An Overview of Power Quality Issues

Bill Rogers, December 2005



The quality of electric power has become of supreme importance to electric utilities and their customers. Modern equipment is more sensitive to power system anomalies than in the past. Microprocessor based controls and power electronics devices are sensitive to many types of disturbances. The non-linear nature of power electronics causes harmonic currents that result in additional thermal loading in equipment, interference with communications systems, and possible misoperation of control systems. And finally, the increased use of shunt capacitors for power factor correction creates resonances that can amplify transient disturbances and harmonic distortion.


Users of sensitive equipment have become increasingly aware of the role played by electric utility disturbances in regards to proper operation of their equipment. If the lights go out and equipment shuts down for a while they call it an “outage”. If the lights blink and equipment is generally unaffected, they’re likely to call it a “surge”. Unfortunately, the terms “outage” and “surge” have no correlation with standard definitions of the phenomena involved. The purpose of this paper is to shed a little light on the nature of electric system disturbances and perhaps help the reader address problems more effectively.


Power quality problems tend to fall into one of three general categories: 


Transient disturbances
Fundamental frequency disturbances
Variations in steady state power


Transient Disturbances


Transients can be either impulsive or oscillatory. An impulsive transient is unidirectional, i.e., the current or voltage wave is primarily of a single polarity, and it’s typically less than 200 microseconds (uS) in duration.


A common impulsive transient is lightning. Currents induced by a lightning strike can be on the order of tens of thousands of amps and they can reach peak amplitude in as little as 2 or 3 uS. The rapid rise has frequency components in the tens or hundreds of megahertz range, so it’s important in the design of grounding and bonding systems to treat lightning as intense radio frequency (RF) energy.


Here’s a typical lightning stroke current waveform:


This one reaches a peak of 20,000 amps in about 8 uS. 


Another example of an impulsive transient is the electrostatic discharge most of us have experienced as we touch another person or object after walking across a carpeted floor (or as we step out of our vehicles). The charge buildup can result in potentials on the order of 15kv. The abrupt release of charge can damage sensitive electronics, thus the invention of the wrist straps that technicians are required to wear when servicing sensitive equipment.


Oscillatory transients are most often caused by switching events on the electric system. An oscillatory transient is a decaying oscillation riding on top of the fundamental 60Hz waveform. Amplitudes are typically 1.5 to 2 times the peak and duration is a few cycles at most.


The frequency of the oscillation gives us a clue as to the source. Frequencies above 3 kilohertz  are associated with power electronics and the energizing of a line or cable. Power electronics (the switching power supply in your computer for example) produce oscillatory transients that repeat several times per 60Hz cycle.


Lower frequency oscillatory transients are associated with capacitor switching. Utilities use capacitor banks to improve power factor (which lowers system losses). Capacitor banks have to be switched in and out of the circuit as the load varies. Utilities employ surge arresters to limit oscillatory transients from capacitor switching, but these don’t always work as advertised.


Here’s a typical capacitor switching transient waveform:




Fundamental Frequency Disturbances


In this category we have voltage sags and swells, long-term variations, and interruptions. These are voltage variations at the fundamental frequency (60Hz in North America).  These problems don’t produce the sudden and sometimes catastrophic damage associated with transients, but they can be very destructive over time.


Voltage sags are due to system faults and the switching of heavy loads (including the startup of large motors) and they can last anywhere from a few cycles to a minute or more. Voltage sags can cause tripping of sensitive equipment such as process controls and variable speed motor drives, and the dropping out of relay contactors. There are no standards for what is acceptable in terms of voltage sag. It’s very important for the customer not to create his own sags through improper wiring and undersized conductors.


Voltage swells are much less common than sags. A voltage swell is most often caused by a line-to-ground fault on a polyphase transmission line or feeder. The ground fault causes a voltage rise on the un-faulted phases. The magnitude of the swell can be as high as 1.73 times the initial voltage. A voltage swell can also be caused by removing a large load or by switching in a capacitor bank that is too large for the prevailing conditions.


Long-term variations include overvoltages, undervoltages, and voltage imbalance. In this area, we do have some useful standards. ANSI C84.1 covers voltage ratings for electric power systems and equipment under steady-state 60Hz conditions. It specifies service voltage ranges and utilization (equipment) voltage ranges for both “normal” and “unusual” system conditions. Here are the essentials:


Equipment must be designed to operate with acceptable performance under extreme steady state conditions of +6% to –13% of nominal voltage.
Nominal voltage is defined in terms of the typical 120/240v 3-wire single-phase service, and is referenced to a 120v base.
Applying the allowable tolerance to the 120v base yields a maximum of 127v and a minimum of 104v at the equipment. This correlates to 254v maximum and 208v minimum for devices connected line-to-line on a 120/240v single phase system.
Tolerance on service voltage is tighter. For the nebulously-defined “normal” conditions, the tolerance on service voltage is +/- 5% (114-126), but for “short duration” or “unusual” conditions it’s +6% to –8% (110-127).


Here’s a handy chart. The “A” range is for normal conditions and the “B” range is for so-called short duration or unusual conditions:



The bottom line is this: For a typical 120/240v single-phase customer, line-to-line voltages at the service entrance between 220 and 254 are considered to be within tolerance. And the voltage at the equipment could be as low as 208. Please keep in mind that this is a minimum standard – equipment manufacturer’s tolerances could be much tighter.  


It’s important to note that nominal voltages such as 110, 115, 220, and 230 are obsolete in the US utility industry. Distribution voltages are referenced to 120v at the substation. It doesn't matter what the actual voltage is on the feeders leaving the station - could be anywhere from 4.16kv to 34.5kv - it's what's on the customer's side of the stepdown transformer(s) at the delivery point that counts. Not every customer needs 208, 240, 277, or 480, but everybody needs 120, so that’s been adopted as the base. Ask a utility man what the voltage is at the substation and you’re likely to hear a number like 126. What that really means is that the voltage on the feeders leaving the station is 5% above nominal.


Voltage imbalance on polyphase systems is caused by unbalanced loads and unbalanced system impedances. It’s important because unbalanced voltages serving a polyphase motor cause significant heating within the motor. ANSI C84.1 addresses voltage imbalance limitations at length. There’s no clear-cut limitation on imbalance that applies across the board, but motor manufacturer’s tend to get nervous when the imbalance exceeds 3% from the mean.


Interruptions are just as the name implies – loss of service. The only protection available against interruptions is a (reliable!) backup power system.


Variations in Steady State Power


In this category we have harmonic distortion, voltage flicker, and noise. Problems in this category can have a disturbing effect on sensitive electronics commonly found in computers, process controls and communications systems.


Harmonic distortion is caused by nonlinear loads such as power electronics and arcing devices. Harmonics are multiples of the fundamental frequency. For a 60Hz system, the first harmonic is 60Hz, the second harmonic is 120Hz, the third is 180Hz, and so on. Switch-mode power supplies commonly found in computers and other office electronics produce odd-order harmonics (3rd, 5th, 7th etc) of decreasing amplitude.


Harmonic distortion causes increased thermal loads in electric machinery, so it’s very important for utilities and utility customers to get a handle on this phenomenon and begin taking steps to limit the effects.


Total harmonic distortion (THD) is expressed as a percentage. It’s the square root of the sum of the squares of all harmonic amplitudes (above 1) divided by the amplitude of the first harmonic. IEEE standard 519 recommends that THD in power systems 69kv and below be limited to 5%. A new revision to the standard has limits on individual harmonics for both voltage and current, and it holds the customer responsible for limiting harmonic current injection to the utility. In turn, the utility is responsible for limiting the resulting harmonic voltage distortion on their systems. (The sharp reader will notice that harmonic voltages and harmonic currents are treated differently. The reasons why are beyond the scope of this paper.)


Here’s an example of a switch-mode power supply current waveform:



And here’s a typical adjustable speed motor drive current waveform:



Voltage flicker is caused by loads that exhibit continuous, rapid variations in load current. Arc furnaces are the most common cause. Voltage flicker appears as a modulation of the 60Hz waveform, similar to an amplitude modulated radio signal. It’s expressed as a percentage by dividing the RMS value of the modulating wave by the RMS value of the fundamental wave. In lighting systems, voltage flicker values as low as 0.5% can be perceptible to the human eye. Flicker values of 3% are downright irritating.


Here’s an illustration of flicker. Notice the periodic variation in voltage peaks:




And finally we have noise. Noise is any continuous unwanted signal on electric power circuits. It can be caused by switching, arcing, electric fields, magnetic fields, and radio waves. Noise is often classified according to the method of coupling to the power system. IEEE standard 518 contains recommendations for installation of equipment to minimize the effects of noise.


As you can see, the grid has grown into a beast that sometimes has a mind of its own. Interconnected systems offer many advantages, but there are times when the grid becomes downright hostile to the loads it serves.