Power Quality Diagnosis

Expert Systems with Applications xxx (2011) xxx–xxx

Contents lists available at ScienceDirect

Expert Systems with Applications

journal homepage: www.elsevier .com/locate /eswa

Power quality diagnosis using time frequency analysis and rule based techniques

M. Faisal, A. Mohamed ⇑, H. Shareef, A. HussainDepartment of Electrical, Electronic and Systems Engineering, University Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia

a r t i c l e i n f o a b s t r a c t

Keywords:Event identificationPower quality diagnosisPower quality monitoringRule-basedS-transform

0957-4174/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.eswa.2011.04.047

⇑ Corresponding author. Tel.: +603 89216006.E-mail address: [email protected] (A. Mohamed).

Please cite this article in press as: Faisal, M., etApplications (2011), doi:10.1016/j.eswa.2011.04

Diagnosing a power quality disturbance means identifying the type and cause of the disturbance. Fastdiagnosis of power quality disturbances is important so as to assist network operators in performingcounter measures and implementing suitable power quality mitigation actions. In this study a novelmethod for performing power quality diagnosis is presented by using the S-transform and rule basedclassification techniques. The proposed power quality diagnosis method was evaluated for its function-ality in detecting the type of short duration voltage disturbances and identifying the cause of the distur-bances which may be due to permanent or non permanent faults. Based on the results, this new methodhas the potential to be used in the existing real time power quality monitoring system in Malaysia toexpedite the diagnosis on the recorded voltage disturbances.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

It is widely believed that power quality problems can only betraced to the incoming power from the utility. However, it is notalways true because the source of such problems can be tracedeither to the customer facility or even up to the equipment insidea facility. Power quality problems that originate from the utility of-ten have the greatest impact on a facility’s operation. Typical util-ity-generated events are such as breaker clearing which canproduce voltage sag and arcing contactors which may generate im-pulses (Yalcinkaya, Bollen, & Crossley, 1998). The power qualityproblems arise within the customer’s own installation may bedue to lose connections, overloaded circuits and transformers,ground loops and wiring errors. In addition, some automated andcomputer-based equipments produce harmonics which may prop-agate into the network and affect other customers.

The initial step taken to solve power quality problems is byinstalling permanent power quality recorders (PQR) in the powersupply networks either at the main substations or the point ofcommon coupling (PCC) between the customers and the powerutility. A power quality monitoring system (PQMS) as shown inFig. 1 is usually required to connect all the PQR via a communica-tion network to transmit the real-time measured power qualitydata to the power utility engineers. The PQMS server immediatelysends a summary of the power quality disturbance data throughshort message system to the utility engineers. However, the exist-ing PQMS does not diagnose power quality disturbances in whichthe sources and causes of the disturbances are not known. The

ll rights reserved.

al. Power quality diagnosis usin.047

sources of disturbance may originate from distribution feeder 1(DF1), transmission feeder 2 (TF2) and other distribution feeders.Usually, the sources and causes of the disturbances are known onlyafter conducting site inspections which is a time consumingprocess.

Many research works focus on automated detection and clas-sification of power quality disturbances but not on diagnosis ofdisturbances. Some studies on power quality diagnosis considerevent identification, stochastic diagnosis, and power qualityindexing and cause identification (Azam, Fang, Pattipati, &Rajaiah Karanam, 2004; Il-Yop et al., 2003; Il-Yop, Dong-Jun,Joong-Moon, Seon-Ju, & Seung-Il, 2007; Kezunovic, 2001). Forevent identification, three functions are considered whichinclude event location, cause identification and solution sugges-tion. An automated software approach for the analysis of voltagesags, their causes and impacts has been developed using theFourier and wavelet-transform signal processing techniques andfuzzy expert system (Kezunovic, 2001). Another method foridentifying the types and causes of disturbances is by usingthe model based approach and decision fusion (Azam et al.,2004). A network based power quality diagnostic system (PQDS)that incorporates a graphical user interface of an existing PQMShas also been developed (Il-Yop et al., 2003, 2007). This PQDShas the functions for event detection, power quality trend anal-ysis, event cause identification and event location. However, inthis PQDS, the source of voltage sag can only be determinedby analyzing the actual power system topology and this presentsa disadvantage to the system.

In this paper a novel approach to perform automatic powerquality diagnosis for identifying the cause of short duration volt-age disturbances is presented. The proposed method is developedby using the S-transform and rule based techniques. The excellent

g time frequency analysis and rule based techniques. Expert Systems with

http://dx.doi.org/10.1016/j.eswa.2011.04.047

mailto:[email protected]


http://www.sciencedirect.com/science/journal/09574174

http://www.elsevier.com/locate/eswa


33 kV

PQRCurrent Input

132 kV

Source XX X X

X

X

DF3

DF1

DF2

X

B

Voltage Input

X

C

D

JTF2

33 kV 415 Volt

E

11 kV33 kV

A

DF4

TF1

PQMSServer

SendSMS

Access usingDesktop computer

Fig. 1. A power quality monitoring system (PQMS) installed in a distribution network.

2 M. Faisal et al. / Expert Systems with Applications xxx (2011) xxx–xxx

time-frequency resolution characteristic of the S-transform makesit an attractive candidate for the analysis of power quality distur-bances under noisy condition and has the ability to detect thesingle and multiple disturbances correctly. Features extracted bythe S-transform are applied to a modular rule based classifierfor automatic diagnosis of the disturbances as shown in Fig. 2.The process starts with the recording of disturbance data usingthe on-line PQMS. These data are then processed using theS-transform to extract the features that characterize the distur-bances. These features are then applied to three rule basedclassifiers (RBC). The first RBC detects the type of short durationvoltage disturbances which are either voltage sag or swell, whilethe second RBC diagnoses the cause of sag or swell as either dueto permanent or non permanent faults. The third RBC than furtherclassifies the non permanent fault as either transient or incipientfault. In this way, the cause of voltage sag or swell can be deter-mined as to whether it is due to a permanent fault, transient faultor incipient fault.

Data(disturbancewaveforms)

Featureextractiontechnique

Dataacquistion

system

S-transformPower Quality

Recorders

Incipientfault

START

END

3rd Rule BaClassifier

Categories faults

YES

YES

NOTransient

fault ?

Fig. 2. Diagnosis of short dura

Please cite this article in press as: Faisal, M., et al. Power quality diagnosis usinApplications (2011), doi:10.1016/j.eswa.2011.04.047

2. Categories of network faults and causes of short durationvoltage disturbances

Faults that occur in a power network may be categorized aseither permanent or non permanent faults as shown in Fig. 3.Faults in a power network generally cause large increase in shortcircuit current to flow in the network which in turn gives rise tolarge voltage drops across the impedances of the supply system.These short circuits are caused by external interferences termedas permanent faults. Examples of causes of permanent fault aredue to underground cable faults and flashover at medium voltagecircuit breakers. Permanent faults that occur in a power networktend to produce short duration voltage disturbances such as volt-age sag and swell as specified in IEC/TR 61000-2-8.

Another type of network fault is termed as a non permanentfault which occurs at random moments for a finite period of time.The non permanent faults can be further categorized as either tran-sient or incipient faults. A transient fault is defined as a fault that is

Nonpermanent

fault

Categories offaults

2nd Rule BasedClassifier

sed

of

NO

Permanentfault ?

Shortdurationevent?

YES

NO

1st Rule BasedClassifier

tion voltage disturbances.



Network Faults

Permanent FaultsNon Permanent

Faults

Transient Faults Incipient Faults

Fig. 3. Categories of network faults.

M. Faisal et al. / Expert Systems with Applications xxx (2011) xxx–xxx 3

no longer present if power is disconnected for a short time (White,Transient faults, & IEEE Aerospace Conference, 2004). Typicalexamples of causes of transient faults are due to momentary treecontact, bird or other animal contact, lightning strike and conduc-tor clash.

Incipient faults are on the other hand are considered as non per-manent faults caused by partial damage or contamination that pro-gressively weakens the integrity of the network components overtime and leads to insulation failure (Weeks & Steiner, 1982). Suchfaults are intermittent in nature and are considered as self-clearingarcing faults which occur very frequently prior to failing perma-nently. These events typically last for one half-cycle and extinguishat the first natural zero crossing of the current. The magnitude ofthe half-cycle event is primarily dependent on the location of thefault on the feeder. But is also dependent on the point on the volt-age waveform where the fault starts. Operational experience sug-gests that it is beneficial to isolate a feeder suspicious ofincipient cable faults very early after detecting the first symptomof an incipient fault. Isolating the feeder can limit the overall en-ergy at the point of fault and it also limits the often-repeated volt-age transient seen by the system. However, due to short durationof an incipient fault and the inability to achieve selectivity via timecoordination, designing an incipient fault protection function be-comes challenging. Incipient faults are predictable and avoidableif the degradation processes are known. Therefore, the detectionand isolation of incipient faults in the power supply system isessential for guaranteeing safe, reliable, and efficient operation ofcustomers’ installations.

3. S-transform and its application for extracting disturbancefeatures

Short duration voltage disturbances can be easily detected byusing advanced signal processing technique such as the S-trans-form which give time frequency representation of a signal (Pinne-gar & Mansinha, 2003). The S-transform of a function h(t) can bedefined as a continuous wavelet transform (CWT) multiplied by aphase factor e�i2pfs and it is given by,

Sðs; f Þ ¼ e�i2pfsWðs;dÞ ð1Þ

where W(s, d) is the CWT, s is time, and the dilation factor, d is theinverse of frequency, f.

The CWT is a series of correlations of the time series with afunction called as wavelet which is defined as,

Wðs; dÞ ¼Z 1

�1hðtÞxðt � s; dÞdt ð2Þ

where x(t � s, d) is the mother wavelet, t and s are both time.The mother wavelet is then given as,


xðt; f Þ ¼ jf jffiffiffiffiffiffiffi2pp e�

t2 f 2

2 e�i2pft ð3Þ

Substituting (2) and (3) into (1), the new S-transform equationis derived as,

Sðs; f Þ ¼ jf jffiffiffiffiffiffiffi2pp

Z 1

�1hðtÞe�

ðt�sÞ22 e�i2pftdt ð4Þ

The phase factor in (4) is a phase correction of the definition ofthe CWT. It eliminates the concept of wavelet analysis by separat-ing the mother wavelet into two parts, the slowly varying envelope(Gaussian function) which localizes in time, and the oscillatoryexponential kernel e�i2pft which selects the frequency being local-ized (Pinnegar & Mansinha 2003). It is the time localizing Gaussianthat is translated while the oscillatory exponential kernel remainsstationary. By not translating the oscillatory exponential kernel,the S-transform localizes the real and the imaginary componentsof the spectrum independently, localizing the phase spectrum aswell as the amplitude spectrum, and is thus directly invertible intothe Fourier Transform Spectrum (H). This characteristic of thephase is referred to as absolutely referenced phase informationwhich is given by,Z 1

�1Sðs; f Þds ¼ Hðf Þ ð5Þ

The output of the S-transform is an N �M matrix called the S-matrix whose rows pertain to the frequency and columns to time.Each element of the S-matrix is complex number and can be usedas features to classify the non stationary multiple disturbances.The information in the S-matrix can be plotted as time-frequencycontours which can be used to identify the existence of powerquality disturbances.

3.1. STMV and STFV Indices

Two indices called as S-transform magnitude-time voltage(STMV) and S-transform frequency–time voltage (STFV) are de-rived from the S-matrices. The STMV indices for all the three phasevoltages are given as follows:

STMV for the red phase voltage is derived from,

SRVi;j ¼jf jffiffiffiffiffiffiffi2pp

Z 1

�1vRðtÞe�


where SRVi;j is the S-matrix for voltage, vR, i = 1 . . . N (number of col-umns) and j = 1 . . . M (number of rows).

The STMV for the red phase voltage is the maximum value forall the columns in the S-matrix which is given by,

VSTMVR ¼ maxðSRViÞ ð7Þ

Similarly, the STMV for the yellow and the blue phase voltagesare,

VSTMVY ¼ maxðSY ViÞ ð8Þ

VSTMVB ¼ maxðSBViÞ ð9Þ

The STFV indices for all the three phase voltages are next givenas follows:

STFV for the red phase voltage is derived from,

SRVi;j ¼jf jffiffiffiffiffiffiffi2pp

Z 1

�1vRðtÞe�


where SRVi;j is the S-matrix for voltage, vR, i = 1 . . . N and j = 1 . . . M.The STFV for the red phase voltage is the maximum value for all

the rows in the S-matrix and is given by,

VSTFVR ¼maxðSRVjÞ ð11Þ



Fig. 4. Examples of the STMV and STFV plots for a non permanent fault.

Fig. 5. Examples of the STMV and STFV plots for a permanent fault.


Similarly, the STFV for the yellow and the blue phase voltagesare,

VSTFVY ¼maxðSY VjÞ ð12Þ

VSTFVB ¼maxðSBVjÞ ð13Þ


The STMV and STFV are used as input features to the rule basedsystem for identifying the network faults. Examples of STMV andSTFV plots for a non permanent fault, permanent fault and incipi-ent fault can be visualized as shown in Figs. 4–6, respectively. Inthe first row of Fig. 4, the per unit (pu) values of the STMV areshown. The STMV plots are almost similar to that of the root mean



Fig. 6. Examples of the STMV and STFV plots for an incipient fault.


square plots in which the plots show the existence of voltage sagsin the red and blue phases. The STFV plots shown in the second rowof Fig. 4 are the variations in the system frequency before, duringand after a voltage disturbance. Referring to the STFV plots, the fre-quency oscillations are seen more severe for the phases whichexperience the voltage disturbances.

The main difference between the plots shown in Figs. 4 and 5 isin the oscillations of the system frequency. The permanent faultgives a more severe frequency oscillation and inflicted more severevariations in the system frequency during the initiation and endingof the faults as compared to a non permanent fault. The STMV plotin Fig. 6 shows short duration increase in the yellow phase voltagevalues which signifies the existence of high frequency current inone of the cable cores and hence indicates the existence of anincipient fault which is usually initiated by a partial discharge. Thishigh frequency current will partially bridge the gap between phaseinsulation to ground or phase-to-phase insulation. Normally, whenpartial discharge is initiated, high frequency transient currentpulses will appear and persist for nano-seconds to a milli-second,then disappear and reappear repeatedly (Weeks & Steiner 1982).The event may be detected as a very small change in the current

Table 1Features selected for detecting short duration voltage disturbances.

Feature Index Description

F1 MaxSMT Max values of STMV in per unitF2 MinSMT Min values of STMV in per unitF3 Dmax Duration of max values of STMV above 1.10 per unitF4 Dmin Duration of min values of STMV below 0.90 per unitF5 FR5 Values of frequency resolution from 0.0061 to 0.022F6 FR6 Values of frequency resolution from 0.022 to 0.04F7 FR7 Values of frequency resolution from 0.04 to 0.08F8 FR8 Values of frequency resolution from 0.08 to 0.40F9 FR9 Values of frequency resolution from 0.40 to 0.50


drawn by the sample under test. A significant increase either inthe PD level or in the developing rate of PD activity can providean early indication of an incipient fault condition.

3.2. Features for detecting short duration voltage disturbances

From the STMV and the STFV indices, several features are ex-tracted and used for detecting the short duration voltage distur-bances as depicted in Table 1.

3.3. Features for diagnosing the causes of the short duration voltagedisturbances

To perform automatic diagnosis of the causes of the voltage dis-turbances, four features (F1, F2, F3 and F4) are chosen from Table 2and two new features, F10 and F11 are selected from the STFV in-dex. The overall features shown in Table 2 are used to identify thecause of voltage disturbances which may be due to permanent andnon permanent faults. The respective feature values for classifyingpermanent and non permanent faults are given as depicted in Ta-ble 3.

The feature values for classifying transient and incipient faultsare explained in Table 4. If all these features meet the specific dataranges for incipient fault, then it is considered as an incipient fault.

Table 2Features selected for performing power quality diagnosis.

Feature Index Description

F1 MaxSMT Max values of STMV in per unitF2 MinSMT Min values of STMV in per unitF3 Dmax Duration of max values of STMV above 1.10 per unitF4 Dmin Duration of min values of STMV below 0.90 per unitF10 STDSFT Standard deviation of STFVF11 DFSFT Difference of max and min of STFV in per unit



Table 3Feature values for classifying permanent faults and non permanent faults.

Feature Index Description Permanent fault Non-permanent fault

F1 MaxSMT Max valuesof STMV inper unit

STMV(1P) < 0.90&STMV(2P) > 1.10STMV(2P) < 0.90&STMV(1P) > 1.10

STMV(1P) < 0.90 &STMV(2P) > 1.10STMV(2P) < 0.90 &STMV(1P) > 1.100.90 < STMV(2P) < 1.10& STMV(1P) > 1.10

F2 MinSMT Min valuesof STMV inper unit

STMV(1P) < 0.90STMV(2P) < 0.90STMV(3P) < 0.90


F3 Dmax Duration ofmax valuesof STMVabove 1.10per unit

Dmax > 10 ms Dmax > 10 ms

F4 Dmin Duration ofmin valuesof STMVbelow 0.90per unit

Dmin > 10 ms Dmin > 10 ms

F10 STDSFT Std of STFV STD > 8 STD < 8F11 DFSFT Difference

of max andmin ofSTFVvalues inper unit

DF > 30 DF < 30

Table 5The first level RBC rules for detecting the short duration voltage disturbances.


If the data range does not meet the limits of incipient fault, thenthe fault is categorized as a transient fault.

Rules Description

R1 If (0.90 < F3 < 1.10), (0.90 < F4 < 1.10), (F5 = 0), (F6 = 0), (F7 = 0),(F8 = 0) and (F9 = 0) then data is only C1 (Signal and only containssinusoidal waveform)

R2 If (F3 < 0.90) & (0.90 < F4 < 1.10), then data is C2 (Voltage sag)R3 If (0.90 < F3 < 1.10) & (F4 > 1.10), then data is C3 (Voltage swell)R4 If (F5 > 0) and (F9 = 0) then data is C4 (Harmonic)R5 If (F6 > 0), (F9 = 0) then data is C5 (Notch)R6 If (F7 > 0), (F9 = 0) then data is C5 (Notch)R7 If (F8 > 0), (F9 = 0) then data is C5 (Notch)R8 If (F5 = 0), (F8 > 0) & (F9 = 0) then data is C6 (Oscillatory Transient)R9 If (F5 = 0) and (F9 > 0) then data is C7 (Impulsive Transient)

4. Application of the rule based classifiers for implementingpower quality diagnosis

Three rule based classifiers (RBC) have been developed forimplementing power quality diagnosis in a structured manner inwhich the first level RBC is used to detect the short duration volt-age disturbances such as voltage sag or swell, while the second le-vel RBC is used to diagnose the cause of sag or swell as either dueto permanent or non permanent faults. The third level RBC is then

Table 4Feature values for classifying transient and incipient faults.

Feature Index Description Transient fault Incipient fault

F1 MaxSMT Max valuesof STMV inper unit

STMV(1P) < 0.90&STMV(2P) > 1.10STMV(2P) < 0.90&STMV(1P) > 1.10

0.90 < STMV(2P) < 1.10& STMV(1P) > 1.10

F2 MinSMT Min valuesof STMV inper unit


Not applicable (N/A)

F3 Dmax Duration ofmax valuesof STMVabove 1.10per unit

Dmax > 10 ms Dmax > 10 ms

F4 Dmin Duration ofmin valuesof STMVbelow 0.90per unit

Dmin > 10 ms Dmin > 10 ms

F10 STDSFT Std of STFV STD < 8 STD < 8F11 DFSFT Difference

of max andmin ofSTFVvalues inper unit

DF < 30 DF < 30


used to classify the non permanent fault as either transient orincipient fault. Thus, three levels of classifications are required indiagnosing the short duration voltage disturbances. The overallprocedure for implementing power quality diagnosis using thethree rule based classifiers is as shown in Fig. 1. The rules for thethree classifiers are developed based on the features defined in Ta-ble 1, Table 3 (permanent and non permanent faults) and Table 4(incipient and transient faults). The details of the rules for the first,second and third RBC are described as in Tables 5–7, respectively.

5. Results and discussion

The performance of the proposed power quality diagnosis usingthe S-transform and the rule based classifiers was evaluated bytesting the RBC with 342 sets of voltage disturbance data obtainedfrom the existing PQMS in Malaysia. The power quality monitoringwas performed at the 33 kV network before the first step downtransformer as shown in Fig. 1. The voltage waveforms recordedundergone two transformations before being recorded by thepower quality recorder.

From the 342 sets of data shown in Table 8, the type and causeof short duration voltage disturbances are known. The causes ofthe disturbances were validated by referring to the existing data-base from 2001 to 2008 provided by the national power utility,

Table 6The second level RBC rules for classifying permanent and non permanent faults.

Rules Description

R1 If F1 and F3 meet the requirement as in Table 3, F10 > 8 and F11 > 30then the voltage sag or swell is caused by a permanent fault

R2 If F1 and F3 meet the requirement as in Table 3, F10 < 8 and F11 < 30then the voltage sag or swell is caused by a non permanent fault

R3 If F2 and F4 meet the requirement as in Table 3, F10 > 8 and F11 > 30then the voltage sag or swell is caused by a permanent fault

R4 If F2 and F4 meet the requirement as in Table 3, F10 < 8 and F11 < 30then the voltage sag or swell is caused by a non permanent fault

Table 7The third level RBC rules for classifying incipient and transient faults.

Rules Description

R1 If F3, F4, F10 and F11 meet the requirement as in Table 4 andF1 = (0.90 < STMV(2P) < 1.10 & STMV(1P) > 1.10) and F2 (No event)then the voltage sag or swell is caused by an incipient fault

R2 If F3, F4, F10 and F11 meet the requirement in Table 4 andF1 = (STMV(1P) < 0.90 & STMV(2P) > 1.10, STMV(2P) < 0.90 &STMV(1P) > 1.10) &F2 = (STMV(1P) < 0.90, STMV(2P) < 0.90, STMV(3P) < 0.90) then thevoltage sag or swell is caused by a transient fault



Table 8Description of the power quality data to be diagnosed.

Type of short duration voltagedisturbance

Cause of voltagedisturbance

Number ofdata sets

Voltage sag and swell Permanent fault 121Voltage sag Non permanent fault

(transient fault)111

Voltage swell Non permanent fault(incipient fault)

110

Table 9Results of first level RBC for detecting short duration voltage disturbances.

Type of short durationvoltage disturbance

Number ofdata sets

Correctdetection

Wrongdetection

% Accuracyin detection

Voltage sag and swell 121 121 0 100%Voltage sags 111 111 0 100%Voltage swell 110 110 0 100%

Table 10Results of the second and third level RBCs for diagnosing the cause of voltagedisturbance.

Type of network fault Number ofdata sets

Correctdiagnosis

Wrongdiagnosis

% Accuracy indiagnosis

Permanent fault 121 115 6 95.0%Non permanent fault

(transient fault)111 104 7 93.7%

Non permanent fault(incipient fault)

110 103 7 93.6%


Tenaga Nasional Berhad. The data comprises of 121 sets of voltagedisturbances caused by permanent faults and 221 sets of voltagedisturbances caused by non permanent faults which are eithertransient or incipient fault. The source of the permanent and nonpermanent faults originates from either the transmission, distribu-tion or customers own internal networks. By using the 342 sets ofdata, the RBC are evaluated in which the results for detecting theshort duration voltage disturbances are as shown in Table 9. Theresults for diagnosing the cause of voltage disturbances as towhether the disturbance is caused by a permanent fault, non per-manent fault which is either incipient or transient fault are asshown in Table 10.

Based on the results in Table 9, the first level RBC has success-fully detected the various short duration voltage disturbanceswhich are voltage sag, voltage swell and combined sag and swellwith an accuracy of 100% (Faisal & Mohamed, 2009). The resultsin Table 10 showed that the new technique successfully diagnosedthe causes of short duration voltage disturbances with an average


accuracy of 94.1%. Overall, there are 6 errors in diagnosing perma-nent faults, 7 errors in diagnosing transient faults and 7 errors indiagnosing incipient faults. The errors in the diagnosis of perma-nent faults are due to low values of the standard deviation of STFV(F10) in which the values for F10 for all the 6 error cases are lessthan 8. By referring to the rules in Table 6, these faults are consid-ered as non permanent faults.

6. Conclusion

The new power quality diagnosis method provides diagnosticfunctions that can give sufficient information for power utilityengineers to identify the cause of short duration voltage distur-bances. The main advantage offered by the new method is theuse of the S-transform to decompose the power line signals intoa set of time-frequency components and evaluate the changes inthe system frequency in which simple feature extraction can beperformed. The features extracted from the S-transform are thenused by the three rule based classifiers to diagnose the type andcauses of the short duration voltage disturbances. The numericalresults obtained with actual power quality data recorded in apower distribution system indicated that the new method is effec-tive in diagnosing the type and causes of the disturbances whichmay be due to permanent faults and non permanent faults catego-rized as either incipient or transient faults.

References

Azam, M. S., Fang, T., Pattipati, K. R., & Rajaiah Karanam, R. (2004). A dependencymodel-based approach for identifying and evaluating power quality problems.IEEE Transactions on Power Delivery, 19(3), 1154–1166.

Faisal, M. F., & Mohamed, A. (2009). Identification of multiple power qualitydisturbances in three phase industrial power systems using s-transform andrule based expert system. Journal of Applied Sciences, 9(15), 2688–2700.

IEC/TR 61000-2-8, ELECTROMAGNETIC COMPATIBILITY (EMC) – Part 2–8:ENVIRONMENT –Voltage dips and short interruptions on public electric powersupply systems with statistical measurement results. International TechnicalReport IEC 61000-2-8, pp 8–9.

Il-Yop, Chunga., Dong-Jun, Won b., Joong-Moon, Kimc., Seon-Ju, Ahnd., & Seung-Il,Moond. (2007). Development of a network-based power quality diagnosissystem. Electric Power Systems Research, 77, 1086–1094.

Il-Yop, Chung., Dong-Jun, Won., Joong-Moon, Kim., Seon-Ju, Ahn., Seung-Il, Moon.,Jang-Cheol, Seo., et al. (2003). Development of power quality diagnosis systemfor power quality improvement. IEEE Power Engineering Society General Meeting,2, 12–17.

Kezunovic, M. (2001). Automated analysis of voltage sags, their causes and impacts.IEEE Power Engineering Society Summer Meeting, 2, 1113–1117.

Pinnegar, C. R., & Mansinha, L. (2003). The S-transform with windows of arbitraryand varying shape. Geophysics, 68(1), 381–385.

Weeks, W. L., & Steiner, J. P. (1982). Instrumentation for the detection and locationof incipient faults on power cable. IEEE Transactions on Power Apparatus andSystems, 101(7).

White, A.L., 2004. Transient faults and network reliability. In IEEE AerospaceConference Vol. 1, 6–13 March 2004, pp 1121–1128.

Yalcinkaya, G., Bollen, M. H. J., & Crossley, P. A. (1998). Characterization of voltagesags in industrial distribution systems. IEEE Transaction on Industry applications.,34(4), 682–688. Publisher Item Identifier S 0093-999498)04909-3.



Documents

Power Quality Diagnosis