Change of Dielectric Parameters of XLPE Cable due to Thermal Aging Martin GERMAN-SOBEK, Roman CIMBALA, Jozef KIRÁLY∗ Abstract The change of dielectric parameters points to the aging or defect of HV equipment insulation. Cross-linked polyethylene (XLPE) is the globally preferred modern insulation for power cables, both for distribution and transmission system applications. Many studies and experiments show an influence of aging to degradation of XLPE insulation. It is subsequently reflected by changing of dielectric parameters of insulation. The measurement of dielectric parameters of XLPE cable sample was carried out by the method of dielectric relaxation spectroscopy (DRS) in frequency domain. DRS represents method suitable for measuring the dielectric parameters of polymeric composites during aging. The results are compared and it was observed a change of parameters in consequence of additional aging. Keywords: insulation, XLPE cable, aging of XLPE, capacitance, dissipation factor, thermal aging, dielectric relaxation spectroscopy 1. Introduction The insulation system is an important and sensitive part of any high-voltage electric equipment. Power cables are very important and sensitive devices in the power system, and they play an important role in the safety of the power load and reliable transmission of electricity. Nowadays, the majority of power cables are insulated with polymeric materials. Cross-linked polyethylene (XLPE), as the main polymeric insulation, is widely used as electric insulation material for high-voltage distribution power cables. This insulation material provides excellent physical, chemical and electric properties. During operation, the power cables may be exposed to high currents and voltages and they are critical parts of the transmission infrastructure. Therefore, is ∗ Martin, GERMAN-SOBEK: Eng., Technical University of Košice, Faculty of Electrical Engineering and Informatics, Department of Electric Power Engineering, Mäsiarska 74, 04120 Košice, Slovak Republic; martin.germansobek@tuke.sk Roman CIMBALA: Univ.Prof., Technical University of Košice, Faculty of Electrical Engineering and Informatics, Department of Electric Power Engineering, Mäsiarska 74, 04120 Košice, Slovak Republic; roman.cimbala@tuke.sk Jozef KIRÁLY: Eng., Technical University of Košice, Faculty of Electrical Engineering and Informatics, Department of Electric Power Engineering, Mäsiarska 74, 04120 Košice, Slovak Republic; jozef.kiraly@tuke.sk expected their high resistance against possible failures [1]. Damage of insulation can lead to equipment failure and other disorders. The insulation degradation is inevitable during the operation and the failure rate of XLPE cable increases with the service time [2]. The cables are permanently exposed to thermal ageing during operation. It may cause change in dielectric parameters of cables and also irreversible damage of cable insulation. The primary initiators for the degradation of cable insulation are high currents, voltage, mechanical, chemical and thermal stress and pollution of environment. Interaction various factors together may significantly speed up the degradation processes. Process of aging of insulation is the most acting on the parameters and quality of insulation and it is a phenomenon that is essentially cannot be affected. In most cases, failures in the insulation are related to integral degradation of the insulation like water treeing in XLPE cables. It is also known, that the insulation failures may be caused by lower dielectric strength due to aging processes or by internal defects in the insulation. [1] It is necessary to assess degradation and insulating state of cables. 48 ELECTROTEHNICĂ, ELECTRONICĂ, AUTOMATICĂ,vol. 62 (2014), Nr. 3 2. About aging of XLPE insulation Polyethylene (PE) is a long chain semi crystalline polymer manufactured through the polymerization of ethylene gas. PE was very popular, compared to paper insulation, as insulation for cables because of its low cost, good electric properties (low dielectric constant, low dielectric loss, and high breakdown strength), processability, mechanical toughness and flexibility, good resistance to chemicals and moisture. [3] XLPE is a material produced by the compounding of low density polyethylene (LDPE) with a crosslinking agent such as dicumyl peroxide. XLPE has good dielectric properties for high voltage applications. However, aging of XLPE material cannot avoidable after long time in operation under various stress conditions. XLPE insulated cables for high voltage applications have been studied and investigated in order to evaluate a function of service stresses and aging time. Many researchers are studied improved XLPE properties in order to improve dielectric performance of XLPE material. [4, 5] During the production of XLPE, this is molten and cooled serveral times. The material properties are influenced by the manufacturing process, heating and cooling temperatures and time. During the cable production the material is also heated and cooled several times during the different steps. The cooling process influences significantly the morphology and thus the electric properties of XLPE cable. The electric breakdown strength is depending on the density and it is also depending on the increase of the amorphous structures inside the polyethylene. [6] XLPE is composed of crystalline phase and amorphous phase. Defects such as submicrovoids and microvoids in XLPE may be formed and developed at the interface of crystalline and amorphous area, which can be regarded as weak point of insulation. [7]. Defects in XLPE may be introduced in the process of transporting or laying, although cable manufacturing is improved. In practical service condition, electric field deformations around these defects are very sharp and exert negative effects on the XLPE insulation breakdown. The breakdown of XLPE is closely related to electric trees and partial discharges and it was found that the breakdown voltage decreased with an increase in frequency. [8] In general, during the operation an insulation system is subjected to one or more stress that causes irreversible changes of insulating material properties with time. This process is called aging and ends when the insulation is no more able to withstand the applied stress. The relevant time is the time-to-failure or time-to-breakdown, alternatively called insulation life time. In the case of electric insulation, the stresses most commonly applied in operation are electric field (due to voltage) and temperature (due to loss), but also other stresses, such as mechanical stresses (vibration) and environmental stresses (pollution, humidity) can be present. [5] Stability of microstructure and composition of the insulating material is changing due to degradation processes. These changes result to changing behaviour of insulation material from view of polarization processes. Aging in a polymer changes the electric, physical, mechanical and morphological properties of the insulation [9]. All these properties are influencing the dielectric parameters and characteristics of insulation [10]. During thermal aging, several structural changes occur such as variation in crystalline, chain scission and variation in heat of fusion and in melting point. [7, 11] Integral ageing of XLPE cables changes the morphological properties of the insulation. It is well known that the aged XLPE cable insulations have many microvoids whose number increases with the distance from the cable conductor. Their dimensions and number depend on the technology and the kind of cable insulation. It is generally assumed that during production, microvoids, impurities, water and residual products from crosslinking will be collected in amorphous regions of the insulation. Increasing of aging temperature, the microvoids are of larger size. [12] XLPE insulation is a low-mobility material in which the mean free path length is quite small. However, the injected high energy electrons may easily cause bond dissociation in insulation. With the development of bond dissociation, some low ELECTROTEHNICĂ, ELECTRONICĂ, AUTOMATICĂ, vol. 62 (2014), Nr. 3 density regions, low cohesive energy density regions and thus free volume can be generated. In these regions, impact ionization is more likely, and thus leading to avalanches, partial discharge and eventually breakdown of the XLPE insulation. Such a process can be accelerated when there are weak points on the interface of void and insulation. [8] Aging of XLPE cables is related to the temperature of the insulation. All XLPE cables contain antioxidants which protect the XLPE from oxidation during the extrusion and cross-linking process, and also during the service life of the cable. The rate at which the antioxidant is used up is dependent on temperature. The normal maximum operating temperature of XLPE cables is 90 °C. Tests have shown that XLPE cables can operate at a temperature of 105 °C for a limited time without significantly reducing the service life of cables. At temperatures in excess of 105°C deformation of XLPE readily occurs, particularly at positions where the insulation is under mechanical stress. The maximum overload temperature of XLPE is limited to 105 °C. [3, 13] The rate of consumption of anti-oxidant has been calculated to provide a cable life of a minimum of 30 years at normal maximum operating temperature. Increasing the operating temperature of the cables will increase the rate at which the anti-oxidant is used up and hence reduce the service life. Small increase in temperature has a significant impact on the ageing of the XLPE. The XLPE will start to oxidize and become brittle, once the antioxidant in the cable is used up. This then leads to the cable will be subject to stress cracking and electric failure at positions of mechanical stress. [13] Quantitative assessment of ageing or damage of the XLPE insulations can be separated assessment of the relaxation mechanisms. XLPE cable ageing has been studied for nearly 40 years and many methods have been proposed to evaluate the properties of XLPE. [7]. 3. Measurement Method One of the methods for detection of insulating state of XLPE insulation is measurements of dielectric parameters of 49 insulation using the dielectric relaxation spectroscopy (DRS). DRS is a one of the nondestructive measurement methods and uses polarization as the response of the sample on a time-dependent electric field. For electrolyte solutions, polarization essentially originates from the oriented fluctuations of permanent dipoles (solvent molecules, ion pairs), from intramolecular polarizability and from ion motion. It can be investigated either in the time domain or as a function of the frequency of a harmonic field. Thus, principle of this method is based on examination of molecular dynamics of polarized and polar materials. Generally speaking, the behavior of the dielectric material is characterized DC conductivity, dielectric response function and high-frequency component of the relative permittivity. DRS is widely applied in the characterization of ion-conducting solids and polymers. [14, 15] In the frequency domain, the polarization phenomena follow the alternating electric field. DRS is focused on the measurement of the frequency dependence of real and imaginary components of the impedance of the investigated samples. DRS evaluates the dielectric response function in the frequency domain using the dielectric dissipation loss factor tan δ and complex capacitance C(ω). Using DRS in frequency domain can be evaluated the effect of temperature and thermal aging on dielectric parameters and change of properties of investigated insulation material after aging and degradation processes. [15] The total current flowing across the material when exposed to voltage U(ω) can be expressed as [15]: σ I (ω) = iωC0 ε ∞ + χ′(ω) − i 0 + χ′′(ω) U (ω) ε ω 0 (1) Complex electric induction D(ω) is proportional to the complex dielectric permittivity ε(ω) according to the relation [15]: D (ω ) = ε 0 ε (ω )E (ω ) = ε 0 [1 + χ′(ω ) − iχ′′(ω )]E (ω ) (2) where ε(ω) = ε′(ω) − iε′′(ω) = (1 + χ′(ω)) − iχ′′(ω) (3) 50 ELECTROTEHNICĂ, ELECTRONICĂ, AUTOMATICĂ,vol. 62 (2014), Nr. 3 Actual measurements of this dielectric response in the frequency domain are difficult to perform, if the frequency range becomes very large. The measured relative dielectric permittivity is defined from the following relation [15]: j (ω ) = iωε 0 ~ε r (ω )E (ω ) (4) Therefore ~ε (ω ) = 1 + χ ′(ω ) − i χ ′′(ω ) + σ 0 r ε 0ω (5) The dielectric dissipation factor: ε ′r′ (ω ) + tan δ (ω ) = σ0 ε 0ω ε ′r (ω ) (6) The real part of (5) represents the capacitance of a test object, whereas the imaginary part represents the losses. Both quantities depend on frequency. It should be noted that all dielectric quantities are more or less dependent on temperature and dissipation factor tanδ is strongly dependent on temperature. Any comparison or measurement of these quantities must take this into account. Increased interfacial polarization also produces the increase in dissipation factor, mainly in the low and very low frequency range. The value of dissipation factor is related to the dielectric loss and its value can be regarded as a measure of the quality of the insulation system. Measurements in the frequency domain need voltage sources of variable frequencies and, for applications related to HV power equipment, output voltages up to at least some hundreds of volts. Such measurements become quite lengthy if very low frequencies are considered. [15, 16] Figure 1. Frequency dependence of capacitance before and after aging process 4. Experimental measurement The experimental measurement was performed on two cable sample by the DRS method in frequency domain: • sample A - XLPE cable of operationally unknown technical condition, • sample B – operationally unaged, undamaged XLPE cable. The cable samples were a power cables with an aluminium core and XLPE insulation. Sample A of cable was approximately 25 cm long and protective cable jacket and semiconducting layer have been removed (2 cm at both ends) and shielding taken out. Sample B of cable was approximately 1 m long and protective cable jacket and semiconducting layer have been removed (10 cm at both ends) and shielding taken out. On sample A were recorded changes of capacitance and dielectric dissipation loss factor depending on frequency and the measurement was repeated after an accelerated thermal aging of sample at 90 °C for 360 hours in an air oven. On sample B were recorded changes of capacitance and dielectric dissipation loss factor depending on frequency at various temperatures. The frequency range of measurements was 20 Hz to 2 MHz and the measurements were performed using the precision LCR meter Agilent E4980A. The increase of frequency was decimal. Aim measuring has been comparison of measured frequency dependencies of capacitance and dissipation factor. In Figure 1 and Figure 2 are shown comparison of frequency dependencies of capacitance and dissipation factor before and after aging process, respectively. Figure 2. Frequency dependence of dissipation factor before and after aging process ELECTROTEHNICĂ, ELECTRONICĂ, AUTOMATICĂ, vol. 62 (2014), Nr. 3 Due to the large fluctuations in the measured values at low frequencies from 20 Hz to 100 Hz, the measured values are not within the zone of termination capable and therefore are not shown. Figure 1 shows a decrease of the values of capacitance after aging process. The change of capacitance is very small, order of 10-11 Farad. From frequency circa 200 kHz occurs a noticeable decrease of capacitance in measurement after aging process. Such measurement is sensitive to disturbances and parasitic capacitance. Figure 2 shows a change of the values of dissipation factor after aging Figure 3. Frequency dependence of dissipation factor at various temperature The measured values at low frequencies from 20 Hz to 100 Hz are not shown due to the large fluctuations. Figure 3 shows a change of the values of dissipation factor at higher temperatures. A significant change of dissipation factor is at high frequency from circa 700 kHz where there was an increase of measured values. In frequency range between 100 Hz and 20 kHz were measured smaller values of dissipation factor at higher temperatures compared to measurement at 20 °C. Thus, similar to the measurement of capacity, there is also here a change of values of dissipation factor with a change of temperature. The trend of frequency dependencies was similar. Figure 4 shows a decrease of the values of capacitance at higher temperatures. The change of capacitance is very small, order of 10-10 farad. With increasing temperature there is also a significant and steeper increase of capacitance from frequency circa 400 kHz. At a temperature 20 °C, there is a slight increase of capacitance from circa 1 MHz. Thus, with a change of 51 process. The change of dissipation factor is significant at high frequency from 10 kHz where there was a significant and steeper increase of measured values. The values of dissipation factor are greater after aging process. The accelerated thermal aging process caused a change of frequency dependence of dissipation factor. In Figure 3 and Figure 4 are shown comparison of frequency dependencies of dissipation factor and capacitance at temperature 20 °C, 60 °C, 70 °C and 80 °C, respectively. Figure 4. Frequency dependence of capacitance at various temperature temperature occurs a change of values of capacitance and also a change the curve of the frequency dependencies, but the trend of dependencies is similar. 5. Conclusions It is well known that temperature and thermal aging of insulation influences on the parameters and quality of insulation. Using DRS in frequency domain can be evaluated the effect of temperature and thermal aging on dielectric parameters and change of properties of investigated insulation material after aging and degradation processes. Aim of this paper was to observe a change of dielectric parameters of XLPE cable due to thermal aging. The comparison and evaluation of measured frequency dependencies of capacitance and dissipation factor of two samples of XLPE cable show that temperature and aging process change dielectric properties of samples of XLPE cable. Such changes can be related to structural changes in the 52 ELECTROTEHNICĂ, ELECTRONICĂ, AUTOMATICĂ,vol. 62 (2014), Nr. 3 polyethylene morphology due to thermal aging. It is possible to confirm that degradation led to changes in the polarization processes occurring in insulation. For an overall assessment of the effect of aging is needed to know changes of measured values and their development of previous measurements. The measurement confirmed the influence of thermal ageing to dielectric parameters of XLPE insulation. The measurement confirmed the suitability of DRS for investigate cable insulation system. It is not possible exclude the surrounding of a disturbance during the measurement, which could affect the measured results. It is important to continue to investigate the process of ageing and their impact on the insulation system with use modern and suitable diagnostic methods and measuring equipment. 6. Acknowledgment We support research activities in Slovakia /Project is cofinanced from EU funds. This paper was developed with support of operating program Research and development for the project: "Univerzitný vedecký park Technicom pre inovačné aplikácie s podporou znalostných technológií" (University Science Park Technicom for innovative applications with support of knowledge technologies), code ITMS: 26220220182, co-financed from European funds. 7. References [1]PETZOLD F. et al. "Advanced solution for onsite diagnosis of distribution power cables". in: Electricity Distribution - Part 1, 2009, CIRED 2009, 20th International Conference and Exhibition; 2009; p.1-4. [2]SHUVALOV M. et aI. "Analysis of Water Trees in Power Cable Polymeric Insulation". in: Journal Applied Polymer Science. 2003; vol.88: p. 1543-1549. [3]METWALLY I.A, "The Evolution of Medium Voltage Power Cables". in: Potentials, IEEE. 2012; vol.31(no.3): p. 20-25. [4]HAMPTON N. et al. "Long-life XLPE Insulated Power Cable". in: Jicable 2007. [5]RAWANGPAI A., et al. "Artificial Accelerated Ageing Test of 22 kV XLPE Cable for Distribution System Applications in Thailand". in: World Academy of Science, Engineering and Technology 65. 2010, p. 220-225. [6]WALD D., HAMPTON, N., "How much does studying Polyethylene tell us about XLPE?" in: Electrical Insulation (ISEI), Conference Record of the 2012 IEEE International Symposium; 2012; p.250-254. [7]JIANYING L. et al. "The Effect of Accelerated Water Tree Ageing on the Properties of XLPE Cable Insulation". in: IEEE Transaction on Dielectrics and Electric Insulation. 2011; vol. 18: p. 1562-1569. [8]WEIWEI LI et al. "Frequency dependence of breakdown performance of XLPE with different artificial defects". in: Dielectrics and Electrical Insulation, IEEE Transactions. 2012; vol.19(no. 4): p.1351-1359. [9]HOFF G., KRANZ H.G. "Correlation Between Return Voltage and Relaxation Current Measurements on XLPE Medium Voltage Cables". in: 11th ISH 1999; London. [10]HOFF G., KRANZ H.G. "Interpretation of Dielectric Response Measurement Data from Service Aged XLPE-Cables". in: IEEE 7th International Conference on Solid Dielectrics,; 2001 June 25-29; Eindhoven; p. 381-384. [11]THOMAS A.J., SAHA T.K. "Statistical Analysis of Diagnostic Indicators during an Accelerated Ageing Experiment". in: IEEE Transactions on Dielectrics and Electrical Insulation. 2012; vol. 19(no. 1): p. 274-282. [12]NIKOLAJEVIC S.V. "Accelerated aging of XLPE and EPR cable insulations in wet conditions". in: Electrical Insulation, 1998. Conference Record of the 1998 IEEE International Symposium,; 1998; vol.1, p.9396. [13]Ageing of XLPE Compounds, Brochure, General Cable Australia Pty Ltd. [14]BUCHNER R. “Dielectric spectroscopy of solutions”. in: Novel Approaches to the Structure and Dynamics of Liquids: Experiments, Theories and Simulations,; 2004, p. 265–288. [15]ZAENGL W.S. “Dielectric Spectroscopy in Time Domain for HPV Power Equipment, Part I: Theoretical Consideration”. in: IEE Transaction on Dielectrics Electrical Insulation. 2003; vol. 19(no. 5): p. 5-19. [16]CIMBALA R. [Aging of high-voltage insulation systems]. Košice: TUKE; 2007. [17]FOTHERGILL J.C., DODD S.J., DISSADO L.A., LIU T., NILSSON U.H. “The Measurement of Very Low Conductivity and Dielectric Loss in XLPE Cables,” in: IEE Transaction on Dielectrics Electrical Insulation. 2011; vol. 15(no. 5): p. 1544-1553. [18]JIANDONG Wu et al. "The condition assessment system of XLPE cables using the isothermal relaxation current technique". in: Properties and Applications of Dielectric Materials, 2009. ICPADM 2009. IEEE 9th ELECTROTEHNICĂ, ELECTRONICĂ, AUTOMATICĂ, vol. 62 (2014), Nr. 3 International Conference; 2009; p.11141117. [19]FEI Liu et al. "Insulation ageing diagnosis of XLPE power cables under service conditions". in: Condition Monitoring and Diagnosis (CMD), 2012 International Conference; 2012; p.647-650. [20]OYEGOKE B.S. et al. "Condition Assessment of XLPE Insulated Cables Using Isothermal Relaxation Current Technique". in: Power Engineering Society General Meeting, IEEE; 2006; p. 6. [21]ABDERRAZZAQ M.H., HUSSIN M.S., ALHAYEK K., "The effect of high frequency, high voltage supply on the growth of electric trees on cross linked polyethlyne insulation of power cables," in: Solid Dielectrics (ICSD), 2013 IEEE International Conference; 2013 June 30-July 4; p.812,815. [22]UEYAMA T., MORI T., SUENAGA K., "Experiments of power cable diagnosis by using isothermal relaxation current method," in: Condition Monitoring and Diagnosis (CMD), 2012 International Conference; 2012 Sept 23-27; p.72,75. [23]FRECHETTE M.F. et al. "Dielectric response of polyethylene modified with nanosilica," in: Electrical Insulation Conference (EIC), 2013 IEEE; ,2013 June 2-5; p.276,280. 8. Biography Martin GERMAN-SOBEK was born in Košice (Slovakia), on July 26, 1987. He graduated the Technical University of Košice, Faculty of Electrical Engineering and Informatics in Košice (Slovakia), in 2011. 53 He is PhD student at the Technical University of Košice, Faculty of Electrical Engineering and Informatics in Košice (Slovakia). His research interests concern: diagnostic of high-voltage solid insulation system, highvoltage equipment, power cables. Roman CIMBALA was born in Košice (Slovakia), on April 28th, 1962. He graduated the Technical University of Košice, Faculty of Electrical Engineering and Informatics in Košice (Slovakia), in 1989. He received the PhD degree in electrical engineering from the Slovak University of Technology in Bratislava (Slovakia), in 1994. He is Professor at the Technical University of Košice (Slovakia). His research interests concern: dielectric spectroscopy, diagnostics of HV systems. Jozef KIRÁLY was born in Košice (Slovakia), on May 3th, 1987. He graduated the Technical University of Košice, Faculty of Electrical Engineering and Informatics in Košice (Slovakia), in 2011. He is PhD. student at the Technical University of Košice (Slovakia). His research interests concern: dielectric spectroscopy, diagnostics of HV systems, ferrofluids.