A Non-Destructive Method For Testing Underground Distribution Cables Up To 25kV

By Claude Kane, Westinghouse Electric Corporation
In an endeavour to reduce the number of in-service cable failures, maintenance personnel have historically used high voltage DC testing. The use of high voltage DC on older Paper Insulated Lead Cable (PILC) was perfectly acceptable and it was assumed that this would also apply to testing the newer polyethylene extruded cables. This assumption has turned out to be false.

A common problem found in Cross-linked Polyethylene (XL-PE) cables is the formation ofelectrochemical or water trees in the insulation. Contaminates, protrusions and irregular surfaces commonly found in extruded cables, coupled with moisture, provide an ideal environment for growth of these trees. It has been found that the AC breakdown strength of treed cables is greatly Laboratory testing of both new and service-aged XLPE cables have revealed two interesting facts:

DC maintenance testing at current recommended DC voltages is inadequate to make a judgement as to the cables condition.
Testing at 4-5 times currently recommended DC levels is required to expose even the weakest insulation.

This explains why many cables have "passed" a high voltage DC test, but fail shortly after being returned to service.

An alternative to high voltage DC testing of cables has been developed that is non-destructive in nature and not only determines the extent of deterioration but also determines the location of the deterioration.

A van has been manufactured that houses all the necessary equipment to perform the tests. There are two primary diagnostic tests performed they are:

Testing for the presence of partial discharges. This test will detect the magnitude and thelocation of insulation defects that have active insulation degradation at operating voltage. Voids in the cable insulation, in terminations and in splices will all be identified.
Measurement and location of voltage conditioned induced impedance anomalies. Anomalies are items such as water trees. An anomaly is a change in the cable surge impedance after applying a conditioning voltage. Anomalies are calculated in percent of surge impedance change before and after conditioning.

A specific pattern of voltage conditioning is applied to the cables. At selected points during the testing process either partial discharges are detected or a time-delay reflectragram (TDR) isperformed to measure the cable surge impedance.

Test Procedure
An initial reflectragram (RG) is made. This is the baseline condition for measurement of anomaliesand also identifies the beginning and the end of the cable as well as location of all splices.

Then 8 kV AC is applied to the cable for a specified period and partial discharges (PD) aremeasured. 12 kV AC is applied for approximately 10 seconds to try to induce partial dischargesand then the voltage is reduced down to 8 kV and PD is measured once again.

The next step involves applying 4 kV AC with a low power-12 kV superimposed pulse on thepositive half cycle of the waveform. Both partial discharge and reflectragrams are made.Finally an 8 kV AC conditioning voltage is applied and final measurements are made.Controlled experiments have shown a high correlation of anomaly detection and location withactual AC breakdown of the insulation. Figure 2 shows a graphic of a section of cable whereanomalies were identified and corresponding failures.

The key here is that where anomalies were identified, there were failures. Where no anomalieswere not found, there were no failures.

Since May of 1994, 310 cables have been tested. The table below shows a breakdown of type ofcable tested:

Breakdown of Type Of Cable Tested
Type of Cable Quantity

PILC 30

XLPE 265

PILC/XLPE 6

EPR/XLPE 9

**Breakdown of Type Of Cable Tested**
Type of Cable	Quantity
PILC	30
XLPE	265
PILC/XLPE	6
EPR/XLPE	9

Typical Results Of Cable Testing
Legend for charts (left):
PD [pCl] - Level of partial discharge in pico-coulombs
an [%] - Anomaly location and the percent change in surge impedance
- Splice Location
=END - End of cable
Feeder: PMH11-264 to Disconnect # 1212
Cable Length: 647 m (2,122 ft.)
Cable Type: XLPE/500/AL

Measurement And Location of Anomalies and PD Sites

Conclusions Phase A:
Small surge impedance anomalies at 1,263 feet. Splice at 1,095 feet has a very high level of partialdischarge. Recommend replacement of splice.

Conclusions Phase B:
Surge impedance anomalies were not observed. Partial discharges were noted at 24 feet. Werecommend this section of cable be replaced.

Conclusions Phase C:
Several small surge impedance anomalies were observed. Partial discharges were also observed.Recommend replacement of cable PMH11-264 to splice at 1,095 feet.

A summary of results to date show some enlightening facts.

90 per cent of PD sites involve Splices and Terminations.

XLPE cables with large water trees do not necessarily produce PD at the test voltages used.

Relatively new XLPE cable have no impedance anomalies.

Head-end sections or sections connected to overhead lines have more impedance anomalies thanother sections.

Summary
It has been found the testing XLPE cables with high voltage DC does not necessarily prove thereliability of the cable. Often, improper DC testing can cause in-service failure of a cable. A newnon-destructive testing service has been introduced by DIACS International and WestinghouseElectric Corporation that can identify the magnitude and location of active insulation degradationand large water trees.

Several utilities have benefited from this type of cable testing. Not only are they able to determinethe extent and location of defects and assist in reducing unplanned outages to key customers, theynow have a basis to proactively plan a cable replacement program. In the past decisions weremade based on age, past failures and other subjective data.

Today, these decisions can be made using a non-destructive testing method with objective data.By prioritizing cable replacement decisions one can defer certain cable replacement costs whichcan add up to millions of dollars and help ensure reliable service to their customers.

Claude Kane is with Westinghouse Electric Corporation, located in Minnetonka, MN.