21, rue dArtois, F-75008 PARIS http : //www.cigre.

org

(CIGRE-119)

CIGR Canada Conference on Power Systems Vancouver, October 17- 19, 2010

Rejuvenation of Insulation in Aged Power Cables

Wayne J. Chatterton, Ph.D. 1 , Jean Dionne, Eng, MBA2 UtilX Corp1(USA), Transelec Common inc2(CAN)

SUMMARY
Cable rejuvenation fluids have been safely injected into cables for more than 20 years. This process increases the dielectric strength of the cable and removes the water trees plaguing older polyethylene cables. The increased dielectric strength and the removal of water trees leads to a considerable increase in the remaining life of the cable. This paper first looks at the early stages of the injection process and provides comprehensive information about the available fluids. A look at how long it takes the fluid to move from the conductor into the PE insulation is calculated. Next, the basic chemistry of the water tree filling and fluid oligomerization is addressed. The resultant fluid is then tracked by diffusion in the polyethylene to give an estimate of how long the fluid stays in the insulation. This can give a cable owner comfort in the length of time the cable should maintain in reliable service. This cable is looked at from the start of injection to the end of its useful life. One short case study will be documented to highlight the efficacy of silicone rejuvenation technology. The first is a 115 kV class transmission cable from Duke Power,

KEYWORDS
Aging solid dielectric cable, Cable injection, Cable insulation rejuvenation, Electrical tree Injection Technology, Water-reactive material, Water tree

wchatterton@utilx.com, jdionne@transelec.com

1. INTRODUCTION
Injection technology, otherwise known as cable insulation rejuvenation, is a well established option to cable replacement[1]. CableCure is a process where fluids are injected (Dow Corning solutions) inbetween the conductor strands of medium-high voltage cables. The water-reactive fluid than rapidly diffuses into the insulation of that cable. Once inside the insulation, the fluid repairs the damage caused by existing water trees and other dielectric defects. CableCURE/XL also retards the growth of future water trees and extends the life of even badly-aged cable by more than 20 years. CableCURE is a corrective process where insulation material of cables is rejuvenated by injecting CableCure/XL in-between the conductor strands. The fluid then diffuses into the insulation and rejuvenates the insulation (repairs the damage caused by existing water trees and other dielectric defects) over a period of time. The cable once injected is guaranteed for 20 years against dielectric failures. Over the past 20 years more than 85 million feet of medium-high voltage cables have been successfully injected by our fluids and 99% still remain in service and failure free.

2. CABLE INJECTION TECHNOLOGY & PROCESS, AN OVERVIEW

For decades, utilities have used polyethylene-insulated cables for construction of underground electrical distribution and transmission lines. Other types of insulated cables such as PILC (paper-oil-lead cables) are very reliable but higher costs and environmental concerns make the option less attractive. Earlier vintages (pre-1980) have experienced a higher than expected failure rate. Some of these failures are the consequence of inappropriate installation, dig-ins, power surges or other operating related incidents. These types of cable faults can be reduced or prevented with stringent quality controls as well as increased customer and contractor awareness. The degradation of the polyethylene insulation used in most of these cables is by far, the single most important source of cable faults. This premature aging process is caused by a phenomenon known as water treeing. While modern solid dielectric cables are likely to maintain their performance for much longer, the pre-1980 vintage cables are at or near the end of their useful lives. Over the last 20 years an alternative technology has been developed and proven in field installations, which provides a more cost effective approach than replacement of these aging circuits. A silicone fluid injection process has been developed and deployed in medium and high voltage solid-dielectric cables and has provided failure-free performance in over 99.% of the applications. In their infancy, water trees originate from voids, contaminants and imperfections on the inner and outer surface or from the inside of the insulation layer. The degradation is similar to the performance of a steel pipe. As oxidation reduces the burst pressure of a water main, water trees reduce the AC break down (ACBD) strength of polyethylene-insulated cables. As the pressure exceeds the strength of the pipe small cracks appear and the final result is a broken water main. In underground cables, water trees are influenced by high voltage and water. As water trees grow, the ACBD is reduced. In time the electrical stress exceeds the

ACBD and water trees evolve into electric trees. This final state of degradation is irreversible and cable failure is imminent. A fault will occur in a short period of time. Electric trees are micro voids that are the final stage of water trees. They are the consequence of surges, electrical impulses or partial discharge that increase pressure on permeated water trees and alter permanently the insulation. Routine procedures (such as snapping a capacitive charge, bad switching procedures or inappropriate cable testing) if not performed properly may also produce electric trees. These micro-faults cannot be rejuvenated.

Figure 1. Electric Tree

2. 1 DIELECTRIC ENHANCEMENT
Typically AC breakdown performance increases after the injection at a rate of about 0.5% each day. Peak performance is reached after about 24 months at between 200% and 400%. The peak performance is typically very close to what virgin polyethylene cables could be expected to have and falls in the 30- to 40-kV/mm range. Ultimate performance is realized only after the fluid has had sufficient time to diffuse through the entire width of the insulation. While the time that this takes is dependent upon the insulation thickness, the temperature profile of the cable, the morphology of the solid dielectric, and the amount of water present, peak performance is typically realized about two years after the injection is performed. Injection Process Dielectric enhancement fluid is injected into the conductor strands. The fluid rapidly diffuses from the conductor strands into the solid dielectric material. Once inside the insulation, the fluid repairs the damage caused by existing water trees and other dielectric defects. This fluid also retards the growth of future water trees and extends the life of even badlyaged cable by more than 20 years. After years of cable rejuvenation experience it is known that the dielectric strength of a cable increases with the silicone injection process, in fact the AC breakdown strength is moving higher within days after the injection and continues to so for at least two years[2]. This paper will look at a typical 1/0 cable and investigate the cable injection from start to finish as well as the typical chemistry involved. Diffusion, permeability, solubility and retention times are all important in an effective chemical treatment of an aged water-treed polyethylene cable.

3 Cable Injection
For the purposes of this paper, a 1000 foot (304.8 m) run of 1/0 AWG cable with 175 mil thick insulation will be used. It has 19 strands in the conductor and it is assumed to have a normal geometry in the core. The insulation will be assumed to have been extruded over the conductor to make the final cable. Based on this information and using standard AWG dimensions the cable would have an interstitial volume of 2,205.60 cm3. This is the volume of the interstitial region of this cable minus the area of the conductor. The silicone fluid employed in most cable restorations has a density of 0.95 g/cm3. Thus this cable can hold up to 2095.3 g of Silicone. The silicone used most frequently for cable injection is phenylmethyl dimethoxy silane (PhMDMS), Fig. 1. This material has been injected for more than 20 years and into more than 85 million feet (25,900 km) of cable. It has been studied for years and its diffusion properties in polyethylene are known at varying temperatures.
CH3 Si OCH3 OCH3

Fig. 1 PhMDMS Diffusion Properties: 23oC 40oC 50oC 70oC 6.54x10-10 cm2/sec 5.64x10-9 cm2/sec 6.30x10-8 cm2/sec 2.40x10-7 cm2/sec

Based on these diffusion coefficients and armed with the surface area inside the interstitial region of the cable it is possible to calculate how long it takes for the fluid to leave the conductor and be in the insulation. Figure 2 shows the time for the fluid to diffuse into the insulation from the conductor under perfect conditions. This would assume a dry conductor and that only monomer is in the conductor area. If there was bulk water or other contaminates then it will take longer for the conductor to be dry. Thus, at 23oC it would take slightly longer than 142 days for the fluid to leave the conductor and diffuse into the insulation. For this same cable, it would only take 110.2 days at 40oC and only 7.3 days at 50oC to dry the conductor. At 70oC all of the silicone fluid will diffuse into the insulation in less than 1 day. This shows how important cable operating temperature is to fluid diffusion and ultimately cable rejuvenation.

Fluid in 1/0 Cable Versus Tem perature 1000' Section

2500 Fluid in Cable (grams) 2000 1500 1000 500 0 0 50 Time (Days) 100 150 23oC 40oC 50oC 70oC

Fig. 2

4 Rejuvenation Chemistry
Once the fluid has left the conductor and has entered the insulation it can start to do what it was engineered to do react with and remove water. The PhMDMS fluid is polar and diffuses to the amorphous regions of the polyethylene insulation. These are the areas that the water trees will be present. The fluid hydrolyzes in the presence of the water removing the water, drying the cable and leaving behind an oligomer to treat water in the future [3]. The first step is hydrolysis the reaction and removal of a molecule of water, Figure 3:
CH3 Si O" H OCH3 CH3

TiPT Catalyst
OCH3 H

Si

OCH3

+ +
H3C OH

O"H

Fig. 3 Hydrolysis of PhMDMS

This produces a molecule of a compound called a Silanol because it has a Si-OH moiety. A molecule of the methanol by-product is also produced for every hydrolysis reaction that takes place and this MeOH (CH3-OH) diffuses out of the cable. The silanol can then react with another molecule of water to form a silane diol (Si-(OH)2 only under very wet conditions) or react with another PhMDMS molecule to start the dimerization process, Figure 4. The dimer is the first step toward making a large oligomer that will fill the water tree void in the polyethylene insulation.
CH3 Si OCH3 CH3

O'H CH3 Si OCH 3 OCH3

Slower than hydroylsis due to Steric Hindrance

O'

OCH 3 Si

Si

OCH 3 H3C

+
H3C OH

Fig. 4 Dimerization

This dimer is bulkier than the corresponding monomer (PhMDMS) and hence its diffusion is slower in the polyethylene insulation. The dimer also has latent and reactive sites to remove more water if it is still present in the water treed region. These sites are the same Si-OCH3 as in the monomer or they can be the Si-O-Si bridge between the entities. This dimer can be hydrolyzed just like in Figure 3 to form a dimer silanol and then react with another molecule of PhMDMS as in Figure 4 to form the trimer. This process can continue until all of the water is removed from the region the cable rejuvenation fluid is in. The trimerization step is shown in Figure 5 the methanol produced is excluded to make the drawing clearer:
CH3 Si OH O' Si OCH3 H3C CH3 Si OCH3 H 3C OCH3 OCH3 Si OCH3 O' O CH3 Si

CH3 Si

Fig. 5 Trimerization of PhMDMS

The rate of diffusion of these oligomers can be measured in polyethylene at 50oC, for example. These values are shown below. Monomer (1) 6.30x10-8 cm2/sec Dimer (2) 1.10x10-8 cm2/sec Tetramer (4) 6.0x10-9 cm2/sec The diffusion coefficients from above can then be used to calculate diffusion time in the 1/0 AWG cable with 175 mil thick insulation and a 1000 foot (304.8 m) run. The following retention times in the insulation would be expected at 50oC: Monomer Dimer Tetramer 16.8 Days 5.00 Years 32.99 Years

These were calculated from the diffusion rates of the entities listed. This would also assume that 100% of that individual species was in the insulation and calculates how long all of the chemical would take to diffuse out of the polyethylene at 50oC. Mixtures of the different oligomers would diffuse faster or slower depending on the mix ratio. It is also known from Infrared (IR) studies of the polyethylene insulation that higher oligomers are observed. In some cases a degree of polymerization (DP) of up to 8 has been observed (Octomer). These higher oligomers would diffuse even slower than the DP 4 tetramer shown above. The hydrolysis of the PhMDMS was monitored at ambient temperature to determine the level of oligomerization that could be observed by IR or Gas Chromatography (GC) analysis. Thus, one mole of PhMDMS was reacted with 1.9 moles of water (this is below the

stoichiometric 2 moles of water expected for complete reaction) and the following oligomers were observed, Fig. 6.

Time 22 hrs 46 hrs 166 hrs

Monomer 5% 2% 1%

Dimer 7% 6% 7%

Tetramer 27% 27% 20%

Hexamer 10% 12% 16%

Fig. 6 Ambient reaction of PhMDMS with Water

Hence, the presence of the observed oligomers in the 1/0 cable implies that the cable is dry. If there was still water present then only silanols (Fig. 3) or silane diols would be present. The higher oligomers would be hydrolyzed to a silanol if excess water (greater than 2 moles per mole of PhMDMS) were present. The oligomers are much slower diffusers as shown above. It is this slow diffusion that gives the effective long term treatment of the cable. The oligomers stick around for a long time in the water tree void.

5 Electrical Breakdown Strengths

There have been numerous examples of cables tested before and after cable rejuvenation [4]. Cables have been injected for more than 20 years and in all cases the dielectric strength has shown great improvement. An example of a Weibull plot of AC breakdown probability is shown in Fig. 7.

Figure 7: Weibull plot before (red) and after (blue) Chemical Rejuvenation

The cable on the right in Figure 7 (blue) had been rejuvenated for the equivalent of 2 years in the lab prior to having this Weibull plot generated. As can be seen there is a significant improvement in the AC breakdown strength of this cable segment. As the fluid diffuses into

the insulation and reacts with and removes water the dielectric strength of the cable rapidly improves to levels usually seen with new cable of the same vintage.

6 Case Study Duke Power

The Little Rock Retail tap line is a segment of a double circuit line feeding a Duke Power retail substation which crosses the glide path of one of the Charlotte Douglas Airport Runways. The portion of the line that directly crosses the glide path was placed underground. It represents approximately 1,800 double circuit feet of 115 kV class 500mm2 Al, XLPE insulated cable. The original underground installation took place in 1979. The cables were direct buried in a flat configuration 305 mm on center. They were positioned in an envelope with controlled backfill 305 mm above and below the cables. Six phase conductors and one spare cable were installed. The cable had a fabric separator tape, extruded conductor shield, 22 mm of XLPE insulation and an extruded insulation shield. A 51 mm wide semi-conducting rubberized fabric tape was applied over the insulation shield. The metallic shield consisted of a 38 mm-wide, 0.25 mm thick tinned copper tape applied helically. It also included 24 number 14 AWG tinned copper wires. A rubberized fabric separator/binder tape was applied over the wires, and the cable was finished with a 4.3 mm insulating PVC jacket. In 1994, Duke incurred a cable failure on one of the two circuits. The cable was removed from service and sections of the cable were sent to Cable Technology Laboratories (CTL) for failure investigation and an evaluation of the dielectric strength of the cable that remained in service. It was noted that the cable adjacent to the failure had a very low ac voltage breakdown level. It was suggested in that the low breakdown strength was due to a high moisture level within the insulation. The average breakdown voltage was reported to be 278kV (316 MV/m) across 3 samples. Duke Power decided to replace all three phase conductors in the failed circuit and investigate the possibility of injection for the remaining circuit. The new installation involved installing a duct bank for the failed circuit with sufficient conduit for installation of a new second circuit should the need arise in the future.

6.1 Initial Investigation of Injection Efficacy

After the cable evaluation was performed, it was decided to inject the cables with a fluid developed and patented by Dow Corning and injected by cableCure. In order to study the potential efficacy of the injection technology, a section of the cable, as described above, was injected with restoration fluid. The complete cable was then maintained in 68 water for 30 days in order to a ccelerate the fluid diffusion into the cables C insulation. After this conditioning period the cable was cut into three 10m sections. AC voltage withstand testing was performed on each section and compared to the values obtained for the untreated cable. In all three cases the voltage was first raised to an equivalent of 100 MV/m and maintained at that level for 10 minutes. The voltage was then stepped in increments equivalent to 23MV/m (20 kV) every ten minutes until reaching 348kV (395 MV/m). Terminal flashovers at this level required a voltage reduction to 308kV (350 MV/m) where the cables were maintained in each case at that level for 13 hours, after which the test was discontinued. As a result of the voltage limitation it was not possible to break

down the cable and hence to obtain an understanding of the full improvement of the cables performance. In addition to the ac breakdown testing, a 76mm long piece of the treated cable was sent for fluid penetration analysis. The Cable was sectioned into 8 wedge shaped pieces each of which was microtomed to 250 m thick. These sections were analyzed by infrared microscopy in order to analyze the depth to which fluid had penetrated the cables insulation as a result of the accelerated diffusion scheme. Measurements were taken sequentially from just beyond the conductor shield to the outer diameter of the cable. Treatment profiles showed complete penetration through the cable with remarkable radial symmetry. This culmination of data was sufficient to convince all parties to move forward with a cable treatment option as opposed to cable replacement. 550 circuit meters of cable (2,200 cable m) were injected in 1995. This included 3 phases and a spare cable. These cables remained in service until 2004. In 2004 a termination failure occurred, and it was decided to reroute a new circuit through the duct bank which had been constructed after the original failure in 1994. Duke Power made the abandoned 3 phase circuit available for follow up research. The spare cable remains energized and underground for future research.

6.2 Analysis Results

Five 12m sections of the cable were further evaluated. The cable was tested for insulation tensile strength, dimensional coherency, shield stripping strength, dissipation factor, volume resistivity of extruded shields, ac voltage breakdown, water tree presence, and partial discharge detection. Only the ac voltage breakdown is described here as the other data has been previously published. The data that is most relevant towards providing an overall assessment of the effectiveness of the injection is the ac voltage breakdown information. See Table 1. Four of the 5 samples were subjected to a complete ac voltage breakdown test. One of the 5 samples was damaged in termination preparation and failed as a result of that damage. Each of the remaining 4 cables was subjected to the maximum voltage available for the test (440kV) achieving a nominal stress of 500MV/m. None of the cables failed under this maximum available voltage stress level. It is worth noting that after 10 years of life after injection and a total operational lifespan of 25 years this cable demonstrated an ac breakdown strength approaching a new cable value of approximately 700 V/mil.

TABLE 1: AC VOLTAGE BREAKDOWN

Time Under Test (minutes) Voltage Applied Voltage, kV 88 123 158 194 230 265 300 335 370 405 440 Stress (MV/m) 100 140 180 220 260 300 340 380 420 460 500 Sample 1 10 10 10 10 10 10 10 10 10 10 7* Sample 2 10 10 10 10 10 10 10 10 10 10 7* 10 10 10 10 6* Sample 3 10 10 10 10 10 10 10 8** Sample 4 10 10 10 10 10 10 10 10 10 10 9* Sample 5 10 10 10 10 10 10

No breakdown. Maximum possible voltage reached. Test interrupted due to numerous flashovers. ** Failure located at cable insulation, inadvertent cut made during cable preparation for tests.

Infrared microscopy was employed to analyze the diffusion distribution of the injected fluid in the cable insulation (Figure 8). The demonstrated concentration and distribution pattern is extremely good. The fluid concentration contained in the cables insulation is nearly exactly at the target levels. The distribution of the fluid was also nearly perfect with close to steady concentration values from the inner diameter all the way out to its outer diameter. This cable was very well treated and shows fluid is still present even after more than 10 years of treatment.

% CableCURE fluid

Target Concentration

00 0. 00

10 0. 00

20 0. 00

Fig.8 % Injected fluid/XL to Polyethylene vs. Position from I.D. to O.D.

30 0. 00

Location from Strand Shield (inches)

40 0. 00

50 0. 00

60 0. 00

70 0. 00

80 0. 00

90 0. 00

00 1. 00

7. CONCLUSIONS
Although a significant amount of transmission cable footage has been injected with rejuvenation fluid over the last decade, the amount of footage is small compared to the footage of distribution-class cable that has been injected. This paper reports some results of injection of transmission class cables.. The results agree well with injection results of the well studied distribution cables. Transmission cables can clearly benefit by injection technology in the same proven way as distribution class cables. The unique chemistry of phenylmethyldimethoxysilane makes it an ideal candidate for cable rejuvenation. The silane seeks out water treed regions of cables and desiccates the water. The chemistry is safe and effective while at the same time enhancing the dielectric strength of the cable. A small amount of fluid can treat a long length of cable in a short time. The higher the operating temperature of the cable the faster the diffusion of the fluid in to the insulation and hence, faster reaction with the water in the amorphous regions. Once the fluid is in and has started reacting with the water present, oligomers form which stick around for a long time. This longevity of the oligomers in the insulation provides effective long term treatment of the cable.
References
[1] [2] [3] [4] See for example, Proceedings JiCable 2007, Papers C.7.1.1 to 7.1.13 (pages 714-768 for state of the art studies on Water Treeing and Reliability, Versailles, France (June, 2007) Mokry S, Chatterton W, Sibbald P, et al. Cable Fault Prevention Using Dielectric Enhancement Technology. Proceedings of Jicable, 1995. W. Stagi and W. Chatterton Cable Rejuvenation-Past Present and Future Proceedings JiCable 2007, paper C7.2.14, Versailles, France (2007), pages 858-861 Those interested should contact www.utilx.com for examples in the US or http://www.transelec.com/act-cure-eng.php for examples in Canada.

BIOGRAPHY
William R Stagi BSME is V.P. of Engineering for UtilX Corporation. He is an IEEE member, and active as an ICC sub committee member. He has 12 years of experience in the field of injection technology. Wayne J. Chatterton, Ph.D. is Director of Technology of UtilX Corp. He is an IEEE member, ICC Sub Committee member and is a member of ASTM. He is formally from Dow Corning Corp. and spent 14 years on the cable rejuvenation team. He is one of the original pioneers in cable restoration. Jean Dionne, Eng, MBA. is Director of Special Projects at Transelec Common Inc. in charge of cable rejuvenation and characterization for the Canadian market.

10