1 Introduction

The great demands for crosslinked polyethylene (XLPE) in the worldwide market is ascribed to its excellent electrical insulating properties, outstanding physicochemical characteristics, easy processing and low cost. Nowadays, the medium and high voltage power cables with the XLPE insulation have been widely used in underground transmission and distribution power system [1,2,3]. However, degradation of XLPE insulation is still one of the most important problems that causes cable failure during its service life.

Under normal service conditions, the operating temperature of cable insulation is around 90 °C due to the large ampacity. Sometimes the maximum temperature up to 150 °C caused by short circuit or overload is allowed in a short term [4, 5]. Therefore, electric-thermal ageing becomes a main origin of irreversible damage of the cable insulation. For most polymer materials, degradation process is often dominated by oxidation reaction when running or ageing in an oxidizing atmosphere [6]. Afterwards, chain scission and carbonyl groups are produced [7]. The impact of oxidation reaction on XLPE insulation has been studied for decades. It is reported that the lamellas and spherulites could be destroyed, and free volume could be formed in XLPE insulation, resulting in a reduced crystallinity and degraded electrical and mechanical characteristics [8, 9]. Some additives, e.g., dicumyl peroxide (DCP) and phenolic or sulfur-type antioxidants, have been used to mitigate the electric-thermal ageing of XLPE [10, 11]. Actually, antioxidants can play a role in restraining the acceleration of oxidation process, but it can't totally prevent it. In the service of power cables, the effect of electrical filed applied on the XLPE insulation can not be neglected. He D. et al. reported electrical-thermal degradation of 10 kV XLPE insulation aged under an AC voltage of 26.1 kV and 103–135 °C for 150–375 days [12]. The mechanical strength and electrical breakdown field of the XLPE samples were decreased sharply with ageing time when the applied temperature is at 135 °C. Montanari et al. investigated the electrical behavior of the cable models with a thin insulation layer under a long-term electrical-thermal aging, which can give some guidance in the cable design [13]. Wang X. et al. studied the space charge characteristics of XPLE films modified with nanoparticles under a DC voltage of 40 kV and 60 °C for 30–90 days [14].

Electrical-thermal ageing is the main factor causing insulation degradation when the XLPE cables serve in a dry environment. However, water can be introduced into the XLPE insulation during manufacture, storage and installation. In the practical service condition, water might also enter into insulation from the joints or terminals, inducing growth of water tree in the XLPE insulation. Water tree is a common ageing phenomenon that can impair XLPE insulation, which can convert into electrical tree subsequently, increasing failure rate of cables. It becomes another important cause for insulation failure of in-service XLPE power cables. Furthermore, some microscopic defects such as impurities, micro cracks, and scratches in cable insulation might be generated during the manufacturing, transportation, constructing and operating processes [15]. These defects tend to become the starting point for the growth of water trees for cables served in a moist environment. Many measures such as water barriers, semicon cleanliness, and laying practices were adopted to make cables more resistant to water treeing [16]. XLPE cable with water tree retardant additives (often referred to simply as TR-XLPE) is now widespread [17]. However, even though great achievements have been acquired on cable quality in recent years, the power cable running underground is still vulnerable to water tree degradation.

Electric-thermal ageing and water tree degradation are two significant issues endangering the normal operation of XLPE cables. Therefore, accelerated electrical test (AET) and accelerated water tree tests (AWTT), as two important ways to study the degradation processes of XLPE were widely reported [6, 8, 17,18,19]. However, they have been studied separately, not highlighting the discrepancy caused by water. There is still a lack of comparative investigation on the accelerated electrical test and accelerated water tree tests for XLPE cable. In this work, AET and AWTT are conducted on 10 kV XLPE cables. The differences in physicochemical and dielectric properties caused by those two ageing processes are displayed.

2 Samples and Experiments

2.1 Sample Preparation

Based on the recommendations of standard AEIC CS8-07 and ICEA S-97–682, AET and AWTT on 10 kV XLPE cables were carried out. The ageing conditions of the XLPE samples are shown in Table 1.

Table 1 The ageing conditions of XLPE cable samples
Full size table

For the accelerated electrical test, the cables are put in ambient air and heated by inductive current generated in the conductor loop. The cables were heated up by a straight-through transformer (BCX-15KVA) and kept at between 95 and 100 °C for 8 h and then cooled down to room temperature for 16 h to imitate a load circle every day. During the ageing process, a power frequency voltage of 27.5 kV (TNS(SVC)100KVA) is applied. During the AWTT process, the cables were put in tubes filled with tap water, and the schematic and object diagrams were shown in Fig. 1a and b. The ageing test was divided into three stages at regular intervals for 1440 h (phase 1), 2880 h (phase 2) and 4320 h (phase 3), and the samples were denoted as AET-1440 h, AET-2880 h, AET-4320 h for accelerated electrical ageing, and AWTT-1440 h, AWTT-2880 h, AWTT-4320 h for accelerated water-tree ageing. Three cables are selected to experimental test under each condition in this work.

Fig. 1
figure 1

The AWTT schematic a and object b diagrams for 10 kV XLPE cable

Full size image

2.2 Physicochemical and Electrical Experiments

For the physicochemical and dielectric properties test, the XLPE slices were cut from the cable insulation layer by using J/Q XLPE cable slicer. FTIR spectrometer (model IR Prestige-21) was employed to analyze the variation of XLPE molecular structure in the ageing process. All the samples mentioned above were measured in the range of 500–4000. 1 cm × 1 cm XLPE samples were prepared for X-ray diffraction test, and were scanned with a DX-1000 X-ray diffractometer from 15° to 30° with a scan step rate of 0.026 °. TGA/SDTA851 from Mettler Toledo was used to analyze the thermal decomposition behavior of XLPE samples. 10–20 mg samples were prepared and put in an alumina crucible, and the test temperature went from room temperature to 600 °C by rate of 20℃/min. During the whole test process, argon gas flow of 50 ml/min was used for protection. 35 mm × 35 mm square samples with thickness of 0.2 mm were cut from the cable insulation layer for the insulating strength test in accordance with the IEC 60243-1-1998 (“Methods of Test for Electric Strength of Insulating Materials”). In the test, 50 Hz AC voltage was applied, and the voltage increasing speed was 500 V/s. At least 10 breakdown points were carried out for each kind of sample. Round-disk XLPE samples with diameter of 10 mm and thickness of 0.5 mm were prepared for the dielectric test, and gold was sputtered onto both sides of the samples as conductive electrodes. The dielectric properties were measured with Novocontrol broadband dielectric spectrometer (Concept 80, Germany) in the frequency range from 1 to 105 Hz.

3 Results and Discussion

3.1 FTIR Results

The FTIR spectra of both unaged and aged XLPE samples are shown in Fig. 2a and b. We can see that the absorption peaks at the values between 1700 and 1750 are obviously intensified after AET and AWTT ageing. As well known, carbonyl groups (C = O) is a typical oxidation product in XLPE, which can be an indicator of thermo-oxidative ageing. The absorption peaks of carbonyl groups appear between 1700 and 1800, which mainly contains ester groups (-COO-) at 1760, aldehyde groups (-CHO) at 1738, and keto-groups (-CO-) at 1720, respectively [19]. Carbonyl index (CI) was calculated based on the ratio of absorption peak at 1720 cm−1 and 2010 cm−1, which is commonly used to quantify the oxidation degree [12, 20, 21]. The value of CI can be calculated as follows:

$$CI = \frac{{I_{{1720}} }}{{I_{{2010}} }}$$
(1)

where I1720 and I2010 are the absorption peak intensity of at 1720 cm−1 and 2010 cm−1, respectively. As shown in Fig. 3, the carbonyl index is about 0.53 for the unaged XLPE sample, which means there are just a few carbonyl group residuals during the manufacturing process [22]. However, the values of carbonyl index of XLPE increased to 0.87 and 1.10 for the AET-4320 h and AWTT-4320 h samples, respectively. AWTT induces a higher carbonyl index, indicating that water or moist environment act as degradation catalysts to accelerate the decomposition process of XLPE.

Fig. 2
figure 2

FTIR Spectra of different XLPE samples. a Accelerated electrical test. b Accelerated water tree test

Full size image
Fig. 3
figure 3

Variation of the Carbonyl index with the ageing time

Full size image

3.2 XRD Results

The X-ray diffraction (XRD) patterns of both unaged and aged XLPE samples are shown in Fig. 4a and b. A typical XRD curve of the unaged sample shows an amorphous halo at 2θ = 21.1° (peak 1) and two major crystalline peaks at 2θ = 21.7° (peak 2) and 2θ = 24.1° (peak 3), corresponding to the (110) and the (200) lattice planes [22, 23]. Another tiny peak (peak 4) appeared at around 2θ = 26.8° is related to the plane lattice of (020) as Miller indexes [22]. The results present that the XRD curves do not show any new crystalline peaks or obvious changes in the position of crystalline peaks. It indicates that AET and AWTT process do not induce any new crystalline phases in the materials structure, but the major crystalline peak at 21.7°is greatly decreased after AET and AWTT ageing.

Fig. 4
figure 4

X-ray diffraction patterns of a AET XLPE samples, b AWTT XLPE samples, c the Gaussian Fitting for the XRD spectra of the amorphous regions and the crystalline regions of the typical unaged XLPE sample

Full size image

As shown in Fig. 4c, the crystallinity of XLPE samples were calculated using Hinrichsen’s method [24] that is based on fitting the XRD spectra into three different Gaussian functions by ORIGIN 9.0 software with the following relationship:

$$x(\% ) = \frac{{area2 + area3 + area4}}{{area1 + area2 + area3 + area4}} \times 100\%$$
(2)

where χ(%) is the crystallinity percentage, area 1 corresponds to the amorphous halo, areas 2, area 3 and area 4 correspond to the crystalline peaks at 21.7° (peak 2), 24.1° (peak 3) and 26.8° (peak 4), respectively.

Figure 5 presents the variations of crystallinity with the ageing time. For the AET sample, the crystallinity increases a little firstly when the ageing time reaches around 1500 h, and then goes down. Li et al. found the similar phenomenon in the thermo-oxidative aged XLPE, and the increase of crystallinity in the initial ageing stage is attributed to recrystallization of XLPE with favors of the alignment of the chains that were imperfectly crystallized in the manufacturing process [22]. However, the crystallinity of AWTT samples is reduced during the whole ageing time.

Fig. 5
figure 5

The crystallinity of XLPE samples with the different AET and AWTT

Full size image

3.3 Thermogravimetric Analysis

Figure 6 displays typical thermogravimetric analysis (TGA) and derivative thermogravimetric (DTG) curves of AET and AWTT XLPE samples. It is find that the decomposition process starts at the temperature of 450 °C and ended at the temperature around 500 oC. In the inset of Fig. 6a, the DTG curve of the AET sample presents a peak that corresponds to the fastest decomposition temperature Td-1. The peak shifts to lower temperature, i.e. from 492 to 488 °C with the increases of ageing time, which demonstrated that the thermal stability of the samples was deteriorated with the increase of the ageing time. Activation energies for the XLPE decomposition were determined using the Coast-Redfern approach [25], which shows that the activation energy increases a little from 3.44 to 3.48 eV in the first stage of AET sample (before 1440 h) and then decreases to 3.37 eV with the further ageing period, as listed in Table 2. In accordance with the XRD result, the crystallinity increases due to recrystallization of XLPE, which may make crosslinked XLPE molecules difficult to decompose in the initial ageing stage. However, the DTG curves of AWTT samples present an abnormal multi-peak phenomenon and several fastest decomposition temperature Td-1, Td-2, and Td-3 can be obtained, indicating that there exist different main stages during the thermal decomposition process. The activation energies corresponding to Td-1, Td-2, and Td-3 were calculated based on the Coast-Redfern approach and presented in the Table 2. Taking the sample of AWTT-4320 h for example, the activation energies of the three decomposition processes (Td-1, Td-2, and Td-3) were 3.38, 3.16 and 3.04 eV, respectively. The activation energies of the decomposition related to Td-2, and Td-3 is much lower.

Fig. 6
figure 6

The TGA and DTG curves of XLPE samples with the different a AET and b AWTT

Full size image
Table 2 Parameters obtained from TGA and DTG measurement
Full size table

As shown in Fig. 5, the crystallinity of AWTT samples decreases with the ageing time. Therefore, it is believed that a mass of microscopic defects may be generated because some crystalline region (lamellae) in XLPE can be broken under the multiple effects of electric, thermal oxidation reaction, and water. Especially, the dielectrophoresis force caused by water molecules can significantly promote degradation of XLPE molecular chains [26]. Furthermore, some low-density areas can be formed in the cable insulation which may be more easily decomposed and thus resulted in the multi-peak phenomenon in the DTG curves.

3.4 Insulating Strength

Figure 7 shows the Weibull plots of insulating strength under power frequency voltage. The results in Table 3 present the calculated Weibull scale and shape parameters (α and β). α is the breakdown voltage in the confidence interval of 63.2% and β displays the data scatter. A higher β represents a narrower range in breakdown voltage. For the AET samples, the breakdown voltage shows an initially increase a little, and then decrease from 23.03 kV to 22.49 kV with ageing time. It has a similar evolution trend with that of the crystallinity of XLPE. The similar phenomenon was reported in other reports [12, 19], which indicates that the insulating strength of XLPE may be closely related with the crystalline morphology. However, for the AWTT samples, the breakdown voltage is monotonically decreases from 23.03 kV to 21.74 kV with ageing time the for AWTT samples. The insulating strength of the samples were not reduced greatly after the AET and AWTT in this work.

Fig. 7
figure 7

a Weibull plots and b insulating strength of AET and AWTT samples

Full size image
Table 3 The parameters of Weibull distribution under power frequency voltage
Full size table

3.5 Dielectric Results

Figure 8 depicts the frequency dependence of real and imaginary parts of permittivity (ε′, ε″) for the aged and unaged XLPE samples. As presented in Fig. 8a, the ε′ of unaged sample is ~ 2.5, however, the value is enhanced to 2.9 and 3.25 for the AET and AWTT samples (AET-4320 h and AWTT-4320 h), respectively. In Fig. 8b, the ε″ of AET samples was significantly increased at low frequency (< 50 Hz), which can be attributed to the contribution of enhanced dc conductivity according to the Debye Eq. (3).

$$\begin{aligned} \varepsilon ^{*} \left( \omega \right) = &\ \varepsilon ^{\prime } \left( \omega \right) - j\varepsilon ^{ \prime\prime } \left( \omega \right) \\ =&\ \varepsilon _{\infty } + \frac{{\varepsilon _{s} - \varepsilon _{\infty } }}{{1 + \omega ^{2} \tau ^{2} }} - j\left( {\frac{{\sigma _{0} }}{{\omega \varepsilon _{0} }} + \frac{{(\varepsilon _{s} - \varepsilon _{\infty } )\omega \tau }}{{1 + \omega ^{2} \tau ^{2} }}} \right) \\ \end{aligned}$$
(3)

where εs and ε∞ denote the static and optical permittivity, respectively, ω is the angular frequency, τ is the dipolar relaxation time, σ0 is the direct current (dc) conductivity.

Fig. 8
figure 8

Temperature dependences of ε′ a and ε″ b of XLPE samples with the different AET and AWTT

Full size image

For the AWTT sample, the ε′ shows an apparent decline trend with applied frequency from 1 to 100 Hz, and an obvious relaxation peak can be found at around 100 Hz, which was reported and considered as the interfacial polarization due molecular chains segments or microdefect caused by AWTT process [18, 27,28,29]. Below 10 Hz, the dielectric loss was dominated by dc conductivity. In addition, at the same ageing stage, the permittivity and dielectric loss of the AWTT samples is obviously higher than that of the AET samples in the whole applied frequency. It indicates that AWTT process can induce a more serious degradation in the dielectric properties of XLEP insulation due to the incorporation of moisture ingress and water tree [30].

3.6 The Micro-Level Analysis of Defects

It is known that the microstructure of XLPE insulation generally consists of spherulites (crystalline lamellar ribbons) and amorphous phase as shown in Fig. 9a. The interface between the crystalline and amorphous area is generally regarded as weak point, where the microdefects such as voids in XLPE may easily formed and developed during the AET or AWTT ageing processes. In the amorphous region between lamellar ribbons, the molecular chains are linked via inter-molecular Van der Waals bonds, which is vulnerable to the continuous electric-thermal ageing as well as oxidation reaction [31,32,33], leading to free volume rearrangement and formation of submicrocavities. In Fig. 9b, two adjacent lamellas from spherulite are selected to illustrate the change of crystalline structure of XLPE insulation during the AET ageing process. Thermal swelling force (σ1) is applied on the vertical direction of lamellas due to thermal expansion of amorphous inter-lamellar phase, leading to a growing lamella space. Thus oxygen can diffuse into spherulite more easily, promoting oxidizing reaction and breaking the molecular chains that connects lamellas. Moreover, under the applied electrical filed, space charge such as electrons and charged particles can be accumulated in the interface between the crystalline and amorphous area or in the submicrocavities of the XLPE insulation, which may cause seriously electric field distortion and energy transfer in the XLPE insulation [12, 34]. Electrons can be accelerated under the distorted electrical filed and break molecular chains. After a long-term AET process, the folded chains between lamellas can be broken, and the lamella structure may be destroyed, resulting in a low crystallinity of XLPE.

Fig. 9
figure 9

a The microstructure of spherulites and lamellas, the development of XLPE degradation during b AET and c AWTT ageing processes

Full size image

As shown in Fig. 9c, water may immerse into the cable insulating materials in AWTT process so that another dielectrophoresis force σ2 may be introduced. With the increase of ageing time, water molecules will permeate into amorphous region in depth and enlarge the angle between two lamellar ribbons [35, 36], thus cause more severe degradation of XLPE insulation. Since tap water was used in the AWTT process, some impurities like metal ions and carboxylates may work as carriers in the interface between amorphous and lamellar zones, which was noticed and reported by several researchers in the research of water tree [37, 38] and electrical tree [36, 39]. It was found by TEM observation that the initial electrical tree grew along the lamellar layer [36, 39]. These results support that lamellar/amorphous interface in the insulation layer may work as channels of water during AWTT ageing and consequently lead to the change of physical, chemical and electrical performance. L.A. Dissado et al. [40] proposed that extended states of carriers can lie in between the chains in the crystals or in the region of unfulfilled lattice sites in the amorphous region. Moreover, interfacial polarization caused by the carriers can happen in the interface between crystalline and amorphous area or the submicrocavities, and results in a high dielectric loss in low frequency range [41]. The impurity introduced with water can also enhance the electric conduction and lead to thermionic polarization process [42]. As a consequence, AWTT process gave rise to a more severe degradation in the XLPE insulating materials.

4 Conclusion

This work studied the effect of accelerated electrical test and accelerated water tree test on the physicochemical and dielectric properties of 10 kV XLPE cable. The AWTT combines the functions of water and electricity, inducing higher carbonyl index, lower decomposition activation energy and lower crystallinity than AET process. During AWTT process, oxygen-containing groups which have various decomposition temperatures may cause the multi-peak phenomenon in DTG curve. The dielectric loss of AWTT samples is higher than that of AET samples, and a new dielectric relaxation can be observed due to polarization relaxation at around 100 Hz. All in all, AWTT process can lead to a more severe degradation due to superimposition of moisture and electrical stress. A schematic diagram is proposed to illustrate the microstructure evolution of XLPE insulation during AET and AWTT process. Thermal swelling force and dielectrophoresis force are two significant factors to break up molecular chains and the spherulites.