21, rue d'Artois, F-75008 Paris

http://www.cigre.org B1-111 Session 2004

© CIGRÉ

Higher-stress Designed XLPE Insulated Cable in Japan

A. TOYA K.KOBASHI Y. OKUYAMA

Tokyo Electric The Kansai Electric Chubu Electric
Power Co. Power Co., Inc. Power Co., Inc.

S. SAKUMA* S. KATAKAI K. KATO

VISCAS Corp. J-Power Systems Corp. EXSYM Corp.

(Japan)

1. Introduction
In Japan, the realization of high density power transmission in major city areas requires the
effective utilization of limited underground space, available, for example, in tunnels and ducts.
Thus, for many years, particular efforts have been directed towards reducing the construction
period, cutting the cost, and improving the reliability of extra-high voltage lines. This has
necessitated a reduction in the number of joints, and thus long cable manufactured in continuous
length had to be transported from the factory to the installation site, an undertaking that required
the development of technology to installation the cable directly from the unloading point. At the
same time, it has become essential to maximize the transportation length of the cable and to
reduce its weight.
In response to these needs, cross-linked polyethylene insulated cable with a reduced insulation
thickness and the joints suited to it has been developed in Japan. This cable is capable of handling
currents of up to 500 kV, and has been successively used in practical applications.
This paper will discuss Japan’s XLPE cable insulation technology for reducing insulation
thickness, namely the design technology, the materials technology, the manufacturing quality
control technology, and the quality assurance technology that has permitted the conversion to high
stress cables, as well as operation records of these high stress cables. It will touch upon the
adequacy of the test conditions and concepts involved in the development of the tests for verifying
that the products could be applied in actual systems. It will also present the efforts made to
ascertain approximately what future design stresses could be achieved by determining through
experimentation the ultimate stress that could be tolerated without degradation.
2. The XLPE cable insulation design technique
The dispersion of the dielectric breakdown strength versus AC voltage and lightening impulse
voltage of XLPE cable are considered to conform to the Weibull distribution. Consequently, the
probability of an XLPE cable breakdown due to electric field strength E (kV/mm) is shown by the
following expression. ⎧ ⎛ E − E L ⎞⎫
P = 1 − exp ⎨− ⎜⎜ ⎟⎟ ⎬m
⎩ ⎝ ES ⎠⎭

* s-sakuma@viscas-j.co.jp 1
 Here, E: Breakdown electric field, EL: Location parameter, ES: Scale parameter and m: shape
parameter
In Japan, the minimum insulation breakdown stress for AC and lightening impulse voltage is
determined based on numerous breakdown data and the insulation for XLPE cables is then
designed based on this EL. Specifically, insulation thickness corresponding with the required AC
withstand voltage level, lightening impulse withstand voltage level derived from their respective
EL are calculated and the thicker is adopted. Since improvements in the EL value are made at the
quality control level, the system is set up such that when periodic reviews are made and the
manufacturers adopt new design stresses, approvals are obtained from users in view of that data. It
should be highlighted that these stresses are derived as average stress, while maximum stress is
strictly considered for allowable defect level and quality control in factory and on-site.
(1) Insulation thickness determined based on the AC voltage
Here, E: Maximum system voltage, k1: Temperature coefficient (1.2),
E 3 × k1 × k 2 × k 3
t ac = k2: Degradation coefficient
E Lac (2.3, when degradation coefficient n = 15),
k : Uncertainty tolerance (1.1)

(2) Insulation thickness determined based on the lightening impulse voltage

LIWV × k1 × k 2 × k 3 Here, LIWV: Lightening impulse withstand voltage,
t imp = k1: Temperature coefficient (1.25),
E Limp
k2: Degradation coefficient
due to repeated impulse overvoltage (1.0),
k : Uncertainty tolerance (1.1)

Actually, insulation thickness is determined by taking into account the electrical strength at the
starting point of stress relief cone in the joint in addition to the above mentioned considerations.
Table 1 shows an example of the insulation design of the XLPE cable for each currently applied
voltage in Japan. Ultimately, the insulation thickness is determined by the electrical strength at
the starting point of stress relief cone in the joint, and we therefore know that it is possible to
make it even more compact by improving the characteristics.
The advantages of reducing cable size by reducing insulation thickness
Table 2 shows the effect and advantages of reduced insulation thickness. Fig. 1 shows an
example of the potential for increasing capacity using existing ducts by reducing insulation
thickness. Though the laying configuration also plays a part, transmission capacity can be

Table 1. Example of the insulation design of the XLPE cable

Nominal Voltage (kV) 66 154 275 500
kV 72 168 300 550
AC withstand voltage kV 35 35 35 40
AC(mm) 3.6 8.5 15 24.3
Lightening impulse withstand LIWV kV 350 750 1050 1425
voltage imp kV 75 75 75 80
(mm) 6.4 13.8 19.3 24.5
Insulation thickness determined by the electrical strength 9 10 15 17 23 27
*1)
at the starting point of stress relief cone in the joint
Insulation thickness (mm) 9 10 15 17 23 27
AC design stress at the starting point of stress relief cone (kV/mm) ------ TJ: 17, EMJ· PJ: 24
Impulse design stress at the starting point of stress relief cone (kV/mm) --TJ: 53, EMJ: 54, PJ: 56

2
 Table 2. Advantages of reduced insulation thickness
Effect of reduced
Advantages resulting from reduced insulation thickness
insulation thickness
• Effectively takes advantage of Increases the transmission
existing space. capacity of existing ducts
Smaller
Reduces the number of joints by Cuts cost,
cable diameter
increasing the transportation length Reduces construction periods,
Improves line reliability
Reduces the number of joints by Cuts cost,
increasing the transportation length Reduces construction periods
Less cable weight Improves the working conditions Improves line reliability
for laying the cable Reduces potential problems
during laying

3000
2500
can b e laid with in duct (mm2)

2500

Trans potation length (m)

154kV 17mm
M aximum s ize that

17mm 2000
2000
275kV
23mm
1500
154kV 1500
23mm 23mm
1000
11mm
275kV 77kVi 3c ore )- 11mm 1000
500 13mm 23mm
2‚Vmm
77kVi 3c ore )- 13mm
27mm
0
500
100 120 140 160 180 200 220 77kV3core 154kV 275kV
0 1 2 3 4
In ner-diameter o f d uct (mm) 250mm2 ‚P 000mm2 1000mm2
Fig.-1 Relationship between of the insulation thickness Fig.-2 Relationship between of the insulation
and the maximum size that can be laid within the duct thickness and transportation length

Fig.3 Example of long length cable

(500kV1000mm2, Alminum corrugated
sheath, insulation thickness = 27mm,
in-land transportation)
Cable length 1800m, GIS to GIS with
no joint

increased by some 10 to 30%. Moreover, as shown in Fig. 2, reducing the insulation thickness
can sometimes increase the transportation length by 50% or more. In such cases, major
advantages result, such as a reduction of the number of joints by up to two third and also shorter
construction periods.
4. Approaches and results of compacting XLPE cable
4-1 Identification of the key factors
Developing higher stress XLPE cables, requires identification of the key factors governing the
breakdown of the insulation. Accordingly, the pre-breakdown partial discharge detection method
(PPD method) that stopped the progression of the electrical tree before it resulted in the

3
 breakdown of a cable was developed, and the main causes of
breakdowns were identified [1]. Fig. 4 shows an example of this.
Results of these tests clearly showed that the origin of the
majority of these breakdowns was impurities in the insulation[2] .
By analyzing these impurities, it was possible to identify the
primary cause of the contamination and to devise appropriate
countermeasures. It is also important to clearly define the size of
the permissible impurities in the insulation. To do this, the
electrical field that generates electrical tree for the XLPE
insulation was reassessed experimentally. Based on PPD
method using such samples as model cables with artificial Fig-4 Example of origin of
protrusions, the electrical field of the defect point (electric field breakdown of cable detected
that generates electrical tree) was determined by means of the by PPD method.
finite element method from the shape of the defect and the cut off voltage. From these results it
was verified that the electric field that generates electrical tree of the XLPE cable was Ei=300
kV/mm and if 10 µm was considered to be the minimum value of the radius of curvature of the tip
then it was essentially adequate.[3] From the results of the concerned study it was decided that,
allowing for a certain tolerance, the permissible size of impurities was 50 µm for a 500 kV XLPE
cable (with an insulation thickness of 27 mm). Moreover, by assuring this level, it could be
expected that a cable design stress EL of EL(AC)=40 kV/mm for AC and EL(imp)=80 kV/mm for
lightening impulse would be adequate.
4-2. High performance materials and defect removal technology
The following technologies have been utilized to the maximum degree in currently produced
XLPE cables to assure the above mentioned permissible defect level. From the resin
manufacturer stage to the cable production stage, we can point to such developments as the
reduction of impurities in insulation through such measures as meticulous cleanliness, completely
closed systems, the development of finer screen meshes and compound manufacturing control; the
reduction of protrusions through similar measures for semiconductor compounds and three-layer
simultaneous extrusion methods; and, in addition, the reduction of voids through dry curing.
4-3. Manufacturing control technology
Along with these exhaustive contamination prevention measures, efforts have also been directed
toward verifying that there are no defects that could affect raw materials and then lead on to
problems in the products. In order to control impurities in raw materials, examinations, such as
inspections by the resin manufacturer at the shipment stage and laser inspections at the
compounding stage, are being performed. The manufacture of products at the prescribed level is
being verified through the introduction of an impurities inspection device for 500 kV[4] cables, and
through the adoption of devices that estimate the maximum defect on completed cable by a
branching method[2]. In addition, pressure and temperature controls are exerted on the cable
manufacturing process and controls are being performed at the manufacturing processing level to
see that harmful voids do not occur.
4-4 Reassessment of the design parameters
Cable design parameters include such things as the temperature coefficient K1, and the
degradation coefficient K2, as discussed in Section Two. These important parameters are
periodically reexamined in the course of the analysis and evaluation of various data that has been
collected from assessments of actual cables at public forums whose membership includes electric
companies, manufacturers, and distinguished professors. The results are then applied to designs.
Table-3 shows various parameters that have been reassessed and the basis for it.
Of these, n=9 was adopted as the conventional (longevity index) for cable, premised on partial

4
 Table-3 Applied design parameters
Former Present Reason for change
value value
AC degradation factor Application of perfect dry curing system,
4.0 2.3 water-proof layer (n=9 → n=15)
Repeated impulse degradation factor 1.1 1.0 Application of experimental data
AC breakdown strength temperature Application of experimental data
1.1 1.2
coefficient (RT/90 )
Impulse breakdown strength temperature Application of experimental data
1.25 1.25
coefficient (RT/90 )

discharge degradation.[6] [7] However, as

is clearly apparent from Fig. 5, the prime

Applied stress[kV/m m ]
6
100
causes of cable breakdown are
predominantly impurities and protrusions; 20
moreover, because the voids in the cable
manufactured by the dry curing system are
not, for practical purposes, of a size that 15

generates partial discharges, the value of 10

the degradation index for XLPE cable can 0.1 1 10 100 1000 10000
be expected to be a large value. From Voltage applied tim e[hrs]
this viewpoint, recent long-term tests (V-t Fig-5 V-t test results on model cable
tests) conducted on cable and evaluations
of the life index have made it clear that, as shown in Fig. 6, close to n=20 may be expected. This
result demonstrates that n=15 can be adopted as a safe level.[2] Furthermore, despite concerns
over degradation of the water tree due to the affect of moisture on the XLPE cable, it is now clear
that the adoption of a metal sheath prevents the penetration of water and the initial amount of
moisture does not play a part in long-term degradation.
4-5. Improvement of accessory performance
XLPE cable lines require not only reliable cables but also reliable joints. Whereas extruded
molded joints (EMJ) and prefabricated joints (PJ) have previously been utilized for high stress
joints, recently rubber block joints (RBJ) have been adopted. Here, we will discuss the improved
performance of the PJ, for a representative joint.
With joint, it is the electrical field at the starting point of the stress relief cone that determines
insulation thickness of the cable. For this reason, the finishing method of the insulation screen in
jointing and the slope angle of the semi-conductive layer of the rubber stress cone are being
studied to improve the electrical performance[8] and to reduce the stress in this point.[9] Stress
was calculated for slope angles of 7º and 5º, as shown in Fig. 6, and it was confirmed that the 5º
angle was able to reduce up to 10% more stress than the 7º angle. Then, based upon this result, a
stress cone with a larger arc (radius of curvature 150 mm) was adopted.

Semi-conductive rubber

Insulating rubber

Slope angle Circular arc R

θ
Conventional shape Improved shape

Fig-6 Differences in electric field at the starting point of stress relief cone depending on cone shape[9]

5
4-6 Historical trend of the insulation thickness and operating stress
Fig-7 depicts the operating records for a high stress cable system developed through tests similar
to those described above. The XLPE cable service for 66 kV (t = 15 mm: 2.5 kV/mm) and 77 kV
(t = 17 mm: 2.6 kV/mm) began 38 years ago in 1965, and today 12 kV/mm, or approximately 4.7
times the stress, is being applied. These values are average stresses used for design and the
maximum stress reaches as much as 19 kV/mm at 800 mm2, which is the minimum size for 500
kV.
14.00
14.
Ave r a ge e le ctr ica l str e ss (k V/m m )

500kV
12.
12.00
Trend of the
Average electricalstress (kV/m m )

Voltage
10.
10.00 insulation thickness
66kV 15→13→11→9mm
220kV
8.
8.00
275kV 77kV 17→15→13→11mm
154kV 23→19→17mm
6.
6.00
220kV 23→20mm
154kV
275kV 27→23mm
4.
4.00
500kV 35→32→ 27mm
66kV
2.0
2.0
77kV
0
0.0
1965
1965 1970
1970 1975
1975 1980
1980 1985
1985 1990
1990 1995
1995 2000
2000

Fiscalyear
Fisca l ye a r

Fig-7 Trend of the AC operating stress and cable insulation thickness

5. Performance verification and operating results for high stress cable

5-1. Concepts on design verification programs in Japan
Modifying design stress, such as reducing the insulation thickness of the cable, requires
comprehensive verification of the cable line system, including the accessories. XLPE cable and
joint test methods in Japan are verified to withstand 30 years of system voltage and overvoltage,
and they are studied with the objective of confirming that their performance will be stably
maintained. The test program has been standardized as JEC 3408-1996.
The tests are divided into the four classifications listed below. The test conditions for the
development test and the type test reflect the insulation design concepts and are determined by
equivalency conversion of the 30 years of projected use, representing the design (longevity), by
V-t rules. Moreover, given the intimate relationship between the defect control level and design
stress, impurities, void, and protrusion tests are implemented by means of type tests and
acceptance tests.
[1] Development tests: In order to verify the (longevity) of the developed cable systems under
the thermo-mechanical conditions in which they are expected to operate, a long term test and
voltage withstand test after long term test are implemented; this to corroborates whether the
design of the developed product, its manufacturing method, and the execution of its
manufacturing procedure could be put to practical use.
[2] Type tests: The lifetime of certain products against a prescribed commercial operation
voltage is verified by means of a one month long-term test or one hour voltage withstand test, and
then, upon completion of the test, a voltage withstand test is performed. For verification of quality
control, detailed investigations on impurities, voids, and protrusion tests are conducted on the
sample to check that the quality of the cables and joints is within the appropriate range. Because
of the critical effect of the jointing method on product quality, tests are performed on assembled
samples under simulated local conditions.
[3] Routine tests: Pre-shipment voltage withstand tests, partial discharge tests, impurities, void,
and protrusion tests are performed on completed cables and testable joint parts, to confirm that the

6
 products to be shipped have been stably manufactured in the same manner as the certified
products.
[4] Site tests: Voltage withstand tests are performed to screen for fabricating defects.
Fig-8 shows the effect of the test voltage on the 275 kV XLPE cable system.

525kV(RT)
Temperature Coeff. 1.2 Type test
440kV(90 )
Safety factor 1.1 Type test (Alternative)

n=15 260kV Development

log Voltage

230kV

System voltage E0(=300kV)/√3=173kV

log Time
1hr 1month 0.5year 30years as Design life
Fig-8 Testing voltage in accordance with V-t characteristics
5-2. Operating results
Fig-8 shows actual operating records for high stress cable systems developed under such tests.
Expressed in terms of average stress, we have 38 years at 2.6 kV/mm, 23 years at 6 kV/mm, and
13 years even at 10 kV/mm. Expressed in terms of maximum stress for the minimum conductor
size, they are respectively 5.4, 10.1, 16.9 kV/mm. As a result of the stringent quality control levels
at the factory and during installation, extremely high reliability has been achieved for actual lines
with just 0.4 incidents for a cable of 66 kV or more/ 100 km circuit/ year; a very low accident rate
of 0.002 incidents/ 100 joints/ year and 0.004 incidents/ 100 terminations/ year for joints and
terminations, respectively, was also recorded. These results appear to corroborate the adequacy
of the design and of the test program.
Aiming at the limits of operating stress
As is apparent from the above discussion, marked improvements in such areas as cable material
technology, manufacturing quality control technology, and accessory design technology have led,
year after year, to reductions in insulation thickness. Nevertheless, it has also been shown that, if
we are able to further reduce the defects present within the insulation and the interface, it will be
possible to further raise the design stress and, if appropriate joints are developed, it will become
possible to make cables even more compact. Research is also now underway to experimentally
determine the stress where degradation does not occur. We would like to introduce some of those
results and the design stress that will be possible in the future.
In XLPE cables that have not been penetrated by water, it is assumed that the degradation of the
insulation occurs at an area of electrical field concentration caused by impurities or protrusions.
A theory has been advanced to the effect that the paramount degradation mechanism is oxidation
caused by the injected charge, the acceleration, and the severing of the molecular chain within the
free volume of the crystal. Consequently, the advance of oxidation degradation within an
electrical field concentration area of the XLPE insulation was investigated and the threshold of
degradation was studied, using XLPE block specimen inserted with needle of various tip curvature.
According to the previously obtained data, which is shown in Table 5, the stress threshold is
thought to be at 60 kV/mm and above. The maximum local stress obtained from the insulation
thickness and impurities control level (approximately 250 µm) at that time in 1965 (now, 38
years) was about 61.0 kV/ mm, considering a minimum size of 80 mm2, and therefore we know

7
that the experiment results shown in Table 5 are appropriate.
It would thus be desirable to develop new technology aimed at further improving the impurities
control levels of purchased materials and the quality control levels in manufacturing plants, as
well as at improving defect control technology at the starting point of stress relief cone in joint.
Table-5 Test condition and test result
Aging Temperature RT 60 90
Already verified Equivalent Years Max. 21.4 years Max. 3.4 years Max 1.10 years
140 V/mm Breakdown Breakdown
120 V/mm Electrical Tree Breakdown Breakdown
Initiation
100 V/mm Oxidation Electrical Tree Breakdown
Initiation
80 V/mm No change Oxidation Electrical Tree
Initiation
60 V/mm No change No change

7. Conclusion
This paper discussed the development of high stress cable in Japan and the test program for
corroborating its effectiveness, and also indicated its appropriateness and actual operating records.
The development of 500 kV XLPE cable has already made the utilization of approximately 12
kV/mm cable with average stress a veritable reality. As a result, we are convinced that it will be
possible for cable to be further compacted through the application and expansion of the
technology that has been cultivated in manufacturing quality control and that is greatly reducing
impurities and protrusions, as well as through the improvement and development of joints.

References
(1) H.Suzuki, T.Endo, et.al., "Experimental Study on Causes of Electrical Breakdown of EHV-XLPE
Insulation using 'Pre-breakdown Partial Discharge Detection Method'", 1990 CEIDP 5-2, 1990
(2) T.Kubota, Y.Takahashi, S.Sakuma, M.Watanabe, M.Kanaoka, H.Yamanouchi, "Development of
500-kV XLPE Cables and Accessories for Long Distance Underground Transmission Line –Part I:
Insulation Design of Cables-", IEEE Trans. on Power Delivery, Vol.9, No.4, pp.1741-1749, 1994
(3) M.Fukawa, T.Kawai, Y.Okano, S.Sakuma, S.Asai, M.Kanaoka, H.Yamanouchi, "ibid. -Part III:
Electrical Properties of 500kV Cables-",IEEE Trans. on Power Delivery, Vol.11, No.2, pp.627-634, 1996
(4) H.Ohno, S.Sakuma, S.Fukunaga, K.Osozawa, H.Yamanouchi, "Construction of the world's first
long-distance 500 kV XLPE cable line", CIGRE 2000, No.21-106, 2000
(5) Electric Technology Research Association's Report, Vol.51, No.1, 1995 [in Japanese]
(6) J.M.Oudin, Ch.A.Flamand, "The use of thermoplastic insulated materials in the manufacture of DC
and AC-Cables", CIGRE 1962, No.209, 1962
(7) F.H.Kreuger, "Endurance tests with polymeric insulated cables, materials and criteria", CIGRE
1968, No.21-02, 1968
(8) T.Kubota, Y.Takahashi, T.Hasegawa, H.Noda, M.Yamaguchi, M.Tan, "Development of 500-kV
XLPE Cables and Accessories for Long Distance Underground Transmission Line –Part II: Jointing
Techniques -", IEEE Trans. on Power Delivery, Vol.9, No.4, pp.1750-1756, 1994
(9) M.Okada, M.Sato, Y.Yamashita, M.Hara, H.Niinobe, "Development of prefabricated joint for
500kV XLPE cable", Jicable 99, A5.2, 1999