MENU
Cable Fault Location can sometimes be a very simple task, taking only minutes to perform. But it can also be a very complicated, time-consuming activity, requiring the use of multiple pieces of test equipment and procedures. What primarily determines the complexity of the fault location process is the electrical characteristics of cable fault itself. Other secondary issues include the voltage rating, construction, and network configuration of the cable circuit.
There is no “one method that fits all” approach and there is also no one tool that will locate all your cable faults. Other things to consider are the amount of energy and voltages you are willing to apply to the faulted cable – these have the potential to cause additional damage to the faulted cable and/or other healthy cables running close to the faulted cable. This paper goes through the practical processes that were followed on one of these difficult faults that was located on a MV cable the field.
A utility customer in the Carolinas requested our support in detecting a MV Cable fault that they could not locate using a local testing contractor. The 3-phase cable feeder in question consisted of 2 cables in parallel per phase, with a total length of 7600 ft. The cables had a tape shielded neutral and were rated at 15kV and operated at 13.8kV. The cable was installed in duct banks with several manholes along its route. The customer had allegedly had a history of cable faults on this circuit. From all accounts this was mainly attributed to poor workmanship during the installation of the cable.
Prior to HV Diagnostics being contacted to provide the fault location service, a VLF/TD (Very Low Frequency / Tan Delta) maintenance test was done on the cable by a testing contractor using a model HVA28TD, and integrated VLF TD test instrument - See Figure 1. The test reported that C Phase had failed 150 seconds into the VLF TD Test. The Tan Delta report showed TD values in excess of 500E-3 before the cable finally failed. TD values of this magnitude fall squarely into the “Action Required” category of IEEE400.2 – Table 3.
Due to other energized cables running parallel to the faulted cable, the customer wanted to minimize the duration, energy and voltage levels applied to the cable to prevent possible additional damage to the cable itself and the adjacent cables. Cable fault location “Thumpers”, often effective in finding many types of cables faults, and are frequently combined with a TDR (Time Domain Reflectometer), can generate significant amounts of energy. Even though a thumper was available, due to the above, and the fact that this was a tape shielded cable, where TDR often have limitations, it was decided to use the thumper approach only as a last resort.
The first step was to ensure that the cable in question was de-energized and isolated from any power sources. An inspection and verification of both ends of the circuits confirmed the cables were safely de-energized and ready for testing. There were 6 cables terminations inside the switchgear cubicle, 2 cables per phase and color-coded Brown, Orange, and Yellow. See Figure 2. The VLF/TD report provided had only identified “C Phase” as the failed cable. There was no confirmation in the report on which color corresponded to “C” phase and which conductor of the two had faulted. The cables therefore had to be retested to verify which phase and cable segment was faulted.
Cable End Identification had to be performed using a continuity test to “ring out” the cables to properly identify both ends and label each phase and conductor as Brown 1 and 2 (B1/B2), Orange 1 and 2 (O1/O2) and Yellow 1 and 2 (Y1/Y2).
After verifying proper identification on both ends, a quick 6kV insulation resistance test was performed using the Digi-Bridge instrument in DC mode to establish which of the cable segments were faulted. It was found that the Yellow 2 (Y2) was the original C phase that had faulted in the original report. It had a low resistance of 7.5MΩ to ground. All other phases / conductors showed high insulation resistance - Figure 3.
It should be noted that passing a DC high voltage insulation test does not necessary mean that the MV cables are fit for service or healthy and a VLF AC withstand test is the preferred method of testing MV cables to achieve this, however in this case we wanted to quickly and conveniently find and isolate the faulted cable. If the DC test had not found the faulted cable, the more effective VLF test would have been employed.
Due to the high resistance of the fault, and the type of shielding on the cables, a TDR would not be able to pre-locate the fault.
To allow for other fault location techniques to be used, the resistance of the fault needed to be decreased even further. The VLF test instrument was then used in Fault Condition Mode at 6kV to reduce the resistance of the fault. The faulted cable was able to hold 6kV voltage, however the diagnostic TD values were very elevated with TD numbers coming well over 400E-3. The Voltage was then ramped up to the IEEE400.2 recommended withstand voltage for 15kV Cable (16kV) to assist in lowing the resistance of the fault. This is typically done by injecting current (few mA) locally into the already low resistant faulted section of the insulation, thereby carbonizing the area at the fault and lowering the resistance even further. This normally only takes a few AC cycles / few minutes to perform.
After a short stint of fault conditioning, the fault resistance was reduced to 90kOhms. Following the successful conditioning of the fault, the Digi-Bridge was used in Bridge Fault Location Mode to pre-locate the fault.
The Digi-Bridge uses the classical but still highly effective traditional wheatstone bridge to locate faults in cables. There are essentially two main modes of operation for the Bridge method of fault location depending on the numbers of healthy adjacent phases that are present next to the fault phase.
In this case we had 5 other healthy adjacent cables that we could have used, since there were 2 cables per phase. Having 2 adjacent cables is more accurate than having one, so we used adjacent cable segments B2 and O2 to test faulted segment Y2.
The Digi-Bridge quickly and accurately found the fault. A total of six tests were performed to confirm the results with the the first three tests conducted from the near end of the cable.
A result of 97.3% was measured, indicating the fault to be at about 97 of the length of the cable from the test set.
The last three tests were done from the far-end of the cable. These results showed 3%, indicating the fault was the complimentary 3% from the far end, thus confirming the results of the first three tests.
If the length of the cable is known, it can be entered into the Digi-Bridge, and you will receive a distance in feet or meters.The total length of the cable was 7,600ft, therefore the distance to the fault from the test set was 7372ft or 228ft from the far-end. See test results below – Figure 5 and 6 respectively.
The route of the cable circuit was known and there was a manhole with splices at approximately 230ft from the far-end. Poorly installed cable accessories (like splices and terminations) are often a source of cable circuit failures and the likely candidate in this situation was the splice/joint at this location.
After obtaining the necessary confined space permits etc, a qualified crew entered the manhole and pumped out some water to get better access. Inside the manhole the crew was confronted with several runs of cables, all unmarked. See Figure 7.
Since these VLF units typically only produce very low energy, there is often no external physical damage visible at the failure point on the surface of the cable jacket. Even if there appears to be evidence of a cable failure, it is always prudent to securely verify the correct cable before cutting / spiking into a cable. At the very least you may cut into the wrong cable, sacrificing a perfectly good cable and the addition of an unnecessary splice etc, but worse, you may well cut into an energized cable with potentially lethal consequences to both people and equipment.
Therefore, the correct faulted cable and its associated suspected splice had to be identified in the bundle of cables present in the manhole duct section.
The model EZ-Cable ID was then used to identify the correct faulted cable circuit. The EZ-Cable ID consists of a Transmitter (Tx) and a Receiver (Rx). The Tx produces a unique pulse train that travels down the cable and allows for positive identification by the handheld receiver Rx – see Figure 8.
Once the transmitter of the EZ-Cable ID is connected to the identified cable at one end, the far end is grounded. The operator then uses a flexible coil around each cable / splice in the manhole, until he receives a positive ID that he/she is on the correct cable. See Figure 8.
In this above situation, the faulted cable was wrapped along with the other two healthy phases in a fire-retardant tape – Figure 9. This tape made it difficult to place the Flexible Coil around the cable due to the large bundle of all 3 phases. The Flexible Coil was swapped out with the Handheld Pickup Sensor “PUC”.
The PUC is designed for identification of 3 conductor cables and worked perfectly for this application as it not only identified the correct wrapped circuit it was also then able to identify the correct phase once the tape as removed in this section. The fault was ultimately located in a splice– wrapped under the fireproofing tape shown in Figure 9 on Y2.
Fault location is both an art and a science, and with experience, they all play an important part in how quickly and how effectively a cable fault is detected and located with the minimum amount of impact on the system.
There is simply no one tool that will find all your cable faults and an investment needs to be made in equipment and training.
The above case study shows how the use of multiple pieces of test equipment and a clear understanding of when and what to use, depending on the type of fault, cable construction and topology, all can play an important role in locating the fault.
REFERENCES
[1] IEEE 1234-2019 IEEE Guide for Fault-Locating Techniques on Shielded Power Cable
[2] IEEE-PES-ICC Presentation 2022 Orlando FL: A Multi-tool Approach to MV/HV Cable Testing in the Field - C. Goodwin
For more information about the devices we used in this case study, please see the below links.
https://www.hvdiagnostics.com/hva28td
https://www.hvdiagnostics.com/digi-bridge
https://www.hvdiagnostics.com/ez-cable-id
For service or support contact HV Diagnostics, Inc. here
