Transformer Ageing E-Book

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Serie: Wiley - IEEE
• Sprache: Englisch

Seitenzahl: 734

Veröffentlichungsjahr: 2017

Beschreibung

A one-stop guide to transformer ageing, presenting industrially relevant state-of-the-art diagnostic techniques backed by extensive research data * Offers a comprehensive coverage of transformer ageing topics including insulation materials, condition monitoring and diagnostic techniques * Features chapters on smart transformer monitoring frameworks, transformer life estimation and biodegradable oil * Highlights industrially relevant techniques adopted in electricity utilities, backed by extensive research

Leseprobe

Table of Contents

Cover

Title Page

Preface

Acknowledgments

Contributing Authors

1 Transformer Insulation Materials and Ageing

1.1 Introduction

1.2 Solid Insulation – Paper, Pressboard

1.3 Liquid Insulation – Oil

1.4 Insulation Ageing

1.5 Scope of the Book

References

Further Reading

2 Overview of Insulation Diagnostics

2.1 Introduction

2.2 Dissolved Gas Analysis

2.3 Furan Analysis

2.4 DP Measurements

2.5 Traditional Electrical Techniques of Insulation Diagnosis

2.6 DDF, Capacitance, and Power Factor

2.7 Partial Discharge Measurements

2.8 Conclusion

References

3 Dielectric Response Measurements

Part A: Time Domain Polarization‐Based Dielectric Response Measurements and Interpretations

3A.1 Basic PDC Measurement Procedure [1]

3A.2 Basic Theory of Dielectric Response [1–4]

3A.3 Practical PDC Measurement Issues [7]

3A.4 Interpretation of PDC Measurement Results

3A.5 PDC Measurement Results on Field Transformers

3A.6 PDC Measurement Equipment

3A.7 Summary of PDC Measurements

3A.8 Recovery Voltage Measurements (RVM)

3A.9 PDC and RVM Modeling

3A.10 Final Comments on PDC and RV Measurements

References

Part B: Frequency Domain Dielectric Spectroscopy

3B.1 Theory of FDS [1]

3B.2 FDS of Oil–Paper Insulation

3B.3 FDS Measurement Procedure for Transformers

3B.4 Factors Affecting FDS Measurement on Transformers

3B.5 Interpretation of FDS Results

3B.6 Modeling FDS of Transformer Insulation

3B.7 Frequency Domain to Time Dielectric Response Measurement

3B.8 Summary

References

4 Dissolved Gas Analysis Interpretation and Intelligent Machine Learning Techniques

4.1 Introduction

4.2 CIGRE Works Related to DGA

4.3 Advancement of New Gases for Fault Diagnosis

4.4 Uncertainty in Dissolved Gas Analysis

4.5 Statistical Learning‐Based Intelligent Diagnostic Techniques

4.6 Review on Pattern Recognition Techniques for Transformer Insulation Diagnosis

4.7 Hybrid Algorithm for Improving Power Transformer Insulation Diagnosis

4.8 Synthetic Minority Oversampling Technique

4.9 Integrated SMOTEBoost Algorithm

4.10 Hybrid of SMOTEBoost and Bootstrap

4.11 Collection of Pattern Recognition Algorithms Used for DGA

4.12 Case Studies and Results

4.13 Numeric Experiment Setup

4.14 Discussion on Classification Accuracy

4.15 Generalization Capability Validation

4.16 Summary

References

5 Advanced Signal Processing Techniques for Partial Discharge Measurement

5.1 Partial Discharge in Power Transformer

5.2 Overview of PD Analysis

5.3 PD Measurement Methods

5.4 Advanced Signal Processing Techniques

5.5 Application of Advanced Signal Processing Techniques for PD Analysis

5.6 PD Source Classification

5.7 Acoustic and UHF Methods for PD Signal Detection and Localization

5.8 Summary

References

6 Frequency Response Analysis Interpretation for Winding Deformation of Power Transformers

6.1 Fundamentals of Frequency Response Analysis

6.2 International Practice and Standards

6.3 Other Interpretation Schemes

6.4 Case Studies

6.5 Sensitivity of Non‐mechanical Factors

6.6 Summary

References

Further Reading

7 Impact of Moisture and Remaining Life Estimation

7.1 Introduction

7.2 Measurement of Water Content

7.3 Consequence of Wet Insulation on Transformer Operation

7.4 Calculating Water Content of Paper in an Operating Transformer

7.5 Case Studies

7.6 Summary

References

Further Reading

8 Biodegradable Oils and their Impact on Paper Ageing

8.1 Introduction

8.2 Comparison of the Properties of Insulating Liquids with Mineral Oil

8.3 Ageing of Biodegradable Oil

8.4 Comparison of Physiochemical and Dielectric Properties

8.5 Guide for Assessing the Quality of Ester Liquids in Service Unit

8.6 Case Studies

8.7 Use of 2‐FAL as an Ageing Indicator in Ester‐Based Insulation Systems

8.8 Stray Gas Generation in Ester Insulation Liquids

8.9 Analyzing of Moisture in Solid Insulation of an Ester‐Filled Transformer Using FDS Methods

8.10 Field Experience in Condition Monitoring of Natural Ester‐Filled Transformers

8.11 Summary

References

9 Smart Transformer Condition Monitoring and Diagnosis

9.1 Introduction

9.2 Intelligent Framework for Transformer Condition Monitoring and Diagnosis

9.3 Advanced Signal Processing for Transformer Condition Monitoring

9.4 Feature Extraction and Pattern Recognition for Transformer Diagnosis

9.5 Data and Information Fusion for Transformer Diagnosis

9.6 Determining Health Index for Transformer Insulation System [30, 43]

9.7 SmartBox for Smart Transformer [4]

References

10 Conclusions and Future Research

Index

End User License Agreement

List of Tables

Chapter 01

Table 1.1 Desirable qualities for electrical grade fibers

Table 1.2 Relative costs of insulating oils [6]

Table 1.3 Properties of mineral oil

Table 1.4 Oil properties in various standards

Table 1.5 Limits for new mineral oil

Chapter 02

Table 2.1 Fault abbreviations

Table 2.2 Typical faults in power transformers

Table 2.3 Dissolved gas concentration limits as per IEEE

Table 2.4 Concentration limits of dissolved gas as per Doernenberg (ppm by volume)

Table 2.5 Ratio limits of key gases as per Doernenburg

Table 2.6 Ratio limits of key gases as per Rogers ratio method

Table 2.7 DGA interpretation table: ratio limits of key gases

Table 2.8 Comparison of fault types of two types of Duval pentagon

Table 2.9 Possible causes of specific furanic compound presence

Table 2.10 Typical values of 2‐FAL concentration (Conc., expressed in mg/kg) and rate of increase (RoI, expressed in mg/kg/y) for different families of equipment and oil type

Chapter 03

Table 3A.1 Geometric details of test transformer model

Table 3A.2 Oil conductivity [22]

Table 3A.3 Transformer details

Table 3A.4 Insulation condition of A1 and A2

Table 3A.5 Insulation condition of B1 and B2

Table 3A.6 Insulation condition of C1 and C2

Table 3A.7 Summary of polarization spectra results

Table 3A.8 Details of transformers under investigation

Table 3A.9 Summary of RVM results

Table 3A.10 Oil and paper moisture content before and after oil reclamation

Table 3B.1 Transformers under test

Table 3B.2 Summary of FDS results

Table 3B.3 Details of transformers under test

Table 3B.4 Service details of transformers tested

Table 3B.5 Maximum and minimum dissipation factors with respect to varying temperature

Chapter 04

Table 4.1 Advantages and limitations of different pattern recognition algorithms

Table 4.2 Dataset configurations and results of some representative pattern recognition algorithms in the literature

Table 4.3 Configuration of eight datasets

Table 4.4 Sample distribution of original dataset

Table 4.5 Mean value and standard deviation of five DGA datasets

Table 4.6 Mean value and standard deviation of three oil characteristics datasets

Table 4.7 Mean value and standard deviation of three oil characteristics datasets

Table 4.8 Sample distribution after SMOTE and bootstrapping

Table 4.9 Comparison of classification accuracy of SVM over eight datasets (%)

Table 4.10 Comparison of classification accuracy of

k

NN over eight datasets (%)

Table 4.11 Comparison of classification accuracy of RBF over eight datasets (%)

Table 4.12 Comparison of classification accuracy of C4.5 decision tree over eight datasets (%)

Table 4.13 Comparison of generalization ability of different pattern recognition algorithms (trained on dataset 4, tested on dataset 1)

Table 4.14 Comparison of generalization ability of different pattern recognition algorithms (trained on dataset 6, tested on dataset 5)

Chapter 05

Table 5.1 Types of noise in PD measurements

Table 5.2 PD measurement methods

Table 5.3 Comparisons of measures on PD signals acquired from PD source models

Table 5.4 Information on feature datasets

Table 5.5 Classification results (%) of algorithms (with statistical operators) [79]

Table 5.6 Classification results (%) of algorithms (with DWT) [79]

Table 5.7 Classification results (%) of algorithms (with PCA) [79]

Table 5.8 Classification results (%) of algorithms (with KPCA) [79]

Table 5.9 Classification results (%) of algorithms (with SNE) [79]

Chapter 06

Table 6.1 Comparison of main electrical diagnostic techniques for winding deformation

Table 6.2 Suitable power transformer diagnostic test according to faulty condition

Table 6.3 Relative factor and degree of deformation

Table 6.4 Frequency regions and corresponding influencing factors

Table 6.5 Types of transformer and the number of recommended tests

Table 6.6 CC range used by a utility according to method of comparison

Table 6.7 Statistical indicators on all phases for the selected frequency region

Table 6.8 Statistical indicators on all phases for the selected frequency region

Table 6.9 The estimated percentage of winding damage

Table 6.10 Statistical indicators on all phases for the selected frequency region

Table 6.11 The estimated percentage of winding damage

Chapter 07

Table 7.1 Freundlich isotherm coefficients calculated by Fessler

et al

. [6]

Table 7.2 Comparison of

A

and

B

oil solubility coefficients for mineral oil, both measured during this investigation and published by other researchers

Table 7.3 Comparison of

A

and

B

oil solubility coefficients for service‐aged mineral oil (values used from Ref. [8])

Table 7.4 Comparison of paper water content at room temperature [10]

Table 7.5 Assumptions implied in the use of cellulose adsorption isotherms in conjunction with KF titration, and the resulting inaccuracies if the assumptions are invalid [12]

Table 7.6

A

and

B

coefficients proposed by Koch and Tenbohlen [15]

Table 7.7 Estimated life, in years, of paper aged in low oxygen

Table 7.8 Estimated life, in years, of paper aged in high oxygen

Table 7.9 Transformers investigated during study

Table 7.10 Conditions used in calculations, and the estimated water content of paper

Chapter 08

Table 8.1 Fatty acids composition of different vegetable oils [11, 15, 16]

Table 8.2 Comparison of typical properties of commercially available insulating liquids [15]

Table 8.3 Separation distance between outdoor fluid‐insulated transformer and building

Table 8.4 Outdoor fluid‐insulated transformer equipment distance

Table 8.5 Overview of limiting values of dielectric and chemical parameters of insulating oils

Table 8.6 Measured oil parameter under two different ageing conditions [47, 48]

Table 8.7 Correctly identified faults for DGA results of case studies A to C

Table 8.8 Correctly identified faults for DGA results in Refs [61, 65, 68]

Table 8.9 Activation energy for moderately wet pressboard insulation

Chapter 09

Table 9.1 Typical online sensor‐based measurement and offline measurement for transformer condition monitoring and diagnosis

Table 9.2 Oil characteristics database for demonstrating mRMR feature selection technique

Table 9.3 Feature subsets with different number of features

Table 9.4 Prediction result obtained by mRMR and SVM algorithm (using MID search scheme, results in %)

Table 9.5 Prediction result obtained by mRMR and SVM algorithm (using MIQ search scheme, results in %)

Table 9.6 Bayesian fusion of DGA test

Table 9.7 Dempster–Shafer fusion result

Table 9.8 Oil test record of eight transformers digested from historic records [38]

Table 9.9 Examples of determining health index of transformer insulation for the construction of oil characteristics database [38]

Table 9.10 Examples of health index of transformer insulation systems [38]

Table 9.11 Comparison of prediction accuracy (%) on health index of three algorithms [38]

List of Illustrations

Chapter 01

Figure 1.1 Chemical structure of glucose.

Figure 1.2 Chemical structure of cellulose (polymer chains of glucose).

Figure 1.3 Schematic diagram of LFH [3].

Figure 1.4 Temperature range of normal usability of various classes of insulation oil [6].

Figure 1.5 Relative biodegradability of vegetable oil.

Figure 1.6 Factors influencing performance and degradation of transformer oil–paper insulation and resulting breakdown mechanisms.

Chapter 02

Figure 2.1 (a) Thermal decomposition of oil products include C

2

H

4

and CH

4

, together with smaller quantities of H

2

and C

2

H

6

. Traces of C

2

H

2

may be formed if the fault is severe or involves electrical contacts. Principal gas: ethylene. (b) Thermal overheating and decomposition of cellulose produces large quantities of CO

2

and CO. Hydrocarbon gases, such as CH

4

and C

2

H

2

, will be formed if the fault involves an oil‐impregnated structure. Principal gas: carbon monoxide. (c) Low‐energy electrical discharges – partial discharges produce H

2

and CH

4

, with small quantities of C

2

H

6

and C

2

H

4

. Comparable amounts of CO and CO

2

may result from discharges in cellulose. Principal gas: hydrogen. (d) High‐energy electrical discharge such as arcing produces large amounts of H

2

and C

2

H

2

, with minor quantities of CH

4

, C

2

H

6

, and C

2

H

4

. CO and CO

2

may also be formed if the fault involves cellulose. Principal gas: acetylene.

Figure 2.2 Doernenburg’s ratio method flow chart.

Figure 2.3 Duval’s triangle.

Figure 2.4 Coordinates and fault zones in the Duval pentagon method: (a) Duval pentagon type I; (b) Duval pentagon type II.

Figure 2.5 One of the most probable cellulose thermal degradation mechanisms.

Figure 2.6 2‐FAL analysis of four GSU transformers, commissioned in 1992.

Figure 2.7 Variation of the breakdown voltage, humidity, and 2‐FAL in a 75 MVA transformer manufactured in 1981.

Figure 2.8 Structure of cellulose.

Figure 2.9 Number of polymer chains versus ageing time.

Figure 2.10 Gel permeation chromatograms of new and aged insulation paper [22, 27].

Figure 2.11 Tensile strength and DP of thermally upgraded paper in mineral oil (sealed vessel at 130°C, 150°C, and 170°C). Data at 160°C is included for comparison with [28].

Figure 2.12 Relationship between tensile strength and DP

w

for artificially aged insulation paper.

Figure 2.13 Water evolved during ageing of paper.

Figure 2.14 Variation of predicted insulation life with choice of final DP values for oil‐impregnated paper.

Figure 2.15 Variation of predicted insulation life with the choice of initial and final DP values for dry Kraft paper in dry oil.

Figure 2.16 Change of DP with time and temperature.

Figure 2.17 Effects of water and oxygen on DP change during ageing of Kraft paper in oil at 140°C.

Figure 2.18 Scatter plot for DP versus total furan concentration in oil.

Figure 2.19 Total furan concentration versus DP calculated by number average molecular weight (DP

n

).

Figure 2.20 Tensile index and DP plot for Kraft paper versus time at 3% moisture.

Figure 2.21 Correlation between insulation ageing determined by furfural and DP measurements.

Figure 2.22 Correlation between insulation ageing determined by CO

2

and DP measurements

Figure 2.23 Correlation between tensile strength and DP value for non‐upgraded Kraft paper.

Figure 2.24 Correlation between tensile strength and DP.

Figure 2.25 Residual life estimation – laboratory experiment, Chendong equation.

Figure 2.26 2‐FAL and other related furanic compounds in oil versus DP: collected laboratory data.

Figure 2.27 Simple vector diagram for loss factor test.

Chapter 03

Figure 3A.1 Basic circuit for measuring polarization and depolarization currents.

Figure 3A.2 Polarization and depolarization currents.

Figure 3A.3 Different transformer terminal connections for PDC measurements: (a) single‐phase two‐winding transformer; (b) three‐phase transformer with star‐connected LV winding with neutral and delta‐connected HV winding; (c) three‐phase transformer with star‐connected tertiary winding; (d) three‐phase autotransformer with tertiary winding; (e) three‐phase autotransformer without tertiary winding.

Figure 3A.4 Relaxation currents at different excitation voltages (scaled to the same level for comparison).

Figure 3A.5 Crossing over of depolarization and polarization currents; inset shows the initial period on an enlarged scale.

Figure 3A.6 Dielectric response function

f

(

t

) at different ratios of charging and discharging times

t

p

and

t

d

.

Figure 3A.7 (a) Polarization and depolarization currents with and without adequate pre‐discharging. (b) Polarization and depolarization currents of a dry paper sample as a function of charging time.

Figure 3A.8 Polarization and depolarization currents in the presence of noise.

Figure 3A.9 Polarization and depolarization currents after filtering out noise.

Figure 3A.10 Polarization and depolarization current plots for transformer under downward transition of temperature.

Figure 3A.11 Polarization and depolarization current plots for transformer under upward transition of temperature.

Figure 3A.12 Higher magnitude of depolarization current over polarization current due to temperature transient; inset shows the initial period on an enlarged scale.

Figure 3A.13 Triboelectric effect.

Figure 3A.14 Paper moisture conditioning setup: 1, oil container; 2, heater and stirrer for oil; 3, pressure gauge for oil; 4, temperature controller for the conditioning vessel; 5, vacuum pump for oil; 6, thermocouple probe for conditioning vessel temperature control; 7, conditioning vessel with heat‐reserving jacket; 8, heating block for conditioning vessel; 9, pressure gauge for conditioning vessel; 10, vacuum pump for conditioning vessel.

Figure 3A.15 Internal winding arrangement of the “pancake” model.

Figure 3A.16 Internal construction details of the “pancake” model.

Figure 3A.17 “Pancake” model.

Figure 3A.18 Structure schematic of ageing test device.

Figure 3A.19 Measurement cell in climatic chamber.

Figure 3A.20 Test arrangement for the “PDC” measuring technique.

Figure 3A.21 Polarization and depolarization currents of unaged oil‐impregnated pressboard samples with different moisture content (m.c.).

Figure 3A.22 Scatter in relaxation current measurements for different pressboard samples with same moisture content.

Figure 3A.23 Variation of polarization currents with paper conductivity.

Figure 3A.24 Variation of polarization currents with oil conductivity.

Figure 3A.25 Influence of board moisture on the polarization current.

Figure 3A.26 Variation of depolarization currents with paper conductivity.

Figure 3A.27 Variation of depolarization currents with oil conductivity.

Figure 3A.28 Polarization currents for different paper ageing and moisture but with the same oil.

Figure 3A.29 Depolarization currents for different paper ageing and moisture but with the same oil.

Figure 3A.30 Polarization currents under the influence of moisture.

Figure 3A.31 Depolarization currents under the influence of moisture.

Figure 3A.32 Variation of polarization currents with paper conductivity: (1)

σ

paper

 = 1.5 pS/m; (2)

σ

paper

 = 0.15 pS/m; (3)

σ

paper

 = 0.015 pS/m.

Figure 3A.33 Variation of depolarization currents with paper conductivity: (1)

σ

paper

 = 1.5 pS/m; (2)

σ

paper

 = 0.15 pS/m; (3)

σ

paper

 = 0.015 pS/m.

Figure 3A.34 Relaxation currents of different multi‐layer samples: (a)

σ

oil

 = 0.5 pS/m, (b, c)

σ

oil

 = 0.3 pS/m, (d)

σ

oil

 = 0.1 pS/m.

Figure 3A.35 Measured relaxation currents before and after oil exchange.

Figure 3A.36 Variation of polarization currents with oil conductivity.

Figure 3A.37 Variation of depolarization currents with oil conductivity.

Figure 3A.38 Variation of polarization currents with oil condition.

Figure 3A.39 Variation of depolarization currents with oil condition.

Figure 3A.40 Polarization currents for different oil ageing and moisture.

Figure 3A.41 Depolarization currents for different oil ageing and moisture.

Figure 3A.42 Relaxation currents of a pressboard sample with an initial moisture content of 0.2%, before and after ageing.

Figure 3A.43 Relaxation currents of a pressboard sample with an initial moisture content of 1.0%, before and after ageing.

Figure 3A.44 Polarization current at different ageing conditions (with 1.2% moisture content).

Figure 3A.45 Polarization current at different ageing conditions (with 3.48% moisture content).

Figure 3A.46 Comparison of PDC for new and aged pressboard with 4% moisture content.

Figure 3A.47 Relaxation currents of 2 mm pressboard with 0.2% moisture content measured at different temperatures.

Figure 3A.48 Relaxation currents of a multi‐layer sample in dependence on temperature.

Figure 3A.49 Variation of polarization current with temperature.

Figure 3A.50 Variation of depolarization current with temperature.

Figure 3A.51

X–Y

arrangement structure of oil and paper [27].

Figure 3A.52 Series arrangement structure of oil and paper [27].

Figure 3A.53 Variation of oil conductivity with temperature.

Figure 3A.54 Variation of paper conductivity with temperature.

Figure 3A.55 Influence of temperature on polarization current.

Figure 3A.56 Polarization currents for 4% moisture content at different temperatures.

Figure 3A.57 Relaxation currents of an aged pressboard sample as a function of charging voltage.

Figure 3A.58 Relaxation currents at different excitation voltages (scaled to the same level for comparison).

Figure 3A.59 Crossing over of depolarization and polarization currents; inset shows the initial period on an enlarged scale.

Figure 3A.60 Relaxation currents as a function of charging time of an aged pressboard sample and depolarization currents for 5, 50, and 500 s charging time as calculated from equivalent circuits obtained from relaxation currents at 2000 and 20000 s charging time.

Figure 3A.61 Relaxation currents as a function of charging time of an aged multi‐layer sample and calculated depolarization curves for 5, 50, and 500 s charging time from equivalent circuits obtained from relaxation currents at 2000 and 20000 s charging time.

Figure 3A.62 Polarization and depolarization currents of a dry pressboard sample as a function of charging duration.

Figure 3A.63 Relaxation currents of different multi‐layer samples: (a, b, d) ratio oil/board = 4, (c) ratio oil/board=1.

Figure 3A.64 Relaxation currents measured on four different configurations of the “pancake” model. The measurements have been performed for a voltage to electrode spacing ratio of 3 V/mm.

Figure 3A.65 Influence of insulation geometry on polarization currents at 0.6% moisture.

Figure 3A.66 Relaxation currents of two identical single‐phase 150 kV, 21 MVA power transformers.

Figure 3A.67 Measured polarization and depolarization currents on a new three‐phase 245/16 kV 230 MVA power transformer manufactured in 2000 and a used three‐phase 251/15.5 kV 220 MVA unit manufactured in 1967.

Figure 3A.68 Variation of polarization and depolarization currents for A1 and A2. 1 and 2, Pol/Depol current for A1; 3 and 4, Pol/Depol current for A2.

Figure 3A.69 Variation of polarization and depolarization currents for B1 and B2. 1 and 2, Pol/Depol current for B1; 3 and 4, Pol/Depol current for B2.

Figure 3A.70 Variation of polarization and depolarization currents for C1 and C2. 1 and 2, Pol/Depol current for C1; 3 and 4, Pol/Depol current for C2.

Figure 3A.71 Schematic diagram of RV measuring circuit.

Figure 3A.72 Timing diagram of RVM procedure.

Figure 3A.73 RVM parameters.

Figure 3A.74 RV profiles with increasing values of charging times (dots indicating maximum value of RV in each RVM cycle).

Figure 3A.75 Plot of polarization spectra: joining peaks of individual RVM cycles plotted against the corresponding charging times.

Figure 3A.76 Plot of initial slope spectrum.

Figure 3A.77 Peak time spectrum.

Figure 3A.78 Polarization spectra curves relative to paper moisture content at 25°C.

Figure 3A.79 Polarization spectra curves relative to paper moisture content.

Figure 3A.80 Polarization spectrum of a transformer before and after drying treatment.

Figure 3A.81 Polarization spectra curves relative to ageing at 120°C.

Figure 3A.82 Variation of polarization spectra with ageing by‐product.

Figure 3A.83 Polarization spectra for new and aged transformers.

Figure 3A.84 Polarization spectra between two different windings of the same transformer.

Figure 3A.85 Polarization spectra of different transformers under different operating conditions.

Figure 3A.86 RV spectra plotted against charging time at different temperatures.

Figure 3A.87 Temperature dependence of polarization spectra curves.

Figure 3A.88 Return voltage spectra dependence on temperature for an average‐loss transformer.

Figure 3A.89 RVM polarization spectra of maximum voltage and initial slope of the pancake model for two different geometrical arrangements, measured at 2000 V.

Figure 3A.90 Examples of RVM: (a) and (b) show “standard” spectra; (c) and (d) show “non‐standard” spectra.

Figure 3A.91 Local moisture ingress, as area of insulation paper of different DP value and bad oil quality can cause a local maximum in the RVM spectrum.

Figure 3A.92 Localized moisture ingress during onsite repairs on a transformer. The second curve shows the RVM spectrum after drying.

Figure 3A.93 Areas of differing DP value of the paper can cause multiple peaks.

Figure 3A.94 RVM spectrum of the test sample with bad‐quality oil.

Figure 3A.95 Equivalent circuit to model a linear dielectric.

Figure 3A.96 Polarization and depolarization currents before and after oil reclamation.

Figure 3A.97 RV spectra before and after oil reclamation.

Figure 3A.98 Modified Debye model.

Figure 3A.99 Structure of ES.

Figure 3B.1 Variation of loss factor of transformer insulation with frequency.

Figure 3B.2 Basic FDS measurement circuit using the IDA 200 system.

Figure 3B.3 FDS measurement configuration with guard electrode.

Figure 3B.4 Temperature variation profile during FDS measurements.

Figure 3B.5 FDS measurement between HV and LV.

Figure 3B.6 FDS measurement between HV and ground.

Figure 3B.7 Variation of real and imaginary part of complex permittivity at different moisture contents, at constant temperature of 20°C.

Figure 3B.8 FDS plots for paper with different moisture levels.

Figure 3B.9 Dissipation factor as a function of frequency at different moisture contents.

Figure 3B.10 Dissipation factor as a function of frequency at different moisture contents.

Figure 3B.11 Frequency‐dependent loss factor of a transformer (A) before and (B) after oil refurbishment.

Figure 3B.12 Dielectric loss tan

δ

variations over frequency range 0.1 mHz to 1 kHz for transformers A and B.

Figure 3B.13 Dielectric loss tan

δ

variations over frequency range 0.1 mHz to 1 kHz for transformers C and D.

Figure 3B.14 Effect of ageing on Kraft paper with 0.2% initial moisture content: (A) unaged; (B) 94 hours aged DP 840; (C) 1029 hours aged DP 360.

Figure 3B.15 Effect of ageing on Kraft paper with 4% initial moisture content: (A) unaged; (B) 72 hours aged DP 1000; (C) 2005 hours aged DP 280.

Figure 3B.16 Frequency response of aged and unaged impregnated pressboard samples: (a) initial moisture content 4%; (b) initial moisture content 0.6%. A and C, unaged; B and D, aged.

Figure 3B.17 Dissipation factor spectra for two differently aged transformers: (A) less aged; (B) more aged.

Figure 3B.18 FDS results: (A)

C

′ for transformer T1; (B)

C

′ for transformer T2; (C)

C

″ for transformer T1; (D)

C

″ for transformer T2.

Figure 3B.19 Variation of loss with frequency for transformers A1, A2, and A3.

Figure 3B.20 Frequency‐dependent DDF for transformers of different service times.

Figure 3B.21 Frequency response of permittivity of paper with acids: (A) without acid; (B) with light acid; (C) with heavy acid.

Figure 3B.22 Measured dielectric response at different temperatures.

Figure 3B.23 Tan

δ

plots at different temperature.

Figure 3B.24 (a) Real capacitance at different temperatures. (b) Imaginary capacitance at different temperatures.

Figure 3B.25 Dissipation factor tan

δ

at different temperatures.

Figure 3B.26 Dissipation factor tan

δ

at different temperatures of a field‐installed unit.

Figure 3B.27 Tan

δ

plots at different geometrical structure of insulation for oil/pressboard ratios 0:100 and 85:15.

Figure 3B.28 Cross‐section of transformer major insulation structure.

Figure 3B.29

X–Y

model of transformer insulation geometry.

Figure 3B.30 Effects of barriers on dissipation factor when

Y

 = 0.52.

Figure 3B.31 Effects of barriers on real capacitance when

Y

 = 0.52.

Figure 3B.32 Effects of barriers on imaginary capacitance when

Y

 = 0.52.

Figure 3B.33 Effects of spacers on dissipation factor when

X

 = 1.00.

Figure 3B.34 Effects of spacers on real capacitance when

X

 = 1.00.

Figure 3B.35 Effects of spacers on imaginary capacitance when

X

 = 1.00.

Figure 3B.36 Equivalent circuit to model a linear dielectric material.

Chapter 04

Figure 4.1 DGA interpretation scheme including C3 hydrocarbons.

Figure 4.2 Framework of pattern recognition.

Figure 4.3 Schematic diagram of SMOTE algorithm,

X

denotes a sample and

X

1

, …,

X

5

are its five nearest neighbors. SMOTE generates new data

m

i

along the line between

X

and one of five neighbors.

Figure 4.4 Architecture of hybrid SMOTEBoost and bootstrap algorithm.

Figure 4.5 Schematic of radial basis function network.

Figure 4.6 Sample distribution of dataset 1 in Table 4.3: (a) original dataset; (b) after processing by SMOTE and bootstrap. DS = discharge faults, OT = thermal faults, PD = partial discharge, normal deterioration.

Figure 4.7 Classification error rate of the SVM with integration of SMOTEBoost using oil characteristics dataset 6.

Figure 4.8 Classification error rate of the

k

NN classifier with integration of SMOTEBoost and bootstrap.

Chapter 05

Figure 5.1 Major steps of PD signal analysis.

Figure 5.2 Conventional PD measurement method (defined in IEC 60270).

Figure 5.3 An example of an acquired PD signal using Omicron MPD 600, which complies with IEC 60270.

Figure 5.4 Frequency responses of Rogowski coil and HFCT (model: 140/100HC from HVPD).

Figure 5.5 Types of structure element: (a) flat; (b) sinusoidal; and (c) triangular structure elements.

Figure 5.6 The MM‐based signal decomposition method.

Figure 5.7 Schematic models of SISO BE and EVA systems.

Figure 5.8 TF sparsity map for multiple PD source separation: (a) conceptual diagram and (b) flowchart.

Figure 5.9 Sparsity trends for time (top) and frequency (bottom) domains of decomposed signals. Each gray line represents a sparsity trend of each PD pulse. Solid gray lines represent the average of three different tendencies of the trends.

Figure 5.10 (a) Parameters of PS analysis and (b) the recently developed PRPS diagram.

Figure 5.11 Interpretation of spectral kurtosis [62].

Figure 5.12 AMT method for thresholding.

Figure 5.13 Energy values of (a) upper and (b) lower envelopes.

Figure 5.14 Results of AMT for random pulses [average upper and lower envelopes (gray dots) and optimal thresholds (dark gray lines)].

Figure 5.15 The signal decomposition‐based (EEMD‐based) PD signal de‐noising method.

Figure 5.16 De‐noising results of internal discharge. (a1)–(e1) Original signal, noise‐corrupted signal (SNR = −5 dB), and de‐noised signals of DWT (db5), EMD + AMT, and EEMD‐based method. (a2)–(e2) Corresponding PRPD diagrams.

Figure 5.17 BE‐based PD signal de‐noising and pattern representation (first) method.

Figure 5.18 Results of discharge in transformer oil using (a1)–(g1) CT1 and (a2)–(g2) CT2. (a) Original signals, (b) noise‐corrupted signals, (c) optimal equalized signals, (d) results of AMT on equalized signals, (e) de‐noised signals, (f) PRPD diagram of original signals, and (g) PRPS diagram of de‐noised signals.

Figure 5.19 Results of PRPD and PRPS diagrams for a distribution transformer using (a) CT1 and (b) capacitive measurement.

Figure 5.20 BE‐based PD signal de‐noising and pattern representation (second) method.

Figure 5.21 Results of corona. (a) Original signal, (b) noise‐corrupted signal, (c) pre‐whitened signal with magnified portion, (d) de‐noised (equalized) signal, (e) kurtogram of original signal, and (f) kurtogram of de‐noised signal.

Figure 5.22 Results of a 10 MVA transformer. (a) Original signal, (b) pre‐whitened signal, (c) de‐noised (equalized) signal, and (d) kurtogram of de‐noised signal.

Figure 5.23 Graphics‐based PD signal de‐noising method.

Figure 5.24 De‐noising results of internal discharge. (a) Original signal, (b) fractal dimension, (c) original PRPD diagram, and (d) de‐noised PRPD diagram.

Figure 5.25 De‐noising results of corona. (a) Original signal, (b) original signal (magnified), (c) original PRPD diagram, and (d) de‐noised PRPD diagram (magnified).

Figure 5.26 De‐noising results of PD signals acquired from a transformer. (a) Original signal, (b) original signal (magnified), (c) original PRPD diagram, and (d) de‐noised PRPD diagram.

Figure 5.27 Results of multiple PD sources combining corona, surface discharge, and discharge in transformer oil. (a) Result of conventional TF map, (b1) result of TF sparsity map, and (b2) PRPD diagram based on the result of TF sparsity map.

Figure 5.28 Results of multiple PD sources combining corona, surface discharge, and internal discharge. (a) Result of conventional TF map, (b1) result of TF sparsity map, and (b2) PRPD diagram based on the result of TF sparsity map.

Figure 5.29 Results for a 100 kVA transformer. (a) TF sparsity map and (b) PRPD diagram.

Figure 5.30 PD measurement results of a 10 kVA distribution transformer: (a) measured PD signals; (b) DWT decomposition.

Figure 5.31 Setup of acoustic PD measurement of a transformer.

Figure 5.32 Possible sound wave travel paths from PD source to three acoustic sensors.

Figure 5.33 Tested transformer tank and AE sensor installation.

Figure 5.34 The results of spatial intersectional PD localization: (a)

X–Y–Z

axis view; (b)

Y–Z

axis view; (c)

X–Y

axis view; and (d)

X–Z

axis view.

Figure 5.35 The final PD source localization result.

Chapter 06

Figure 6.1 Flux lines and force directions on HV and LV windings.

Figure 6.2 Buckling of an inner winding.

Figure 6.3 FRA measurement signals and typical transfer function plot for a power transformer.

Figure 6.4 Shorted parallel strands.

Figure 6.5 FRA test configurations for three‐phase transformer: (a) end‐to‐end open‐circuit test; (b) end‐to‐end short‐circuit test; (c) capacitive inter‐winding test; (d) inductive inter‐winding test.

Figure 6.6 Example of measured frequency response from each test configuration: (a) end‐to‐end open‐circuit test; (b) end‐to‐end short‐circuit test; (c) capacitive inter‐winding test; (d) inductive inter‐winding test.

Figure 6.7 End‐to‐end short‐circuit test response of the HV windings.

Figure 6.8 End‐to‐end open‐circuit responses of the HV winding with noises at low decibels.

Figure 6.9 Nyquist plot representation of the responses.

Figure 6.10 End‐to‐end short‐circuit responses of the HV winding.

Figure 6.11 Nyquist plot representation of the responses.

Figure 6.12 Frequency responses of a 16 kVA transformer subject to accelerated ageing process during laboratory study.

Chapter 07

Figure 7.1 Moisture equilibrium charts.

Figure 7.2 Oommen curves for low‐moisture region of moisture equilibrium for paper–oil system.

Figure 7.3 MIT‐developed curves for water equilibrium in cellulose–mineral oil system for a wide range of moisture concentrations.

Figure 7.4 Plot of water activity and WCO for mineral oil. The black line shows the ideal relationship between water activity and WCO when using equation (7.3).

Figure 7.5 Comparison of solubility for differently degraded mineral oils.

Figure 7.6 Example water activity probe.

Figure 7.7 Comparison of measuring WCP using probes at different locations in transformer, corrected for temperature difference.

Figure 7.8 Bubbling inception temperature. The triangle dots are calculated from equation (7.10). The orange trace [equation (7.11)] has a 10 °C safety margin added – the original line of best fit included 171.92, however, this was reduced to 160 to take into account a 10 °C margin of safety. ‘Water emission’ and ‘Water droplets’ taken from Ref. [16], Koch data from Ref. [15].

Figure 7.9 Dependency of

A

value on oxygen content of oil and water content of paper. Low oxygen is <7000 ppm, medium oxygen 7000–14,000 ppm, and high oxygen 16,500–25,000 ppm. A high oxygen level is typical for a free‐breathing transformer.

Figure 7.10 Dielectric response of transformers.

Figure 7.11 Temperature and water activity of oil in TR1.

Figure 7.12 Temperature and water activity of oil in TR2.

Figure 7.13 WCP calculated using water activity data for TR1.

Figure 7.14 WCP calculated using water activity data for TR2.

Figure 7.15 TR1 calculation of WCP from WCO.

Figure 7.16 TR1 oil temperature on sampling.

Figure 7.17 TR2 calculation of WCP from WCO.

Figure 7.18 TR2 oil temperature on sampling.

Figure 7.19 TR1 paper quality estimation.

Figure 7.20 TR2 paper quality estimation.

Figure 7.21 Temperature and water activity of the oil.

Figure 7.22 WCP for TR3.

Figure 7.23 Comparison of temperature measurements in transformer.

Figure 7.24 Calculation of WCP using oil measurements.

Figure 7.25 Temperature of oil when sampled.

Figure 7.26 Long‐term WCP for TR3.

Figure 7.27 Analysis of oil furan content, calculated and measured DP.

Chapter 08

Figure 8.1 Molecular structures of different hydrocarbons in mineral insulating oil [6, 8].

Figure 8.2 Molecular structure of triglycerides [15].

Figure 8.3 Generic molecular structure of synthetic ester [17].

Figure 8.4 Major step of oxidative degradation of natural ester‐based insulating oil [20–22].

Figure 8.5 Major steps of NE hydrolysis degradation.

Figure 8.6 Major steps of SE oxidative degradation.

Figure 8.7 (a) Moisture saturation graph for new natural and synthetic ester insulating liquids in comparison with typical new mineral oil [14, 28–30]. (b) Equilibrium chart for mineral oil. (c) Equilibrium chart for natural ester. (d) Equilibrium chart for synthetic ester.

Figure 8.8 Biodegradability of synthetic ester insulating liquids in comparison with mineral oil.

Figure 8.9 Average winding temperature rise and top oil differentials between mineral oil and natural ester.

Figure 8.10 Life of thermally upgraded insulation in mineral and natural ester.

Figure 8.11 DP measurement results of case study [47, 57].

Figure 8.12 Relationship between DP and 2‐FAL concentration of oil [47].

Figure 8.13 DGA results after 100 electrical breakdowns.

Figure 8.14 Stray gas in oil after ageing over 72 h at 150 °C: (a) combustible gases; (b) CO

2

.

Figure 8.15 Combustible gas dissolved after ageing for 2800 h with pressboard and copper [47, 57].

Figure 8.16 Comparison of FDS for different types of oil‐impregnated unaged pressboard insulation at 55 °C: (a) mineral 0.8%, NEA 0.8%, NEB 0.8%, SE (MIDEL 7131) 0.6%; (b) mineral 2.4%, NEA 2.1%, NEB 2.7%, SE (MIDEL 7131) 3%.

Figure 8.17 Comparison of FDS for different types of aged oil‐impregnated pressboard insulation: (a) mineral 0.8%, NEA 0.8%, NEB 0.8%; (b) mineral 3.6%, NEA 2.8%, NEB 2.9%.

Figure 8.18 Comparison of moisture analysis of paper insulation using KFT and FDS‐based tool (MODS).

Figure 8.19 DGA and physicochemical analysis of natural esters.

Figure 8.20 Change of oil properties over ageing.

Figure 8.21 Plot of gases dissolved in vegetable oil of Sydney transformer unit: (a) change in gases around the time of energization; (b) gases after a 6‐month settling in period.

Chapter 09

Figure 9.1 Intelligent framework for power transformer condition monitoring and diagnosis.

Figure 9.2 Multi‐scale thresholding. The horizontal axis denotes the cumulative number of signals and the vertical axis denotes the amplitude of signals (mV).

Figure 9.3 Results on PD signal de‐noising using simulated signals: (a) simulated noise‐corrupted PD signals (SNR = −18.6 dB); (b) de‐noised PD signals by single‐amplitude‐based thresholding wavelet transform; (c) de‐noised PD signals by multi‐scale thresholding wavelet transform (signals having higher probability indexes are depicted in the darkest shade).

Figure 9.4 Results of de‐noising PD signals obtained from experimental model of corona: (a) noise‐corrupted PD signals (PD signals are in red, noise‐corrupted signals are in blue) with SNR = −28; (b) de‐noised PD signals by single‐amplitude‐based thresholding wavelet transform; (c) de‐noised PD signals by multi‐scale thresholding wavelet transform (signals having higher probability indexes are depicted in a darker color).

Figure 9.5 Results of de‐noising PD signals obtained from a 10 MVA (33 kV/11 kV) substation transformer: (a) measured signals in 25 AC cycles; (b) de‐noised PD signals by single‐amplitude‐based thresholding wavelet transform; (c) de‐noised PD signals by multi‐scale thresholding wavelet transform (signals having higher probability indexes are depicted in a darker color).

Figure 9.6 PCA visualizations for DWT feature extraction on PD dataset (Model 1 – corona, Model 2 – discharge in oil, Model 3 – surface discharge, Model 4 – internal discharge, Model 5 – discharge due to floating particles).

Figure 9.7 Multiple PD sources obtained from internal discharge and discharge due to floating particles.

Figure 9.8 PRPD diagram of a 5 MVA power transformer in a substation.

Figure 9.9 SOM clustering results on two sets of samples: sample A (collected from bottom) and sample E (collected from Buchholz). The transformer conditions include 1 – normal operating condition, 2 – discharge fault, 3 – lower to medium‐range thermal fault, 4 – high‐range thermal fault, and 5 – partial discharge. (The numbers in brackets are the number of data points with that fault type in the training dataset, which consists of 390 DGA records with known transformer condition provided by a utility company).

Figure 9.10 Conceptual diagram of pattern recognition‐based transformer health index prediction [38].

Figure 9.11 SmartBox – configurable hardware and software platform: (a) hardware modules; (b) system software architecture.

Figure 9.12 Software platform based on a multi‐agent system for transformer condition monitoring and assessment.

Guide

Cover

Table of Contents

Begin Reading

Pages

iii

iv

vii

viii

ix

xi

xii

xiii

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

Transformer Ageing

Monitoring and Estimation Techniques

Edited by

Tapan Kumar Saha

The University of QueenslandSt. Lucia, Brisbane, Australia

Prithwiraj Purkait

Haldia Institute of TechnologyWest Bengal, India

This edition first published 2017© 2017 John Wiley & Sons Singapore Pte. Ltd

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permision to reuse material from this title is available at http://www.wiley.com/go/permissions.

The right of Tapan Kumar Saha and Prithwiraj Purkait to be identified as the authors of the editorial material in this work has been asserted in accordance with law.

Registered OfficesJohn Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USAJohn Wiley & Sons Singapore Pte. Ltd, 1 Fusionopolis Walk, #07‐01 Solaris South Tower, Singapore 138628

Editorial Office1 Fusionopolis Walk, #07‐01 Solaris South Tower, Singapore 138628

For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.

Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats.

Limit of Liability/Disclaimer of Warranty

Limit of Liability/Disclaimer of Warranty: While the publisher and author(s) have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising herefrom. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Library of Congress Cataloging‐in‐Publication Data

Names: Saha, Tapan Kumar, 1959– editor. | Purkait, Prithwiraj, 1973– editor.Title: Transformer ageing : monitoring and estimation techniques / edited by Tapan Kumar Saha, Prithwiraj Purkait.Other titles: Transformer agingDescription: Chichester, West Sussex : John Wiley & Sons, Inc., 2018. | Includes bibliographical references and index.Identifiers: LCCN 2017003580 (print) | LCCN 2017004685 (ebook) | ISBN 9781119239963 (cloth) | ISBN 9781119239994 (Adobe PDF) | ISBN 9781119239987 (ePub)Subjects: LCSH: Electric transformers--Maintenance and repair--Handbooks, manuals, etc. | Electric lines--Maintenance and repair--Handbooks, manuals, etc. | Electric insulators and insulation--Testing. | Electric power distribution.Classification: LCC TK2551 .T7635 2018 (print) | LCC TK2551 (ebook) | DDC 621.31/4--dc23LC record available at https://lccn.loc.gov/2017003580

Cover image: © gjp311/GettyimagesCover design: Wiley

Preface

The transformer is one of the most important pieces of equipment in a power grid. The condition of large power transformers has a significant impact on the reliability of the power grid. Large power transformers are expensive and complex in design and operation. Transformer condition monitoring and assessment of their remaining life is an important task for transformer owners/operators. Transformer condition monitoring covers many areas closely related to transformer structure and operation. The condition of the insulation system plays a major role in determining the life of a transformer. Similarly, winding/core integrity, bushing, and tap changer health are also important in maintaining the overall reliable operation of a transformer.

Throughout the life of transformer operation, the insulation system degrades and the degradation mechanism depends on the operating conditions inside the tank. Thermal, hydrolytic, and oxidation processes are the main causes of ageing of a transformer. Many diagnosis techniques have been in use for several decades, and their interpretation tools have always been the focus of improvements over the years. Many new diagnosis tools are being investigated continuously by researchers and engineers in the field. New insulation systems for solid and liquid insulations are being proposed and investigated for power transformers. This book will provide fundamental knowledge of transformer insulation materials, their ageing mechanisms, traditional as well as advanced condition monitoring techniques, and interpretation techniques. Basic knowledge of the transformer will be a prerequisite for readers.

The research work presented in this book was conducted with funding support from the Australian Research Council and the Australian electricity supply industry, with collaboration from several transmission and distribution utilities in Australia over a period of 25 years. Our expectation is that this book will provide state‐of‐the‐art knowledge about transformer ageing, condition monitoring, and fault diagnosis. No single book is currently available that provides such an important knowledge base for transformer condition monitoring and life assessment. The authors hope that this book will be a “one‐stop” information provider for engineering students, practicing engineers, and researchers. We believe anyone working in transformer condition monitoring – particularly engineers working in electricity utilities, graduate or senior undergraduate students and researchers, postdoctoral fellows, and academics – will benefit from this publication.

The authors of this book have published scores of journal and conference articles via the IEEE, IET, and CIGRE. Many of these will provide an additional knowledge base resource for the reader. Many diagnostic algorithms have been developed throughout this journey, and they are currently available from the University of Queensland research team.

This book is organized into ten chapters. Chapter 1 discusses the sources, properties, and applications of insulating materials used in transformers, along with an overview of ageing of oil–paper insulation systems. This chapter also provides a foundation for understanding insulation diagnostic tools, which can assist the reader in relating the diagnosis with the cause of insulation ageing. Chapter 2 explains comprehensively the dissolved gas analysis (DGA), furan analysis, and degree of polymerization (DP), and their relevant international standards. In addition, this chapter introduces electrical‐based traditional diagnoses, which include insulation resistance (IR), polarization index (PI), dielectric dissipation factor (DDF), capacitance and power factor, dispersion factor, and partial discharge (PD).

Chapter 3 provides theoretical explanations of polarization–depolarization current (PDC), recovery voltage measurements (RVM), and frequency domain dielectric spectroscopy (FDS), along with their interpretation schemes. The effects of moisture, ageing, temperature, and insulation geometry on the interpretation of PDC, RVM, and FDS measurements are also described in this chapter.

Chapter 4 outlines commonly used interpretation techniques of dissolved gas analysis (DGA), with a comprehensive review and illustration of machine learning‐based DGA interpretation techniques, with specific focus on artificial neural networks (ANNs), fuzzy logic systems, expert systems, decision‐making algorithms, and support vector machine (SVM) and population‐based algorithms. This chapter also provides some insights into training dataset construction and data quality improvement, and discusses approaches to classification accuracy and generalization capability validation.

Chapter 5 provides a detailed analysis of partial discharge (PD) measurement and interpretation tools for transformer condition monitoring. This chapter primarily highlights advanced signal processing techniques, with focus on wavelet transform (WT), empirical mode decomposition (EMD), ensemble EMD (EEMD), and mathematical morphology (MM) methods. Special techniques developed for multiple‐PD source separation and their PD feature extraction and recognition are also discussed in this chapter.

Chapter 6 concentrates on frequency response analysis (FRA) for transformer winding mechanical deformation/displacement analysis. A number of international standards are discussed in this chapter, along with a novel statistical approach.

Chapter 7 primarily explains moisture measurements by online sensors, with a comprehensive guideline for practicing engineers to estimate the remaining life of insulation as a function of the water content of paper.

Chapter 8 presents biodegradable oil fundamentals and their impact on paper insulation ageing. Then, oil chemical/physical measurements and PDC/FDS interpretation schemes for biodegradable oil‐filled transformers are presented, with a comparison of currently used condition monitoring interpretation techniques for mineral oil‐based transformers.

Chapter 9 provides an intelligent framework for transformer condition monitoring using online sensors, along with the importance of numerical modeling to assist fault detection in transformers, statistical learning for dealing with measurement uncertainties, and data and information fusion for transformer condition assessment. A hardware and software platform for implementing a smart transformer condition monitoring system and a concept of health index and their interpretation are also discussed in this chapter.

Chapter 10 highlights the limitations of current condition monitoring techniques and the need for future research.

Acknowledgments

Many people have supported this work, directly or indirectly, throughout our involvement with transformer research. We would like to acknowledge some of the key personnel without whose contributions this publication would never have reached this point.

Emeritus Professor Mat Darveniza for introducing the topic of transformer insulation ageing and life assessment during Tapan Saha’s Ph.D. research.

Honorary Reader David Hill and Dr. Tri Li from the School of Chemistry and Molecular Biosciences, University of Queensland for helping to understand chemistry of insulation materials and some chemical‐based diagnosis concepts.

Mr. Richard Marco, Mr. Brian Williams, and Dr. Zheng Tong Yao during the initial hardware/software design of PDC‐RVM equipment at the University of Queensland.

A number of research fellows who worked at the University of Queensland with Tapan Saha during the last 25 years need to be mentioned specifically: Dr. Abbas Zargari, Dr. Prithwiraj Purkait, Dr. Manoj Pradhan, Dr. Chandima Ekanayake, Dr. Hui Ma, and Dr. Dan Martin.

Tapan Saha has been fortunate to advise numerous Ph.D. students in this area. Their contributions are worthy of note: Dr. Zheng Tong Yao, Dr. Karl Mardira, Dr. Jing Haur Yew (Kelvin), Dr. Raj Jadav, Dr. Mohd Fairouz, Dr. Jeffery Chan, Dr. Yi Cui, and Dr. Kapila Bandara. Thanks to a number of Masters by research students for their contributions to this publication.

Dr. David Allan from Powerlink Queensland, who provided extensive industry collaboration throughout Tapan Saha’s research.

Mr. Bryce Corderoy, Mr. Vic Galea, and Dr. Frances Mitchell from TransGrid New South Wales for providing industry‐oriented transformer research opportunities.

Numerous undergraduate and Masters students for their contributions through their thesis projects.

The Australian Research Council for providing several funding supports through the ARC Linkage Project Scheme, without which this volume of work would never have been possible to conduct.

Industry support from Powerlink Queensland, Energex, Ergon Energy, TransGrid, Ausgrid, and Aurecon through extensive industry collaborations.

CIGRE Australian Panel Members A2 and D1 for providing extensive knowledge in the area of transformer and insulation diagnostics.

The authors of many papers and books, from which we have continuously benefitted in our journey. If we have inadvertently missed any referencing or acknowledgment of these authors, we sincerely apologize.

Special thanks to Mr. Steven Wright for his help during many experiments in the intelligent equipment condition monitoring laboratory and for proofreading the book.

Special thanks to Dr. Hui Ma for reading many chapters of this book throughout the last 12 months of manuscript preparation.

Sincere thanks to the University of Queensland for providing the facilities and opportunities to carry out research in this area.

Thanks to our families for understanding and support throughout our research career.

Contributing Authors

A number of our current colleagues and Ph.D. students at the University of Queensland have contributed directly in preparing the manuscript of this book. Their contributions are greatly appreciated.

Chapter 1: Prof. Tapan Saha & Prof. Prithwiraj Purkait

Chapter 2: Prof. Tapan Saha & Prof. Prithwiraj Purkait

Chapter 3 Part A: Prof. Tapan Saha & Prof. Prithwiraj Purkait

Chapter 3 Part B: Prof. Tapan Saha, Prof. Prithwiraj Purkait, & Dr. Chandima Ekanayake

Chapter 4: Dr. Yi Cui, Prof. Tapan Saha, & Dr. Hui Ma

Chapter 5: Dr. Jeffery Chan, Prof. Tapan Saha, & Dr. Hui Ma

Chapter 6: Dr. Mohd Fairouz, Prof. Tapan Saha, & Dr. Chandima Ekanayake

Chapter 7: Dr. Dan Martin & Prof. Tapan Saha

Chapter 8: Dr. Kapila Bandara, Prof. Tapan Saha, & Dr. Chandima Ekanayake

Chapter 9: Dr. Hui Ma & Prof. Tapan Saha

Chapter 10: Prof. Tapan Saha & Prof. Prithwiraj Purkait

Research colleagues: Dr. Chandima Ekanayake, Dr. Hui Ma, & Dr. Dan Martin.

Former Ph.D. students: Dr. Jeffery Chan, Dr. Mohd Fairouz, Dr. Yi Cui, & Dr. Kapila Bandara.

1Transformer Insulation Materials and Ageing

1.1 Introduction

The primary and secondary coils of a transformer are the key components in performing its basic function of transforming voltage and current. Materials are used to insulate the primary and secondary coils. In transformers, in addition to the primary and secondary coils, there are several other important components and accessories. The insulating material is one of the most critical components of a transformer. Sufficient insulation between different active parts of the transformer is necessary for its safe operation. Adequate insulation is not only necessary to isolate coils from one another, or from the core and tank, but also ensures the safety of the transformer against accidental over‐voltages.

The insulation system in a transformer can be categorized as follows.

Major insulation:

– between core and low‐voltage (LV) winding;

– between LV and high‐voltage (HV) winding;

– between top and bottom of winding and yoke;

– between HV and tank;

– bushings.

Minor insulation:

– between conductors;

– between turns;

– between layers;

– between laminations;

– between joints and connections.

The insulation material commonly used between the grounded core and the LV coil to ground, and also between HV and LV coils, is oil‐impregnated solid pressboard. Solid insulation, including pressboard or paper, can have small internal air voids. This reduces the insulating strength of the solid insulation as well as reducing its heat dissipation capacity. When transformer oil is used to impregnate solid insulation, the voids are filled with oil, resulting in an improvement of both the insulation strength and the heat dissipation capacity of the solid insulation. In larger transformers, cellulose‐based paper tape is usually wrapped over individual conductors. Layer‐to‐layer or disc‐to‐disc insulation is mostly provided by oil‐impregnated Kraft paper or even thick pressboard or transformer board in case of higher‐rating transformers.

1.2 Solid Insulation – Paper, Pressboard