Development Of A Laboratory Experiment To Demonstrate Power Quality Issues

Less than 15 years ago, harmonics were not even mentioned in an articlel listing all of the problems with electrical power that could cause malfl.mctions or damage to electronic equipment. However, the widespread application of electronic devices in business and industry is causing new problems for power systems. Nonlinear loads, such as the power supplies for electronic devices, introduce harmonic currents into the power system, which can cause failures in power system equipment as well as in other loads. Evidence of harmonic problems include circuit breakers tripping when they shouldn’t or failing to trip when they should, overheated neutral conductors or transformers, erratic operation or tripping of adjustable speed drives, blown power factor correction capacitors, and communication interference. The problems are different, but their causes are related. Since every user of the power system contributes to the problem, I believe all electrical engineers and technicians need to have a basic understanding of power quality issues. All undergraduates in the EET curriculum at Purdue are being provided with such a background in the form of lecture material and hands-on laboratory experience. Because of budget constraints, emphasis was placed on developing an experiment that could be performed with inexpensive loads and use equipment that is available in most electronics and power laboratories (e.g., oscilloscope, true-RMS voltmeter). During the development of the experiment, our department was fortunate enough to receive a gifi from the Fluke Corporation of Power Harmonic Analysis Meters, which greatly enhanced the students’ laboratory experience. This paper discusses some of the basic theory of harmonics and their effects on the power system, followed by a description of the laboratory experiment. I N T R O D U C T I O N What are harmonics? The voltage waveform received from the power company normally consists of a single frequency sinusoid. For linear loads on the power system (resistors, inductors, and capacitors) the current will also be a single frequency sinusoid. However, some loads are nonlinear loads and cause a nonsinusoidal current when a sinusoidal voltage is applied. Nonlinear loads often contain some type of switching device that causes noncontinuous operation. Examples include the power supplies for electronic devices including computers, programmable controllers, and oflice equipment; variable frequency motor drives; and electronic ballasts for fluorescent lights. Although the current is not sinusoidal for a nonlinear load, it is periodic, assuming the load is in a steady-state operating condition. {tixi$~ 1996 ASEE Annual Conference Proceedings ‘@lly’: P ge 156.1 The mathematician Fourier showed that any periodic 100 waveform could be represented by a series of sinusoids whose fi-equencies are integral multiples of the frequency of the original waveform.2 Thus nonsinusoidal currents due to : 50 > nonlinear loads will contain harmonic components whose ~ frequency is an integral multiple of the power system :0 frequency (60 Hz in North America). Figure 1 shows a 60 ~ Hz sine wave, having an amplitude of 100. Also shown are <-50 third (180 Hz) and fifth (300 Hz) harmonics with different amplitudes. The actual amplitude of the harmonics and their phase relationships would of course depend on the shape of -1oo the original nonsinusoidal waveform. Harmonic currents can o 5.55 11.1 16.6 time (msec) cause a variety of problems in the power system. Figure 1: Example of first, third, and fifth harmonics Effects of harmonics Harmonic currents can affect many devices in the power system including transformers, conductors, motors, and circuit breakers.3 Overheating and erratic operation are possible when harmonics are present. Since these effects have been previously described, they will not all be explained in detail here. However, two key areas which formed the basis of the power quality experiment will be described after a discussion of why they were chosen. These are the effects of harmonics on certain types of metering equipment and the effects of harmonics on neutral conductors in a three-phase system. THEORY OF A POWER QUALITY EXPERIMENT Before developing objectives and procedures for an experiment to illustrate harmonic phenomena, several factors had to be considered. The first consideration was the existing equipment in the laboratory and what could be done with it. In our lab, each station has a two-channel oscilloscope and a true-RMS digital multi meter (DMM). While we had a variety of motors, transformers, and R-L-C loads, we did not have a nonlinear load that the students could measure conveniently. Thus it was clear that additional devices would have to be acquired. This, of course, brought up the issue of cost, since finds are very limited. While deciding what type of equipment to obtain, it was necessary to consider what phenomena could be reasonably observed. For example, attempting to measure the additional heating of conductors due to harmonics would require fairly elaborate equipment and might not contribute much to the students’ learning. On the other hand, using the oscilloscope and true-RMS meter, the students could measure the peak value of a waveform and its RMS value. Those measurements would allow them to calculate the crest factor of the waveform. They also could measure the waveform with an average responding meter, which would allow them to compare the response of different types of meters when harmonics are present. Since many students have a tendency to believe whatever a piece of test equipment tells them, I felt this would be a valuable lesson for thelm. Another area that could be easily investigated was the problem of high neutral currents in a three-phase system that is feeding single-phase nonlinear loads. Again, this could be easily done with the existing test equipment, although new loads would be required. Because this is a real problem in industrial and commercial facilities, I felt this would also provide a good learning experience for the students. Following the selection of these two phenomena for the laboratory exercise, we were fortunate to obtain a donation of FlukeF41 Harmonic Analysis Meters for each station in the lab. They were incorporated into the ?$iiia’-’ } 1996 ASEE Annual Conference Proceedings ‘..*,El@#.$ . P ge 156.2 experimental procedures, so the description that follows includes them. These meters provide a wide variety of itiormation to the students, including the harmonic content of a waveform, total harmonic distortion (THD), and crest factor. However, the majority of the lab could be accomplished with the oscilloscope and true-RMS DMM. Before discussing the equipment setup, I will briefly discuss the theory behind these two phenomena. Meter response to harmonics As previously mentioned, most students believe you just hookup a meter and whatever it says must be correct. Unfortunately that may be a bad assumption if more than one frequency is present in the waveform. Since voltages and currents in the power system traditionally were solely 60 hertz, many meters were designed to take advantage of that fact, resulting in a cheaper meter. Most AC meters actually rectifi the AC waveform 1 .-. on the scale. Since the meter responds to the average, but Figure 2: Rectified sine wave and harmonic containing wave reads out the RMS, it is called an average responding, RMS calibrated meter. This works fine, as long as the waveform is a single-frequency sinusoid. The second waveform in figure 2 is the rectified sum of several harmonics: [ jlt)=100 sin377t+ sin 1131t + sin 1885t + sin 2639t 3 5 7 ) These terms are, of course, the first four terms of the Fourier series for a square wave, and the waveform in Figure 2 is in fact beginning to approximate a rectified square wave. An average responding meter would recti~ the waveform, as shown in figure 2, and would respond to the average, which can be shown to be 74.6. The meter would then multiply by 1.111 to yield a meter reading of 82.9. In fact the RMS of the waveform can be readily calculated as 76.5, so in this case, the meter would be reading too high. In other cases, the average responding meter may read too low. $iiii’ } 1996 ASEE Annual Conference Proceedings ‘O.yllyy’: . P ge 156.3 THE EXPERIMENT SETUP AND RESULTS Meter readings in the presence of harmonics To illustrate the performance of average responding meters when harmonics are present, requires a load with variable harmonic content. To keep the cost low, I chose a commercial incandescent lamp dimmer switch with a light bulb as its load. A dimmer mounted in a plastic receptacle box costs less than $5.00. The dimmer switch operates like a triac; adjusting the dimmer “chops” out part of the sine wave, changing the RMS voltage to the load. Figure 5 (obtained with the Harmonic Analysk Meter) shows an example of the dimmer switch output voltage. ?@X&j 1996 ASEE Annual Conference Proceedings ‘..+,~yy’,: P ge 156.4 Students were instructed to measure the output voltage of the dimmer using both average-responding and true-RMS voltmeters. By varying the voltage, they could observe how the accuracy of the average-responding meter was affected by the harmonics. The percent error was calculated from the measured voltages as a percent of the average-responding meter reading. Crest factor is obtainable fi-om an oscilloscope and true-RMS meter; however, the Harmonic Analysis Meters provided that information directly. The crest factor for a sine wave is the square-root of two (1.4 14). Observation of Figure 5 indicates that reducing the voltage will lead to a higher crest factor as the delay angle is increased toward 90°. Using the Fluke F41 meter, it was found that as the voltage was decreased both the crest factor and the THD increased. To provide the students with a visual indication of the effect of harmonics on the averageresponding meter, I had them plot th