The ICC Website is changing. As this transition is made, our new format can be found at www.pesicc.org/ICCWP. Please send any comments or questions to mvh@voncorp.com.
John Tanaka, Chairman, University of Connecticut, reported that the Insulations Subcommittee met at 10:10 am on Wednesday, April 8, 1998 with 154 members and guests signing the attendance sheets.
Chairman Tanaka called the sub-committee 5 meeting to order at 10:10 am. He asked whether there were additions or corrections to the minutes of the November 5 St. Pete meeting. Hearing none, he declared the minutes to be approved as distributed.
Tanaka also announced that Joe Dudas, the assistant sub-committee chairman could not attend and that John Densley would be helping with the meeting in Joe's place. He also announced that the IEEE has imposed a $10 fee for joining the Standards Society in order to be able to vote on any of the Standards. In effect, the IEEE has imposed a poll tax.
Paul Caronia of Union Carbide made the first presentation on "XLPE Materials for Extruded High/Extra High Voltage Cables." He pointed out that there have been many new HV/EHV installations worldwide. He then listed the many electrical, technical, political, social and environmental reasons for the growth of the HV/EHV cable installations. (See Appendix 5-G-1). In Appendix 5-G-2, the differences between medium voltage cable and high voltage cable are listed along with the cross section of the construction and the electrical stress across the insulation. This same page lists the advanced cable manufacturing technology needed for the high voltage cable production. Appendix 5-G-3 shows the slides listing the key properties needed for the insulation and semicon materials, the trends in the insulation cleanliness, the characteristics needed for the conductor shields, and the comparison in the semicon surface smoothness for the EHV cable in comparison to conventional semicons used for medium voltage cables. Appendix 5-G-4 shows two slides indicating the laboratory tools developed to characterize materials to be used for high voltage cables.
Serge Pelissou of Hydro-Quebec asked how Union Carbide managed to produce material with the needed properties. Caronia replied that the techniques were proprietary.
Laurence Jones of Nova-Barealis commented that Borealis manufactured nearly all the materials for high voltage cables produced around the world, but he was not aware of a semicon shield on EHV cables. Which cable design was the presenter showing? Caronia said that these were cables produced in the U.S.
Jim Medek, consultant, asked for recommendations on how supersmooth could be specified. Caronia said that the material could be specified, in general, to a specific manufacturer or that possibly the ICC could work on a definition.
Steve Swingler, National Grid, commented that, in a discussion of intrinsic strength of polyethylene, it is important to carefully define the polyethylene being measured. One polyethylene is not always the same as the next polyethylene. Dielectric strength is also a function of the electrode used and the area and volume of the sample. Because of the differences in polymer morphology, a sample 75 microns thick can give results quite different from a sample several mm thick. In addition, the electric strength value quoted by Caronia seems reemarkably low.
Steve Boggs, University of Connecticut, commented that Swingler's observation were quite valid. The breakdown strength exhibits a statistical distribution and varies from 200kV/mm to 330 kV/mm.
Ray Bartnikas, Hydro-Quebec, asked to go back to the question on the smooth semicon. Was acetylene black used? Caronia said "yes." Bartnikas then asked about the diameter of the particles of carbon in the carbon black. Caronia said that, off the top of the head, he did not know. Bartnikas further asked whether the carbon had a spherical geometry. Caronia said that he could not answer the question. He commented that the smoothness of the compound was not solely dependent on the carbon black.
Evangeline Cometa, AT Plastics, asked whether acetylene black-based semicons are necessary when in Europe, most EHV cables are made with conventional furnace blacks. Some calbes are even made with reduced wall designs. Is it because they already use compounds with comparable cleanliness to acetylene black-based compounds? She asked if the European engineers could comment on this.
Lars Westling, Nova Borealis, commented that the European 110 kV cables have commercial sheaths. The higher voltage cables have acetylene black semicons.
Ray Bristol, Hendricks, asked whether the development of high voltage cables will lead to cleaner materials for medium voltage cables. The answer was yes.
Eugene Favrie, SAGEM Cables Division, commented that his company had been manufacturing E. H. V. cables of their own design with conventional (furnace black) semicons for a long time with good results. Currently a 500 kV cable with the conventional semicon is being installed in China. Tests are in progress on 9 km of this cable.
Walter Zenger, EPRI, asked whether there is a different compound number for the materials intended for EHV use. The answer was that the compound number is the same, but the following letter is different.
Larry Kelly, consultant, said that he, for one, did not understand how the plaque data relates to cable breakdown in the graph discussed. Caronia replied the calculations in the graph were based on cable dimensions and not plaque data.
John Cooper, PDC, said that 11-25 was being reconvened. The moisture measurement in SF6 standard either needs to be reaffirmed or, if there are changes to be made, will need to request a PAR.
Harry Orton, consultant, announced a proposed effort to reactivate a utilities survey. This will help maintain reliable cable systems. Such data will help improve communications between users and manufacturers. He noted that the AEIC survey is no longer being conducted. The NELPA survey still exists. He pointed out the survey successes such as the clear understanding of the importance of cable performance provided by jackets for direct buried cables. There are unanswered questions today such as the performance of some of the new insulating materials and some of the new cable designs. The group Harry is proposing to carry out this function consists of Al Kong, Push Patel and George Austin in addition to himself. Austin is the statistics expert. In order for the survey to be successful, an annual input of data is required.
Since the last ICC meeting, 83 utilities have been contacted, but are unwilling to provide funding. This is due to the business climate the utilities are facing today. An alternative is to get funding from independent funding agencies. Proposals are being written in order to attempt to secure funding.
Eric Marsden, Nova-Borealis, gave a presentation on the "Mechanisms of Water Tree Retardancy - An Overview." He started his presentation by defining and characterizing bow tie trees and vented trees. In a second slide, he initiated a discussion on the mechanism of water treeing. The basic conditions for water tree growth are (1) the presence of water, (2) a highly divergent electrical stress, and (3) a high divergent mechanical stress. In a third slide found in Appendix 5-H-1, it was pointed out the the mechanisms for tree retardancy is to use some method of dissipating mechanical and electrical stresses. The strategies for dissipating mechanical stresses is to add a rubbery phase to LDPE, either by making a physical blend with a rubbery material of making a blend in the reactor to incorporate a rubbery component. Electrical stress dissipation can be accomplished in several ways. If water is rapidly absorbed, the electrical stress grading effect will be maximized. This can be accomplished by adding hydrophilic materials. Electrical stress can also be accomplished by increasing the dissipation factor. The localized electrical stress can then be dissipated as heat. This is accomplished by adding polar or easily polarizable materials or semiconductive materials. For example, clays or other complex silicates behave in this mode. Finally, it is important to reduce the effect of ions. This can be done by adding complexing agents such as certain silicas or materials containing carboxyl groups. It can also be done by adding hydrophilic materials to reduce the ionic concentration. In conclusion, Marsden pointed out that in order to achieve water tree retardancy, several strategies can be used. In general, multifunctional systems are most effective. Polar ethylene copolymers and/or hydrophilic materials provide some of the multifunctional capabilities. When considering additives, the general considerations are to maintain health and safety, maintain insulation properties, maintain physical toughness, maintain processability, maintain oxidative stability, and maintain compatibility with shields. The slides used by Marsden are found in Appendix 5-H-1 through 5-H-3
Evangeline Cometa, AT Plastics, asked how new additives change the insulation characteristics such as stress grading. If the insulation is uniform, how is stress grading obtained? Marsden said that if the material is uniformly produced, the physical property will, in effect, be smeared out.
Steve Boggs, commented that Marsden's presentation fits well with the theory of water treeing outlined by Zeller in his CEIDP Whitehead Lecture in which he proposes that water treeing is caused by a chemical potential which is proportional to the derivative of the energy stored in the electric field with respect to the conductivity of the water. This chemical potential, which can reach thousands of eV, drives electro-oxidation of the polymer which converts the water tree channels from hydrophobic to hydrophilic. This causes water to condense from the bulk of the PE to the hydrophilic channels and results in water trees being self-propagating. However, the high field at a tree tip can also cause high field conduction, space charge effects (both from ion motion within the water and charge injection into the solid), and electromechanical forces. As a result, a water tree can vary from essentially electromechanical at very high fields, in which relatively little electro-oxidation takes place prior to rupture of the polymer by electromechanical forces, to essentially electrochemical at typical operating fields, where a great deal of electro-oxidation induced damage of the polymer occurs before the relatively small electromechanical forces can rupture the polymer. Early understanding of water treeing was greatly impeded by the fact that most laboratory water trees are grown at very high fields at which relatively electro-oxidation takes place, so that electro-oxidation was not closely associated with water treeing until effort was focused on field-grown water trees. Work by Pace on electrical signal generated by gowing water trees suggests that his water trees grew in increments of 10 to 100 nm. However, Pace did not measure all the parameters necessary for a good estimate.
John Densley, Ontario Hydro Technologies, gave a presentation on the "Partial Discharge Characteristics of Interfaces in Extruded Cable Systems - Influence of Contaminants."
The objectives of the study were to (1) develop diagnostic tools to assess the condition of the interface and predict remaining life, (2) determine critical parameters that affect performance, and (3) understand physical phenomena involved. The equipotential distributions along the interface of EPR and epoxy as affected by air bubble, water drop, paper fiber, and metal whisker were shown as well as the electric field distribution along these same interfaces without contamination and with the same set of four different contaminates. The test sample geometry is shown on the fourth slide in Appendix 5-I-2. Voltages were chosen o model a 275 kV joint. The partial discharge data are shown in Appendix 5-I-3. Pulse rise time, pulse fall time, pulse width, amplitude, and frequency were measured for pulses 1-2 nanoseconds wide. From these measurements, it could be identified as coming from the solid material or from the interface. The data shown in Appendix 5-I-4 and 5-I-5 indicate the difference in PD pulse depending on the nature of the contaminant. That the breakdown path is dependent on the nature of the contaminant is indicated on the diagram in appendix 5-I-6.
In conclusion, the PD pulse characteristics are significantly different in the presence of contaminants at the interfaces. The contaminants significantly reduce the times to failure. The pulse fall times and frequency are slightly greater for the paper fiber than the metal whisker, but the rise times and widths are similar. The amplitudes of the PD pulses are about 50% larger for the metal whisker than the paper fiber. This could explain the shorter times to failure for the metal whisker. The additional studies of EPR and XLPE interfaces are needed.
Ray Bartnikas, Hydro-Quebec Institute of Research, asked about the bandwidth of the oscilloscope and whther it was analog or digital. He also wondered about the nature of the short pulses. Densley replied that two oscilloscopes were used, an analog with a bandwidth of 1Ghz and a digital with a 2 gigasamples/s sampling rate and a bandwidth of 500 Mhz. The short pulses were less than 2 ns duration and were partial discharges in the materials, not along the interfaces. The long tail found in partial discharges in large voids due to the drift of the slower positive ions do not occur. In narrow channels the positive ions move rapidly to the channel walls and this results in a fall time of the same order as the rise time and a pulse duration of approximately 2 ns. Partial discharges along interfaces have similar rise times as those occurring in electrical tree channels but have much longer fall times. Bartnikas then asked whether there had been an examination of light emission. Densley replied that light emission had not been examined in these tests but that it had been measured in tests to studies of electrical tree initiation, i.e., before channels were formed. Bartnikas further asked about the effect of overvoltage on the partial discharge characteristics. Densley said that this had not been examined.
Steve Boggs, University of Connecticut, commented that while he and Greg Stone were both at Ontario Hydro they had measured the light emission and partial discharge characteristics of electrical trees from needle tips in epoxy resin. There was correlation between the elecrical and light pulses that showed that the partial discharges cascaded down the tree channels. The investigation was confined to electrical trees within epoxy and did not involve interfaces with other materials.
Steve Swingler, National Grid., asked about the comparison between paper fiber and metal fiber. With paper fiber, the discharge pulses are broader and more frequent than for metal fiber, but the samples with paper fiber lasted longer. Densley said that he did not think the difference in the pulse widths was significant and the metal fiber gave higher amplitude pulses, which were probably more localised than those produced with a paper fiber.
Chinh Dang, Hydro-Quebec commented that the model is not appropriate for certain constructions where both electrodes are in direct contact with the cable insulation. Densley replied that there is a difference when dealing with transmission class joints and idstribution class joints. Dang then said that the other joint design, with no embedded electrode, was in fact used for transmission class joints as well. He further asked whether one would expect a difference in pulse shape for discharges initiated at the electrode interface as compared to those initiated at the insulation interface? Densley said No. Dang then asked whether the phase resolved partial discharge distribution measurements had been done. Densley siad No.
The Sub-5 meeting was adjourned at 11:50 am.
Return to the SC 5 Homepage
