In this third part of my series on the most common power quality phenomena, I group several IEEE 1159 categories together under the general description of “distortion.” Though one might argue that all power quality phenomena result in distortions of the voltage sine wave, these are the more steady-state variety, including harmonics, interharmonics, notching, unbalance (or imbalance), voltage fluctuations (or light flicker) and noise.
While we relied on the rms plots of voltage and current along with waveform graphs of the factors in the first two articles covering rms variations and transients, these distortion phenomena often require some unique tools to help determine the source. The rms values will not provide much information at harmonic and interharmonics because the rms value can be similar with and without the distortion, as it is repetitive each cycle by nature (see Figure 1). The voltage distortion looks like an oscillatory transient occurring four times per cycle on all three phases, with the distortion on two phases being coincident each time they occur. But we need to look at the current waveform to get a better idea of what is really happening.
The voltage waveform for Phase A is displayed along with all three-phase currents in Figure 2. The current waveforms have the “double camel hump” waveshape that may be familiar to some readers as being the current drawn by three-phase rectified power supplies. If you look closely at where the notches occur, you can see that the A phase current changes magnitude abruptly at the same instant as the voltage notch, along with one of the other phase currents also changing abruptly. This is the commutation period, when the rectifying device controlling Phase A’s current (often a silicon-controlled rectifier, or SCR) doesn’t turn off at the instant when the control signal says so but overlaps for an instant while the next phase rectifier is turned on. This creates a short circuit between the two phases for an instant until the initial one turns completely off. A short circuit results in a lot of current flow, which means a significant voltage drop across the source impedance and subsequent abrupt reduction in the voltage at the load, the notch. It “rings” or oscillates because of the inductance and capacitance in the power supply circuit and the supply circuit.
The highly distorted current waveforms are very rich in harmonic content, and use of the fast (or discrete) fourier transform (FFT or DFT) gives us the clues to the source. The dominant harmonics in Figure 3 are 5th and 7th, 11th and 13th, 17th and 19th, etc., which follows the h = n × p +/- 1 rule, where h is the harmonic number, n are integers starting at 1, and p is the number of poles or paths of conduction. A three-phase rectified circuit has 3 × 2, or p = 6. That yields the results above. So both distortions, the voltage notches and current harmonics point to the same source.
Looking just at the waveforms and the magnitude of the harmonic distortion wouldn’t make much sense of figures 4a and 4b, which have the exact same magnitude of each of the harmonic current yet look quite different in shape. Figure 4a shows the “Batman” waveform from decreasing amounts of 3rd, 5th, 7th, 9th and 11th harmonics without any phase-angle shift. Figure 4b shows the same amounts, but there are different phase-angle values, including 180 degrees for the 5th to produce a more typical single-phase rectified power supply, such as when using a laptop or other IT-equipment battery chargers.
While finding unbalance using rms values is possible for some situations, it doesn’t work in all cases, especially when there is a phase unbalance and/or distorted signals. The “sequence components” technique works in all cases, rather than the NEMA MG-1 extracted definition of the maximum rms deviation divided by the average of the three phases. Figure 5 shows an example on a phasor plot where the current unbalance is both a magnitude and phase-angle unbalance. Few, if any, people could see the phase unbalance by looking at the waveforms alone.
In some cases, the voltage fluctuations that result in light flicker can be observed in the voltage rms and waveform plots. But it is impossible to know how severe the light flicker effect is. The frequency of the voltage’s modulation is the key element that time-domain plots don’t reveal, and they cannot model the human eye/brain response to it. Flicker meters or power quality analyzers that comply with IEEE 1453 or IEC 61000-4-15 have the capability to calculate a parameter called perceptibility short term (Pst). If that value is greater than 1, most people will perceive it. The parameter doesn’t indicate at what level some types of process control equipment may have problems, such as extruders of fiber, but it is a reasonable number for humans. And noise is a somewhat random distortion that requires a high-speed sampling instrument and sometimes a spectrum analyzer to determine the frequency content to lead to the source.
The knowledge of these three articles can give you the edge in solving the majority of power quality-related problems. Just these few simple rules can go a long way to happy customers (and extra cash).