor distribution panel. The root mean squared (RMS) plot of the voltage shows a pretty uneventful graph, until around 9 a.m. of the second day of monitoring.
A closer examination of the data at that time shows severe notching of the voltage waveform. Many of the transients go down to zero volts. The point-on-wave pattern is random and only on the positive half. At the same time, though not pictured here, there was no change in the current waveform captured on the same circuit.
Ohm’s and Kirchoff’s Laws would suggest that a load-side-generated disturbance would need a very, very large increase in current to result in zero volts. And if it were a source-side problem, one would expect the load current to show some change, although the type of load would determine the effect. Instead, there was no change at all in the current.
There was some prolonged head scratching as people tried to find a logical explanation for the recorded data. It turns out that the user was able to correlate the fact that large horsepower motors started up that day in the plant at 9 a.m. in a location right below the point of monitoring.
The clip leads of the voltage probes had been placed on oxidized screw heads. When the motors started up, the vibration caused the clip leads to lose electrical connection with the screw head, resulting in zero volts at the voltage probe for the monitor’s inputs. The notches weren’t real as far as the power system was concerned; they were just a reflection of the vibration of the motor. The oxidized contacts are alleged to produce a diode-type effect, which conducts in only one direction; hence, the effect was only on the positive half of the waveform.
So remember: If you have data that tends to require you to rewrite the laws of physics to explain it, then apply the test of reasonableness before attempting to capture the Nobel Prize. EC
BINGHAM, manager of products and technology for Dranetz-BMI in Edison, N.J., can be reached at 732.287.3680.