Interpretation and Analysis of Power Quality Measurements Christopher J. Melhorn Mark F. McGranaghan Electrotek Concepts, Inc. Electrotek Concepts, Inc. Knoxville, Tennessee Knoxville, Tennessee • Offer examples of different methods for ABSTRACT presenting the results of power quality measurements. This paper describes advances in power quality monitoring equipment and software tools for analyzing • Describe tools for analyzing and presenting the power quality measurement results. power quality measurement results. Power quality monitoring has advanced from strictly problem solving Analysis tools for processing measurement data will to ongoing monitoring of system performance. The be described. These tools can present the information as increased amount of data being collected requires more individual events (disturbance waveforms), trends, or advanced analysis tools. Types of power quality statistical summaries. By comparing events with libraries variations are described and the methods of of typical power quality variation characteristics and characterizing each type with measurements are correlating with system events (e.g. capacitor switching), presented. Finally, methods for summarizing the causes of the variations can be determined. In the same information and presenting it in useful report formats manner, the measured data should be correlated with are described. impacts to help characterize the sensitivity of end use equipment to power quality variations. This will help INTRODUCTION identify equipment that requires power conditioning and provide specifications for the protection that can be developed based on the power quality variation Power quality has become an important concern for characteristics. utility, facility, and consulting engineers in recent years. End use equipment is more sensitive to disturbances that arise both on the supplying power system and within the CATEGORIES OF POWER QUALITY customer facilities. Also, this equipment is more VARIATIONS interconnected in networks and industrial processes so that the impacts of a problem with any piece of equipment It is important to first understand the kinds of power are much more severe. quality variations that can cause problems with sensitive loads. Categories for these variations must be developed The increased concern for power quality has resulted with a consistent set of definitions so that measurement in significant advances in monitoring equipment that can equipment can be designed in a consistent manner and so be used to characterize disturbances and power quality that information can be shared between different groups variations. This paper discusses the types of information performing measurements and evaluations. An IEEE that can be obtained from different kinds of monitoring Working Group has been developing a consistent set of equipment and methods for analyzing and presenting the definitions that can be used for coordination of information in a useful form. measurements.[1] Important objectives for the paper include the Power quality variations fall into two basic categories: following: 1. Disturbances. Disturbances are measured by • Describe important types of power quality triggering on an abnormality in the voltage or variations. the current. Transient voltages may be • Identify categories of monitoring equipment detected when the peak magnitude exceeds a that can be used to measure power quality specified threshold. RMS voltage variations variations. (e.g. sags or interruptions) may be detected 1 Table 1. Summary of Power Quality Variation Categories Example Power Quality Method of Power Conditioning Variation Category Characterizing Typical Causes Solutions Peak magnitude, Lightning, Surge Arresters, Rise time, Electro-Static Discharge, Filters, Impulsive Transients Duration Load Switching Isolation Transformers Waveforms, Line/Cable Switching, Surge Arresters, Peak Magnitude, Capacitor Switching, Filters, Oscillatory Transients Frequency Components Load Switching Isolation Transformers Ferroresonant Transformers, RMS vs. time, Energy Storage Technologies*, Sags/Swells Magnitude, Duration Remote System Faults UPS System Protection Energy Storage Technologies*, (Breakers, Fuses), UPS, Interruptions Duration Maintenance Backup Generators Undervoltages/ RMS vs. Time, Motor Starting, Voltage Regulators, Overvoltages Statistics Load Variations Ferroresonant Transformers Harmonic Spectrum, Filters (active or passive), Total Harm. Distortion, Nonlinear Loads, Transformers (cancellation or Harmonic Distortion Statistics System Resonance zero sequence components) Variation Magnitude, Intermittent Loads, Frequency of Occurence, Motor Starting, Voltage Flicker Modulation Frequency Arc Furnaces Static Var Systems * Note: Energy Storage Technologies refers to a variety of alternative energy storage technologies that can be used for standby supply as part of power conditioning (e.g. superconducting magnetic energy storage, capacitors, flywheels, batteries) when the RMS variation exceeds a specified Table 1 summarizes the different categories and lists level. possible causes and power conditioning equipment solutions for each category. 2. Steady State Variations. These include normal RMS voltage variations and harmonic Steady State Voltage Characteristics distortion. These variations must be measured There is no such thing as steady state on the power by sampling the voltage and/or current over system. Loads are continually changing and the power time. The information is best presented as a system is continually adjusting to these changes. All of trend of the quantity (e.g. voltage distortion) these changes and adjustments result in voltage over time and then analyzed using statistical variations that are referred to as long duration voltage methods (e.g. average distortion level, 95% variations. These can be undervoltages or overvoltages, probability of not being exceeded, etc.). depending on the specific circuit conditions. Characteristics of the steady state voltage are best In the past, measurement equipment has been expressed with long duration profiles and statistics. designed to handle either the disturbances (e.g. Important characteristics include the voltage magnitude disturbance analyzers) or steady state variations (e.g. and unbalance. Harmonic distortion is also a voltage recorders, harmonics monitors). With advances in characteristic of the steady state voltage but this processing capability, new instruments have become characteristic is treated separately because it does not available that can characterize the full range of power involve variations in the fundamental frequency quality variations. The new challenge involves component of the voltage. characterizing all the data in a convenient form so that it can be used to help identify and solve problems. Most end use equipment is not very sensitive to these voltage variations, as long as they are within reasonable limits. ANSI C84.1 [7] specifies the steady 2 C C C C C C C state voltage tolerances for both magnitudes and angles of each individual harmonic component. It is also unbalance expected on a power system. Long duration common to use a single quantity, the Total Harmonic variations are considered to be present when the limits are Distortion, as a measure of the magnitude of harmonic exceeded for greater than 1 minute. distortion. For currents, the distortion values must be referred to a constant base (e.g. the rated load current or demand current) rather than the fundamental component. This provides a constant reference while the fundamental can vary over a wide range. Table 2. Example current waveforms for various nonlinear loads. Current Weighting Type of Load Typical Waveform Distortion Factor (Wi) 1.0 0.5 Single Phase 80% 2.5 0.0 Power Supply (high 3rd) -0.5 -1.0 0 10 20 30 40 1.0 0.5 high 2nd,3rd, Semiconverter 4th at partial 2.5 0.0 loads -0.5 -1.0 0 10 20 30 40 1.0 Figure 1. Example 24 hour voltage profile illustrating 6 Pulse Converter, 0.5 long duration voltage variations. capacitive smoothing, 80% 2.0 0.0 no series inductance -0.5 -1.0 0 10 20 30 40 Harmonic Distortion 1.0 6 Pulse Converter, 0.5 Harmonic distortion of the voltage and current results capacitive smoothing 40% 1.0 0.0 with series inductance > 3%, from the operation of nonlinear loads and devices on the -0.5 or dc drive -1.0 power system. The nonlinear loads that cause harmonics 0 10 20 30 40 1.0 can often be represented as current sources of harmonics. 6 Pulse Converter 0.5 with large inductor 28% 0.8 The system voltage appears stiff to individual loads and 0.0 for current smoothing -0.5 the loads draw distorted current waveforms. Table 2 -1.0 10 20 30 40 0 illustrates some example current waveforms for different 1.0 types of nonlinear loads. The weighting factors indicated 0.5 12 Pulse Converter 15% 0.5 0.0 in the table are being proposed in the Guide for Applying -0.5 Harmonic Limits on the Power System (Draft 2)[2] for -1.0 10 20 30 40 0 preliminary evaluation of harmonic producing loads in a 1.0 varies with 0.5 facility. ac Voltage firing angle 0.7 0.0 Regulator -0.5 Harmonic voltage distortion results from the -1.0 10 20 30 40 0 interaction of these harmonic currents with the system impedance. The harmonic standard, IEEE 519-1992 [2], Harmonic distortion is a characteristic of the steady has proposed two way responsibility for controlling state voltage and current. It is not a disturbance. harmonic levels on the power system. Therefore, characterizing harmonic distortion levels is accomplished with profiles of the harmonic distortion over 1. End users must limit the harmonic currents time (e.g. 24 hours) and statistics. Figure 2 illustrates a injected onto the power system. typical profile of harmonic voltage distortion on a feeder circuit over a one month period. 2. The power supplier will control the harmonic voltage distortion by making sure system resonant conditions do not cause excessive magnification of the harmonic levels. Harmonic distortion levels can be characterized by the complete harmonic spectrum with magnitudes and phase 3 T H D ( % ) V l t ( V ) Time Trend for Voltage THD crossing. Magnification of the transient can be avoided 1.8 by not using low voltage capacitors within the end user 1.6 1.6 facilities. The actual equipment can be protected with 1.4 filters or surge arresters. 1.4 1.2 1.2 Short Duration Voltage Variations 1.0 1.0 Short duration voltage variations include variations in 0.8 the fundamental frequency voltage that last less than one minute. These variations are best characterized by plots 0.6 05/26 06/02 06/09 06/16 06/23 06/30 07/07 06/16 06/23 Time of the RMS voltage vs. time but it is often sufficient to Time describe them by a voltage magnitude and a duration that Figure 2. Example Profile of Harmonic Voltage the voltage is outside of specified thresholds. It is usually Distortion on a Distribution Feeder Circuit. not necessary to have detailed waveform plots since the RMS voltage magnitude is of primary interest. Transients The term transients is normally used to refer to fast The voltage variations can be a momentary low changes in the system voltage or current. Transients are voltage (voltage sag), high voltage (voltage swell), or loss disturbances, rather than steady state variations such as of voltage (interruption). Interruptions are the most harmonic distortion or voltage unbalance. Disturbances severe in terms of their impacts on end users but voltage can be measured by triggering on the abnormality sags can be more important because they may occur much involved. For transients, this could be the peak more frequently. A fault condition can cause a momentary magnitude, the rate of rise, or just the change in the voltage sag over a wide portion of the system even waveform from one cycle to the next. Transients can be though no end users may experience an interruption. This divided into two sub-categories, impulsive transients and is true for most transmission faults. Figure 4 is an example oscillatory transients, depending on the characteristics. of this kind of event. Many end users have equipment that may be sensitive to these kinds of variations. Solving Transients are normally characterized by the actual this problem on the utility system may be very expensive waveform, although summary descriptors can also be so manufacturers are developing ride through developed (peak magnitude, primary frequency, rate-of- technologies with energy storage to handle these voltage rise, etc.). Figure 3 gives a capacitor switching transient variations on the end user side. waveform. This is one of the most important transients that is initiated on the utility supply system and can affect the operation of end user equipment. 2.0 34.5 kV Bus Voltage Capacitor Switching Transient 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 0 20 40 60 80 100 Time (mS) Figure 4. Voltage Sag Caused by a Remote Fault Figure 3. Capacitor Switching Transient. Condition. Transient problems are solved by controlling the transient at the source, changing the characteristics of the system affecting the transient or by protecting equipment so that it is not impacted. For instance, capacitor switching transients can be controlled at the source by closing the breaker contacts close to a voltage zero 4 Table 3. Methods for Measuring Voltages and Currents with Multi-Meters. Meter Type Circuit Sine Wave Square Wave Distorted Light Triangle Wave Wave Dimmer 2 1.5 1.5 3 2 1.5 1.5 1 1 2 1 1 0.5 0.5 1 0.5 0.5 0 0 0 0 0 -0.5 -0.5 -0.5 -0.5 -1 -1 -1 -1 -1 -2 -1.5 -1.5 -1.5 -1.5 -3 -2 -2 Peak Peak / 1.414 100 % 82 % 184 % 113 % 121 % Method Average Sine Avg. x 1.1 100 % 110 % 60 % 84 % 96 % Responding True RMS RMS Converter 100 % 100 % 100 % 100 % 100 % sinusoidal signal, this average value is related TYPES OF EQUIPMENT FOR to the RMS value by the constant, k=1.1. This MONITORING POWER QUALITY value k is used to scale all waveforms measured. Multimeters or DMMs 3. True RMS. The RMS value of a signal is a After initial tests of wiring integrity, it may also be measure of the heating which will result if the necessary to make quick checks of the voltage and/or voltage is impressed across a resistive load. current levels within a facility. Overloading of circuits, One method of detecting the true RMS value is under- and over-voltage problems, and unbalances to actually use a thermal detector to measure a between circuits can be detected in this manner. These heating value. More modern digital meters use measurements just require a simple multimeter. Signals to a digital calculation of the RMS value by check include: squaring the signal on a sample by sample basis, averaging over a period, and then taking • phase-to-ground voltages the square root of the result. • phase-to-neutral voltages These different methods all give the same result for a • neutral-to-ground voltages clean, sinusoidal signal but can give significantly different answers for distorted signals. This is very important • phase-to-phase voltages (three phase system) because significant distortion levels are quite common, • phase currents especially for the phase and neutral currents within the facility. Table 3 can be used to better illustrate this point. • neutral currents The most important factor to consider when selecting Each waveform in Table 3 has an RMS value of 1.0 pu and using a multimeter is the method of calculation used in (100.0%). The corresponding measured value for each the meter. All of the commonly used meters are calibrated type of meter is displayed under the associated to give an RMS indication for the measured signal. waveforms, per-unitized to the 1.0 pu RMS value. However, a number of different methods are used to calculate the RMS value. The three most common Oscilloscopes methods are: An oscilloscope is valuable when performing real time 1. Peak Method. The meter reads the peak of the tests. Looking at the voltage and current waveforms can signal and divides the result by 1.414 (square tell a lot about what is going on, even without performing root of 2) to obtain the RMS. detailed harmonic analysis on the waveforms. You can get the magnitudes of the voltages and currents, look for 2. Averaging Method. The meter determines the obvious distortion, and detect any major variations in the average value of a rectified signal. For a clean signals. 5 There are numerous makes and models of oscilloscopes to choose from. A digital oscilloscope with data storage is valuable because the waveform can be saved and analyzed. Oscilloscopes in this category often have waveform analysis capability (energy calculation, spectrum analysis) also. In addition, the digital oscilloscopes can usually be obtained with communications so that waveform data can be uploaded to a PC for additional analysis with a software package. Disturbance Analyzers Disturbance analyzers and disturbance monitors form a category of instruments which have been developed specifically for power quality measurements. They Figure 5. Graphics Based Analyzer Output typically can measure a wide variety of system disturbances from very short duration transient voltages Spectrum Analyzers and Harmonic Analyzers to long duration outages or under-voltages. Thresholds Many instruments and on line monitoring equipment can be set and the instruments left unattended to record now include the capability to sample waveforms and disturbances over a period of time. The information is perform FFT calculations. The capabilities of these most commonly recorded on a paper tape but many instruments vary widely and the user must be careful that devices have attachments so that it can be recorded on the accuracy and information obtained is adequate for the disk as well. investigation. The following are some basic requirements for harmonic measurements used to investigate a problem: There are basically two categories of these devices: • Capability to measure both voltage and current 1. Conventional analyzers that summarize events simultaneously so that harmonic power flow with specific information such as over/under information can be obtained. voltage magnitudes, sags/surge magnitude and duration, transient magnitude and • Capability to measure both magnitude and duration, etc. phase angle of individual harmonic components (also needed for power flow 2. Graphics-Based analyzers that save and print calculations). the actual waveform along with the descriptive • Synchronization and a high enough sampling information which would be generated by one rate for accurate measurement of harmonic of the conventional analyzers. components up to at least the 37th harmonic (this requirement is a combination of a high It is often difficult to determine the characteristics of a sampling rate and a sampling interval based on disturbance or a transient from the summary information the 60 Hz fundamental). available from conventional disturbance analyzers. For instance, an oscillatory transient cannot be effectively • Capability to characterize the statistical nature described by a peak and a duration. Therefore, it is almost of harmonic distortion levels (harmonics levels imperative to have the waveform capture capability in a change with changing load conditions and disturbance analyzer for detailed analysis of a power changing system conditions). quality problem (Figure 5). However, a simple Harmonic distortion is a continuous phenomena. It conventional disturbance monitor can be valuable for can be characterized at a point in time by the frequency initial checks at a problem location. spectrums of the voltages and currents. However, for proper representation, measurements over a period of time must be made and the statistical characteristics of the harmonic components and the total distortion determined. 6 V o l t a g e ( % ) V o l t a g e ( % V o l t a g e ( p u ) C u r r e n t ( A m p s ) A m p s IBM503E April 22, 1992 at 21:43:21 Local RGE_BETA November 19, 1991 at 06:07:56 PQNode Local IBM503M May 09, 1992 at 13:04:07 Local Phase C-A Voltage Trigger Phase A Voltage Trigger Phase A Current RMS Variation Wave Fault SS Wave 115 1.5 600 110 Duration Max1.094 400 Fund 267.5 105 200 100 1 0.150 Sec Min-1.280 RMS 281.0 0 95 -200 90 Min CF 1.772 -400 85 0.5 80 -600 81.38 Min -495.6 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 10 20 30 40 50 Ave 0 Max 498.0 Time (Seconds) Time (mSeconds) 96.77 THD 30.67 150 100 -0.5 100 80 Max HRMS 86.19 50 60 0 101.4 -1 TIF/IT70249 40 -50 20 -100 -150 -1.5 0 0 25 50 75 100 125 150 175 0 10 20 30 40 50 60 70 0 10 20 30 40 50 Time (mSeconds) Time (mSeconds) Harmonic BMI/Electrotek Uncalibrated Data Figure 6. Output From Combination Disturbance and Harmonic Analyzer. Power Combination Disturbance and Harmonic Quality Monitoring Voltage In Plant Digital Fault Analyzers Instruments Recorders Recorders Monitors The most recent instruments combine limited harmonic sampling and energy monitoring functions with Data Data Data Data complete disturbance monitoring functions as well (Figure Translator Translator Translator Translator 6). The output is graphically based and the data is PQ Data remotely gathered over phone lines into a central Interchange Format database. Statistical analysis can then be performed on the data. The data is also available for input and PQ Data manipulation into other programs such as spreadsheets CHARACTERIZER and other graphical output processors. Event ANALYZING POWER QUALITY Information MEASUREMENT DATA PQ Monitoring Site Database Information Analyzing power quality measurements has become increasingly more sophisticated within the past few years. Database Manager It is not enough to simply look at RMS quantities of the voltage and current. Disturbances that occur on the power system have durations in the milli-second time Event Data Protection Energy Statistical frame, equipment is more sensitive to these disturbances, Viewer Trending Analysis Analysis Analysis and there is more equipment connected to the power Applications systems that cause disturbances or power quality Figure 7. Example Data Analysis System. problems. For these reasons, it is often necessary to continuously monitor system performance and Different types of power quality variations require characterize possible impacts of disturbances. different types of analysis to characterize system performance. Some examples are given in the following The data analysis system must be flexible enough to sections. With a flexible system, these applications can be handle data from a variety of monitoring equipment and customized to individual user needs. maintain a database that can be used by many different applications. The concept is illustrated in Figure 7. Transients Transients are normally characterized by the actual waveform, although summary descriptors can also be developed for: • peak magnitude 7 C t C u m u l a t i v e P r o b a V l t M i t d C t C u m u l a t i v e P r o b a • primary frequency and duration is shown. Figure 10 illustrates this type of plot. • time of occurrence • rate of rise RMS Variation Magnitude vs Duration vs Count An example of this data in statistical form is presented 800 in Figure 8. 700 600 Transient Magnitude 250 100 Samples: 561 500 Min: 1.05014 Count 400 Max: 1.59 200 Range: 0.53986 80 300 Mean: 1.14418 160 Std Dev: 0.0825617 140 200 Variance: 0.00681643 120 100 Std Err: 0.00348576 150 60 100 80 Mode: 1.11834 60 Magnitude(%) Skewness: 1.73749 0 40 Kurtosis: 4.45834 1-10 20 100 40 10- 60 60- 180- 0 180 300- 300 >3000 3000 Duration(Cycles) 50 20 Figure 10. Three Dimensional RMS Variation Bar Graph. 0 0 0.0 0.5 1.0 1.5 2.0 Magnitude (per-unit) Harmonics Figure 8. Bar Chart for Transient Peak Voltage. Harmonics are characterized by individual snapshots RMS Variations of voltage and current with the associated spectrums. It is important to understand that the harmonic distortion RMS variations are generally characterized by the levels are always changing and these characteristics RMS value vs. time or by the minimum magnitude of the cannot be represented with a single snapshot. Therefore, voltage during the event vs. the duration of the event. time trends and statistics are needed. An example time Figure 1 was an illustration of a plot of magnitude vs. Time trend plot for one month was included in Figure 2. Figure for a 24 hour period. 11 shows the statistics of the harmonic current level. This would be good for comparison with IEEE-519 limits. This method is fine for looking at single sites and single events. But when a whole system is involved, Histogram for Harmonic RMS Current 1500 100 either customer or utility, it may be preferable to look at a Samples: 6691 Min: 4.93836 range of events (e.g. one month, one year, etc.) for Max: 20.9626 1250 Range: 16.0242 80 Mean: 6.95771 multiple sites. This would give an indication as to what Std Dev: 1.9269 1000 Variance: 3.71296 type of RMS events are occurring on a given system. The Std Err: 0.0235567 60 Mode: 6.06852 magnitude duration plot in Figure 9 illustrates the minimum Skewness: 1.53821 750 Kurtosis: 1.2563 voltage (in percent) during the event and the duration of 40 500 the event (time in cycles that voltage was out of the 20 thresholds). 250 0 0 RMS Magnitude vs Duration 0 5 10 15 20 25 150 Current (A) Total Events: 149 Events Below: 141 Events Above: 8 Figure 11. Histogram for Harmonic RMS Current for 125 Below CBEMA: 26 Above CBEMA: 0 Approximately Four Months. 100 75 SUMMARY 50 Systematic procedures for evaluating power quality 25 concerns can be developed but they must include all 0 levels of the system, from the transmission system to the 0 1 2 3 4 10-2 10-1 10 10 10 10 10 Duration (60 Hz Cycles) end user facilities. Power quality problems show up as Figure 9. Example Magnitude Duration Plot. impacts within the end user facility but may involve interaction between all levels of the system. Another method for displaying this type of data is a three dimensional bar graph where the count, magnitude, 8 A consistent set of definitions for different types of 2. IEEE Std. 519-1992, IEEE Recommended Practices power quality variations is the starting point for and Requirements for Harmonic Control in developing evaluation procedures. The definitions permit Electrical Power Systems, IEEE, New York, 1993. standardized measurements and evaluations across different systems. 3. “Electrical Power System Compatibility with Industrial Process Equipment - Part 1: Voltage A data analysis system for power quality Sags,” Paper by the IEEE Working Group P1346, measurements should be able to process data from a Proceedings of the Industrial and Commercial variety of instruments and support a range of applications Power Systems Conference, 94CH3425-6, May, for processing data. With continuous power quality 1994. monitoring, it is very important to be able to summarize variations with time trends and statistics, in addition to 4. CENELEC Standard CLC/BTTF 68-6 (Sec) 23, characterizing individual events. “Voltage Characteristics of Electricity Supplied by Public Distribution Systems,” June, 1993. BIOGRAPHIES 5. IEC Standard 1000-2-2, “Compatibility Levels for Low Frequency Conducted Disturbances and Christopher J. Melhorn received an ASE from York Signalling in Public Low Voltage Power Supply College of Pennsylvania in 1986 and a BSEET from the Systems.” Pennsylvania State University in 1989. Chris has been employed with Electrotek Concepts, Inc. since 1990. His 6. IEC Standard 1000-4-7, “General guide on experience at Electrotek includes working with EPRI and harmonics and interharmonics measurements and utilities on case studies involving power quality issues. instrumentation, for power supply systems and He was also extensively involved in the EPRI DPQ project equipment connected thereto.” site selection phase. Chris is presently involved in developing new software for the power systems 7. ANSI C84.1-1989, American National Standard for engineering environment and working to increase Electric Power Systems and Equipment - Voltage Electrotek's industrial based clientele. Ratings (60 Hertz). Mark F. McGranaghan received a BSEE and an MSEE 9. M. McGranaghan, D. Mueller, and M. Samotyj, from the University of Toledo and an MBA from the “Voltage Sags in Industrial Plants,” IEEE University of Pittsburgh. Mark serves as Manager of Transactions on Industry Applications, Vol. 29, Power Systems Engineering at Electrotek Concepts, Inc., No. 2, March/April, 1993. Mark is responsible for a wide range of studies, seminars, and products involving the analysis of power quality 10. L. Conrad, K. Little, and C. Grigg, “Predicting and concerns. He has worked with electric utilities and end Preventing Problems Associated with Remote users throughout the country performing case studies to Fault-Clearing Voltage Dips, “ IEEE Transactions characterize power quality problems and solutions as part on Industry Applications, vol. 27, pp. 167-172, of an extensive Electric Power Research Institute (EPRI) January, 1991. project. He has also been involved in the EPRI Distribution Power Quality Monitoring Project which is 11. V. Wagner, A. Andreshak, and J. Staniak, “Power establishing the baseline power quality characteristics of Quality and Factory Automation, “ Proceedings of U.S. distribution systems through a multi-year monitoring the IAS Annual Meeting, vol. 35, no. 6, pp. 1391- effort. Mark was involved in the design and specification 1396. of the instrumentation and software for this project. REFERENCES 1. IEEE Working Group P1159, Recommended Practice for Monitoring Electric Power Quality - Draft 7, December, 1994. 9