indices for comparative assessment of the …apic/papers/power_quality/2012...indices for...

Indices for comparative assessment of the harmonic effect ofdifferent home appliances

Alexandre B. Nassif1*,†, Jing Yong2, Wilsun Xu3 and C.Y. Chung4

1Hydro One Networks Inc., Toronto, ON, Canada M5G 2C92State Key Laboratory of Power Transmission Equipment, Chongqing University, Chongqing, China

3Department of Electrical and Computer Engineering, University of Alberta, Edmonton, AB, Canada T6G 2V44Hong Kong Polytechnic University, Kowloon, Hong Kong

SUMMARY

Most of the modern home appliances use power electronic circuits as a power supply to interface withthe distribution grid. They have become a major source of harmonic distortions in power distributionsystems. This article presents a study of the harmonic current characteristics of such appliance through ameasurement-based comparative approach. Magnitude-related indices are introduced to compare theharmonic effects of the various appliances and to quantify their relative severity. Phase angle–related indicesare introduced to study the diversity of the current harmonics phase angle among the devices and tounderstand its effect on harmonic cancellation. The proposed approach can be used to establish a databaseof harmonic characteristics for home appliances and other small, distributed harmonic-producing loads.Copyright © 2012 John Wiley & Sons, Ltd.

key words: power system harmonics; energy-saving home appliances; waveform distortions; harmoniccancellation

1. INTRODUCTION

Power electronic-based home appliances have proliferated in distribution feeders supplying modernhouses. For example, in the past, lighting loads were predominantly resistive, and heating, ventilation,and air-conditioning were mostly a combination of resistive and motor loads; today, the former arebeing gradually replaced by compact fluorescent lamps (CFLs), and the latter are being graduallyreplaced by variable speed air-conditioning [1]. The main reason behind this fact is the awarenessfor energy conservation. Although being power electronic based and being energy efficient are notsynonymous, most of the energy-efficient loads are based on power electronics. In general, theseappliances consume considerably less power than their older counterparts [1,2]. Newly introducedappliances are adjustable speed drive (ASD)–based fridges, heat pump furnaces, fans, laundry dryers,and washers [1]; computers, monitors, laptops, and TV sets using switch-mode power supplies [3];microwave ovens, home entertainment systems, and CFLs; and many other electronic loads.These devices are harmonic polluters and inject a relatively large amount of harmonics into the

power grid relative to their power demand. When large amounts of such appliances operate in powerdistribution systems, the collective effect on the feeder power quality has become a large concern toutilities. For example, there is a need to understand the harmonic effect of mass deployment of CFLs[4–8]. Besides the CFL, the literature has documented the measurement of TVs, PCs, and variablespeed air-conditioning as well as their measured or estimated harmonic effects on medium-voltagedistribution feeders [2,6–10]. In addition, extensive measurements of other common appliances such

*Correspondence to: Alexandre B. Nassif, Hydro One Networks Inc., Toronto, ON, Canada M5G 2C9.†E-mail: [email protected]

Copyright © 2012 John Wiley & Sons, Ltd.

EUROPEAN TRANSACTIONS ON ELECTRICAL POWEREuro. Trans. Electr. Power (2012)Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/etep.1620

as TVs of different technologies, heat pumps, stereos, CFLs, and refrigerators have been presented byHardie and Watson [11,12]. However, the question “Which are the most significant home appliancesfrom the perspective of harmonics currents?” is still difficult to answer, as no comparative study ofhome appliances harmonics has provided quantitative results that can be used in a practical way.Hence, it is difficult for industry to judge the power quality implication of various appliances.The objective of this article is to present a comparative study on the harmonic-producing

characteristics of common home appliances. Measurements on the harmonic currents of many applianceswere taken for this purpose. The results are then analyzed and compared from both the magnitude andphase angle perspectives. Indices are introduced to ensure a consistent comparison can be made.

2. CHARACTERISTICS OF COMMON HOME APPLIANCES

The measured appliances are listed in Table I. Codes are assigned to them for easy identification inthe figures and tables shown later. Table I also identifies the number of measured appliances ofeach type. Detailed results for all measured appliances are presented in Appendix A along with thedata-processing methods. As shown in Table I, more than one piece of equipment of the same type wasmeasured for most of the appliances. The appliances were measured under close-to-ideal voltageconditions. All portable appliances such as TVs and PCs were supplied by a signal generator(California Instruments 4500Lx-Hv) with voltage total harmonic distortion (THDV)< 0.5%, whereaslarge appliances such as furnace and laundry loads were measured when installed in residential housessupplied with voltage distortion THDV< 3%. Close to ideal (low-voltage distortion) conditions wereadopted because taking into consideration all possible voltage supply conditions would deem suchan analysis impossible because of the large combinatorial nature of supply voltage magnitudes andphase angles to carry out the analysis presented in this article.Table II shows the main electric results obtained for the various measured appliances. Those

results include the displacement power factor (DPF) and total power factor (PF), which were calculatedby using the IEEE definition [13,14]. Some appliances have different operating cycles that exhibitdifferent power consumption and harmonic current characteristics. For example, a washing machinecan have washing, rinsing, and spinning cycles. Other appliances, such as the CFLs, exhibit practicallyconstant power consumption. To provide representative results, various operating cycles weremeasured. Table II presents the characteristics for the full load under normal operating conditions,and the column of “operating power” shows the power consumption under this condition. Therefore,for some appliances, the operating power is less than the rated power and is generally the condition inwhich the device draws the most distorted current.

Table I. Measured home appliances.

Appliance type Code No. tested

Compact fluorescent lamp CFL 12Electronic-ballast fluorescent lamp EBL 3Magnetic-ballast fluorescent lamp MBL 1Incandescent Lamp INC 1Desktop PC PC 5LCD computer monitor LCD 5Laptop LAP 5LCD high-definition television LCD TV 4CRT television CRT TV 1Microwave oven MW 3ASD-based fridge ASD FR 3Regular fridge R FR 2ASD-based washer WSH 2ASD-based dryer ASD DRY 1Regular dryer DRY 2Furnace FUR 1

A. B. NASSIF ET AL.

Copyright © 2012 John Wiley & Sons, Ltd. Euro. Trans. Electr. Power (2012)DOI: 10.1002/etep

The measured appliances are available in a variety of ratings. The tested CFLs are rated 5 to 30W,and the tested LCD TVs are rated 75 to 300W. The shape of the current waveforms and the spectra are,however, similar. A representative rating for each appliance was obtained from survey studies. Themost common rating of CFLs was identified as 15W [4], and therefore this amount of power wasadopted as the operating power for this device. Although the operating power of the ASD-based fridgeis higher than that of the regular fridge, it is important to point out that the measured ASD-basedfridges were of a volume much larger than that of the measured regular fridges. In addition, theseASD-based fridges seem to feature harmonic distortion compensation. More details on how to usethe operating power to determine the appliances’ current models are presented in Appendix A. Table IIconfirms that almost all the electronic home appliances are harmonic producers. Many of the electronicappliances such as CFLs, desktop PCs, and computer monitors have a THDI higher than 100%. Table IIalso shows that most of the appliances have reasonably high DPFs, and some of them have a leadingPF (indicated by “*” in Table II).The appliances’ current phasors at the 3rd to 13th harmonic orders are displayed in Figure 1. The

phasors use the supply voltage angle as a reference, setting the phase angle of the fundamentalfrequency voltage to zero. The currents are normalized to emphasize the current harmonic anglesand to improve visualization. Figure 1 reveals the potential harmonic cancellation that might occurwhen two or more appliances are operated together. For example, the 3rd harmonic current of theLCD monitor is approximately 180� out of phase with that of the furnace. Conversely, the 3rdharmonic of the desktop PC is almost in phase (0�) with that of the CRT TV.

3. RATE OF HARMONIC DECLINE

One way to characterize the harmonic currents from home appliances is by analyzing the decline rateof the current spectrum for these devices. Engineering mathematics books indicate that a squarewaveform has the spectrum characterized by Ih = I1/h, where h = 1, 3, 5, 7. . .. Conversely, triangleand trapezoidal waveforms have the spectrum characterized by Ih = I1/h

2. Many studies have foundthat a first approximation to the harmonic characteristics of an ideal single-phase ASD is Ih = I1/h atall odd harmonics because the waveform is approximated as a square wave [15]. For the homeappliances, however, this type of characterization has not been established. The characterization ofthe harmonic current magnitudes for home appliances is proposed as

Ih ¼ I1=ha (1)

where a is a parameter that determines the decline rate of the current spectrum and is estimated byperforming curve fitting on the normalized (by the fundamental frequency current) spectra of the home

Table II. Harmonic characteristics of the main home appliances.

Appliance THDI (%) DPF PF Operating power (W) Irms (A) I1 (A)

CFL 120 0.9* 0.6 15 0.24 0.15EBL 140 0.96* 0.56 15 0.24 0.14MBL 8 0.38 0.38 33 0.73 0.73INC 3 1 1 60 0.49 0.49PC 112 1 0.69 100 1.30 0.88LCD 110 0.96* 0.64 40 0.48 0.34LAP 130 0.96* 0.58 75 1.14 0.68LCD TV 10 0.99* 0.98 300 2.59 2.58CRT TV 145 1 0.56 70 1.00 0.57MW 41 0.99* 0.9 1200 11.08 10.49ASD FR 7 0.92 0.91 170 1.6 1.6R FR 16 0.94 0.93 150 1.30 1.27WSH 75 0.45 0.35 180 4.20 3.36ASD DRY 55 0.98 0.86 1000 4.90 4.30DRY 1.4 0.84 0.84 4180 20.2 20.2FUR 11 0.84 0.84 500 4.85 4.82

CURRENT HARMONICS OF HOME APPLIANCES


appliances. Such a relationship is proposed because the electronic home appliances, even those basedon the single-phase rectifier topology, do not draw a square current waveform and therefore cannot beapproximated to the known relationship Ih = I1/h. The characterization serves two purposes: (i) it candefine patterns for the appliances’ current spectra for each type of appliance, and (ii) it can be usedto compare the relative distortion of the home appliances’ harmonic currents.The estimated values for a for all appliances are shown in Table III. The appliances are sorted

according to the a values, which increase from left to right. Table III reveals that many of the applianceshave a< 1, which means that the currents drawn by those appliances contain an amount of harmonicsthat are higher than those contained in the ideal square-wave case. This can also be confirmed bycomparing the THD of the square wave, which is 45.7%, to those of the appliances, which werepresented in Table II.The appliances’ current data are displayed in Figure 2, which reveals that most power electronic-

based appliances have a spectrum with harmonics higher than those obtained from the I1/h relationship.Figure 2 shows that the LCD TV, ASD-based fridge, furnace, magnetic-ballasted fluorescent lamp, andregular fridge are the appliances that have harmonics below the I1/h curve. The appliances that have avalues close to 2 are the regular fridge and the furnace. By investigating those appliances currentwaveforms, it was observed that they have triangular-like shapes. For those two appliances, theapproximation to the ratio I1/h

2 agrees well.By using these results, the harmonic magnitude characteristics of home appliances can be obtained

using as a first approximation by using the frequency-dependent relationship Ih = I1/ha. Using the

Table III. Comparing a for all home appliances.

LAP CRT TV EBL CFL PC LCD WSH

a 0.41 0.46 0.47 0.6 0.65 0.66 0.83

ASD DRY MW R FR FUR ASD FR MBL LCD TVa 0.9 1.32 1.99 2.12 2.27 2.43 2.49

30

210

60

240

90

270

120

300

150

330

180 0

30

210

60

240

90

270

120

300

150

330

180 0

30

210

60

240

90

270

120

300

150

330

180 0

30

210

60

240

90

270

120

300

150

330

180 0

30

210

60

240

90

270

120

300

150

330

180 0

30

210

60

240

90

270

120

300

150

330

180 0

MWASD FRWSHLCD TVPCLAPLCDCFLFUREBLMBLASD DRYR FRCRT TV

3rd harmonic

13th harmonic11th harmonic9th harmonic

7th harmonic5th harmonic

Figure 1. Normalized harmonic currents of all appliances using the fundamental voltage phase angle as theangular reference.

A. B. NASSIF ET AL.


estimated values for the parameter a is one way to estimate the magnitudes of the current spectra of allappliances if the rated fundamental frequency current is known. This information is useful as a firstapproximation to perform harmonic studies, as the harmonic current spectra of the appliances arenot provided by the manufacturer. To further illustrate the accuracy of this approximation, Figure 3displays the appliance current spectra along with the respective approximation and shows that anacceptable fitting can be obtained for most of the appliances.To deal with the diversity of the magnitudes and phase angles of the various appliances, two studies

are proposed. The severities of the harmonic currents are studied by using the equivalent CFL index, asintroduced in Section 4. The harmonic compatibility index is proposed to study the harmonic currentscancellation due to phase angles disparity and is introduced in Section 5.

4. THE HARMONIC MAGNITUDE STUDY

This section introduces the equivalent CFL index to assess the relative severity of the current harmonicmagnitudes of the home appliances. To provide a consistent comparison, the index is calculated by

1 3 5 7 9 11 13 15 17 190

0.2

0.4

0.6

0.8

1

Harmonic order

Har

mo

nic

cu

rren

ts [

no

rmal

ized

] CRT TVLAPPCLCDCFLEBLWSHASD DRYMWLCD TVASD FRFURMBLR FR

1/h

Figure 2. Current spectra of the home appliances and curve fitting to obtain a.

3 5 7 9 11 13 15 17 190

0.5

1

MW

3 5 7 9 11131517190

0.05

0.1

AS

D F

R

3 5 7 9 11 1315 17 190

0.5

WS

H

3 5 7 9 11 13 15 17 190

0.05

0.1

LCD

TV

3 5 7 9 11131517190

0.5

1

PC

3 5 7 9 11 1315 17 190

0.5

1

LAP

3 5 7 9 11 13 15 17 190

0.5

1

LCD

3 5 7 9 11131517190

0.5

CF

L

3 5 7 9 11 1315 17 190

0.1

0.2

FU

R

3 5 7 9 11 13 15 17 190

0.5

1

EB

L

3 5 7 9 11131517190

0.1

0.2

MB

L

3 5 7 9 11 1315 17 190

0.5

AS

D D

RY

3 5 7 9 11 13 15 17 190

0.1

0.2

R F

R

3 5 7 9 11131517190

0.5

1

CR

T T

V Normalized spectrum

Fitted spectrum

Figure 3. Current spectra and fitted 1/ha.



calculating the harmonic emissions of an appliance as compared with the emissions of anotherappliance used as a template to facilitate harmonic emission benchmarking. Of all the powerelectronic-based home appliances, the CFLs have been most often studied to characterize them anddetermine their harmonic current effects [6–8]. Because CFLs have been extensively studied andcharacterized, it is reasonable to use their current harmonics as a “relative harmonic injectionbenchmark” for comparing the harmonic injections from the other home appliances.The equivalent CFL index is introduced to provide a quantitative comparison of the harmonic

effects of the appliances. This index quantifies each appliance in terms of its harmonic effect expressedas the number of CFLs it is equivalent to. This index is defined as follows:

1. For each type of appliance measured, a representative harmonic current is established. Thisrepresentative current magnitude is obtained from a weight average of the measurements ofseveral types of each of the various appliances. This procedure is explained in Appendix A.Two examples are displayed in Figure 4a: the representative desktop PC and the representativeCFL. The operating power of the representative appliances is shown in Table II.

2. For each harmonic order h, the ratio of the appliance’s current to that of the representative CFLcurrent is determined by using the following equation:

Ratioh Appliance ¼ Ih appliance

Ih CFL(2)

where Ih_appliance is the appliance’s harmonic current at order h, and Ih_CFL is the representative CFLharmonic current at order h (all in amperes). For example, a Ratioh value of 2 implies that the appliancegenerates twice as much harmonic current at order h as the representative CFL. Figure 4b shows theratios for the desktop PC.

1. To obtain a single index and compare the appliance with a representative CFL, the ratios ofdifferent harmonic orders are aggregated into one value by using a weighted average asfollows:

1 3 5 7 9 11 13 15 17 19 21 23 25 27 290

0.2

0.4

0.6

0.8

1

Harmonic order

Cur

rent

[A

]

IPC

ICFL

a.

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Total0

2

4

6

8

10

Harmonic order

Rat

io-h

PC

Equivalent-CFLPC

b.

Figure 4. Harmonic emissions from a desktop PC and a CFL: (a) harmonic current magnitudes (A) and(b) harmonic current ratios.

A. B. NASSIF ET AL.


Equivalent-CFLAppliance ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiXHh¼3

wh � Ratioh Appliance

� �2

vuut ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiXHh¼3

I2h AppliancePHh¼3

I2h CFL

0BBB@

1CCCA

vuuuuuut (3)

The weighting factorwh ¼ Ih CFL=

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiPh¼3

HI2h CFL

ris the individual harmonic distortion of the CFL

current. This weighting factor is used because if a CFL produces more harmonic at h, the harmonicsfrom other appliances will also be treated with more significance at the same order. The last bar ofFigure 4b (labeled as Total) represents the aggregated ratio (i.e. equivalent CFL) of the desktop PC.

4.1. Basic results provided by the equivalent CFL index

The results obtained from the calculated individual equivalent CFL index of all appliances arepresented in Table IV. The results are arranged in the ascending order of the index. Table IV shows thata desktop PC is equivalent to seven CFLs in terms of harmonic current injection. A microwave ovenhas a harmonic effect equivalent to 26.4 CFLs, and a laundry washer has a harmonic effect equivalentto 16 CFLs. Table IV also lists the power ratios with respect to the CFL and the harmonic current ratiosfrom the 3rd to the 13th harmonics.

4.2. Comparison of all appliance types

The equivalent CFL index compares the harmonic effect of the appliances on the basis of individualunits. The fact that each household may have multiple appliances of the same type is recognized in thisstudy. The implication is that the total harmonic currents generated by each type of appliance (thatcan be 1 or more units) in a typical house must also be compared. For this purpose, the number ofappliances that could be installed in a household is estimated from statistical surveys [4,16]. Theaverage number of TVs is indicated to be 2, and the average number of refrigerators, laundry washers,and dryers is indicated to be approximately 1. As mentioned earlier for the lighting loads, an averagehouse has a total number of between 25 and 30 installed lamps. At present, approximately 20% to 25%of all lamps are CFLs [5], so it is assumed that six CFLs are installed in a typical house. The averagenumber of magnetic-ballasted fluorescent lamps is estimated to be 2 [4]. On the basis of these data, thetotal harmonic effect of each appliance type in a household is determined and quantified as the “totalequivalent CFL” index, which compares the harmonic effect of each type of appliance installed in atypical household.

Table IV. Comparing the harmonic effect of one unit of home appliances.

Appliancetype

Operatingpower (W)

Powerratio

EquivalentCFL

Ratio3 Ratio5 Ratio7 Ratio9 Ratio11 Ratio13

MBL 30 2 0.42 0.56 0.02 0.11 0.08 0.12 0.06CFL 15 1 1.00 1.00 1.00 1.00 1.00 1.00 1.00ASD FR 170 11.3 1.14 1.03 0.28 2.74 1.23 0.89 0.60EBL 18 1.2 1.15 0.99 1.39 1.69 0.60 0.46 1.10LCD TV 300 20 1.32 1.34 1.56 0.88 0.89 0.58 0.95R FR 150 10 1.35 1.22 1.33 1.44 1.92 2.08 1.54LCD 40 2.67 2.35 2.04 3.30 1.24 0.23 1.92 2.43FUR 500 33.33 2.53 2.68 2.72 1.71 1.26 1.67 1.76CRT TV 70 4.67 5.81 4.93 6.77 7.02 7.37 7.15 4.42LAP 75 5 6.15 5.37 5.66 5.57 7.97 11.38 11.05PC 100 6.67 7.05 7.32 8.02 5.47 2.43 2.58 4.62ASD DRY 1000 66.67 12.58 15.23 8.26 5.90 10.85 4.90 8.55WSH 180 12 16.09 12.80 17.83 20.16 22.74 24.57 19.76MW 1200 80 26.42 33.09 16.72 9.14 7.20 6.13 4.14



The values of the total equivalent CFL index and the associated assumptions are shown in Table V,which can be used for several comparisons. For example, the harmonic effect of CFLs is comparablewith that of the home-office equipment such as desktop PCs. The washer (ASD-based) and the micro-wave oven are identified as more significant harmonic sources. The LCD TV injects little harmoniccontent, whereas the CRT TV injects an amount comparable with that of the CFLs. The collective loadof CFLs in a typical household is a harmonic contributor comparable with the other harmonic sources.

5. THE HARMONIC PHASE ANGLE STUDY

The equivalent CFL index is useful for comparing the harmonic current magnitudes of the homeappliances, but it does not provide information on the harmonic phase angle diversification. Theharmonic current phase angles are important because they indicate how the harmonics from varioushome appliances will either add to or cancel out each other. Although this effect has been studiedby analyzing the diversity effect of single-phase diode bridge rectifiers [17,18], in this article weexpanded the concept by defining compatibility indices among all appliance types. The harmoniccompatibility index (CIh) is introduced in this section to account for the current harmonic phase anglesof the appliances as these are operated together.The adopted definition for the compatibility index is as follows:

1. Like the method used for the equivalent CFL index, for each type of appliance measured, therepresentative harmonic current of each appliance is used for the harmonic compatibility index.The representative harmonic currents are presented in Appendix A.

2. For each harmonic order h, the phasor relationship of the appliance’s current to that of therepresentative CFL current is determined by using the following equation:

CIh Appliance ¼_Ih Appliance þ _Ih CFL eq

�� _Ih Appliance

�� þ _Ih CFL eq

�� ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiI2 þ I2 þ 2I2 cos θð Þp

2I¼ cos

θh2

� �� (4)

where

_Ih CFL eq ¼ _Ih CFL � Ratioh Appliance (5)

and

θh ¼ ang _Ih Appliance� � ang _Ih CFL

� �� (6)

The meaning of this equation is that because the magnitude of the CFL current phasor is scaled to be ofthe same magnitude as the magnitude of that of the appliance, only the phase angles of the appliance

Table V. Comparing the harmonic effect of collective home appliances.

Appliancetype

Operatingpower(W/unit)

No.units

Total power(W)

Powerratio

Totalequivalent

CFL

MBL 30 2 60 0.67 0.15LCD TV 300 1 300 3.33 0.20ASD FR 170 1 170 1.89 0.21R FR 150 1 150 1.67 0.28FUR 500 1 500 5.56 0.58LCD 40 2 80 0.89 0.78CFL 15 6 90 1.00 1.00CRT TV 70 1 70 0.78 1.00LAP 75 1 75 0.83 1.05PC 100 2 200 2.22 2.18ASD DRY 1000 1 1000 11.11 2.76WSH 180 1 180 2.00 2.87MW 1200 1 1200 13.33 4.05

A. B. NASSIF ET AL.


and the CFL are taken into consideration in the phasor operation. As θh varies between 0 and p radians,CIh_Appliance varies between 0 and 1. Figure 5 shows representative values for the index as it is used toexplain the index at each harmonic order h. Assume the CFL is being compared with Appliance 1,Appliance 2, Appliance 3, and Appliance 4; Figure 5a reveals that the index calculated for Appliance 1means solely harmonic addition, resulting in CIh_Appliance 1= 1 because both components are in phase;that is, θ= 0�. Conversely, the index calculated for Appliance 2 (Figure 5b) is CIh_Appliance 2= 1/√2because θ= 90�. The index calculated for Appliance 3 (Figure 5c) is found to be CIh_Appliance 3= 1/2because the value of CIh is the same of that of the magnitudes of the individual phasors; that is,θ= 120�. Finally, the index calculated for Appliance 4 (Figure 5d) is CIh_Appliance 4= 0 because bothcomponents are in opposite phase angles; that is, θ= 180�.

3. An aggregation similar to that proposed for the magnitude-based equivalent CFL index isintroduced for the harmonic compatibility indices, also by using the same weighting factor,which results in the following aggregated index:

CIAppliance ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiX

wh � CIh Appliance� 2q

(7)

The resulting aggregation of the individual CIh_Appliance carries the information about the combinedharmonic compatibility into a single index. The calculation of the compatibility between the PC andthe CFL is used as an example. The individual current phasors are displayed in Figure 6. The indicesCIh_PC and CIPC are shown in Figure 6, where they have all been normalized by the CFL’s thirdharmonic, to show the relative effect of each individual CIh_PC on the total CIPC.The phase angle–based harmonic compatibility index can reflect the measure of the harmonic

cancellation between two appliances should there be no difference in magnitude, that is, thecompatibility index clearly represents potential cancellation should the magnitudes be of comparablesizes. In other words, if both appliances’ current harmonics were of the same magnitude, the indexCIh_Appliance would reflect the amount of harmonic cancellation. Suppose case 1, θh= 0, resulting inCIh_Appliance = 100%, whichmeans zero harmonic cancellation, and case 2, θh=p/2 and CIh_Appliance = 71%,which means 29% harmonic cancellation as compared with the first case.

5.1. Basic results for the harmonic compatibility index

The results for the compatibility index for all appliances are calculated and presented in Table VI, in theascending order of CIAppliance.

30

210

60

240

90

270

120

300

150

330

180 0

a

30

210

60

240

90

270

120

300

150

330

180 0

b

30

210

60

240

90

270

120

300

150

330

180 0

c

30

210

60

240

90

270

120

300

150

330

180 0

d

_ 1h ApplianceI

_ _h CFL eqI

_ 2h ApplianceI

_ _h CFL eqI

_ _h CFL eqI

_ 3h ApplianceI

_ _h CFL eqI

_ 4h ApplianceI

0o 90o

120o 180o

1hCI 1 2hCI

1 2hCI 0hCI

Figure 5. Phasor diagram to illustrate the compatibility index.



To further illustrate the general behavior of the appliances, Figure 7 shows the distribution of theCIh_Appliance indices by using histograms. The following conclusions can be drawn from the resultspresented Table VI and Figure 7:

1. More appliances show higher compatibility (have higher CIh_Appliance) at the 3rd harmonic thanat other orders. This finding agrees with expected harmonic cancellation that can occur indistribution systems [19,20] and indicates that the harmonic cancellation at such an order issmaller than that observed at higher orders. The appliances that show notably low compatibilityat the 3rd harmonic are the fridges and the ASD-based dryer.

2. The harmonic compatibility is generally low for harmonics at the 5th and higher orders. Thisfinding implies that the harmonic cancellation at these orders is higher than that observed forthe 3rd harmonic.

For the magnitude-based index, the equivalent CFL gives absolute results by using the CFL as abenchmark. These results are straightforward for assessing the relative harmonic severity among theappliances. For the phase angle assessment, the harmonic compatibility index using the CFL as atemplate does not provide directly the relative harmonic current angle difference between any two

0.5

1

30

210

60

240

90

270

120

300

150

330

180 0

0.5

1

30

210

60

240

90

270

120

300

150

330

180 0

0.5

1

30

210

60

240

90

270

120

300

150

330

180 0

0.5

1

30

210

60

240

90

270

120

300

150

330

180 0

0.5

1

30

210

60

240

90

270

120

300

150

330

180 0

0.5

1

30

210

60

240

90

270

120

300

150

330

180 0

CI3 = 0.82 CI

5 = 0.39 CI

7 = 0.12

CI9 = 0.33 CI

11 = 0.89

CI13

= 0.94

CIPC

= 0.64

Figure 6. Normalized current phasors along with CIh for PC and CFL.

Table VI. Harmonic compatibility index results using CFL as a template.

CIAppliance CI3 CI5 CI7 CI9 CI11 CI13

ASD FR 0.27 0.03 0.31 0.23 0.07 0.99 0.79ASD DRY 0.35 0.29 0.34 0.32 0.75 0.53 0.15R FR 0.61 0.27 0.98 0.14 0.53 0.64 0.7CRT TV 0.63 0.82 0.39 0.22 0.3 0.72 0.56WSH 0.64 0.74 0.4 0.61 0.7 0.93 0.53PC 0.64 0.82 0.39 0.12 0.33 0.89 0.94FUR 0.65 0.5 0.8 0.71 0.76 0.66 0.54MW 0.74 0.94 0.34 0.73 0.01 0.13 0.87LCD 0.81 0.93 0.73 0.54 0.82 0.65 0.63LCD TV 0.88 0.89 0.86 1.00 0.72 0.53 0.99MBL 0.89 0.85 0.97 1.00 0.84 0.07 0.27EBL 0.92 0.97 0.92 0.86 0.98 0.3 0.08LAP 0.93 0.98 0.91 0.87 1.00 0.16 0.88CFL 1.00 1.00 1.00 1.00 1.00 1.00 1.00

A. B. NASSIF ET AL.


given appliances. To obtain such a relationship, the indices between any two appliances can beobtained by using the CFL-calculated CIh_Appliance as follows:

CIh Appliances 1-2 ¼_Ih Appliance 1 þ _Ih Appliance 2

�� _Ih Appliance 1

�� þ _Ih Appliance 2

�� ¼ cosθh Appliance 1 � θh Appliance 2

2

� �� (8)

where the appliances currents are normalized to have the same magnitude. The results for the harmoniccompatibility indices CIh_Appliances 1–2 for each appliance in relation to all the other appliances at the3rd–5th harmonic orders are shown in Table VII. Higher-order indices are omitted because of spacelimitations. Table VII provides a direct measure of the harmonic compatibility between any two givenappliances.

5.2. Results for the aggregated harmonic compatibility index

The indices CIh_Appliances 1–2 such as those presented in Table VII can be combined as a weight average.The aggregated indices can be obtained by using a weight average similar to that of Equations (3) and

0 0.2 0.4 0.6 0.8 10

2

4

0 0.2 0.4 0.6 0.8 10

2

4

Fre

quen

cy o

f CI 5

0 0.2 0.4 0.6 0.8 10

2

4

Fre

quen

cy o

f CI 7

Fre

quen

cy o

f CI 11

0 0.2 0.4 0.6 0.8 10

2

4

Fre

quen

cy o

f CI 9

0 0.2 0.4 0.6 0.8 10

2

4

0 0.2 0.4 0.6 0.8 10

2

4

Fre

quen

cy o

f CI 13

Fre

quen

cy o

f CI 3

Figure 7. Distribution of the CIh_Appliance indices.

Table VII. Harmonic compatibility index results for 3rd and 5th harmonic.

3rd MW ASDFR

WSH LCDTV

PC LAP LCD CFL FUR EBL MBL ASDDR

RFR

CRTTV

5th

MW 1.00 0.36 0.48 0.99 0.96 0.99 1.00 0.94 0.18 0.99 0.63 0.59 0.06 0.96ASD FR 1.00 1.00 0.65 0.49 0.60 0.25 0.41 0.03 0.85 0.26 0.50 0.97 0.95 0.60WSH 0.72 0.74 1.00 0.35 0.22 0.58 0.43 0.74 0.95 0.57 0.98 0.42 0.85 0.23LCD TV 0.78 0.75 0.12 1.00 0.99 0.97 1.00 0.89 0.04 0.97 0.51 0.70 0.20 0.99PC 0.73 0.75 1.00 0.13 1.00 0.93 0.97 0.82 0.09 0.93 0.39 0.79 0.33 1.00LAP 0.07 0.10 0.74 0.58 0.73 1.00 0.99 0.98 0.29 1.00 0.71 0.50 0.06 0.93LCD 0.39 0.42 0.92 0.28 0.91 0.95 1.00 0.93 0.13 0.99 0.59 0.63 0.11 0.98CFL 0.34 0.31 0.40 0.86 0.39 0.91 0.73 1.00 0.50 0.97 0.85 0.29 0.27 0.82FUR 0.28 0.32 0.87 0.38 0.87 0.98 0.99 0.80 1.00 0.29 0.88 0.68 0.97 0.09EBL 0.05 0.09 0.73 0.59 0.73 1.00 0.94 0.92 0.97 1.00 0.71 0.50 0.05 0.93MBL 0.55 0.52 0.18 0.95 0.17 0.80 0.56 0.97 0.64 0.80 1.00 0.26 0.74 0.40ASD DR 0.77 0.79 1.00 0.19 1.00 0.69 0.89 0.34 0.83 0.68 0.11 1.00 0.84 0.79R FR 0.51 0.48 0.22 0.94 0.22 0.82 0.59 0.98 0.68 0.83 1.00 0.15 1.00 0.32CRT TV 0.73 0.75 1.00 0.13 1.00 0.73 0.91 0.39 0.87 0.72 0.17 1.00 0.22 1.00



(7) and by using the same weighting factor. The aggregated index is obtained as

CIAppliances 1-2 ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiX

wh � CIh Appliances 1-2� 2q

(9)

Table VIII shows the aggregate compatibility indices for each appliance in relation to all the otherappliances. Table VIII also shows that the lowest compatibility levels are obtained when comparingthe pulsed loads (PCs, laptops, and LCD monitors) with the load from the fridge.As a further illustration, Figure 8 shows the distribution of the aggregated index for all appliances by

using histograms. Figure 8 is useful to identify the general compatibility trends of each individualappliance. For example, Figure 8 shows that generally, the microwave oven is more compatible thanthe ASD-based fridge with all other appliances.

5.3. Effect of the phase angle on the harmonic current cancellation

A case study is presented to illustrate the importance of the phase angles of the appliances’ currentharmonics. The harmonic currents of selected groups of appliances are analyzed. The combinationsof appliances were determined by using survey results, as presented in Table IX. The harmonicinjections of these groups are compared by both considering and ignoring the phase angle disparity.This study is useful to show the harmonic current cancellations that occur when appliances of differenttypes are connected together. The indices proposed in this article are used to explain the results.The main results from these studies are presented in Figure 9, which shows the total power of the

combination of appliances in the x-axis and the current harmonics in the y-axis. A linear data fittingis also provided to show the rate of harmonic current increase as the total power increases. The upperline is the arithmetic sum, and the lower one is the phasor sum. Figure 9 reveals the following:

• As the total power of each set of appliances increases, the trends of harmonic currents increaseaccordingly, for both cases (arithmetic and phasor sum). Although the trends increase, increasingthe power does not necessarily increase the net current harmonics. As explained earlier, this resultis expected because of a large amount of harmonic cancellation, especially at levels higher thanthe 3rd harmonic order.

• The slopes of the linear data fitting show larger difference as the harmonic orders increase. For thecase presented in Figure 9, the angles of the slopes for 3rd, 5th, 7th, and 9th harmonics are 69�, 54�,38�, and 27�, respectively, for the case of the arithmetic sum. For the case of the phasor sum, theangles of the slopes for 3rd, 5th, 7th, and 9th harmonics are 60�, 25�, 14�, and 5�, respectively.The results confirm that (i) the rate of increase of the harmonic currents is more prominent at lowerorder harmonics and that (ii) the differences of the slope angle increases for higher-orderharmonics.

Table VIII. Harmonic compatibility index results for aggregated harmonics.

Total MW ASDFR

WSH LCDTV

PC LAP LCD CFL FUR EBL MBL ASDDRY

RFR

CRTTV

MW 1.00 0.72 0.66 0.87 0.84 0.80 0.83 0.74 0.47 0.79 0.62 0.70 0.41 0.83ASD FR 0.72 1.00 0.72 0.57 0.72 0.31 0.52 0.27 0.70 0.33 0.49 0.89 0.82 0.73WSH 0.66 0.72 1.00 0.34 0.67 0.69 0.72 0.64 0.93 0.69 0.75 0.74 0.69 0.66LCD TV 0.87 0.57 0.34 1.00 0.72 0.83 0.76 0.88 0.34 0.82 0.75 0.54 0.56 0.71PC 0.84 0.72 0.67 0.72 1.00 0.79 0.92 0.64 0.57 0.78 0.30 0.87 0.49 0.97LAP 0.80 0.31 0.69 0.83 0.79 1.00 0.95 0.93 0.72 0.99 0.77 0.63 0.50 0.77LCD 0.83 0.52 0.72 0.76 0.92 0.95 1.00 0.81 0.70 0.95 0.57 0.80 0.47 0.88CFL 0.74 0.27 0.64 0.88 0.64 0.93 0.81 1.00 0.65 0.92 0.89 0.35 0.61 0.63FUR 0.47 0.70 0.93 0.34 0.57 0.72 0.70 0.65 1.00 0.72 0.79 0.76 0.83 0.56EBL 0.79 0.33 0.69 0.82 0.78 0.99 0.95 0.92 0.72 1.00 0.77 0.64 0.52 0.77MBL 0.62 0.49 0.75 0.75 0.30 0.77 0.57 0.89 0.79 0.77 1.00 0.29 0.79 0.38ASD DRY 0.70 0.89 0.74 0.54 0.87 0.63 0.80 0.35 0.76 0.64 0.29 1.00 0.69 0.85R FR 0.41 0.82 0.69 0.56 0.49 0.50 0.47 0.61 0.83 0.52 0.79 0.69 1.00 0.51CRT TV 0.83 0.73 0.66 0.71 0.97 0.77 0.88 0.63 0.56 0.77 0.38 0.85 0.51 1.00

A. B. NASSIF ET AL.


• The observed behavior can be explained by using both indices. The equivalent CFL index is wellsuited for explaining the arithmetic sum. Between 900 and 2100W, for example, as themicrowave oven is connected, an increase occurs in the 3rd harmonic current that is equivalentto Ratio3 = 33 CFLs, whereas the increase in the 9th harmonic is equivalent to that of onlyRatio9 = 7 CFLs. On the other hand, the compatibility index is useful for explaining the phasorsum. For the same case, the microwave oven has high compatibility at the 3rd harmonic withthe computing loads and CFL but low compatibility with the furnace and fridge. For this reason,the increase in the 3rd harmonic current is not as high as that shown for the arithmetic sum but isstill high because of the relatively high harmonic compatibility. At the 9th harmonic, however, themicrowave has low compatibility with all loads, except those from the PC and furnace, resultingin significant harmonic cancellation. Therefore, there is barely any increase for that powerdemand step.

6. CONCLUSIONS

A comparison of the harmonic levels among home appliances was carried out. The issues relating tothe harmonic current magnitudes and phase angles were studied by using developed indices. Theequivalent CFL index was introduced to quantify the relative severity of the magnitudes of the harmoniccurrents from home appliances. The harmonic compatibility index is an indication of how much currentcancellation can occur between the harmonic currents of any two given appliances and therefore dealswith the phase angle cancellation issue. Both indices provide results that can be used as a benchmark forharmonic studies at a residential customer level. The key conclusions are as follows:

1. The current spectrum decline rate study was performed to characterize the magnitudes of thecurrent spectra of all appliances. The estimated parameter a is unique for each appliance or setof electrically similar appliances.

0 0.5 10

5

CI-

MW

0 0.5 10

5

CI-

AS

D F

RD

0 0.5 10

5

10

CI-

WS

H

0 0.5 10

5C

I-LC

D T

V

0 0.5 10

5

CI-

PC

0 0.5 10

5

CI-

LAP

0 0.5 10

5

CI-

LCD

0 0.5 10

5

CI-

CF

L

0 0.5 10

5

CI-

FU

R

0 0.5 10

5

CI-

EB

L

0 0.5 10

5C

I-M

BL

0 0.5 10

5

CI-

AS

D D

RY

0 0.5 10

5

CI-

R F

R

0 0.5 10

5

CI-

CR

T T

V

Figure 8. Distribution of the CIAppliance_1–2 indices for all appliances.



2. The equivalent CFL index compares the harmonic effect of appliances on the basis of individualunits. On the basis of the index, the appliances can be ranked in the following ascending order onan individual unit basis by taking into consideration their harmonic severities: MBL, CFL, ASDFR, EBL, LCD TV, R FR, LCD, FUR, CRT TV, LAP, PC, ASD DRY, WSH, and MW.

3. The total equivalent CFL index compares the harmonic effect of each type of appliances installedin a typical household. The appliances can be ranked in the following ascending order on thebasis of the total number of units by taking into consideration their harmonic severities: MBL,ASD FR, LCD TV, R FR, FUR, LCD, CFL, CRT TV, LAP, PC, ASD DRY, WSH, and MW.As the number of CFLs increases in a house, their harmonic-producing role may become moreprominent (if the use of other appliances does not increase).

4. The phase angle-based harmonic compatibility index is a direct measure of the harmoniccancellation between two appliances should there be no difference in magnitude. It can be usedto evaluate whether the appliances’ harmonic currents will mainly add to or cancel out each other.It was found that the compatibility was highest at the 3rd harmonic order and low at the higher-order

Table IX. Appliance usage patterns.

Appliances combinations

Case 1 Power: 663W Case 2 Power: 903W

•FUR•FUR •R FR •PC+LCD •R FR

•CRT TV •2 CFLs

Case 3 Power: 1011W Case 4 Power: 2294W•FUR

•FUR •R FR •PC+LCD •R FR•PC+LCD •4 CFLs •LAP •6 CFLs•LAP •CRT TV •MW •CRT TV

Case 5 Power: 2632W Case 6 Power: 4049W•FUR •R FR

•FUR •2PCs + 2 LCD •8 CFLs•PC+LCD •R FR •LAP •2 TVs•LAP •8 CFLs •MW •WSH+ASD DRY•MW •2 CRT TVs •2 MBLs

0 1 2 3 4 50

5

10

15

0 1 2 3 4 50

2

4

6

5th h

arm

onic

[A]

0 1 2 3 4 50

1

2

3

4

Power demand [KW]0 1 2 3 4 5

0

1

2

3

9th h

arm

onic

[A]

Power demand [KW]

Arithmetic sum

Phasor sum

Arithmetic sum

Arithmetic sumArithmetic sum

Phasor sum

Phasor sum

Phasor sum

3rd h

arm

onic

[A]

7th h

arm

onic

[A

]

Figure 9. Effect of the harmonic current phase angles.

A. B. NASSIF ET AL.


harmonics. Usually, the appliances had high compatibility to those of similar construction. Forexample, diode-bridge rectifier-based appliances had harmonic currents of high compatibility.

Although the results are based on a relatively small sample of appliances, the findings have beenquite useful. For example, the CFL is not as harmonically harmful as one may expect even if one houseadds 5 to 10 of them. Another value of this work is the establishment of a framework for comparing theharmonic effect of various appliances consistently. The next effort is to increase the population of theappliances measured so a useful database can be established.

7. LIST OF SYMBOLS AND ABBREVIATIONS

ASD Adjustable Speed DriveASD FR Fridge Using Adjustable Speed Drive Technologya Decline Rate of the Current SpectrumCFL Compact Fluorescent LampCIh Compatibility Index at Harmonic Order hCRT Cathode Ray TubeDRY Laundry DryerDPF Displacement Power FactorEBL Electronic Ballast Fluorescent LampFUR Furnaceh Harmonic OrderI1 Current at Fundamental FrequencyIh Current at harmonic hIh_appliance Appliance’s Harmonic Current at Order hIh_appliance Representative CFL Harmonic Current at Order hINC Incandescent LampLAP LaptopLCD Liquid Crystal DisplayMBL Magnetic Ballast Fluorescent LampMW Microwave OvenPC Desktop Personal ComputerPF Total Power FactorR FR Conventional FridgeTHDI Current Total Harmonic DistortionTHDV Voltage Total Harmonic Distortionθh Angle Between Voltage and Current at Harmonic Order hTV TelevisionVSAC Variable Speed Air Conditioningwh Weighting Factor at Harmonic Order hWSH Laundry Washer

REFERENCES

1. Sulfstede LE. Applying Power Electronics to Residential HVAC-The Issues. IEEE Transactions on IndustryApplications 1993; 29(2): 300–305.

2. Emanuel AE, Janczak J, Pileggi DJ, et al. Voltage Distortion in Distribution Feeders with Nonlinear Loads. IEEETransactions on Power Delivery 1994; 9(1): 79–87.

3. Gruzs TM. A Survey of Neutral Currents in Three-Phase Computer Power Systems. IEEE Transactions on IndustryApplications 1990; 26(4): 719–725.

4. IEA International Energy Agency. Cool Appliances – Policy Strategies for Energy Efficient Homes. IEA -International Energy Agency: Paris, France, 2003.

5. NEMA. CFL Share of Household Lamps Reaches New High, Nov. 10, 2008, available online on January 23, 2009at: http://www.nema.org/media/pr/20081110a.cfm



http://www.nema.org/media/pr/20081110a.cfm

6. Verberder RR, Morse OC, Alling WR. Harmonics from Compact Fluorescent Lamps. IEEE Transactions onIndustry Applications 1993; 29(3): 670–674.

7. Pileggi DJ, Gulachenski EM, Root CE, Gentile TJ, Emanuel AE. The Effect of Modern Compact Fluorescent Lightson Voltage Distortion. IEEE Transactions on Power Delivery 1993; 8(3): 1451–1459.

8. Dwyer R, Khan AK, McGranaghan M, et al. Evaluation of Harmonic Impacts from Compact Fluorescent Lights onDistribution Systems. IEEE Transactions on Power Delivery 1995; 10(4): 1772–1779.

9. Browne N, Perera S, Ribeiro PF. Harmonic Levels and Television Events. Proceedings of the IEEE PES GeneralMeeting 2007, June 24–28, 2007.

10. Katic V, Dumnic B, Mujovic S, Radovic J. Effects of Low Power Electronics & Computer Equipment on PowerQuality at Distribution Grid-Measurements and Forecast, Proceedings of the IEEE International Conference onIndustrial Technology, 2004, pp. 585–589.

11. Hardie S, Watson N. The effect of new residential appliances on Power Quality. Australasian Universities PowerEngineering Conference (AUPEC) 2010, Christchurch, 5–8 December 2010.

12. Hardie S, Watson N. Power Quality implications of new residential appliances. Electricity Engineers’ Association(EEA) Conference, Christchurch, 17–18 June 2010.

13. IEEE Trial-Use Standard Definitions for the Measurement of Electric Power Quantities under Sinusoidal,Nonsinusoidal, Balanced or Unbalanced Conditions, IEEE STD 1459–2000.

14. IEEE Working Group on Nonsinusoidal Situations: Effects on Meter Performance and Definitions of Power.Practical Definitions for Powers in Systems with Nonsinusoidal Waveforms and Unbalanced Loads: a Discussion.IEEE Transactions on Power Delivery 1996; 11(1): 79–101.

15. Yacamini R. Power System Harmonics. Part 1. Harmonic Sources. Power Engineering Journal 1994; 8(4): 193–198.16. Statistics Canada, Selected dwelling characteristics and household equipment, available online at: http://www40.

statcan.gc.ca/l01/cst01/famil09b-eng.htm17. Mansoor A, Grady WM, Chowdhury AH, Samotyj MJ. An Investigation of Harmonic Attenuation and Diversity

among Distributed Single-Phase Electronic Loads. IEEE Transactions on Power Delivery 1995; 10(1): 467–473.18. Mansoor A, Grady WM, Staats PT, Thallam RS, Doyle MT, Samotyj MJ. Predicting the Net Harmonic Currents

Produced by Large Numbers of Distributed Single-Phase Computer Loads. IEEE Transactions on Power Delivery1995; 10(4): 2001–2006.

19. Emanuel AE, Orr JA, Cyganski D, Gulachenski EM. A Survey of Harmonic Voltages and Currents at theCustomer’s Bus. IEEE Transactions on Power Delivery 1993; 8(1): 411–421.

20. Hegazy YG, Salama MMA. Calculations of Diversified Harmonic Currents in Electric Distribution Systems. IEEProceedings Generation and Distribution 2003; 150(6): 651–658.

APPENDIX A

Measurements were conducted for more than one device within the same appliance type. Each devicewill yield a set of harmonic spectra data. From these data, representative harmonic characteristics ofthe devices were derived by using a weighted average of the individual device data according to thefollowing steps:

1. Compute the per-watt (pw) waveform data for each device i of the same appliance type:

Ipw�i tð Þ ¼ Ii tð Þ=Pi rms (A1)

where Pi_rms is the measured operating power of the device.

2. Average the previously mentioned pw waveform data over the number of devices measuredwithin the same brand:

Ipw-brand�b tð Þ ¼XNbrand

i¼1

Ipw�i tð Þ=Nbrand (A2)

where Nbrand is the number of devices of the same brand

3. Average the previously mentioned pw waveform data over the numbers of brands measured:

Ipw tð Þ ¼XBb¼1

Ipw-brand�b tð Þ=B (A3)

4. Select a representative operating power level for the type of applications whose data is beingprocessed, Ptypical. Some of the representative operating power levels are determined from thesurvey data and others are based on common sense estimates. Different operating conditions

A. B. NASSIF ET AL.


http://www40.statcan.gc.ca/l01/cst01/famil09b-eng.htm

http://www40.statcan.gc.ca/l01/cst01/famil09b-eng.htm

for each type of appliance were also considered. Although devices such as the CFL have anapproximately constant behavior over all operating conditions, other devices such as the washingmachine will operate under several different conditions. We did measure them under differentconditions but only considered for this analysis the conditions where the largest harmonicinjections (which coincided with largest load demand) were present; those were also conditionsthat lasted longer (such as the washing machine “washing cycle”). If all conditions were to betaken into consideration, the large combinatorial nature of the problem would likely prohibit thisanalysis.

5. The representative waveform for the type of appliances under study is finally determined as

IAppliance tð Þ ¼ Ipw tð Þ � Ptypical (A4)

The previously mentioned data have two applications. One is to establish the harmonic currentsource model, which is obtained by calculating the spectra of the obtained waveforms, for theappliance type. The other application is to compute the equivalent CFL and harmonic compatibilityindices. The reason for using steps 2 and 3 is because it has been observed that appliancesmanufactured by the same brand usually have very similar current waveforms. Consequently, if thenumber of measured devices of brand X is much higher than that of brand Y, the final model willbe biased by the devices of brand X. Therefore, a model for each brand is obtained and finally auniversal model that fits many brands is proposed. This model can be scaled by the appliance ratedpower and this can be used for harmonic analyses. The pw models for the home appliances arepresented in Figure 10.

0.000

0.002

0.004

0.006

0.008

0.010

0.012

0.014

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Harmonic order

PC

LCD

LAP

LCD TV

MW

ASD FR

WSH

FUR

CRT TV

ASD DRY

EBL

MBL

-200

-150

-100

-50

0

50

100

150

200

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Harmonic order

PC

LCD

LAP

LCD TV

MW

ASD FR

WSH

FUR

CRT TV

ASD DRY

EBL

MBL

pw

mag

. [A

/W]

pw

an

gle

[d

egre

es]

Figure 10. General pw current source model for all appliances.



indices for comparative assessment of the …apic/papers/power_quality/2012...indices for...

Documents