Partial Discharge Inception Voltage Measurement

for an Artificially Created Void Inside the Solid

Dielectric
Mahboob Hassan, Nabila Elbeheiry, Shady S. Refaat
Texas A&M University at Qatar, Doha, Qatar
Emails: mahboob.alg@gmail.com, nabila.elbeheiry@qatar.tamu.edu, shady.khalil@qatar.tamu.edu

Abstract— All the electrical equipment are provided with thickness and the larger the ratio of void diameter to void
insulation material for their safe and reliable operation; hence thickness the field within a void becomes larger [4],[11],
insulation material quality contributes to the lifespan of the however the effect of varying the void thickness or the
equipment. Partial Discharge (PD) test on solid insulation is one thickness of the insulation is not studied yet. Two important
of the methods that determine the quality of insulation. PD parameters that are used in PD tests are the inception and
activity is a physical phenomenon, in which solid insulation extension voltages. The inception voltage is the lowest voltage
experiences a partial breakdown in their electrically weak at witch PD pulses first appear as the voltage increases
regions. PD often starts from voids or cavity and lead to gradually from zero. While the extension voltage is the one
insulation material failure the insulating material. Better
where recurring PD pulses occur when the voltage across the
understanding of PD, helps in diagnosing insulation defects and
predict flashover before they occur. Therefore, this paper
void gradually decreases from a value higher than the
investigates the impact of partial discharge as a consequence of inception voltage [12]. Doing such research on PD and
a single cavity in solid dielectrics. The proposed algorithm is investigating the breakdown voltage is trivial as one can
demonstrated by mathematical model and laboratory develop the cables' industry to prevent further discharges [13].
experiments. The paper evaluates the behavior of partial Also fixing defects before cable failure is another benefit [14].
discharge for insulations with different thicknesses and void Additionally, since PD is contributing hugely in the aging of
size. The theoretical behavior was computed using Townsend power cables, knowledge of PD patterns, in general, helps in
breakdown criterion along with distribution law for doing need-based maintenance instead of regular based ones
capacitance. An online monitoring system was used to detect and doing quality assurance tests [15]-[16].
partial discharge in the lab. The obtained theoretical and
experimental results were compared. Many techniques have been presented to figure out inception
voltage inside solid dielectric materials such as in research
Keywords—Partial Discharge (PD), Partial Discharge held by Parent and his team that uses Paschen’s law to
Inception voltage (PDIV), Partial Discharge Extinction voltage compute the breakdown voltage [17]. Paschen’s law can be
(PDEV), Insulation used if only two of the qualities describing the electric field
are known. Another established approach is to measure the
I. INTRODUCTION peak voltage in the first cycle with a PD pulse of duration
Partial discharge, a phenomenon that results in deteriorating higher than five picovolts. Then an average of several
insulation, occur due to long term operation, inaccuracy measurements can be taken [18].
during manufacturing and installation, insufficient grounding, The contribution in this paper is to investigate the effect of the
adverse environmental conditions or poor contact. Etc [1]. PD void thickness and the thickness of the insulation on PD. To
is defined as the appearance of electrical sparks inside the do so, both mathematical simulations and laboratory
insulation that partially bridges the path between conductors measurements were conducted. In the paper, both the
[2]. A full-bridge will cause the breakdown of the system. PD inception and extension voltages are measured to evaluate the
occurs in solid, liquid and gaseous insulating materials. behavior of PD. The characteristics of PD due to the existence
Depending on the location of occurrence, partial discharges of a single void are the main focus of this paper along with
are classified into three categories: Internal, Surface and many others.
Corona Discharges [3]-[4]. Internal Discharge takes place
within a void, cavity or impurity inside a solid insulating The paper is composed of four sections: procedure, results,
material. Surface Discharge occurs at the insulation surface discussion, and conclusion. In the procedure, the process taken
caused by a high tangential field [5]-[6]. Corona Discharge is to perform the laboratory experiments is explained step by
caused by the ionization of a fluid surrounding a conductor step, including all the precautions taken. The results section
due to the high electrical field [7]. starts by presenting the collected measurements, then an
explanation of how Townsend breakdown criterion was used
Internal discharge occurs due to the change in the dielectric to calculate the PDIV, followed by the calculated values.
properties as a result of the presence of a void; the dielectric Lastly, the discussion section contains a comparison between
constant of the air is less than that of the insulating material measured and calculated results.
[8]. These changes cause a non-uniform distribution of the
electric field around the void and then, at a specific voltage - II. PROCEDURE FOR PDIV MEASUREMENT
inception voltage- PD occurs [9]. Different void parameters, This section starts with an explanation of how samples that
including size, type, and shape, affect the inception voltage. models insulation containing a void were modeled in the lab.
Previous research in the field has found that the applied
Next, how electrodes are prepared and used to stimulate the
voltage waveform impact PD pulses [10]. Additionally, while
investigating the effect of different void parameters, it was PD events by applying high voltage stress is described. Then
found that the smaller the ratio of void thickness to insulation the used digital tool to measure PD is shown, followed

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 05,2020 at 09:13:56 UTC from IEEE Xplore. Restrictions apply.
 (a) Before casting (b) Casted in epoxy adhesive

Fig. 2. Ground electrode

Fig. 1. Sample constituted by three perspex sheets The ground electrode was casted using the epoxy adhesive
(Araldite), as shown in Fig. 2(b), to prevent any surface
by the system, composing of the test cell, measuring device discharge during the experiment, as it acts as insulation. Both
and voltage supply, setup demonstration. Lastly, the of these electrodes are fixed to set up the test cell as shown in
approach taken to measure partial discharge inception voltage Fig. 3.
is described.
C. Digital Method of PD measurement
A. Preparation of Samples
Modern innovative technology has introduced digital PD
To do the experiment, three samples, with different void measuring devices which can be used both in the laboratory
and insulation thickness, were prepared in the lab to model and on site. The new computerized measurement system
insulation with a void inside. Each sample is made of three 80 mainly focuses on the identification of various PD patterns
mm diameter circular Perspex discs where one of the discs and their analysis.. An online monitoring system that enables
was drilled with a hole at the center. The discs surfaces were automatic PD data analysis can contribute to the improvement
cleaned using benzene, and then the benzene trace was of the new system [21]. The new monitoring system reduce
carefully removed using alcohol. Next, soap water and tap the requirement of personnel expertise and improves
water were used to wash and rinse the discs respectively. accuracy, but it also provides economic benefits [22]. This
Lastly, cotton wool was used to wipe the surfaces of the discs. method promises to be more reliable to access the condition
This was done carefully to ensure that fibers of cotton do not of dielectric but needs to be used in conjunction with modern
stick to the surfaces of the discs. techniques [22].
Afterward, the disc with the hole was placed between the The MPD 540 PD analysis system is a highly precise
other two discs. Then this arrangement was tapped using system that detects, records, and analyses PD events. It
Adhesive tape around its edges, in order to form the sample consists of a coupling capacitor (CPL 542), a fiber optical
with an artificial cavity as displayed in Fig. 1. Before joining USB controller (US B502), an acquisition unit and software
the three discs, it was made sure that there are no air gaps in used to record the PD events as per international standards
between. Finally, an oil-tight joint was made by tapping the [23]. MPD 540 acquisition unit is capable of working in
edges all-around of the discs using an adhesive tape which was conjunction with many sensors including capacitive sensors
resistant to oil. and inductive sensors. as shown in Fig. 4. It can be connected
to 960 channels to perform multi-channel PD measurement.
B. Electrode Configuration To Evaluate PD process
The test setup needed to stimulate PD event was prepared
in the lab and displayed in Fig. 3. As the figure shows the
experimental setup is composed of a metallic (conducting)
base fixed with the ground electrode and a non-conducting
clamp that holds the high voltage electrode. It can be altered
right or left at any angle and moved upwards or downwards to
fix the high voltage electrode at any desired vertical position.
Electrodes, an integral part of the test setup, are used to
stress the sample under high voltage to stimulate PD activity.
Many electrodes were designed in Europe and US to examine
PD [19, 20]. The high voltage electrodes were provided using
a solid cylindrical rod of 7 mm diameter with semispherical
ends. Additionally, the ground electrodes were made of
copper circular discs having 4.95 cm diameter and 12.94 mm
width. The discs were prepared using the lathe machine which
was used to chisel copper round bars. Sharp edges at the ends
were then eliminated, to prevent the corona discharge. One of
the electrodes faces was provided with threading at the center
up to nearly half of the depth, to fix it at the screw as shown Fig. 3. Experimental set-up
in Fig. 2(a) [19, 20].

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 05,2020 at 09:13:56 UTC from IEEE Xplore. Restrictions apply.
 E. Inception Vvoltage Measuring Approach
After the connections were made as shown in Fig. 6., the
voltage was gradually increased from zero to the value at
which discharge starts appearing. Then the voltage is
gradually decreased to its minimum. Advanced PD measuring
system stores the values of partial discharge inception voltage
(Vi) and partial discharge extinction voltage (Ve)
automatically.
III. RESULTS OF PDIV

Fig. 4. MPD540 acquisition units

A. Experimental Results
Doing the above-mentioned connections, and following
D. System Setup the procedure enabled measuring the inception and extinction
The system consists of one fiber optics controller USB502, voltage for the three samples. In an attempt to obtain an
one or more MPD540 acquisition units and a coupling unit accurate result, the measurements were taken three times for
MPD540CPL or CPL542 (also referred to as quadripole). As each sample and the mean was calculated. Table I lists the gap
shown in Fig 5. A fibre optical cable is used to connect the and insulation thickness for each sample along with the
MPD540 to the USB502 f. These cables are useful for making measured voltages. The mean values of measured partial
a connection between test set up and computer. discharge inception voltage (Vi) are recorded in Table II.

TABLE I. MEASURED VALUES OF INCEPTION VOLTAGES (VI) AND

EXTINCTION VOLTAGES (VE)

Thickness of Inception Extinction

Voltage (Vi) Voltage (Ve)
Gap (t') Insulation (T)
(mm) (mm) (kV) (kV)
16.01 7.179
2.526 7.578 16.31 8.617
15.50 7.885
18.54 9.736
3.630 8.901 18.50 10.950
18.62 10.020
Fig. 5. MPD540 with coupling capacitor and PC arrangement 22.79 7.396

Additionally, MPD540 enables the use of up to 2 km of Fiber 5.327 10.379 23.05 7.395
optical cables. The TX (transmitter) must be connected to RX 22.68 6.628
(receiver) and vice versa as in the wiring diagram shown in
Fig 5. MPD540 is connected to a power supply (or battery). TABLE II. MEAN MEASURED VALUES OF INCEPTION VOLTAGES (VI)
Two outputs PD and V of MPD540CPL or quadrupole are
connected to the similar inputs of the MPD540 with the help Thickness of Inception Voltage
of BNC cables of smaller length. Then a high voltage coupling (Vi)
Gap (t') Insulation (T)
capacitor, provided in the laboratory MPD540CPL, is (kV)
connected using two conducting cables. The MPD540 is kept (mm) (mm)
near the test object and the coupling capacitor. The calibrator, 2.526 7.578 15.94
which can be seen in Fig. 6. was disconnected before the high 3.630 8.901 18.55
voltage supply is turned on. 5.327 10.379 22.84

B. MATHEMATICAL THEORY FOR PDIV ESTIMATION

Based on Townsend breakdown criterion [ ( ∙ )−

1] = 1 breakdown voltage Vb at pressure p for the void or gap
is given [25, 26] as:

B ∙ p ∙ t′
V = (1)
1
ln A ∙ p ∙ t´/ ln 1 +
γ

Where p is the pressure, t' is depth of void (space of gap) and

γ denotes coefficient of secondary ionization. For the primary
Fig. 6. Connection of MPD 540 PD measuring system ionization coefficient, Townsend equation [13]-[15] can be
used to evaluate values of constant A and constant B.

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 05,2020 at 09:13:56 UTC from IEEE Xplore. Restrictions apply.
 function of pt' [32]. So the value of k in for air [25, 27, 28] is
α = A ∙ p exp[−Bp/E] (2) given in Table IV as a function of pt'.

A denotes saturation ionization in gaseous material for a TABLE IV. VALUE OF CONSTANTS K FOR AIR
specified electric stress/pressure (E/p) ratio and B is Pt'
corresponds to the energies for excitation and ionization [27, Gas K
(kPa-cm)
28].
0.0133-0.2 2.0583(pt')-0.1724
The values of γ reported in the literature [29], those values are 0.20-100 3.5134(pt')0.0599
calculated by either by analyzing ionization growth curves at Air
a fixed electrode gap or by using Townsend breakdown
100-1400 4.6295
criterion γ[exp(at′) − 1] = 1 after measuring the breakdown
voltage. Second method is commonly used for measurement
of γ if one can measure accurately the breakdown voltage. Corresponding to the value of pt' between 0.20 kPa-cm and
Since the value of breakdown voltages is accurately available 100 kPa-cm, k is given as under k = 3.5134 (pt') 0.0599
for the air (gas considered in our case) for a big extent of the
C. PDIV MATHEMATICAL COMPUTATIONS
values of pt’ so these values can be used [30].
In this section, the theoretical values of the inception
Equation (1) is modified as under: voltage were calculated. The calculations were made for three
samples; Samples with identical parameters to the ones used
V = ( )
(3) in the experimental part. A comparison between calculated
and measured values is then made in the discussion section.
Where,
A. Sample 1
k = ln[A/ln(1 + 1/γ)]
Gap thickness (t') = 2.526 mm
For air the constant A and constant B in (2) are written in
Table I [29, 30]. Total thickness of the sample (T) = 7.578 mm
Equation (3) can be represented by: Substituting the values of variables in (4)
E = = (4) 740
2737.50 × × 101
( )
E = 760
.
740 740
Where, Eg denotes electric strength for separation (gap) for ln ∙ 101 ∙ 0.2526 + 3.5134 × ∙ 101 ∙ 0.2526
760 760
the pressure p and t' is gap thickness where discharges occur.
Eg= 36032.49
Fig. 3 depicts t' (thickness of the gap) and T (overall thickness
of insulation). Now substituting the values of variables in (5)
Considering the voltage distribution law for capacitance and 0.7578 + 0.2526(3.5 − 1)
Fig. 3, PD inception voltage (Vi) can be estimated by the given V = 36032.49 V
3.5
relations for voids discharges [31]:
V = 14.302 kV
T + t′(ε − 1)
V =E (5)
ε B. Sample 2
Where εr is dielectric relative permittivity of the sample.
For Perspex value of relative permittivity is 3.5 Gap thickness (t') = 3.63 mm
Value of Eg is substituted from (4) into (5) to estimate Vi
Total thickness of the sample (T) = 8.9 mm
as below:
∙ ( ) Substituting the values of variables in (4)
V = ∙ (6)
( ∙ )
740
2737.50 × × 101
E = 760
.
For Perspex εr = 3.5 740 740
ln × 101 × 0.363 + 3.5134 × × 101 × 0.363
760 760
For air, the constant A and constant B are found from Table III E = 33959.124 V
[26].
Now substituting the values of variables in (5)
TABLE III. VALUES OF CONSTANTS A AND B FOR AIR
0.89 + 0.363(3.5 − 1)
A V = 33959.124 V
B 3.5
Gas (Ionization /kPa-
(V /kPa-cm) V = 17.44 kV
cm)
Air 112.50 2737.50 C. Sample 3
Gap thickness (t') = 5.327 mm
Breakdown voltages Vb and values of constant A and constant
B are used to compute the value of k and obtained it as a Total thickness of the sample (T) = 10.379 mm
Substituting the values of variables in (4)

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 05,2020 at 09:13:56 UTC from IEEE Xplore. Restrictions apply.
 740
2737.50 × × 101
E = 760
.
740 740
ln × 101 × 0.5327 + 3.5134 × × 101 × 0.5327
760 760
Eg = 32002.39 V

Now substituting the values of variables in (5)

1.0379 + 0.5327(3.5 − 1)
V = 32002.51 V
3.5
Vi = 21.67 KV

IV. Discussion
Fig. 7.Inception voltage (Vi) v/s cavity depth (t’)
A comparison between the observed (measured
experimentally) and mathematically calculated inception
voltage was conducted and some of the values is
displayed in Table V. It was found that both the measured
and calculated inception voltage followed the same
pattern of variation with a little difference in their
corresponding values. This difference may be due to the
varying atmospheric conditions and properties of the
sample material.

TABLE V. OBSERVED AND CALCULATED INCEPTION VOLTAGES (VI)

Depth Mean Observed Calculated

Sample of thickness of inception inception
no cavity sample voltage voltage
(cm) (cm) (kV) (kV) Fig. 8. Inception Voltage (Vi) v/s sample thickness (T)
1 2.526 7.578 15.94 14.302
2 3.63 8.901 18.55 17.44 V. Conclusion
3 5.327 10.379 22.84 21.67 This paper presents a comparison between the results obtained
from the laboratory partial discharge model with the
Analyzing the experimental results as well as mathematical mode. The laboratory model for partial
mathematically calculated reveals that the inception voltage discharge was obtained by preparing a test cell representing a
(Vi) increases as the cavity depth (t’) is increased as shown partial discharge with a single cavity and then stressing it with
in Fig. 7.Also depicted from the results that the inception a high voltage to stimulate the partial discharge event. After
voltage (Vi) increases as the sample thickness (T) is increased conducting the research, it was found that partial discharge
which means dielectric is less prone to breakdown if the occurs in the form of discrete pulses. Additionally, it was
thickness of the sample is increased as shown in Fig 8. found that the size and dimensions of the air void or cavity
play a vital role for the occurrence of partial discharge, which
Comparing the conclusion with similar ones of other work affects the quality of insulation if sustained for a long duration.
proves their validity. In [33] artificial neural network was
used to find the influence of different parameters on the After analysis of the graphs plotted for various parameters it
PDIV. It was found that for a fixed sample thickness as void was observed that the inception voltage (Vi) increases as the
depth increases the PDIV increases, which agree with what cavity depth (t’) increases. This reveals that the dielectric
we obtained in figure 7. Additionally, they found that as medium with longer void size having its void length aligned
sample thickness increases the PDIV increases agreeing with within the direction of field, are less prone to breakdown as
figure 8. The maximum insulation thickness reached in [33] compared to those having comparatively smaller void length
was 3 mm while we started with 7.5 mm insulation thickness. keeping rest of the conditions same.,
However, the trend we found for the influence of the
thickness of the plate on the PDIV also matches the one found Further analysis on the inception voltage (Vi) increases and
in [34] for insulation thickness starting from 50 mm. The fact the sample thickness (T) increases. b of partial discharges in
that the results match two other conducted research, one with comparison with those having a comparatively smaller
lower insulation thickness range and one with a higher thickness with remaining of the conditions as intact.
thickness range proves their accuracy.
ACKNOWLEDGMENT
This publication was made possible by NPRP grant [10-
0101-170085] from the Qatar National Research Fund (a
member of Qatar Foundation). The statements made herein
are solely the responsibility of the authors.

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 05,2020 at 09:13:56 UTC from IEEE Xplore. Restrictions apply.
 REFERENCES [22] B. A. Lloyd S. R. Campbell, and G. C. Stone, “Continuous On-line
Partial Discharge Monitoring of Generator Stator Windings,” IEEE
[1] Waluyo, S.t. Saodah, and Eltha Hidayatullah, “Internal On-line Partial Transaction on Energy Conservation, Vol. 14, No. 4, pp. 1131-1138,
Discharge Analysis of 68.75 MVA Generator Stator Winding December 1999.
Insulation,” International Journal of Electrical and Computer
Engineering (IJECE) vol. 6, no. 5, pp. 2088-2095, 2016 [23] Mtronix: Advanced PD Analysis System MPD 540, “Product Brief and
Technical Specification,” Berlin,Germany, 2004.
[2] M. G. Danikas, “The Definitions Used for Partial Discharge
Phenomena”, Vol. 28, No. 6, pp. 1075-1081, 1993 [24] Marek Florkowski, Barbara Florkowska and Pawel Zydron,
“Observations of Echo Phenomena Acquired from Partial Discharge
[3] M. Hassan, M. Zuberi, F. I. Bakhsh, and E. Husain, “Analysis of partial Measurements in Voids,” IEEE International Conference on High
discharge in solid sheet insulation using CIGRE-II method,” IEEE Voltage Engineering and Application (ICHVE), pp. 1-4, 2016.
international conference on advances in power conversion and energy
technologies, (APCET); August 2-4, 2012. [25] J. S. Townsend, “Electrons in Gases,” Hutchinson’s London, 1947.
[26] M. U. Zuberi, A. Masood, E. Husain, A. Anwar. “Estimation of partial
[4] A.ElFaraskoury, O. Gouda “ Partial Discharge Measurements with
discharge inception voltages due to voids in solid sheet insulation,”
Internal Artificial Cavities Defects for Underground Cables,” Mepcon
IEEE Electrical Insulation Conference (EIC), pp. 124-128, June 2013.
2014.
[5] D. König and R. Y. Narayana, “Partial Discharges in Power [27] L. B. Loeb “Basic processes of gaseous electronics,” University Press,
Apparatus,” Berlin: Vde-Verlag Gmbh, 1993. 316p. Berkeley 1955.
[6] Y. Qiu, A. Sun, and Z. Zhang, “Factors Affecting Partial Discharges in [28] M. S. Bhalla and J.D. Craggs, “Measurement of ionization and
SF6 Gas Impregnated Polymer Film Insulation,” 2nd International attachment coefficients in sulphur hexafluoride in uniform fields,”
Conference on Properties and Applications of Dielectric Materials, Proc. Phy Soc. Vol. 80, pp. 151-160, 1962.
Beijing, China, September 12-16, 1988. [29] C. Raja Rao and G. R. Govinda Raju, “Growth of ionization currents
[7] A. H. Sari and F. Fadaee. “Effect of corona discharge on in dry air at high values of E/N,” J. Phy. D:App. Phy. Vol. 4, 494-503,
decontamination of Pseudomonas aeruginosa and E-coli,” Surface and 1971.
Coatings Technology, Vol. 205, Supplement 1, pp. S385-S390, 25 [30] T. W. Dakin, G. Luxa, G. Oppermann, J. Vigreux, G. Wind and H.
December 2010. Winkelnkemper, “Breakdown of gases in uniform fields-Paschen's
[8] M. S. Hapeez, A. F. Abidin, H. Hashim, N. R. Hamzah, M. K. Hamzah, curves for air, N2 and SF6,” Electra, No. 32, pp 61-82, 1974.
“Analysis and classification of different types of partial discharges by [31] H. C. Hall and R.M. Russek, “Discharge inception and extinction in
harmonic orders”, Elektronika ir Elektrotechnika, vol. 19, no. 9, pp. dielectric voids,” Proc. IEE, 101(II), p 47, 1954.
35–41, 2013. [Online]. Available: [32] E. Husain and R.S. Nema, “Analysis of Paschen curves for Air, N2 and
http://dx.doi.org/10.5755/j01.eee.19.9.2545 SF6 using the Townsend Breakdown Equation,” IEEE Transactions on
[9] E. Jahoda and J. Kúdelčík, “Internal partial discharge in cavity of Electrical Insulation Vol. EI-17 No.4, August 1982.
polyurethane“, Procedia Engineering, vol. 192, 2017, pp. 365-369, [33] N. P. Kolev and N. M. Chalashkanov, “delling of Partial Discharge
2017. Inception and Extinction Voltages Using Adaptive NeuroFuzzy
[10] F. Guastavino and A. Dardano, “Life tests on twisted pairs in presence Inference System (ANFIS),” 7 International Conference on Solid
of partial discharges: Influence of the voltage waveform”, IEEE Trans. Dielectrics, Winchester, vol. 8, no. 13, pp. 605-608, 2007
Dielectr. Electr. Insul., Vol. 19, No. 1, pp. 45-52, 2012. [34] Yamashita, T., Iwanaga, K., Furusato, T., Koreeda, H., Fujishima, T.
[11] F. Guastavino and A. Dardano, “Life tests on twisted pairs in presence and Sato, J. (2016). Improvement of insulation performance of
of partial discharges: Influence of the voltage waveform”, IEEE Trans. solid/gas composite insulation with embedded electrode. IEEE
Dielectr. Electr. Insul., Vol. 19, No. 1, pp. 45-52, 2012. Transactions on Dielectrics and Electrical Insulation, 23(2), pp.787-
[12] P. Cheetham, W. Kim, et al. “Use of Partial Discharge Inception 794.
Voltage Measurements to Design a Gaseous Helium Cooled High
Temperature Superconducting Power Cable”, IEEE Transactions on
Dielectrics and Electrical Insulation Vol. 24, No. 1, February 2017
[13] Rokunohe T, Kato T, Kojima H, Hayakawa N and H, Okubo,
“Calculation model for predicting partial-discharge inception voltage
in a non-uniform air gap while considering the effect of humidity,”
IEEE Trans. Dielectr. Electr. Insul., vol. 24, no. 2, pp. 1123–1130,
2017.
[14] M. Wild, S. Tenbohlen, E. Gulski, and R. Jongen, “Basic Aspects of
Partial Discharge On-site Testing of Long Length Transmission Power
Cables”, IEEE Transactions on Dielectrics and Electrical Insulation,
vol. 24, no. 2, pp. 1077-1087, 2017.
[15] Z, Cheng and et. al, “Partial Discharge Pattern Recognition of XLPE
Cable Based on Vector Quantization IEEE Transactions on Dielectrics
and Electrical Insulation, vol. 55, no. 6, pp. 7202704, 2019.
[16] M. Mahdipor, A. Akbari, and P. Werle, “ Charge Concept in Partial
Discharge in Power Cables,” IEEE Transactions on Dielectrics and
Electrical Insulation, vol. 24, no. 2, pp. 817-825, 2017.
[17] L A.ElFaraskoury, O. Gouda “ Partial Discharge Measurements with
Internal Artificial Cavities Defects for Underground Cables,” Mepcon
2014.
[18] T. Wakimoto, H. Kojima, N Hayakawa, “Measurement and evaluation
of partial discharge inception voltage for enameled rectangular wires
under AC voltage,” I EEE Transactions on Dielectrics and Electrical
Insulation, Vol 23 no 6, 2016.
[19] D. Kind and D. Konig, “AC breakdown of epoxy resins by partial
discharges in voids,” IEEE Transactions on Electrical Insulation, vol.
EI-3, Issue. 2, pp. 40-46, May 1968.
[20] W. R. Kodoll, H. C. Karner, and T. Tanaka, “Internal partial discharge
resistivity testing,” Partial dircharge measurements 1981, 7 IEC
Publication 270, pp. 11504/1-CIGRE, 1988.
[21] P. Pakonen and V. Latva-Pukkila, “On-line Partial Discharge
Measurements on Covered Conductor Lines,” Nordic and Baltic
Workshop on Power Systems, Tampere, Finland, February 4-5, 2002.

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 05,2020 at 09:13:56 UTC from IEEE Xplore. Restrictions apply.