- Email: [email protected]
Electrical field distribution on the cross-linked polyethylene insulation surface under partial discharge testing
Polymer Testing 82 (2020) 106311
Contents lists available at ScienceDirect
Polymer Testing journal homepage: http://www.elsevier.com/locate/polytest
Electrical field distribution on the cross-linked polyethylene insulation surface under partial discharge testing Sayidul Morsalin *, Toan Phung School of Electrical Engineering and Telecommunications, University of New South Wales, NSW, 2052, Australia
A R T I C L E I N F O
A B S T R A C T
Keywords: Electric field distribution XLPE material Partial discharge Very low frequency testing
Cross-linked polyethylene (XLPE) insulated power cables are widely used in transmission and distribution net works. The enhanced localised electrical field/stress can result in surface discharges (SD), i.e. partial discharges on the surface of the dielectric. To measure discharges, very low frequency (VLF-0.1 Hz or lower) excitation has emerged as an attractive alternative to the conventional power frequency (PF- 50/60 Hz) testing as it signifi cantly reduces the required reactive power from the test supply. As the discharge process mainly resulted from the enhanced electric stress; it is necessary to understand how the electric field distributes on the XLPE dielectric surface when it is exposed to a high voltage AC excitation. In this study, the distribution of surface electric field before, during and after a PD event at VLF and PF test voltage is investigated. Finite element analysis (FEA) based numerical simulation shows that the space charge dynamics cause the differences in the field distribution and supports the experimental results.
1. Introduction Cross-linked polyethylene (XLPE) is a common primary insulation in the extruded medium and high voltage (HV) power cables [1,2]. XLPE has excellent dielectric properties; however, the enhanced localised electrical field/stress during the operation may degrade the insulation surface, resulting in partial discharge, especially in the cable termina tions and joints [1,3,4]. For diagnostic purpose, surface discharge (SD) measurement at power frequency (PF) 50/60 Hz voltage has been well explored. On the other hand, the modelling of the electric field distri bution due to partial discharge is lacking [2,5,6], especially for the very low frequency, VLF (typically 0.1 Hz or lower) test voltage. As a diag nostic method, VLF testing offers the advantage of reducing the requirement of reactive power from the test supply [7]. A surface discharge may occur when the localised electric field along the insulation surface exceeds the certain stress level, known as incep tion field, and the corresponding voltage is the inception voltage (SDIV). Depending on the free electrons availability, the SD process usually originated from a triple point junction, which is the interface between the HV electrode edge and the solid/gaseous dielectrics [2,8]. After any SD event, a considerable amount of space charges (usually positive ions) is generated, can decay in the insulation surface via various physical processes (e.g. charge drift recombination, conduction etc.) and
depending on the applied voltage frequency, significantly influence the local surface field [2,9]. Therefore, in this study, the change of surface field distribution due to space charge dynamics at both VLF and PF test voltage is presented. Based on the sample geometry, a model using finite element analysis (FEA) method is developed in Comsol Multiphysics and integrated with MATLAB to determine the electric field distribution along the insulation surface before, after or during a discharge occur rence. The developed model shows that the local surface field drops after a PD event, but the drop is relatively higher at the power frequency. The simulated model also can effectively reproduce the discharge sequences and agrees with the experimental results. 2. Model geometry 2.1. Specimen preparation XLPE material is usually prepared by crosslinking the base polymer, low-density polyethylene (LDPE), with a commercial peroxide (e.g. Dicumyl peroxide -DCP) [1,10]. For this purpose, an LDPE pellet with a melt flow index 1.8 g/10 min and a density of 0.922 was compressed by using a hot-pressure machine at 110 � C and a pressure of 15 MPa. After that the crosslinking was performed for a duration of 25 min under the same pressure at 180 � C. From the crosslinking, the final XLPE test
* Corresponding author. E-mail addresses: [email protected] (S. Morsalin), [email protected] (T. Phung). https://doi.org/10.1016/j.polymertesting.2019.106311 Received 30 September 2019; Received in revised form 8 December 2019; Accepted 23 December 2019 Available online 24 December 2019 0142-9418/© 2019 Elsevier Ltd. All rights reserved.
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
� � ∂ r � σ rV þ ε0 εr ðrVÞ ¼ 0 ∂t
(2)
where, σ is the conductivity of the material, V is the electric potential, and εr is the relative permittivity. 2.3. Experimental setup The experimental setup including the arrangement of test-cell is shown in Fig. 2. As can be seen, a commercial Mtronix MPD 600 system was employed to detect PDs in real time from the test specimen. When the surface discharge occurs, the coupling capacitor transfers charge to it to compensate for the momentary voltage collapse along the XLPE surface. On the other hand, the quadripole unit acts as the measuring impedance where both PD and voltage signal can be extracted. The captured PD signals in the digitized form are transferred to a personal computer so that data can be saved and stored for future analysis.
Fig. 1. The axis-symmetric two-dimensional view of the model.
3. Electric field distribution Fig. 3 and Fig. 4 illustrate the electrical field distribution on the sample surface before and after a discharge event at 50 Hz and 0.1 Hz excitation respectively. Table 1 lists the global parameters that have been used in the numerical simulation to model the physical behaviours. At higher frequencies, the space charges get relatively lesser time to decay because of their slowly mobile nature and can be found available during the moment when the voltage waveform changes its polarity. These free charges produce an electric field in the same direction to the external field and thus increase the local surface field. With the power frequency excitation, the chance of getting initial free electrons also increases since the time interval between two consecutive discharges at higher frequencies is very small and provides the enhanced availability of free electrons to incept a discharge. Therefore, the total electron generation rate (EGR) due to surface emission should have a higher value at 50 Hz. On the other hand, at very low frequency (0.1 Hz) an opposite scenario is likely to be observed and because of the longer time period, the free charges are assumed to cease entirely after a discharge event. As can be observed in Figs. 3 and 4, the critical zone has the highest electric field stress and comparing with VLF, the first PD is incepted at the higher magnitude under 50 Hz applied voltage. Basically, the critical zone is the region where the surface discharge mostly occurs. In either case, the electric field drops after a discharge event. To show how the electric field distributes along the r-axis, Fig. 5 il lustrates the radial (Er) and vertical component (Ez) of the field magni tude on the insulation surface. Here, the radial field Er before and after a PD event is shown in Fig. 5a while the vertical field Ez is in Fig. 5b. As can be seen, the field distribution is not uniform and decreases more or less exponentially with the increasing distance. Before the first PD occurrence, the highest radial stress Er, can be observed in the critical zone, in particular, a short distance after the triple point (i.e. 2.5 mm here). Once PDs occur, the free charge movement across the erosion surface increases, resulting in a significant decrement of the local surface field as shown in Fig. 5. Accumulation of these free charges on the insulation surface influences the occurrence and magnitude of subse quent PDs.
Fig. 2. Block diagram of the experimental setup.
specimen was obtained which is in the disked shape with a dimension of 35 mm diameter and 4 mm thickness. All the chemicals used in the sample preparation were collected from a local manufacturer and used without any further purification. 2.2. Physical model for electric field estimation A physical model as a representation of the test arrangement was developed in an axial symmetric two-dimensional geometry as depicted in Fig. 1. The model consists of a semi-sphere and plane electrode arrangement and a test sample which is placed between these two electrodes. The semi-sphere electrode is a steel high voltage electrode having 7 mm cross-sectional diameter, and the lower curvature diameter that touches the test object is of 5 mm. On the other hand, the plane electrode is at the ground potential. The whole geometry was simulated in the air medium at 6 kV of the applied voltage. The triple point junction and the critical zone are also provided to simulate the surface discharge. An erosion area of 0.5 mm thickness is introduced in the test sample to represent the surface emission where the detrapping/trapping of free charges (space ions and free electrons) occurs and also, the space charges decay through conduction and drift-recombination. The electric field distribution is calculated using some partial dif ferential equations and expressed as !
ε0 εr r � E ¼ ε0 εr r � ð rVÞ ¼ 0
4. Discharge sequence The rate of voltage change (dv/dt) is proportional to the applied voltage frequency. Hence, the power frequency excitation has a higher dv/dt which allows space charges less time to be decayed through the insulation surface. As a result, the surface conductivity of space charge or the space charge decay between two consecutive PD events at power
(1) 2
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
Fig. 3. Simulation of electric field distribution at 50 Hz; (a) 3D view before (left) and after (right) a PD event (b) 2D view before (left) and after (right) a PD event.
3
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
Fig. 4. Simulation of electric field distribution at 0.1 Hz; (a) 3D view before (left) and after (right) a PD event (b) 2D view before (left) and after (right) a PD event.
4
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
Fig. 5. Field distribution along the insulation surface; (a) Along the r-axis (b) Along the z-axis.
Fig. 6. Simulation of surface electric field at 50 Hz.
Fig. 7. Simulation of surface electric field at 0.1 Hz.
frequency is much smaller as compared to the VLF excitation. Based on this principle, the electric field distributions during a PD event for seven consecutive AC cycles under PF and VLF excitation are illustrated in Fig. 6 and Fig. 7 respectively. As discussed earlier, the surface electric field (Esur) is the result of the
external field that comes from the external supply (fmE0) and the field due to space charge effects (Eq). Initially, the space charge field Eq is equal to the external field due to the absence of surface charge deposited from PD events. When the first PD occurs, Esur drops steadily, which causes an increase of Eq in the opposite direction of the surface field. This 5
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
Fig. 8. Experimental measurements of discharge magnitude and number (a) At 50 Hz (b) At 0.1 Hz.
can be seen at the first half-cycle of any voltage excitation (position A). When the polarity of Esur is the same with Eq, then after a discharge event space charges start to cease through conduction, resulting in the increase of surface conductivity and thus, reducing the space charge field (posi tion B). However, when Esur is the opposite polarity of Eq, the charge accumulation along the surface increases and generates larger space charge field magnitude (position C). In the case of 50 Hz, due to the higher dv/dt, the time interval be tween two consecutive PD events is very short. This short duration al lows a higher availability of free charges (free electrons and space ions) to be present in the erosion surface. Due to the lower statistical time, the next PD event is likely to occur almost immediately after the moment when Esur exceeds Ein. On the other hand, due to the lower dv/dt at VLF, the surface charge decay is more significant. Therefore, after a PD event, the space charge and free electrons get enough time to vanish, which in turn results in less PD occurrence. In other words, the space charge field is less dominant at very low frequency as the slowr rate of voltage change at VLF enables majority of free charges to dissipate along the insulation surface.
6. Conclusion A physical model describing the electric field distribution on the surface of XLPE material due to partial discharge is presented. Using finite element analysis (FEA) the discharge behaviours before, during and after a PD occurrence at very low and power frequency are simu lated. The simulated model shows that the increasing field distribution due to the space charges at power frequency causes the differences in PD magnitudes and discharge occurrence as compared to the very low fre quency. The simulated model also supports the discharge occurrence patterns that are obtained from the experimental study. Declaration of competing interestCOI The authors declare that there is no conflict of interest. CRediT authorship contribution statement Sayidul Morsalin: Conceptualization, Methodology, Software, Data curation, Investigation, Writing - original draft. Toan Phung: Visuali zation, Investigation, Supervision, Writing - review & editing.
5. Experimental results The experimental measurement results with this particular test setup at 50 Hz and VLF are shown in Fig. 8a and Fig. 8b respectively. For both frequency levels, surface discharges are periodically measured at the last 10 min of every hour, and the acquisition considered here is for a total ageing time of 6 h. The discharge occurrences are shown here with the phase-resolved PD pattern of an AC cycle. As can be observed, the PD activities occur during both positive and negative half-cycles of the applied voltage; nevertheless, PD occurrence is less in positive half-cycle as compared to the negative half-cycle at both VLF and 50 Hz supply. When comparing with VLF, as expected, the local enhancement of field stress due to free charges causes relatively larger PD magnitudes and the higher number of surface discharges at the power frequency.
Acknowledgements This work was supported in part by the Australian Research Council (ARC) under Grant LP140100782. The authors would like to acknowl edge Mr. Zhenyu Liu for experimental assistance. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.polymertesting.2019.106311. References [1] C. Kim, P. Jiang, F. Liu, S. Hyon, M. Ri, Y. Yu, M. Ho, Investigation on dielectric breakdown behavior of thermally aged cross-linked polyethylene cable insulation, Polym. Test. (2019), 106045, https://doi.org/10.1016/j. polymertesting.2019.106045. [2] S.Z. Dabbak, H.A. Illias, B.C. Ang, Effect of surface discharges on different polymer dielectric materials under high field stress, IEEE Trans. Dielectr. Electr. Insul. 24 (2017) 3758–3765, https://doi.org/10.1109/TDEI.2017.006418. [3] F. Negri, A. Cavallini, D. Fabiani, A. Saccani, M. Toselli, Effect of graphene oxidebased nanostructured coatings on the electrical performance of cross-linked polyethylene, Polym. Test. 59 (2017) 142–151, https://doi.org/10.1016/j. polymertesting.2017.01.021. [4] Y. Wang, C. Wang, K. Xiao, Investigation of the electrical properties of XLPE/SiC nanocomposites, Polym. Test. 50 (2016) 145–151, https://doi.org/10.1016/j. polymertesting.2016.01.007. [5] A. Cavallini, G.C. Montanari, Effect of supply voltage frequency on testing of insulation system, IEEE Trans. Dielectr. Electr. Insul. 13 (2006) 111–121, https:// doi.org/10.1109/TDEI.2006.1593409.
Table 1 Descriptions of global parameters used in the simulation. Description
Relative Permittivity, εr
Air (Domain 1) HV electrode (Domain 2) XLPE dielectric (Domain 3) Erosion surface (Domain 4) Critical zone (Domain 5) EGR at 50 Hz EGR at 0.1 Hz Inception field Einc
1 0 1 4.03 � 106 S/m 2.4 0 2.4 varies 2.4 5 � 10 3 S/m 50/s during PHC and 100/s during NHC 0.05/s during PHC and 0.1/s during NHC 1.25 kV/mm at PF; 5.6 kV/mm at VLF
Conductivity, σ
*PHC-positive half-cycle; NHC-Negative half-cycle. 6
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
[6] H.A. Illias, G. Chen, P.L. Lewin, Modelling of cycle to cycle behaviour for partial discharge events within a spherical cavity in a solid dielectric material by using Finite Element Analysis, 10th IEEE Int. Conf. Solid Dielectr. (2010) 1–4, https:// doi.org/10.1109/ICSD.2010.5567996. [7] M.T. Peschel, Needed changes in medium voltage cable testing. Were you in on it? Welcome to the world of VLF, IEEE Int. Symp. Electr. Insul. (2010) 1–5, https:// doi.org/10.1109/ELINSL.2010.5549515. [8] T. Liu, Q. Li, X. Huang, Y. Lu, M. Asif, Z. Wang, Partial discharge behavior and ground insulation life expectancy under different voltage frequencies, IEEE Trans.
Dielectr. Electr. Insul. 25 (2018) 603–613, https://doi.org/10.1109/ TDEI.2018.006810. [9] C. Nyamupangedengu, I.R. Jandrell, Partial discharge spectral response to variations in the supply voltage frequency, IEEE Trans. Dielectr. Electr. Insul. 19 (2012) 521–532, https://doi.org/10.1109/TDEI.2012.6180246. [10] C. Zhang, C. Li, L. Nie, Z. Jing, B. Han, Research on the water blade electrode method for assessing water tree resistance of cross-linked polyethylene, Polym. Test. 50 (2016) 235–240, https://doi.org/10.1016/j.polymertesting.2016.01.017.
7
Contents lists available at ScienceDirect
Polymer Testing journal homepage: http://www.elsevier.com/locate/polytest
Electrical field distribution on the cross-linked polyethylene insulation surface under partial discharge testing Sayidul Morsalin *, Toan Phung School of Electrical Engineering and Telecommunications, University of New South Wales, NSW, 2052, Australia
A R T I C L E I N F O
A B S T R A C T
Keywords: Electric field distribution XLPE material Partial discharge Very low frequency testing
Cross-linked polyethylene (XLPE) insulated power cables are widely used in transmission and distribution net works. The enhanced localised electrical field/stress can result in surface discharges (SD), i.e. partial discharges on the surface of the dielectric. To measure discharges, very low frequency (VLF-0.1 Hz or lower) excitation has emerged as an attractive alternative to the conventional power frequency (PF- 50/60 Hz) testing as it signifi cantly reduces the required reactive power from the test supply. As the discharge process mainly resulted from the enhanced electric stress; it is necessary to understand how the electric field distributes on the XLPE dielectric surface when it is exposed to a high voltage AC excitation. In this study, the distribution of surface electric field before, during and after a PD event at VLF and PF test voltage is investigated. Finite element analysis (FEA) based numerical simulation shows that the space charge dynamics cause the differences in the field distribution and supports the experimental results.
1. Introduction Cross-linked polyethylene (XLPE) is a common primary insulation in the extruded medium and high voltage (HV) power cables [1,2]. XLPE has excellent dielectric properties; however, the enhanced localised electrical field/stress during the operation may degrade the insulation surface, resulting in partial discharge, especially in the cable termina tions and joints [1,3,4]. For diagnostic purpose, surface discharge (SD) measurement at power frequency (PF) 50/60 Hz voltage has been well explored. On the other hand, the modelling of the electric field distri bution due to partial discharge is lacking [2,5,6], especially for the very low frequency, VLF (typically 0.1 Hz or lower) test voltage. As a diag nostic method, VLF testing offers the advantage of reducing the requirement of reactive power from the test supply [7]. A surface discharge may occur when the localised electric field along the insulation surface exceeds the certain stress level, known as incep tion field, and the corresponding voltage is the inception voltage (SDIV). Depending on the free electrons availability, the SD process usually originated from a triple point junction, which is the interface between the HV electrode edge and the solid/gaseous dielectrics [2,8]. After any SD event, a considerable amount of space charges (usually positive ions) is generated, can decay in the insulation surface via various physical processes (e.g. charge drift recombination, conduction etc.) and
depending on the applied voltage frequency, significantly influence the local surface field [2,9]. Therefore, in this study, the change of surface field distribution due to space charge dynamics at both VLF and PF test voltage is presented. Based on the sample geometry, a model using finite element analysis (FEA) method is developed in Comsol Multiphysics and integrated with MATLAB to determine the electric field distribution along the insulation surface before, after or during a discharge occur rence. The developed model shows that the local surface field drops after a PD event, but the drop is relatively higher at the power frequency. The simulated model also can effectively reproduce the discharge sequences and agrees with the experimental results. 2. Model geometry 2.1. Specimen preparation XLPE material is usually prepared by crosslinking the base polymer, low-density polyethylene (LDPE), with a commercial peroxide (e.g. Dicumyl peroxide -DCP) [1,10]. For this purpose, an LDPE pellet with a melt flow index 1.8 g/10 min and a density of 0.922 was compressed by using a hot-pressure machine at 110 � C and a pressure of 15 MPa. After that the crosslinking was performed for a duration of 25 min under the same pressure at 180 � C. From the crosslinking, the final XLPE test
* Corresponding author. E-mail addresses: [email protected] (S. Morsalin), [email protected] (T. Phung). https://doi.org/10.1016/j.polymertesting.2019.106311 Received 30 September 2019; Received in revised form 8 December 2019; Accepted 23 December 2019 Available online 24 December 2019 0142-9418/© 2019 Elsevier Ltd. All rights reserved.
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
� � ∂ r � σ rV þ ε0 εr ðrVÞ ¼ 0 ∂t
(2)
where, σ is the conductivity of the material, V is the electric potential, and εr is the relative permittivity. 2.3. Experimental setup The experimental setup including the arrangement of test-cell is shown in Fig. 2. As can be seen, a commercial Mtronix MPD 600 system was employed to detect PDs in real time from the test specimen. When the surface discharge occurs, the coupling capacitor transfers charge to it to compensate for the momentary voltage collapse along the XLPE surface. On the other hand, the quadripole unit acts as the measuring impedance where both PD and voltage signal can be extracted. The captured PD signals in the digitized form are transferred to a personal computer so that data can be saved and stored for future analysis.
Fig. 1. The axis-symmetric two-dimensional view of the model.
3. Electric field distribution Fig. 3 and Fig. 4 illustrate the electrical field distribution on the sample surface before and after a discharge event at 50 Hz and 0.1 Hz excitation respectively. Table 1 lists the global parameters that have been used in the numerical simulation to model the physical behaviours. At higher frequencies, the space charges get relatively lesser time to decay because of their slowly mobile nature and can be found available during the moment when the voltage waveform changes its polarity. These free charges produce an electric field in the same direction to the external field and thus increase the local surface field. With the power frequency excitation, the chance of getting initial free electrons also increases since the time interval between two consecutive discharges at higher frequencies is very small and provides the enhanced availability of free electrons to incept a discharge. Therefore, the total electron generation rate (EGR) due to surface emission should have a higher value at 50 Hz. On the other hand, at very low frequency (0.1 Hz) an opposite scenario is likely to be observed and because of the longer time period, the free charges are assumed to cease entirely after a discharge event. As can be observed in Figs. 3 and 4, the critical zone has the highest electric field stress and comparing with VLF, the first PD is incepted at the higher magnitude under 50 Hz applied voltage. Basically, the critical zone is the region where the surface discharge mostly occurs. In either case, the electric field drops after a discharge event. To show how the electric field distributes along the r-axis, Fig. 5 il lustrates the radial (Er) and vertical component (Ez) of the field magni tude on the insulation surface. Here, the radial field Er before and after a PD event is shown in Fig. 5a while the vertical field Ez is in Fig. 5b. As can be seen, the field distribution is not uniform and decreases more or less exponentially with the increasing distance. Before the first PD occurrence, the highest radial stress Er, can be observed in the critical zone, in particular, a short distance after the triple point (i.e. 2.5 mm here). Once PDs occur, the free charge movement across the erosion surface increases, resulting in a significant decrement of the local surface field as shown in Fig. 5. Accumulation of these free charges on the insulation surface influences the occurrence and magnitude of subse quent PDs.
Fig. 2. Block diagram of the experimental setup.
specimen was obtained which is in the disked shape with a dimension of 35 mm diameter and 4 mm thickness. All the chemicals used in the sample preparation were collected from a local manufacturer and used without any further purification. 2.2. Physical model for electric field estimation A physical model as a representation of the test arrangement was developed in an axial symmetric two-dimensional geometry as depicted in Fig. 1. The model consists of a semi-sphere and plane electrode arrangement and a test sample which is placed between these two electrodes. The semi-sphere electrode is a steel high voltage electrode having 7 mm cross-sectional diameter, and the lower curvature diameter that touches the test object is of 5 mm. On the other hand, the plane electrode is at the ground potential. The whole geometry was simulated in the air medium at 6 kV of the applied voltage. The triple point junction and the critical zone are also provided to simulate the surface discharge. An erosion area of 0.5 mm thickness is introduced in the test sample to represent the surface emission where the detrapping/trapping of free charges (space ions and free electrons) occurs and also, the space charges decay through conduction and drift-recombination. The electric field distribution is calculated using some partial dif ferential equations and expressed as !
ε0 εr r � E ¼ ε0 εr r � ð rVÞ ¼ 0
4. Discharge sequence The rate of voltage change (dv/dt) is proportional to the applied voltage frequency. Hence, the power frequency excitation has a higher dv/dt which allows space charges less time to be decayed through the insulation surface. As a result, the surface conductivity of space charge or the space charge decay between two consecutive PD events at power
(1) 2
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
Fig. 3. Simulation of electric field distribution at 50 Hz; (a) 3D view before (left) and after (right) a PD event (b) 2D view before (left) and after (right) a PD event.
3
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
Fig. 4. Simulation of electric field distribution at 0.1 Hz; (a) 3D view before (left) and after (right) a PD event (b) 2D view before (left) and after (right) a PD event.
4
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
Fig. 5. Field distribution along the insulation surface; (a) Along the r-axis (b) Along the z-axis.
Fig. 6. Simulation of surface electric field at 50 Hz.
Fig. 7. Simulation of surface electric field at 0.1 Hz.
frequency is much smaller as compared to the VLF excitation. Based on this principle, the electric field distributions during a PD event for seven consecutive AC cycles under PF and VLF excitation are illustrated in Fig. 6 and Fig. 7 respectively. As discussed earlier, the surface electric field (Esur) is the result of the
external field that comes from the external supply (fmE0) and the field due to space charge effects (Eq). Initially, the space charge field Eq is equal to the external field due to the absence of surface charge deposited from PD events. When the first PD occurs, Esur drops steadily, which causes an increase of Eq in the opposite direction of the surface field. This 5
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
Fig. 8. Experimental measurements of discharge magnitude and number (a) At 50 Hz (b) At 0.1 Hz.
can be seen at the first half-cycle of any voltage excitation (position A). When the polarity of Esur is the same with Eq, then after a discharge event space charges start to cease through conduction, resulting in the increase of surface conductivity and thus, reducing the space charge field (posi tion B). However, when Esur is the opposite polarity of Eq, the charge accumulation along the surface increases and generates larger space charge field magnitude (position C). In the case of 50 Hz, due to the higher dv/dt, the time interval be tween two consecutive PD events is very short. This short duration al lows a higher availability of free charges (free electrons and space ions) to be present in the erosion surface. Due to the lower statistical time, the next PD event is likely to occur almost immediately after the moment when Esur exceeds Ein. On the other hand, due to the lower dv/dt at VLF, the surface charge decay is more significant. Therefore, after a PD event, the space charge and free electrons get enough time to vanish, which in turn results in less PD occurrence. In other words, the space charge field is less dominant at very low frequency as the slowr rate of voltage change at VLF enables majority of free charges to dissipate along the insulation surface.
6. Conclusion A physical model describing the electric field distribution on the surface of XLPE material due to partial discharge is presented. Using finite element analysis (FEA) the discharge behaviours before, during and after a PD occurrence at very low and power frequency are simu lated. The simulated model shows that the increasing field distribution due to the space charges at power frequency causes the differences in PD magnitudes and discharge occurrence as compared to the very low fre quency. The simulated model also supports the discharge occurrence patterns that are obtained from the experimental study. Declaration of competing interestCOI The authors declare that there is no conflict of interest. CRediT authorship contribution statement Sayidul Morsalin: Conceptualization, Methodology, Software, Data curation, Investigation, Writing - original draft. Toan Phung: Visuali zation, Investigation, Supervision, Writing - review & editing.
5. Experimental results The experimental measurement results with this particular test setup at 50 Hz and VLF are shown in Fig. 8a and Fig. 8b respectively. For both frequency levels, surface discharges are periodically measured at the last 10 min of every hour, and the acquisition considered here is for a total ageing time of 6 h. The discharge occurrences are shown here with the phase-resolved PD pattern of an AC cycle. As can be observed, the PD activities occur during both positive and negative half-cycles of the applied voltage; nevertheless, PD occurrence is less in positive half-cycle as compared to the negative half-cycle at both VLF and 50 Hz supply. When comparing with VLF, as expected, the local enhancement of field stress due to free charges causes relatively larger PD magnitudes and the higher number of surface discharges at the power frequency.
Acknowledgements This work was supported in part by the Australian Research Council (ARC) under Grant LP140100782. The authors would like to acknowl edge Mr. Zhenyu Liu for experimental assistance. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.polymertesting.2019.106311. References [1] C. Kim, P. Jiang, F. Liu, S. Hyon, M. Ri, Y. Yu, M. Ho, Investigation on dielectric breakdown behavior of thermally aged cross-linked polyethylene cable insulation, Polym. Test. (2019), 106045, https://doi.org/10.1016/j. polymertesting.2019.106045. [2] S.Z. Dabbak, H.A. Illias, B.C. Ang, Effect of surface discharges on different polymer dielectric materials under high field stress, IEEE Trans. Dielectr. Electr. Insul. 24 (2017) 3758–3765, https://doi.org/10.1109/TDEI.2017.006418. [3] F. Negri, A. Cavallini, D. Fabiani, A. Saccani, M. Toselli, Effect of graphene oxidebased nanostructured coatings on the electrical performance of cross-linked polyethylene, Polym. Test. 59 (2017) 142–151, https://doi.org/10.1016/j. polymertesting.2017.01.021. [4] Y. Wang, C. Wang, K. Xiao, Investigation of the electrical properties of XLPE/SiC nanocomposites, Polym. Test. 50 (2016) 145–151, https://doi.org/10.1016/j. polymertesting.2016.01.007. [5] A. Cavallini, G.C. Montanari, Effect of supply voltage frequency on testing of insulation system, IEEE Trans. Dielectr. Electr. Insul. 13 (2006) 111–121, https:// doi.org/10.1109/TDEI.2006.1593409.
Table 1 Descriptions of global parameters used in the simulation. Description
Relative Permittivity, εr
Air (Domain 1) HV electrode (Domain 2) XLPE dielectric (Domain 3) Erosion surface (Domain 4) Critical zone (Domain 5) EGR at 50 Hz EGR at 0.1 Hz Inception field Einc
1 0 1 4.03 � 106 S/m 2.4 0 2.4 varies 2.4 5 � 10 3 S/m 50/s during PHC and 100/s during NHC 0.05/s during PHC and 0.1/s during NHC 1.25 kV/mm at PF; 5.6 kV/mm at VLF
Conductivity, σ
*PHC-positive half-cycle; NHC-Negative half-cycle. 6
S. Morsalin and T. Phung
Polymer Testing 82 (2020) 106311
[6] H.A. Illias, G. Chen, P.L. Lewin, Modelling of cycle to cycle behaviour for partial discharge events within a spherical cavity in a solid dielectric material by using Finite Element Analysis, 10th IEEE Int. Conf. Solid Dielectr. (2010) 1–4, https:// doi.org/10.1109/ICSD.2010.5567996. [7] M.T. Peschel, Needed changes in medium voltage cable testing. Were you in on it? Welcome to the world of VLF, IEEE Int. Symp. Electr. Insul. (2010) 1–5, https:// doi.org/10.1109/ELINSL.2010.5549515. [8] T. Liu, Q. Li, X. Huang, Y. Lu, M. Asif, Z. Wang, Partial discharge behavior and ground insulation life expectancy under different voltage frequencies, IEEE Trans.
Dielectr. Electr. Insul. 25 (2018) 603–613, https://doi.org/10.1109/ TDEI.2018.006810. [9] C. Nyamupangedengu, I.R. Jandrell, Partial discharge spectral response to variations in the supply voltage frequency, IEEE Trans. Dielectr. Electr. Insul. 19 (2012) 521–532, https://doi.org/10.1109/TDEI.2012.6180246. [10] C. Zhang, C. Li, L. Nie, Z. Jing, B. Han, Research on the water blade electrode method for assessing water tree resistance of cross-linked polyethylene, Polym. Test. 50 (2016) 235–240, https://doi.org/10.1016/j.polymertesting.2016.01.017.
7