• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
Hp3313171323
 

Hp3313171323

on

  • 157 views

International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of ...

International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of Computer Science, Neural Networks, Electrical Engineering, Software Engineering, Information Technology, Mechanical Engineering, Chemical Engineering, Plastic Engineering, Food Technology, Textile Engineering, Nano Technology & science, Power Electronics, Electronics & Communication Engineering, Computational mathematics, Image processing, Civil Engineering, Structural Engineering, Environmental Engineering, VLSI Testing & Low Power VLSI Design etc.

Statistics

Views

Total Views
157
Views on SlideShare
157
Embed Views
0

Actions

Likes
0
Downloads
0
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Hp3313171323 Hp3313171323 Document Transcript

    • Priyanka M. Kothoke, Namrata R. Bhosale, Amol Despande, Dr. Alice N. Cheeran /International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.1317-13231317 | P a g eAnalysis of Partial Discharge using Phase-Resolved (n-q)Statistical TechniquesPriyanka M. Kothoke, Namrata R. Bhosale, Amol Despande, Dr. Alice N.CheeranDepartment of Electrical Engineering, Veermata Jijabai Technological Institute (VJTI), MumbaiABSTRACTPartial discharge (PD) patterns are animportant tool for the diagnosis of high voltage(HV) insulation systems. Human experts candiscover possible insulation defects in variousrepresentations of the PD data. One of the mostwidely used representations is phase-resolvedPD (PRPD) patterns. In order to ensure reliableoperation of HV equipment, it is vital to relatethe observable statistical characteristics of PDsto the properties of the defect and ultimately todetermine the type of the defect. In this work,we have obtained and analyzed phase-resolveddischarge patterns using parameters such asmean, standard deviation, variance, skewnessand kurtosis.Keywords - Partial Discharge, Phase-resolved,Statistical parameters1. INTRODUCTIONPD is a localized electrical discharge thatpartially bridges the insulation between conductorsand which may or may not occur adjacent to aconductor. [1] In general, PDs are concerned withdielectric materials used, and partially bridging theelectrodes between which the voltage is applied.The insulation may consist of solid, liquid, orgaseous materials, or any combination of them. PDis the main reason for the electrical ageing andinsulation breakdown of high voltage electricalapparatus. Different sources of PD give differenteffect on insulation performance. Therefore, PDclassification is important in order to evaluate theharmfulness of the discharge. [2]PD classification aims at the recognition ofdischarges of unknown origin. For many years, theprocess was performed by investigating the patternof the discharge using the well known ellipse on anoscilloscope screen, which was observed crudelyby eye. Nowadays, there has been extensivepublished research to identify PD sources by usingintelligent technique like artificial neural networks,fuzzy logic and acoustic emission. [2]The recent upsurge of research on PDphenomena has been driven in part by developmentof new fast digital and computer-based techniquesthat can process and analyze signals derived fromPD measurements. There seems to be anexpectation that, with sufficiently sophisticateddigital processing techniques, it should be possiblenot only to gain new insight into the physical andchemical basis of PD phenomena, but also to definePD ‘patterns’ that can be used for identifying thecharacteristics of the insulation ‘defects’ at whichthe observed PD occur. [3] One of the undoubtedadvantages of a computer-aided measuring systemis the ability to process a large amount ofinformation and to transform this information intoan understandable output. [4]There are many types of patterns that can beused for PD source identification. If thesedifferences can be presented in terms of statisticalparameters, identification of the defect type fromthe observed PD pattern may be possible. As eachdefect has its own particular degradationmechanism, it is important to know the correlationbetween discharge patterns and the kind of defect.Therefore, progress in the recognition of internaldischarge and their correlation with the kind ofdefect is becoming increasingly important in thequality control in insulating systems. [4]Researches have been carried out in recognition ofpartial discharge sources using statisticaltechniques and neural network. In our study, wehave tested various internal and external dischargeslike void, surface and corona using statisticalparameters such as mean, standard deviation,variance, skewness and kurtosis in phase resolvedpattern (n-q) and classified the partial dischargesource for unknown partial discharge data.2. STATISTICAL PARAMETERSThe important parameters to characterizePDs are phase angle φ, PD charge magnitude q andPD number of pulses n. PD distribution patterns arecomposed of these three parameters. Statisticalparameters are obtained for phase resolved pattern(n-q).2.1. Processing of data (φ, q and n)Statistical analysis is applied for thecomputation of several statistical operators. Thedefinitions of most of these statistical operators aredescribed below. The profile of all these discretedistribution functions can be put in a generalfunction, i.e., yi=f(xi). The statistical operators canbe computed as follows:
    • Priyanka M. Kothoke, Namrata R. Bhosale, Amol Despande, Dr. Alice N. Cheeran /International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.1317-13231318 | P a g e………(1)………(2)………(3)………(4)Standard Deviation = ...……(5)where,x = number of pulses n,f(x) = PD charge magnitude q,μ = average mean value of PD chargemagnitude q,σ = variance of PD charge magnitude qSkewness and Kurtosis are evaluated withrespect to a reference normal distribution.Skewness is a measure of asymmetry or degree oftilt of the data with respect to normal distribution.If the distribution is symmetric, Sk=0; if it isasymmetric to the left, Sk>0; and if it isasymmetric to the right, Sk<0. Kurtosis is anindicator of sharpness of distribution. If thedistribution has the same sharpness as a normaldistribution, then Ku=0. If it is sharper than thenormal then Ku>0 else it is flatter i.e. Ku<0. [5][6]3. RESULTS AND DISCUSSIONSAnalysis involves determining unknownPD patterns by comparing those with known PDpatterns such as void, surface and corona. Thecomparison is done with respect to their statisticalparameters. [7]3.1. Phase resolved patterns (n-q)The phase resolved patterns n-q areobtained for three known PD patterns: void, surfaceand corona (as discussed in 3.1.1) and threeunknown PD patterns: data1, data2 and data3 (asdiscussed in 3.1.2). For all the following plots, x-axis is number of cycles (n) and y-axis ismagnitude of charge (q).3.1.1. 2D distribution of n-q for known PDpatternsFig. 1(a), Fig. 1(b), Fig. 1(c), Fig. 1(d) and Fig.1(e) are the n-q plot of mean, standard deviation,variance, skewness and kurtosis for void dischargerespectively.Fig. 1(a) Mean plot (n-q) of void dischargeFig. 1(b) Standard deviation plot (n-q) of voiddischargeFig. 1(c) Variance plot (n-q) of void discharge
    • Priyanka M. Kothoke, Namrata R. Bhosale, Amol Despande, Dr. Alice N. Cheeran /International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.1317-13231319 | P a g eFig. 1(d) Skewness plot (n-q) of void dischargeFig. 1(e) Kurtosis plot (n-q) of void dischargeReferring to Fig. 1(a), Fig. 1(b) and Fig. 1(c) ofvoid discharge, it can be seen there is a peakoccurring somewhere after 1500 cycle, which is avoid discharge and in Fig. 1(d) and Fig. 1(e) ofskewness and kurtosis, the value decreases at thatcycle where peak occurs.Fig. 2(a), Fig. 2(b), Fig. 2(c), Fig. 2(d) and Fig.2(e) are the n-q plot of mean, standard deviation,variance, skewness and kurtosis for surfacedischarge respectively.Fig. 2(a) Mean plot (n-q) of surface dischargeFig. 2(b) Standard deviation plot (n-q) of surfacedischargeFig. 2(c) Variance plot (n-q) of surface dischargeFig. 2(d) Skewness plot (n-q) of surface dischargeFig. 2(e) Kurtosis plot (n-q) of surface dischargeIn surface discharge, charges are distributeduniformly over all cycles for mean, standarddeviation, variance, skewness and kurtosis asshown in Fig. 2(a), Fig. 2(b), Fig. 2(c), Fig. 2(d)and Fig. 2(e).Fig. 3(a), Fig. 3(b), Fig. 3(c), Fig. 3(d) and Fig.3(e) are the n-q plot of mean, standard deviation,variance, skewness and kurtosis for coronadischarge respectively.
    • Priyanka M. Kothoke, Namrata R. Bhosale, Amol Despande, Dr. Alice N. Cheeran /International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.1317-13231320 | P a g eFig. 3(a) Mean plot (n-q) of corona dischargeFig. 3(b) Standard deviation (n-q) of coronadischargeFig. 3(c) Variance plot (n-q) of corona dischargeFig. 3(d) Skewness plot (n-q) of corona dischargeFig. 3(e) Kurtosis plot (n-q) of corona dischargeReferring to Fig. 3(a), Fig. 3(b) and Fig. 3(c) ofcorona discharge, it can be seen the charges startsoccurring after 500 cycle increasing somewhere upto 1200 cycle and then decreasing after 2000 cycle,and in Fig. 3(d) and Fig. 3(e) of skewness andkurtosis, the value decreases from 500 cycle till2000 cycle.3.1.2. 2D distribution of (n-q) for unknown PDpatternsFig. 4(a), Fig. 4(b), Fig. 4(c), Fig. 4(d) and Fig.4(e) are the n-q plot of mean, standard deviation,variance, skewness and kurtosis for data1respectively.Fig. 4(a) Mean plot (n-q) of data1Fig. 4(b) Standard deviation plot (n-q) of data1
    • Priyanka M. Kothoke, Namrata R. Bhosale, Amol Despande, Dr. Alice N. Cheeran /International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.1317-13231321 | P a g eFig. 4(c) Variance plot (n-q) of data1Fig. 4(d) Skewness plot (n-q) of data1Fig. 4(e) Kurtosis plot (n-q) of data1In Fig. 4(a), Fig. 4(b), Fig. 4(c), Fig. 4(d) andFig. 4(e), the charges are uniformly distributedsimilar to surface discharge. Hence, it can beconcluded that data1 is having surface discharge.Fig. 5(a) Mean plot (n-q) of data2Fig. 5(b) Standard deviation plot (n-q) of data2Fig. 5(c) Variance plot (n-q) of data2Fig. 5(d) Skewness plot (n-q) of data2
    • Priyanka M. Kothoke, Namrata R. Bhosale, Amol Despande, Dr. Alice N. Cheeran /International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.1317-13231322 | P a g eFig. 5(e) Kurtosis plot (n-q) of data2Fig. 5(a), Fig. 5(b), Fig. 5(c), Fig. 5(d) and Fig.5(e) are the n-q plot of mean, standard deviation,variance, skewness and kurtosis for data2respectively. In these figures we can observe thatthe charges are uniformly distributed similar tosurface discharge. Hence, it can be concluded thatdata2 is having surface discharge.Fig. 6(a), Fig. 6(b), Fig. 6(c), Fig. 6(d) and Fig.6(e) are the n-q plot of mean, standard deviation,variance, skewness and kurtosis for data3respectively.Fig. 6(a) Mean plot (n-q) of data3Fig. 6(b) Standard deviation plot (n-q) of data3Fig. 6(c) Variance plot (n-q) of data3Fig. 6(d) Skewness plot (n-q) of data3Fig. 6(e) Kurtosis plot (n-q) of data3In Fig. 6(a), Fig. 6(b) and Fig. 6(c), there is aoccurrence of peak after 1500 cycle and in Fig.6(d) and Fig. 6(e), the skewness and kurtosis valuedecreases at that peak which is similar to voiddischarge. Hence, it can be concluded that data3 isvoid discharge.3.2. Statistical parametersAfter analysis we obtained the resultswhich are summarized in table 1 and table 2.Table 1. Parameters of known PD patternsParameters void surface coronaMean 13320.32 145.706 1.426Standarddeviation7553.716 126.009 1.139Variance 1.64*10817921.01 2.279Skewness 3.66*10-170.809 0.26Kurtosis 0.04878 2.442 0.966Table 2. Parameters of unknown PD PatternsParameters data1 data2 data3Mean 105.119 553.93 13320.32Standarddeviation97.966 698.3 7553.716Variance 16714.23 4.94*1051.64* 108Skewness 0.692 1.939 -3.7*10-17Kurtosis 2.004 6.31 0.04878
    • Priyanka M. Kothoke, Namrata R. Bhosale, Amol Despande, Dr. Alice N. Cheeran /International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622www.ijera.com Vol. 3, Issue 3, May-Jun 2013, pp.1317-13231323 | P a g eFig. 7 is the statistical characteristics of mean,standard deviation, variance, skewness and kurtosisof void discharge against data3. Fig. 8(a), Fig. 8(b)are the statistical characteristics of mean, standarddeviation, variance, skewness and kurtosis ofsurface discharge against data1 and data2respectively.Fig. 7 Statistical Characteristics of data3 againstvoid dischargeFig. 8(a) Statistical characteristics of data1 againstsurface dischargeFig. 8(b) Statistical characteristics of data2 againstsurface discharge4. OBSERVATIONS AND CONCLUSIONThe following observations are made fromthe results: Plotting statistical parameters of void dischargeagainst data3 in Fig. 8 shows data3characteristics overlaps void characteristics, itcan be concluded that data3 is void discharge. Similarly, for surface discharge, data1 anddata2 characteristics (Fig. 8(a) and Fig. 8(b))approximately fits surface dischargecharacteristics, it can be concluded that data1and data2 is surface discharge.The analysis done from statistical parametersare data1 is surface discharge, data2 is surfacedischarge and data3 is void discharge. The analysisusing statistical parameters can be done for varioustypes of PD discharges.From statistical parameters, the PD sourcecannot be concluded accurately so it needs to beapplied to others classification methods such asneural network, Fuzzy logic etc. as a pre-processing parameters for getting accurate PDsource.REFERENCES[1] Partial Discharge Measurements, IECPublication 270, 1981.[2] Nur Fadilah Ab Aziz, L. Hao, P. L. Lewin,“Analysis of Partial DischargeMeasurement Data Using a SupportVector Machine,” The 5th StudentConference on Research andDevelopment, 11-12 December 2007,Malayasia.[3] M. G. Danikas, “The Definitions Used forPartial Discharge Phenomena,” IEEETrans. Elec. Insul., Vol. 28, pp. 1075-1081, 1993.[4] E. Gulski and F. H. Kreuger, “Computer-aided recognition of Discharge Sources,”IEEE Transactions on ElectricalInsulation, Vol. 27 No. 1, February 1002.[5] N.C. Sahoo, M. M. A. Salama, R.Bartnikas, “Trends in Partial DischargePattern Classification: A Survey”, IEEETransactions on Dielectrics and ElectricalInsulation, Vol. 12, No. 2; April 2005.[6] C. Chang and Q. Su, “StatisticalCharacteristics of Partial Discharges froma Rod-Plane Arrangement”[7] Namrata Bhosale, Priyanka Kothoke,Amol Deshpande, Dr. Alice Cheeran,“Analysis of Partial Discharge usingPhase-Resolved (φ-q) and (φ-n) StatisticalTechniques”, International Journal ofEngineering Research and Technology,Vol. 2 (05), 2013,ISSN2278-0181.