Putting Hipot Out to PastureAs DC high-potential testing starts to show its age,test methods recommended by IEEE are ready to fill the voidBy Wally Vahlstrom, Electro-Test, Inc.

20 EC&M October 2003

F or many years, high-voltageDC testing has been the tra-ditionally accepted method tojudge the serviceability ofmedium-voltage cables. DChigh-potential (hipot) tests

worked well as a withstand and condi-tion assessment test for paper- and oil-insulated, lead-covered (PILC) cable.Even when plastic insulated cables werefirst introduced in the ’60s, it was still thepreferred method.

Although applicable industry stan-dards began to change in 1994, manufac-turers, testing firms, and standardsgroups still agreed that DC hipot testequipment could be used to perform fac-tory tests, acceptance tests, and mainte-nance tests. As far as the installer or main-tenance electrician was concerned, cablesthat withstood the DC hipot test were goodenough to be placed—or placed back—into service. But recent research and fielddata is calling that practice into questionand raising serious doubts about whetherDC hipot testing might damage or causeextruded cables—especially field agedcross-linked polyethylene (XLPE) insu-lated cable—to fail prematurely.

After receiving reports in the early ’90sthat DC hipot testing could be to blamefor latent damage experienced by ex-truded medium-voltage cable insulation,the Electric Power Research Institute(EPRI) funded two studies relating toXLPE and ethylene propylene rubber(EPR) cables. These studies, EPRI ReportsTR-101245 and EL-6902, yielded the fol-lowing conclusions regarding XLPE cable:

• DC hipot testing of field-agedcable reduces its life.

• DC hipot testing of field-aged cablegenerally increases water tree growth.

• DC hipot testing before energizing

new medium voltage cable doesn’t causeany reduction in cable life.

As noted in IEEE Standard 400-2001,Guide for Field Testing and Evaluation ofShielded Power Cable Systems, “testingof cables that have been service aged in awet environment (specifically, XLPE)with DC at the currently recommendedDC voltage levels may cause the cablesto fail after they are returned to service.The failures would not have occurred atthat point in time if the cables had re-mained in service and had not been testedwith DC.” This standard also indicatesother testing has shown that “even mas-sive insulation defects in extruded dielec-tric insulation cannot be detected withDC at the recommended voltage levels.”

The shift in opinion of hipot testingthat began in 1994 when applicable indus-

try standards began to change is only justnow gaining industry-wide acceptance.Current versions of some of these stan-dards no longer provide for DC hipottesting of extruded cables as a mainte-nance test. Of those that still do, all havereduced the recommended test durationfrom 15 min to only 5 min. None endorsesDC hipot testing as a factory test for ex-truded cables, but all continue to providefor DC hipot testing as an acceptance teston newly installed extruded cable. Theseindustry standards also no longer endorseDC hipot testing as a maintenance testfor extruded cables that have been in ser-vice for more than five years.

Deterioration mechanisms. Partialdischarge (PD) and water intrusion are thetwo age-related deterioration mechanisms

MAINTENANCE FACTS

An IEEE study on water trees shows electrical trees will likely progress to failure quickly, so PDtesting is more valuable if performed along with dissipation factor/power factor (DF/PF) testing.

October 2003 EC&M 21

of most interest for both laminated(PILC) and extruded (XLPE and EPR)cable designs (Table below). Although thesource of PD and the process by whichwater enters the insulation are differentfor both cable types, these forms of dete-rioration have become a primary focusfor cable manufacturers and owners.

PD can occur primarily in voids inthe insulation of cables and cable acces-sories. Within a PILC cable, voids mayresult when the insulating oil migrates—due to elevation differences, cracks inthe lead sheath, or incorrect assembly ofterminations—away from an area withinthe cable. Voids in extruded cable systems

can be caused by several things, includ-ing extrusion problems, improper han-dling during installation, or errors in ter-mination assembly.

Whether water gets in through a crackin the lead sheath of a PILC cable or per-meates through the outer layers of an ex-truded cable, it can result in deterioration.

Deterioration assessment test methods.

MAINTENANCE FACTSMoisture decreases the dielectric strengthof the insulation and provides a path forleakage current or other forms of dete-rioration within the insulation. However,most cable experts are convinced that wetpaper insulation and water trees—namedfor their characteristic visual pattern—in extruded insulation don’t initially pro-duce PD. In their 1997 IEEE paper,“Mechanism for Impulse Conversion ofWater Trees to Electrical Trees in XLPE,”Steven Boggs, John Densley, and JinboKuang proposed that transient overvolt-ages were responsible for converting wa-ter trees to electrical treeing, or the deg-radation of electrical insulation by theformation of conductive carbonizedpaths, which allows for small electricaldischarges. It’s commonly accepted thatelectrical trees do produce PD.

Condition assessment methods.IEEE Standard 400-2001 establishessix field tests for shielded power cable thatcan be broken down into two categories:

withstand testing and condition assess-ment testing. DC, very low frequency, os-cillating wave, and power frequency are alltypes of tests that fall into the first category.

Withstand tests are go/no-go tests,that involve applying an overvoltage andprovide no trend data. A key concept ofwithstand tests is that if the overvoltagedoesn’t fail the cable, the insulationcondition is then considered adequate.To make a crude comparison, these testsare analogous to putting a patient on atreadmill for a predetermined time at apredetermined speed and pronouncinghim healthy if he doesn’t die. Such a testoverlooks the possibility of smaller prob-lems that don’t seem like a big deal at thetime. All withstand tests require the cableto be de-energized, disconnected, andtested with a special voltage source.

On the other hand, condition assess-ment tests involve the measurement ofcharacteristics of the insulation. A keyconcept of assessment tests is that the datathey compile can be trended over time to

Online partial discharge testing is one PD testthat will detect and measure partial dischargein Pico Coloumbs. It’s beneficial to conduct itin conjunction with DF/PF testing.

help determine whether and to what ex-tent the insulation has deteriorated. Al-though some methods of performing thesetests involve applying an overvoltage, itisn’t intended as a withstand test thatmight cause a weak spot to fail. Condi-tion assessment tests can be further sub-divided into partial discharge and dissi-pation factor testing.

Partial discharge tests. There are sev-eral methods for detecting and measur-ing PD. Some methods involve de-ener-gizing, disconnecting, and powering thecable from a special voltage source, whileother methods allow the cable to remainenergized at normal line voltage. Bothmethods will detect and measure partialdischarge in Pico Coulombs (PC). Theauthors of the IEEE study on water treesstate that electrical trees will likelyprogress to failure quickly, so PD testingwould be more valuable if performed inconjunction with dissipation factor/power factor (DF/PF) testing.

Dissipation factor tests. When AC

voltage is applied to insulation that’s inpristine condition, the cable performsmuch like a capacitor: capacitive, orcharging, current will flow based prima-rily on insulation type, cable length, andinsulation geometry. Virtually no resis-tive current will flow, resulting in a near-zero power factor. Power factor (PF) isthe ratio of resistive current to total cur-rent. Dissipation factor (DF) is the ratioof resistive current to reactive current. Forsmall values of resistive current, PF andDF are about equal. Any form of deterio-ration that results in an increase of resis-tive current will cause a correspondingincrease in PF and DF.

Since moisture in PILC cable decreasesinsulation resistance and allows resistiveleakage current to increase, the result willbe an increase in insulation PF and DF.Similarly, the presence of water trees inextruded cable will increase resistive leak-age current. The manufacturers of DFtesting equipment claim their products candetect moisture in PILC cable and water

trees in extruded cables. All forms ofavailable PF and DF test equipment requirethe cable to be de-energized, disconnected,and tested with a special voltage source.

Now that the industry has soured onDC hipot testing, cable installers andmaintenance workers have had to beginlooking elsewhere to judge the efficacy ofmedium-voltage cables. The IEEE stan-dard for field testing and evaluatingshielded power cable systems presents sixoptions. Each test has distinct advantagesand disadvantages. Of these methods,four are withstand tests. Only two pro-vide trendable condition assessment in-formation: PF/DF and PD. Of these, onlythe PD test can be performed while thecable remains energized and in service. Forthose applications where it’s impractical toremove cables from service to performmaintenance testing, only the online PDtest offers a practical alternative. EC&M

Vahlstrom is director, technical services forElectro-Test, Inc. in San Ramon, Calif.