Is This Normal? Power Quality Data

When people send in data from a power quality monitor to be reviewed, some common questions include “Is my site normal?” and “Does this data look OK to operate my facility?” Such questions usually get the same ambiguous answer: “It depends.” The quality of electric power as supplied and how much is consumed is not one-size-fits-all.

A multiparameter graph used in the international standards community, especially in Europe, shows the susceptibility and immunity levels of equipment, along with the supplied and planning quality levels of power supplied. It is not shown here because it usually takes a half-hour of explanation before someone who hasn’t seen it already can understand what it all means. But it captures, in a single image, the concept of “normal” not being a single value of a single parameter to be used in a pass/fail decision of normal/abnormal.

The first step in analyzing the data is usually to plot the root-mean-square (rms) voltage for all measured phases and neutral-to-ground (if recorded). In a wye circuit where line-to-neutral (L-N) measurements were made, looking at the line-to-line (L-L) values is also useful because these faults will show up much clearer. It is easier than looking at two L-N channels and trying to determine the disturbance’s cause, looking at how well regulated the voltage is, along with the unbalance between the phases, and comparing this to the requirements of the equipment powered by the voltage source. Some equipment can do just fine with a 5 percent voltage regulation and 3 percent unbalance, and some can tolerate more (others not so much). In most grid-supplied distribution systems, we typically find even better regulation and a frequency regulation better than 0.05 hertz. Is that considered normal?

Where significant rms voltage variations are displayed, either sags or swells, plot the rms current against the voltage. When monitoring at the point of common coupling (PCC), if every time the voltage goes down, the current goes up significantly, then the problem is most likely downstream or inside the facility. The figure above shows such an example, after zooming down to show the rms voltage computed in half-cycle steps.

Do the disturbances occur randomly or in a particular pattern? If this usually occurs at the same time of day, such as early in the workshift, what equipment starts up at that same time and what does the current inrush look like? Do the waveforms for the voltage and current look significantly distorted during the rms variation, such as when a transformer core gets saturated from overcurrent or unbalanced current conditions?

If the current decreases or just gets slightly greater during the sag, check the duration of the sag. If it lasts 3–10 cycles, and starts and ends with a step change in the rms voltage, it is likely caused by the operation of circuit protection on the distribution lines. A call to your local energy provider may help correlate the event on its system, though there usually isn’t much that you can do about that unless it becomes a repetitive problem.

Next, plot the voltage and current total harmonic distortion over the monitoring period. IEEE 519 provides recommended limits at the PCC. But this doesn’t mean that these limits are the normal value to expect, nor are these limits applicable within a facility. The last 30 years of monitoring have shown that the system generally doesn’t become unstable when the old 5 percent voltage total harmonic distortion (Vthd) limit is exceeded. Harmonic-based power losses do go up, but whether these result in misoperation or shortening of equipment life depends on the equipment affected. If the Vthd is high and the current total harmonic distortion (Ithd) is low, then the problem likely originates on the utility side. Remember that a high Ithd doesn’t mean much, unless the total harmonic current is a significant part of the total current capacity on the circuit. For example, 0.75A out of 1A of current (75 percent Ithd) on a 20A circuit means not a problem, while 10A of distorted current on a 15A circuit almost always does.

Other parameters, such as real power (W), apparent power (VA), reactive power (var) and power factor (PF=W/VA) are also good to check against system ratings and prior data to see if things are degrading. Also, check the demand and energy profiles at the same time to find ways to reduce the electric bill.

With the addition of local renewable- energy sources, such as solar or photovoltaic (PV) panels on the facilities’ roofs, the process can get a bit more interesting, and “normal” takes on a whole new dimension. In a future issue, I will look at the effects on the quality of the supply and operation of a facility that we recorded the data for two months prior to PV system being energized.