In Part 2 of our three-part series on acceptance and maintenance testing for medium-voltage electrical power cables, we explored high-potential testing techniques, including tests performed with direct current (DC) and alternating current (AC) performed at power frequency and at very low frequency (VLF). We also explored diagnostic testing and reviewed dissipation factor (tan delta) testing.
In Part 3, we peer into the world of on-line and off-line partial discharge (PD) testing.
Partial discharge is defined in IEC 60270 as:
…a localized electrical discharge that only partially bridges the insulation between conductors and which can or cannot occur adjacent to a conductor.
When the voltage stress exceeds the breakdown strength of that portion of the insulating material, a partial discharge begins (Figure 1) and continues to deteriorate that insulation. When partial discharges occur, various physical and chemical changes may also produce emissions that we can detect, localize, and characterize to provide the information needed to prevent insulation failure in medium- and high-voltage electrical equipment.
Partial discharge occurs in solid, liquid, or gaseous insulating mediums. PD can also occur in the form of corona, surface tracking, or floating-electrode metal-to-metal discharge, causing degradation of the insulation. NFPA 70B, Recommended Practice for Electrical Equipment Maintenance, states that insulation breakdown is the number-one cause of electrical failures, and once PD begins, it will always get worse.
There are two sets of resistances: materials and air voids. In some materials, the insulating medium is not completely solid or uniform, so there may be air voids. The resistance of the material is higher than the resistance of the air gap, meaning the insulation has a higher dielectric strength than the air pocket. As voltage is applied, the voltage stress across the air gap is higher than on the insulation. The walls of the gap are stressed and begin to break down or ionize (Figure 2). Eventually, these ions build up, the resistance of the air gap breaks down, and a discharge across the gap occurs. This can potentially leave semi-conductive residue or carbon behind, which is conductive, kind of like a stress crack on the insulation.
Over time, this residue will build, creating a full discharge from one conductive surface to another.
PARTIAL DISCHARGE INCEPTION VOLTAGE AND EXTINCTION VOLTAGE (PDIV/PDEV)
For partial discharge to occur, sufficient voltage must be applied to the system under test to meet the minimum voltage required to initiate partial discharge activity. This is known as the partial discharge inception voltage (PDIV). Once the PDIV has been reached, the voltage may be lowered, and PD will remain present at the lower voltages until they finally extinguish at what is referred to as the partial discharge extinction voltage (PDEV). The PDEV is therefore less than the PDIV.
If the PDEV voltage level is lower than the system operating voltage (phase-to-ground), this implies that an over-voltage surge on the system could initiate PD, and the PD activity could continue even when the system voltage returned to normal. PD that can continue at operating voltage is therefore more likely to result in insulation failure than PD that extinguishes above normal operating voltage. PD that occurs at the operating voltage level can be detected through on-line detection methods. However, to gain a full understanding of PDIV and PDEV, you would need to adjust the voltage source, which can only be accomplished in an off-line test.
PARTIAL DISCHARGE PARAMETERS
Partial discharges are a consequence of local breakdown due to (a) an electric field enhancement within or on the surface of the insulation or (b) a region of low breakdown strength. Partial discharges appear as individual events of very short duration always accompanied by emissions of light, sound, and heat, as well as electromagnetic pulses, and often result in chemical reactions.
The PD parameters usually measured during tests on installed cable systems include:
- PD inception voltage: PDIV (off-line tests)
- PD extinction voltage: PDEV (off-line tests)
- PD location
- PD magnitude (q)
- PD repetition rate (n)
- PD density: density of discharges measured per unit of time and per unit of length (laminated cable only)
- Phase angle of PD pulse (Φ) given by Φ = 360 (ti/T), where ti is the time measured from the preceding positive-going transition of the sinusoidal test voltage through zero to the PD pulse, and T is the period of the test voltage
- Phase-resolved PD plot (n vs. Φ vs. q)
- PD magnitude vs. voltage plot (q vs. V), (off-line tests)
The characteristics of partial discharge parameters depend on:
- Type and location of defects, such as PD sources, in the insulation system
- Insulating material
- Operating conditions such as applied voltage, load, and time
- On-line tests can measure q, n, and Φ at operating temperature, whereas off-line tests are performed on cable systems that have cooled down.
UNDERSTANDING DIELECTRIC MATERIALS
Water treeing is an important form of degradation that can afflict older HMWPE and XLPE extruded cables. At the site of a water tree, the insulation is degraded, i.e., has a higher dielectric constant and lower dielectric strength than the original insulation (Figure 3). Water tree growth under service conditions is a very slow process, usually taking many years to completely penetrate the insulation. Water trees do not generate partial discharge. However, when subjected to high electrical stresses, water trees can lead to electrical trees because of a lightning impulse, a switching or DC over-voltage, a high AC voltage, or when the tip of the water tree approaches a conductor or insulation shield. There is no evidence of water treeing being an important issue with EPR or TRXLPE cables.
It is important to know the materials being tested to better interpret PD data as the resistance to damage by PD depends on the insulating material. The order of PD resistance from least resistant to most resistant is XLPE << EPR << laminated (i.e., fluid-impregnated paper). Cable accessories, often made with filled rubber, may have high endurance to PD activity provided this does not occur adjacent to extruded cable insulation.
Shielded distribution cables fall into two classes: PD-free and PD-resistant. PD-resistant cable can sustain substantial amounts of PD over long periods of time without failure. PD-free cable can be formulated with a range of dielectrics, having low PD resistivity. However, for both types of cable, certain forms of PD will eventually cause failure, whereas other forms of PD can continue almost indefinitely without failure. Knowledge rules, which give objective guidance for the interpretation of measured data for the different materials in use, are necessary.
TESTING FOR PARTIAL DISCHARGE
Partial discharge pulses are very short, typically 1 ns to 5 ns wide, and they can have significant frequency components up to 1 GHz at their source. Two general approaches are available to detect PD pulses in installed cables: off-line and on-line detection. Off-line testing is normally carried out using a separate voltage source after the cable has been removed from service. On-line (in-service) testing is carried out during normal operation of the cable system.
Advantages of off-line PD testing:
- PDIV and PDEV can be measured if a variable voltage source is used.
- PD characteristics can be obtained at various voltages, which can aid in identifying certain types of defects.
Advantages of on-line PD testing:
- PD characteristics can be obtained under varied load conditions, which can aid in the identification of certain types of defects.
- Tests can be performed without having to take an outage.
LOCATING PD IN CABLE SYSTEMS (MAPPING)
When PD activity occurs, high-frequency current pulses are created and will propagate along the cable. These high-frequency current pulses will propagate in both directions along the cable and can be used to locate the PD activity.
When partial discharge occurs in a cable system, two pulses of similar size and characteristics propagate away from the partial discharge site toward the terminations (Figure 4).
Depending on the cable insulation type, shield construction, and other factors, the speed at which the pulses travel is relatively consistent. For instance, partial discharge pulses travel at a speed of approximately 468 feet per microsecond (142.65 meters / u-sec.) in XLPE insulation.
Mapping can be a challenge when conducting on-line tests and may require multiple test points to assist in mapping. Off-line mapping offers the advantage of isolating the cable under test, thus making the mapping process easier and more accurate.
PHASE-RESOLVED PD PATTERNS
Using phase-resolved partial discharge patterns is a method of comparing incoming signals to the AC power frequency or AC test voltage. The PRPD charts (Figure 5) are shown in real-time. The horizontal x-axis is the phase angle of the power system under test. The x-axis is labeled from 0 to 360 degrees. This is the same thing as 0 to 16.667 ms, which is the time it takes for one cycle of the AC power frequency. The vertical y-axis represents the amplitude of the signal. This amplitude depends on the strength of the signal and the distance from the sensor to the PD event.
We know that PD signals are produced as the sine wave reaches a positive or negative peak (or trough) or peak voltage stress. The sign wave peaks and troughs are exactly 180 degrees apart (16.66 / 2 = 8.3 ms). The PD pulses continue to arrive with intervals equal to 8.3 ms or multiples of 8.3 ms (180 degrees apart) with some fluctuation due to PD inception and extinguish points. As the insulation deteriorates, the PD will begin earlier and extinguish later as its ability to withstand the voltage stresses diminishes. Meanwhile, all the other noise and other signals in the detection bandwidth of the sensor will arrive at random time intervals that are not equal to 8.3 ms. Noise arrives at random points along the x-axis, whereas PD pulses only arrive 180 degrees apart.
EFFECTS OF SHIELD TYPE
In most industrial facilities, a large percentage of the cables are of the tape-shield design where overlapping layers of copper tape are wrapped around the insulation shield of the cable. In addition, widespread use of EPR insulation is found in industrial plants. In contrast, the electrical utility distribution sector tends to use concentric neutral wires instead of tape design for the ground shields and XLPE for the insulation.
The concentric neutral wires hold a certain advantage as the cable matures and begins to show signs of service age. As tape-shielded cable ages and corrosion of the copper tape occurs, particularly at the overlap of the consecutive tape layers, the cable starts to attenuate high frequencies. Even with slight corrosion of the overlapping layers, the cable shield starts to behave as a coil or inductor to high frequencies. The net effect of this attenuation is that a PD detection system connected to a cable may not always detect the high-frequency PD pulse as it would be so attenuated by the time it reaches the PD monitoring equipment. The further away the detection equipment is from the active PD site, the more severe this limitation. This limits the effectiveness of partial discharge testing on tape-shielded cable.
Another consideration when comparing on-line and off-line measurements is the effect of noise on the measurement. In the off-line approach, the detection equipment can be calibrated at the time of the test by injecting a known PD pulse level into the specimen under test. This is not possible with the on-line approach. Further, cables can be isolated during off-line testing.
When detecting PD activity using on-line methods, further analysis will typically be necessary to ensure that the suspect activity was not caused by external noise or activity generated upstream or downstream of the cable section under investigation. It will also be advantageous to locate the source of the partial discharge activity and to quantify and assess the severity of the problem. This can be accomplished by measuring and analyzing activity over time to detect deterioration and to raise an alarm or call to action if PD activity reaches a critical level. As an example, if PD activity that is intermittent or possibly influenced by environmental conditions (changing temperature, humidity, or electrical noise) is found, temporarily installed multi-sensor systems that automatically monitor your plant can be utilized.
ONLINE PARTIAL DISCHARGE TEST
As stated earlier, when PD occurs, small current pulses are induced onto the ground shield or case ground. These pulses will travel dozens of meters along the ground grid in the form of high-frequency current pulses ranging from 500 kHz to 50 MHz, usually centered near 10 MHz. Using high-frequency current transformers (HFCT) is a reliable method to measure these high-frequency PD pulses, which spread out onto the ground grid like ripples in water. They are especially useful for quickly testing for internal PD in a large area such as an entire power transformer or an entire cable, or even an entire substation. The HFCT sensor has a split core, so it’s simply clamped around a low-resistance grounding lead. HFCT sensors have the distinct advantage of being able to detect PD signals on cables from long distances up to 1 km away (this limit depends on the type of ground shield and the strength of the PD pulse). This means multiple PD signals and noise signals can also be detected from many points on the ground grid where a grounding lead is exposed.
The test requires the attachment of these HFCTs to the metallic cable shields (Figure 6). This presents several safety risks and is considered energized work. To perform this task, the technician must enter the restricted approach boundary and, depending on the layout of the cable landing, restricted or blind reaching may be required. This is a violation of the NFPA 70E requirements for work involving electrical hazards. The task also requires connecting conductive materials within the restricted approach boundary, which raises the risk of shock or arc flash occurrence.
The challenge in performing the on-line partial discharge test on energized electrical power cables is safely connecting the HFCTs. Perhaps the best solution is to use engineering controls from the hierarchy of risk control methods.
One engineered approach would be to pre-install the HFCTs into the cable landing compartments. The coax cables can be safely routed along the inner panels of the compartment and then brought out through a bulk-head adapter. Testing can then be carried out safely at any time by simply connecting the scanning equipment to the test connector on the bulkhead adapter.
Another approach is to bring the cable shields out of the cable landing compartment. This can be accomplished by punching a small hole in the cabinet and extending the cable shields out of and then back into the cabinet, forming loops outside of the cabinet allowing HFCT attachment. The ground loops are then enclosed in a NEMA compartment mounted to the cabinet.
OFF-LINE PARTIAL DISCHARGE TEST
On-line PD testing has the obvious advantage that it does not require disconnecting or an outage. The main disadvantage when testing cables under operation is that the test is only performed at the operating voltage level and cannot be adjusted. Obviously, if the applied voltage is fixed and cannot be changed, the PDIV and PDEV voltages cannot be determined, and therefore, in comparison to off-line testing where voltages can be adjusted to simulate transients or other over-voltage conditions, a lower percentage of defects in the cable’s insulating system can be detected through on-line methods.
Power Frequency and Alternative Test Voltage Sources
As stated earlier, for partial discharge to occur, sufficient voltage must be applied to the system under test to meet the minimum voltage required to start partial discharge activity. When testing cables, the on-line testing approach uses a system voltage of a constant fixed magnitude. In an off-line approach, a temporary voltage source will be required. Considerations for an off-line voltage source should include:
- The applied voltage should cause partial discharges in the insulating system under test that has characteristics close, if not identical, to those that occur when the insulating system is in service.
- The temporary voltage source should cause no appreciable damage to the insulating system during the time required to perform the measurements.
- The temporary voltage source should have a variable voltage output so that PDIV and PDEV tests can be performed.
- The size and weight of the equipment required to produce the voltage levels required for testing various assets must be considered. Is the equipment to be used in a fixed location or used in a field application?
Voltage sources that are used for commercially available field partial discharge measurement systems will fall into the general categories of power frequency and alternative voltage sources such as very-low frequency (VLF).
VLF Test Voltage
Depending on the type of defect, sinusoidal VLF voltage sources, usually 0.1 Hz, for extruded dielectric systems may require a higher test voltage to generate the same partial discharge level compared with tests performed with power-frequency voltages. For example, the conductivity of the surface of a cavity that has been exposed to PD increases, which allows any charges deposited on the surface by PD to leak away; this lowers the electric field in the cavity. As more charge can leak away between polarity reversals at VLF than at power frequency, the PDIV at sinusoidal VLF will be larger than that at power frequency (Figure 7). If there has been no previous PD activity to increase the conductivity of the cavity surface, the PDIV at sinusoidal VLF and power frequency will be similar (see IEEE Std. 400.3).
A VLF cosine-pulse waveform generates a 0.1 Hz bipolar pulse wave that changes polarity sinusoidally. Since the sinusoidal transitions are in the power-frequency range, the PDIV measurement will be comparable to power frequency. The VLF cosine-pulse voltage works according to the principle of 50/60 Hz slope technology.
This is particularly important for PD diagnosis since reliable evaluation of the measured results requires direct comparability with the power frequency. Partial discharge characteristics change in the case of large frequency differences, making reliable evaluation of power frequency impossible. The 50/60 Hz slope technology ensures comparability for both voltage wave shapes.
Figure 8 shows a typical example of how PD measurement is carried out during the slope of the applied voltage. The steepness of the VLF cosine-pulse slopes in comparison to the 0.1 Hz sine wave can be seen. It is precisely this rise in voltage that is so important for the PD inception voltage. Therefore, the 0.1 Hz sine wave test voltage cannot be directly compared to the 50/60 Hz power frequency, and critical partial discharge defects are therefore not always reliably detected.
Damped AC Voltage (DAC)
Another approach to reduce the size and weight of the test voltage supply from that of a conventional power frequency supply is the damped AC voltage (DAC) technique. For partial discharge analysis, the cable under test is charged to the pre-selected peak value by a direct current high-voltage source within a couple of seconds and then shorted with an electronic switch via a resonance coil.
Thus, a sinusoidal oscillating AC voltage with low damping is created (Figure 9). The frequency is fixed in a range from 50 Hz to several hundred Hz, depending on the capacitance of the test object. Since the frequency of the test voltage is close to nominal service conditions, all measured PD activities can be effectively evaluated and compared to that of power frequency. Due to the decaying amplitude of the test voltage, the partial discharge extinction voltage can be easily determined.
As we had seen in this three-part series, numerous insulation tests are available to assist in assessing the quality and condition of a cable’s insulation. The pass/fail tests provide the means to identify gross defects, while the diagnostic tests provide us with an understanding of the severity of degradation or the extent of contamination in the insulation.
Technology and philosophies toward testing have advanced over the past 30-plus years and push toward predictive maintenance solutions. A key element in predictive maintenance is monitoring the trend of diagnostic test results. Not all tests are appropriate to all circumstances, nor can any single test give you the complete answer. Each type of test serves as a window looking into the condition of the cable, and by putting together several different tests, you build a more complete picture.
Thomas Sandri is Director of Training Services at Protec Equipment Resources, where his responsibilities include the design and development of learning courses. He has been active in the field of electrical power and telecommunications for over 35 years. During his career, Tom has developed numerous training aids and training courses, has been published in various industry guides, and has conducted seminars domestically and internationally. Thomas supports a wide range of electrical and telecommunication maintenance application disciplines. He has been directly involved with and supported test and measurement applications for over 25 years and is considered an authority in application disciplines including insulation system analysis, medium- and high-voltage cable, and partial discharge analysis, as well as battery and DC systems testing and maintenance. Tom received a BSEE from Thomas Edison University in Trenton, New Jersey.