DISTRIBUTION:

Underground lines:
Different problems, practical solutions

1. Test trench layout for full-scale field-testing is designed to confirm analytical predictions of cable temperature. The extreme climatic and soil conditions in Arizona provide ideal worst-case testing conditions

Underground distribution systems cost more than overhead systems to install and replace. From an operational standpoint, the two biggest challenges are knowing the actual load capability of an underground line and taking care of aged and aging underground cables.

Avoid heat-related outages

Lack of actual in-service, thermal information about underground cables was a key cause of several outages in the US during summer 1999, according to the US Dept of Energy's Power Outage Study Team Final Report (POST), March 2000. Without real-time information, engineers must rely on typical book ratings, which use simplified assumptions regarding environmental conditions, mechanical and electrical characteristics, and load-carrying capability. This conventional thermal model is most useful when estimating the temperature of single cables. Calculating the power flow and temperature rise for combinations of conductors under varying types of cover requires a much more sophisticated analysis.

Accurate knowledge of thermal conditions along a full cable route helps identify hot spots, typically the limiting point in an underground circuit. By using fiberoptic fiber, installed along the length of the cable (or in a nearby spare duct), a utility can periodically monitor the cable temperature and create temperature profiles that enhance the information available from calculated estimates or conventional fixed-point temperature sensors. Cable temperature profiling using a fiberoptic line and computer software is now available to utilities wanting to get a temperature profile on underground cable runs. The profiling identifies the location of hot spots--allowing cables to be rated to the hot spot. Further, new underground transmission cables are now available from most major cable manufacturers with fiber already in place for temperature sensing.

The distributed fiberoptic (DFO) temperature sensing system is owned by the Electric Power Research Institute, Palo Alto, Calif, and licensed for use to Power Delivery Consultants (PDC), Ballston Lake, NY. The temperature monitoring equipment used with the DFO system is the DTS-80 from York Sensors Ltd, Chandlers Ford, England. An enhanced optical time domain reflectometer (OTDR) injects a laser pulse down a conventional communications type multimode fiber. The OTDR detects the reflections and specialized software develops a temperature profile accurate to within 1C, at roughly 3-ft intervals along the length of the cable.

DFO system suitable for all

2. Both perpendicular and oblique crossings were tested

Because the thermal environment of underground cables stays relatively constant, only occasional DFO measurements are necessary to identify hot spots. Once located, the hot spots can be continuously monitored with a less expensive thermocouple and data-logger.

Nevada Power Co (NPC), Las Vegas, Nev, recently used the DRO to evaluate a new, 69-kV underground cable feed adjacent to the Aladdin Hotel and Casino in Las Vegas.

The underground feeds consist of three 1500- and three 500-ft runs of 2000-kcmil copper conductor, cross-linked polyethylene (XLPE)-insulated, lead sheathed cable. Each cable includes two single mode and two multimode embedded fibers in hollow steel tubes wrapped around the cable just under the lead sheathing, placing the sensing cable less than 1 in. from the copper conductor. The cables are in a concrete-encased, 6-in. PVC duct bank.

Art Davoren, senior project manager, NPC, says, "The 69-kV system was first tested without load to establish a baseline cable temperature profile."

After being energized for 90 days, a second set of measurements was taken, this time while under load. Davoren adds, "A thermocouple was placed some 15-20 ft away from the duct bank at the same depth to measure the earth's ambient temperature. Hourly loads were recorded for 14 days immediately prior to the cable temperature measurements in order to calculate line load and loss factors.

"We found short lengths (where the two cable runs came together and a distribution circuit was about 5 ft away laterally) showing temperatures higher than the design calculations. Earth ambient--particularly under asphalt--is also somewhat higher. The findings resulted in a slight derating of the cable," remarks Davoren.

He cautions that the load readings were taken in December 2000, a period of relatively light load on the NPC system. Additional testing and analysis are planned for the peak load period expected this summer.

San Diego Gas & Electric Co, San Diego, Calif, has also identified and located hot spot bottlenecks on several distribution feeders using the DFO technology. Book cable ratings based on simplified assumptions about loading, cable characteristics, and operating environment are generally conservative. In some cases, additional load capacity (of critical importance as electric utilities deal with the pressures of a competitive marketplace) is available.

Southern California Edison Co (SCE), Rosemead, Calif, used the DFO system to evaluate a scheduled reinforcement project for a 66-kV transmission circuit over 8000 ft long. A temperature characterization of the circuit determined that a 7-8% load increase was possible without compromising safety or reliability. Based on the results, SCE was able to defer the reinforcement project for about two years.

URD not immune

The temperature issue is not limited to transmission cable. Engineers have long made underground cable loading decisions based largely on calculated estimates and past practices--with little knowledge about the actual performance of cables in the ground.

Thanks to the efforts of the Distribution Systems Testing Application and Research (DSTAR) group, utilities now know more. DSTAR is a cooperative research effort supported by utilities and utility organizations. The DSTAR Program Manager Reigh Walling, General Electric Power Systems Consulting, Schenectady, NY, says, "DSTAR projects are directed at solving problems that will improve applied technology." Research focuses on day-to-day engineering and operational issues--such as cable ampacity.

In situations where cables are in parallel for a short distance or cross one another with a separation of more than a couple of feet, the effects on cable ampacity are typically ignored.

Field tests, hosted by Arizona Public Service Co, Phoenix, Ariz, and administered by PDC circulated 300-400 amp through 750 kcmil and No. 1/0 cables both direct-buried and in duct (Fig 1). Testers tried two configurations; short taps into a switchgear cabinet and padmount transformer; and perpendicular and oblique crossings (Fig 2). Engineers also loaded a single reference cable to establish a temperature baseline.

DSTAR found that a significant elevation of cable temperature can occur at crossings of greater than 2-ft separation, and in situations like parallel taps to padmount transformers where cables are laid parallel for a short distance (Fig 3). Testing program manager, John Cooper, PDC, comments, "Crossings could produce hot spots on fully loaded cables as can differences in surface cover, such as asphalt parking lots or roads."

As a result of the testing, engineers developed a method to calculate cable temperature elevation and subsequent de-rating. The software package called XDerate is available to DSTAR members for use in underground system design.

Find fault

3. Situations where cables are paralleled for a short distance, such as entry to and exit from transformers and switchgear, were also tested

Once buried, some might be tempted to forget about underground lines. They should think again. According the POST study, underground distribution system cable failure was a factor in major outages at Consolidated Edison Co of New York Inc, New York, NY, July 1999; Public Service Electric & Gas Corp, Newark, NJ, July 1999; and Commonwealth Edison Co, Chicago, Ill, July 30-Aug 12, 1999. In each case the failure of individual distribution cables led to a cascade of additional failures magnifying the impact of the single original failure.

Indeed underground residential distribution (URD) cable deterioration is perhaps the most onerous of problems with older underground lines. As an inexpensive and better performing alternative to paper-insulated lead covered (PILC) cable, mile upon mile of high molecular weight polyethylene (HMWPE) cable was installed by utilities across the US during the sixties and seventies. Long before expected, failure rates for cables installed prior to 1980 approached 20 failures per 100 miles of installed cable. That is not reassuring, since an estimated 2-billion ft of pre-1980 underground cable remains in service today.

Newer insulations--such as XLPE and ethyl propylene rubber (EPR) were developed to resolve the reliability problems. Failure rates on newer underground cables improved to about 1 failure per 100 miles of cable. But, even with the newer insulations, problems with treeing and neutral corrosion in aging, nonjacketed cable have not gone away.

Cables in the ground, particularly the aged and aging URD cables installed prior to 1980, are a problem waiting to happen. Visual inspections are impossible and conventional direct current high-potential testing may actually aggravate potential weak spots.

KEMA Diagnostic Services, Chalfont, Pa, uses very low frequency partial discharge (VLF PD) technology on distribution cable systems up to 35 kV to locate defects and give an indication of the risk of failure. The VLF PD method discovers and measures the intensity of partial discharges. The method involves sending a 0.1-Hz, high-voltage pulse, 1.5 to 2.0 times the phase-to-ground voltage, down the cable. A discharge diagram showing partial discharges as a function of cable length is prepared and analyzed by a trained operator to determine the significance of partial discharge to cable health.

The test method is considered nondestructive and can handle branched cable circuits and long cables using a multi-ended, global positioning system (GPS). The branched circuit cable testing method uses a GPS for time reference, a computer for data processing, a communications unit to transfer data, and a master/slave unit to control triggering of the remote terminals. The system identifies potential failure points in the cable and accessories and gives maintenance personnel a chance to target suspected problem areas.

Field-testing is also used by the Utilx Corp, Kent, Wash, to determine the location of faults, splices and corroded neutral wires. Recently, Keith Lanan, Utilx, developed a sub-nanosecond time domain reflectometer (TDR) test method. Using a specially designed impedance transition device (ITD), the TDR measures the amplitude of a returning pulse to determine magnitude and location of splices and neutral corrosion. The objective is to decide whether a cable is suitable for rehabilitation. Based on Utilx statistics, not all cable is worth that investment, but for the 60 to 90% that is, there is a fix that will add about 20 years to cable life.

Fix the problem

The fix option (cable rehabilitation) has established an impressive track record over the past several years. Older cables, even those that have failed repeatedly, can be treated with CableCure, a dielectric enhancement fluid, licensed and marketed exclusively through Utilx. The CableCure process has eliminated failures and prolonged service life on older cables worldwide.

Moisture in insulation voids and high-voltage stress contribute to cable deterioration. The CableCure restoration treatment repairs the damage by injecting a silicone fluid into stranded cable. The fluid drives out the moisture and stops the deterioration. The treatment cost is one-third to one-half that of replacement cost.

Actual life extension is with CableCure is estimated at 20 years. CableCure fluid has successfully restored feeder and distribution cables from 15 to 35 kV and transmission lines from 46 to 115 kV. The process has proved particularly effective on URD systems with direct-buried cable.

This widely accepted alternative to repairing or replacing bad cable is being used at Salt River Project, Phoenix, Ariz, where a project to treat some 600,000 ft of underground cable with CableCure/XL is under way. According to Matt Crucitt, CableCure project manager at Salt River, some 250,000 ft have already been treated.

According to Carol Jaeger, senior engineer, Puget Sound Energy (PSE), Bellevue, Wash, PSE has used CableCure injection on some 42 miles of underground lines. Jaeger comments that PSE is in the process of identifying additional areas with stranded conductors and bare concentric neutrals to determine which cables must be repaired or replaced--rather than treated.

—Bill Koch