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.
Tanaka then asked all those who had made presentations at the various sub 5 task groups to have their materials submitted to the task group chairmen by the deadline set by the Task Group chairmen. The Task Group chairmen have been given a deadline by the Sub-Committee chairman who in turn has a deadline set by the ICC secretary. Presenters were reminded that the slides or overheads be shrunk to four to a page and that the logo should not appear except on the first slide.
Chairman Tanaka then said that he had organized a panel to discuss the points brought up in a letter received from Jack Lawson. The letter from Lawson follows:
Recently I was asked to comment on proposed changes to distribution (medium voltage) cables for a large utility. The proposed changes involved 1) decreasing the insulation wall thickness below the industry standard, 2) increasing the conductor operating temperatures to 105/140 degrees C., and 3) change of insulation material. Also involved was the important consideration of estimated life. It became apparent that a major problem is the lack of technical data available to utility engineers upon which to make prudent decisions of this kind. Since the changes described above are typical of those which are and will be made by utilities for years into the future, this brought into focus the critical questions:
Following are detailed comments on 1) ACLT testing, 2) a proposed "interim" test, and 3) evaluation of "successful cables."
This proposed test category would use the same typical conditions as experienced on utility distribution systems - with one exception that is, the applied test voltage would be higher than the rated cable voltage but low enough to assure that the deterioration mechanism would be the same as experienced under service conditions (e.g. 1.25 - 1.5 x rated voltage). The cables would not be preconditioned and the number of temperature test values would be limited.
No preset time limit is envisioned. Cable samples could be removed and evaluated at appropriate times (as determined by you folks). It would seem, however, that the most useful data would be obtained within about 15 to 20 years. Monitoring these cables would provide a source of insulation performance data with time - providing additional "points on the curve" to compare with initial values and with ACLT test results. For those utilities who presently are satisfied with their cable selection, these cable samples would provide valuable information for future decisions when new or significantly modified insulations are being considered for utility use. For those utilities who wish to make a cable change in the near future, test results on these sample cables would, in the future, provide a reliable "heads up" due to the moderately increased voltage.
This "interim" type testing would fill an obvious large (time) void between the ACLT tests and end of service life.
After several years of field experience, there usually is an interest in reducing the cost by decreasing the insulation wall. The moderately increased voltage would provide valuable data to justify (or not justify) a future insulation wall reduction.
Evaluation of these samples can confirm (or cause to readjust) the original estimated life.
If selected now, test conditions could be standardized, thus avoiding the present problem with ACLT type testing.
If such testing had been initiated 10-15 years ago on TRXLP and EPR insulated cable, information would now be available which would greatly improve confidence of life prediction.
Cost - If restraint is used when selecting the number of test values (matrix size), the added cost could be minimal. For example, on a footage basis, URD (residential) cables are 70-90% (depending on the utility) of the total usage. These cables rarely exceed temperatures above 30-40 degrees C. Only one temperature and one voltage would be required for these type cables. Just room temperature may suffice. Preconditioning would not be required. Feeder type cables could be limited to two temperature test values. Dry tests, of course, would not require tanks. Could some of the test facilities presently used for ACLT testing be used for this type of longer time testing.?
Data obtained from cables removed from service can provide similar information. (e.g. Houston L&P), and is very helpful. If all such existing data were assembled and presented in one document, it could be very informative.
EPRI WO2713-15 is an example of both field aged (four utilities) and long term laboratory simulated aged cable. Such testing should be very informative.
However, relying on data from cables removed from service presents some problems. If the cables are directly buried, they are known to be subjected to some moisture (let's assume only jacketed cables would be of interest), but removal is very costly and the cables would be subjected to damage during removal. If removed from ducts it is much less costly, but the very important long term wet/dry history would not be known. Historically it has been difficult to convince utilities to remove cable - especially if the cables are performing well. The incentive for accepting this removal expense has usually been to estimate remaining life of "bad actors." Large volumes would be required to compensate for the assumed or unknown field conditions.
One of the major reasons for the lack of support to continue this cable failure report was the argument that "we know that the cables are failing, why continue?" As can be seen from attachments 1 and 2, clearly, directly buried unjacketed HMWPE cable was a failure. Just as clearly, jacketed XLP in duct was/is a success. (We learn from our successes as well as from our failures.) The failure rate for example, of jacketed XLPE (especially in duct) is very low. If we assume that the rate had continued for the next six years (to date), the anticipated and desired life of 30 years will have been exceeded (cables of this design were first installed in 1964-65). Why was this design successful? Data from the oldest (30 plus years) cable should be very informative. Has the effectiveness of a jacket been underestimated? Does the jacket (even PVC) act as a sufficiently effective filter? Does the jacket tend to contain the residual by-products of crosslinking which affected life? Does the insulation contain many bow tie and vented trees and yet not fail? If so, why? Does the ACBD voltage plateau? Is there a linear relationship between wall thickness and cable life? How much longer than 30 years will these cables last? If this "old" cable performed this well, how might the greatly improved (cleaner insulation, smoother shields, LLDPE jackets) cables perform? Could a thin non-metallic inexpensive moisture barrier sufficiently extend life? Could a justified thinner insulation wall pay for such a barrier?
The oldest successfully operating cables can be readily identified. For example, thirty year old cable can be located by locating 30 year old shopping centers or industrial parks or off circuit maps.
As has been said many times - the "ideal" aging test is one which involves a large volume of the cable subjected to typical utility environment over a long period of time. Isn't the experience of these cables then the "ideal" aging test?
It is probably unrealistic to expect the Cable Failure Report to be resurrected. However, perhaps some of the utilities which contributed to the AEIC Report may still collect this information and would be interested in contributing to a reduced scale survey.
As a minimum, it should be possible to keep track of the performance of those cable designs which have realized or exceeded their desired life. Collection of this type of information need not be extensive, a sampling approach could be considered. Utilities which are not experiencing failures don't usually keep records - it's costly. However, many small or medium size utilities or divisions of large utilities have this information, not in precise numbers but in knowledge of whether any significant failures are occurring or not (i.e. the person responsible for cable maintenance). For purposes of evaluating the success or failure of a particular cable design, this degree of accuracy should be adequate.
ICC precepts state that standards shall be "developed and disseminated with the purpose of benefiting industry and society as a whole, not as a service to individual commercial interests." Broader base, more reliable data limits speculative claims which tend to encourage "individual commercial interests."v In addition, utility engineers must consider the "practical application" (see ICC Mission Statement) of cables in circuits, not just the cables in isolation. Under a competitive situation, these circuits will be loaded to the limit. Two conditions which must be considered are "thermal runaway" (due to backfill limitations) and splices (due to he reality of field made splices). Extending conductor temperature limits to higher values should be approached with caution. This problem presently exists even with the present 90/130 degree ratings. Heat sensing instruments to detect overheated splices are in common use.
In addition, the following items affect the utility engineer's selection process.
The subject of extruded dielectrics under ac stress is not part of the curriculum for EEs. The accumulation of knowledge of insulated cables is a "learn on the job" process - it takes time. Rotation of job responsibility is prevalent. Downsizing results in fewer engineers dedicated to this subject.
The highly competitive utility environment encourages taking additional risk (better, faster, cheaper). The flip side, of course, is that accepting additional risk increases the chances for excessive maintenance with attendant severe economic consequences.
In short, utility engineers need all the help they can get (we always did).
The forgoing are some suggestions to add to the data base to assist utility engineers in making these critical decisions. It is hoped that these suggestions will generate some discussion by those of you responsible for research in this area.
Carl Landinger of Hendrix Wire and Cable was the first speaker. Carl identified himself as a manufacturers representative with some experience in research and development, but pointed out he had also worked as an electric utility engineer and as chief engineer for a consulting firm providing services to electric cooperatives.
Carl first pointed out that a cable had to have an adequate life for whatever purpose it served. Other than purely for the longest life, cable might be selected for preferences in handling and installation, availability, special application requirements and even commercial arrangements. Cables which have the desired properties at an acceptable price are selected by the utility engineer. Cable manufacturers offer new and improved designs to fit the need. A new design might be offered to overcome poor service experience with a previous design, or a reduced wall design to utilize existing duct space, or a higher temperature design to increase capacity. In other words, new designs are offered to solve problems expressed by the users. Under these circumstances, the luxury of waiting for 15 years of testing is not available.
One has to have some perspective regarding testing. If there is a design which has had a 30 year history of good service, and if under a well selected battery of short term tests an improved design shows some improvement in properties which could increase life, do we have reason to believe this will not translate to some improvement in long term service? The larger question is whether, discounting for the moment the propensity for manufacturing defect or damage, a properly selected series of short term tests have lied to us in the past? As long as the complete design as produced and tested with accelerating factors actually encountered in service is involved, we have indeed had reliable answers.
We also have some excellent long term field aging tests in process and/or reported results. EPRI 2713-02 has been evaluating jacketed and unjacketed XLP, TRXLP and EPR cables for 9+ years, Houston Lighting and Power over seven years on jacketed XLP, TRXLP and EPR designs, Memphis some 11 years on several designs, etc. And, for better or worse, we have some 30+ years of utility service. The utility experience raises a very important point. Utilities have had widely different results with essentially the same design. This points to the fact that no universally acceptable design will ever evolve.
What then about long term 15+ year tests? It is already demonstrated we will obtain widely different results on the same design depending on the parameters used for aging. Which will be correct for the specific utility system involved? The devil is in the details, and the test details will, depending on the results, be challenged by one faction or another, along with the results. In order to attempt to reduce the challenges, the tester will increase the complexity of the monitoring of the test parameters, resulting in a test of the longevity of test equipment rather than the cable tested.
A great deal of both long and short term testing (including field service) suggests that failure due to general aging is no longer a factor within the time of interest to utilities. Rather, manufacturing defects, installation damage, and unusual system events control the failure rate of modern designs. This is supported by the presence of outliers encountered in both short and long term tests. Recognizing that failure rates are measured in failures/100 miles/year, and tests are conducted on "100s" of feet of sample, it is likely that outliers, not general aging now control the failure rate. This raises the question of sample length needed to achieve meaningful results in any "long term sample" test program. A further complication is batch to batch variation and facility to facility variation in both materials and finished cable. How many replicate samples are needed?
The final concern was that there was not even agreement as to what constituted an acceptable failure rate. Cables do not fail in unison after some specific time period, but follow the familiar "bell shaped curve." This curve is greatly influenced by the details of the utility installation, operating, protection, and even replacement policy.
It was Carl's contention that excellent work has been done and there is sufficient information for the informed utility engineer to make his selection on whatever basis he was using to do so. Remembering once again, adequate cable life in the application, being only one factor.
Mark Walton of BICC was then introduced as a representative of a testing facility. Mark noted that there is more needed than accelerated life testing to make sound judgments on cable life expectancy. Insulation type, stress level and temperature levels must be considered. He also noted that there is quasi-accelerated aging going on in the field. (See Appendix 5-N-1)v It was noted that ACLT testing is very complex. There is no general ideal testing condition. Models for different cables are not the same. Questions asked about ACLT are: 1) Does it duplicate the aging mechanism? 2) Can the result be gotten in a short time? 3) Does it duplicate field aging conditions? and 4) Can ACLT be used to predict cable life? The answer to the last question is yes, it is possible to do this. At present the testing has been developed to predict cable life for XLPE. Models for different materials are different. For example, the model for supersmooth semicons is different from the old style semicon. Models for different cable designs are not the same. A single ACLT will not fit all the models.
Despite the fact that ACLT procedures have only been fully developed to predict cable life for XLPE insulated cables, there are other contributions made by ACLT. It has played a role in understanding the aging process. It is capable of making materials comparisons. It is able to evaluate design considerations such as filled strand, jackets, etc. The major advantages to ACLT are that the conditions can be controlled and that the changes can be monitored. Because of the accelerated conditions, results can be had in a relatively short time.
An alternative to the ACLT is the Interim Test Program. This is a great idea, but not really novel. Field tests at slightly elevated voltages are more difficult to monitor than ACLT. However, the tests are less expensive than ACLT. It is less precise because there is no temperature control. It is being used to evaluate elevated stress in the field and/or reduced wall thickness. The existing programs are at BICC and Orange and Rockland, both sponsored by EPRI. At Orange and Rockland, five EPR cables and one TRXLPE cable are being evaluated. v Lawson's suggestion for long term field data is a great idea. The difficulty is to find a way to implement it. In any consideration as to cable specifications, one factor which should be considered is the total cost of owning the cable.
Michael L. Walker of Houston Lighting & Power Company was next in commenting on Lawson's letter. His comments follow:
"I first would like to commend Jack Lawson on his paper. It provides a very good analysis of the issues facing utility engineers when making decisions concerning the most cost effective cable construction or constructions to be used within their company.
"I agree with Mr. Lawson that a lot of cable decisions are made based as much on perception as they are made on technical data. I, also, agree that there is not enough field data available to determine how a particular type of cable construction will perform in service over a long period of time. That is why my company invested in retrieving field aged samples of various types of cable constructions from our service area and have evaluated their performance. This has enabled us to make decisions based on technical data pertaining to our service environment. Unfortunately, the performance of a particular type of cable construction in HL&P's environment does not guarantee the same performance in another utility's environment. With proper planning, field trials can be conducted with minimal cost and enough information can be obtained about the service environment to make comparative assessments. I would like to see all major utilities take the same approach and consolidate that information into one data base. If the utilities attending this meeting took the same approach as HL&P, I think meaningful information could be obtained within five to seven years as opposed to 10 to 15 years with the approach postulated by Mr. Lawson.
"I do not believe it is prudent, at this time, to rely on ACLT type testing as the only measure to predict service life. ACLT type testing does provide a good comparative diagnostic. The EPRI project 2713, I think, has the potential to provide a reliable mathematical model to predict service life of a cable with a high level of accuracy. When the field data is obtained and statistically incorporated into the current mathematical model developed in 2713, at that time, the industry may have a reliable tool to evaluate service life using ACLT type data.
"In summary, I would add to Mr. Lawson's thoughts concerning the challenges facing the utility engineer. Downsizing, deregulation, and mergers are forcing engineers to make decisions in quick time frames with the key driver being up front costs. This environment will require the engineer to maximize the opportunity to gain as much technical information as possible at forums such as the ICC and we have the responsibility to provide that information in a scientifically sound and unbiased format."
Chris Fletcher, Duke Power, opened his talk by saying that a utility engineer needed a crystal ball.
In the past, a utility engineer told the management that jackets are good. Management said "fine, go buy jacketed cables." The engineer told the management that smooth shields are good. Management said "fine, go buy smooth shields." The engineer said that TRXLPE is good. Management said "fine, go buy TRXLPE." Now, they are no longer saying "that's good, go buy . . . " The engineer now has to justify the price of each item to the purchasing agent. If a cheap cable is available, the engineer is asked why it is necessary to pay more. The engineer needs data. There is a need to group together to compile information. There is a need to make sense for the technology requested. There is a need, not only for good test data, but the ability to move quickly. There is only a short time available to make decisions. Utilities no longer have the luxury of long term testing.
Fletcher noted that over the last five years, there are fewer utilities represented at the ICC meetings. Not as many utilities send people as in previous years. This is probably another manifestation of the desire by the management to cut costs. We are now at a point where testing, which is as emotional an issue as the definition of TR in Hartlein's working group, has to be short term to provide information quickly.
Harry Orton presented reactions to Lawson's letter from the consultants point of view. If he were to be asked as a consultant what insulating material should be selected for a cable, he said that he would base his response on the laboratory test data, the field experience data, and the non-technical factors such as perceptions and politics. He pointed out that it is difficult to reproduce field conditions with laboratory tests. Laboratory tests are not always reproducible, are expensive, and time consuming. Sometimes laboratory tests will favor one material or another. For example, silicone rubber does not stand up well in tracking wheel tests.
In general, it is difficult to predict life. There are an array of laboratory tests available. There are tests involving plaques, there are tests involving miniature cable samples, there are tests for full size cables such as ACLT, AWWT, there are termination tracking wheel tests, there is the IEEE Std 404 for splices, and there are electrothermal tests on cables and accessories. Despite all of these tests, life cannot be predicted with a great deal of confidence, only confusion can be achieved.
Lawson is right in indicating that the proof of the pudding is in the eating. The AEIC data and the NELPA data were useful. Neither data base is being continued although WEPI is continuing some of the NELPA activities. Although caution is advised in interpreting data, there have been some notable successes. For example, the advantages of jacketed cables over the unjacketed was made clear. The improved life when there is lower electrical stress seems to be a general rule. The cable in duct seems to perform better than those which are direct buried.
It would be great if EPRI and/or the utilities were to conduct surveys. Both good and bad performance should be reported. This will enable an overview of why some cables perform well. An annual reporting of data will help the utilities to improve reliability. Political, management and cost barriers will have to be overcome.
In conclusion, Lawson's call for a data reporting system is valid and sorely needed.
Comments from the audience followed.
