The secret to not damaging old, in-service cable is to apply the maximum test voltage in small increments and closely monitor the behavior of leakage current.

How do you go about verifying the performance characteristics of an old, in-service, medium-voltage (MV) cable? Do you proof-test it, running the risk (some say) of damaging the cable, or do you do nothing and wait for the inevitable failure? There are pros and cons to both sides of the proof-test argument. If done correctly and with care, this testing can diagnose potential problems without damaging the cable.

The basics. After servicing a cable for some time, many facilities make periodic maintenance tests on it. These firms want to ensure reliability of their distribution feeders and be able to predict potential cable failures. The testing usually involves a knowledgeable electrical contractor, electrical testing company (usually a NETA affiliate), or in-house electrical maintenance personnel performing a DC high-potential (hi-pot) test.

Usually, facilities proof-test every one to five years, depending on the importance of the feeder and the number of in-service and test failures experienced. Proof-testing can weed out weak spots due to:

• Aging through localized corona deterioration and excessive heat;

• Deterioration caused by switching and lightning surges;

• Discharges caused by open or ungrounded shields;

• Ferroresonant overvoltage;

• Rodent damage; and

• Chemical attack.

While this testing doesn't eliminate 100% of potential service failures, it does identify a good portion of them. If service failures continue after testing, it may be the applied voltage test levels weren't high enough. It could also indicate the applied test voltage actually hastened a pending service failure.

You perform proof-testing at lower voltage levels than field acceptance levels, usually 50% to 60% of corresponding factory DC test levels.

If you're proof-testing an existing MV feeder for the first time, make sure you've made preparations for repair, spare cable, and time availability. This is because hi-pot proof-testing can result in cable damage, depending on the age and condition of the cable and test performance.

The secret to successful proof-testing. Beside cable end preparation, nothing else can affect cable testing as much as how you apply the voltage. The best method for voltage application is a slower, step-by-step buildup. This "feeling-out" procedure involves the application of incremental levels of voltage, each held at 1-min intervals. Then, you record the leakage current values for each voltage step. You can use the plotted leakage current values to evaluate the condition of the cable.

Be careful: these step-by-step leakage current readings aren't always proportional to the respective applied voltage levels. This is due to end corona and nonuniform current behavior at these higher voltages.

Interpreting test results. Don't get hung up on the magnitude of leakage current for each step. Instead, focus on its behavior during each holding period. Fig. 2 provides a graphic example of this.

• If the leakage current level remains steady or goes down during each holding period, you can consider the test data normal and acceptable for cable performance.

• If the leakage current rises and continues to rise during any one holding period, there's an unstable condition. You may decide not to continue testing until cable failure.

• If the leakage current rise is slight, hold the applied voltage level until the current levels off or for a maximum of three more minutes.

Preparing cable prior to testing. The more careful you are in preparing the cable ends, the more accurate representation of leakage current magnitude and behavior you'll get. Remember: Corona discharges off the bare cable conductor and shield ends can account for leakage current values several times the current flowing through the insulation. Let's take a look at a couple different preparation methods.

Method 1. Cover the bare cable conductor and adjacent bare shield ends with an airtight application of six or more layers of half-lapped rubber splicing tape. At the end where you're connecting the test lead, the tape covering should include the lead and its clip or connector as well as the lug and conductor of the cable under test. Make sure you extend the tape onto the cable insulation and test lead insulation for at least a couple of inches.

If the shield ends are bare, place a temporary covering of tape about 1 in. from the edge of the cable or stress cone shield in each direction. This eliminates unexpected flashovers and increases safety.

While Method 1 is useful in getting accurate data, it's more costly than Method 2.

Method 2. Cover the bare cable conductor and bare shield ends with a plastic bag and secure it to the cable jacket with rubber tape. While this method won't give you the same accuracy as Method 1, its economy may influence your choice.

Regardless of how you prepare the cable ends, the accuracy and reliability of any generated test data may revolve around how you perform the following tasks:

• Check clearances at cable ends;

• Ground adjacent conductors;

• Ground adjacent objects near cable ends;

• Check test set and test lead at testing voltage level (lead in place but not connected);

• Calibrate the test set for voltage and microamps periodically;

• Use weather protection;

• Determine temperature of test set at time of test; and

• Determine rate of rise on voltage applied during testing.





Sidebar: How Much Leakage Current is Enough?

This is a common question in hi-pot testing. Sometimes, you don't have enough information on the cable you're testing, such as length of circuit, type of splices and terminations, magnitude of voltage held, and conditions of test hookup (whether or not ends were prepared to reduce corona). Instead, all you have are generalities.

In fact, the magnitude may be several times that calculated using Ohm's Law, based on the voltage held and insulation resistance per published guaranteed K value of the insulation. Still, there's no need for alarm.

As long as the current doesn't behave erratically or start rising during the holding period, the test is acceptable and you can deem the cable reliable.