Power factor and tan delta testing, the process of measuring capacitance and dissipation, helps you determine the condition of a system’s insulation in bushings or between windings. As time wears on, insulation of a system will begin to degrade, or impurities like dirt and moisture may start to compromise it. This will create a conductive path to the current, consequently leaking electric current flowing through the system and reducing capacitance. Ideally, the aim is to keep the ratio of resistive component to capacitive component of an insulator low. This ratio is known as tan delta, or dissipation factor (power factor).Advanced Test Equipment Rentals stocks equipment that can measure power factor and tan delta of your system. You can determine capacitance and dissipation at varying frequencies and identify any age degradation or impurities sooner, before the damage becomes irreparable. Advanced Test Equipment Rentals carries power factor and tan delta test equipment from top manufacturers like Tettex, Megger, OMICRON, Doble, and more. Power factor and tan delta test equipment can be applied to power transformers, instrument transformers, bushings, capacitors, circuit breakers, and surge arrestors.
After performing condition assessments, inspections, testing, and repairs of generators I have discovered that the more I learn, the less I realize that I really know. Generators are a very complicated machine and have a unique insulation system in which you manage PD, rather than eliminate it.
I will digress for a moment,
Previously I had mentioned that I ‘chose’ to be an engineer. As a young child growing up in Saskatchewan, my father liked to travel. I remember visiting the WAC Bennet Dam and G.M.Shrum generating station during these travels. This was before the Peace Canyon Dam was built. I have pictures showing how we were allowed into the underground powerhouse during the tour and we were able to look down into the control room when it was still a fishbowl. Times have changed significantly.
During 911 I was managing a power station on the Columbia River and remember hearing about this terrible event when I arrived at work. Very little information was available at the time, but we could see the military flying overhead, directing the local heli-logging operations to land immediately. Those events changed the world forever and we still live in the shadow of the terror that resulted.
Once again, I will pull out that soap box. And so, my tale continues . . .
I thought I would share some experience with conducting testing of generators during condition assessment. There is much fear among experts of sharing information because they think it will take their power away. My own experiences showed that the people I most respected, always freely shared their knowledge. Without this, we would never have left the dark ages and progressed forward with Universities.
Information about the generator is required to prepare a good test plan which would include:
The winding capacitance is required to calculate the correct high voltage power supply size and can typically be obtained during the IR/PI testing.
Previous test information is required to establish good trends. Much information on the resins used, and characteristics of windings is not available. Given the very small statistical sample for hydro-electric generators, and their site-specific designs for larger units, some judgement is typically required during evaluation and this helps the process.
Although I have used a variety of equipment over the years during testing, I can suggest that the following equipment has performed satisfactorily from my perspective,
Other equipment you will need would include:
Qualitrol can provide equipment for the following:
This is a suggested approach to diagnostic testing and condition assessment of a 13.8 kV hydro-electric generator stator winding. It is not the only way, but it will provide very comprehensive results. The principles used here can also be applied to turbo-generators and motors. I prefer using instruments which have recording capabilities and use datasheets to summarize the field test results and act as a checklist for some procedures. I’ve also included some information on what I can learn when analyzing the results.
Electrical diagnostic testing for the stator winding and core include:
A visual inspection of the winding should be performed to assess the condition, note anomalies, and document visual evidence of problems.
Personnel performing the testing should be trained and qualified for performing HV testing and knowledgeable of visual inspection. All measuring and test equipment should be in calibration. Test anomalies should be investigated at the time of testing. Calibration certificates should be National or International bodies and available upon request.
Rotating machinery testing involves the application of high voltages to machine windings. These windings have capacitive and inductive properties that can lead to hazards that may not be readily apparent. Test personal should be familiar with the hazards associated with the testing. The test setup should conform to all local safety codes.
The following general requirements should be considered when performing testing:
The following should be considered when performing DC testing:
Consideration should also be given to the following by the equipment owner:
Prior to starting testing, consideration should also be given to the following:
Check the insulation resistance (IR) of the stator windings in accordance with IEEE standard 43. Testing is a requirement to ensure a minimum insulation resistance prior to proceeding with diagnostic testing. Table 4 of this standard indicates the minimum recommended insulation resistance, in MΩ at 40°C, of the entire machine winding (all phases) of 100 MΩ as a prerequisite for further testing.
Both single phase and three phase testing are required for assessment of the condition of the insulation system. Comparison of test results can indicate if problems with respect to ground or between phases exist. Typically, the slot section can have problems to ground while the end arm region can have phase to phase problems.
The following procedure is suggested as an approach to Insulation resistance measurements:
Plotting of the insulation resistance over the life of the machine (after correction to 40°C) will allow recognition of some problems.
Both single phase and three phase testing are required for assessment of the condition of the insulation system.
Tan δ testing evaluates various factors that influence the performance of the overall electrical insulation. Bridge techniques should be employed to properly measure and assess the extent of pulseless and pseudo glow discharge activity. Testing is performed in accordance with the requirements of IEEE 286 on each phase at up to rated line to neutral voltage.
The power factor tip-up is defined as the difference in the power factor measured at two voltages. The test plan below is shown in Figure 1. A typical test setup is shown in Figure 2.
Figure 1: C & DF Test
The following procedure is suggested as an approach to dissipation factor (and capacitance) voltage characteristic measurements:
Figure 2: Typical C&DF Test Setup using an Omicron System
The basic tip-up characteristic is calculated from the change in dissipation factor between operating voltage the lower test voltage. The 8 kV measurements will include pseudo-glow, internal PD, and corona losses, while the measurement at 2 kV should only be related to the dielectric loss and leakage currents.
Plotting of this information over the life of the machine will show how the dielectric characteristics and non-linear losses change over time. Cleanliness of the machine will also have a substantial effect on the results of the testing due to how leakage currents in the end arm regions are affected.
Comparing differences between the single phase and three-phase losses will allow examination of effects within the end arm region of the winding, in a similar fashion to PD testing.
The original factory test results prior to assembly should be consulted for guidance in interpretation of the measurements.
Careful comparison of the curve with the PD inception and extinction voltages will show the total losses relate to sparking type discharges. During off-line testing, PD measurements using capacitive couplers are typically limited to PD measurements at the machine terminal while dissipation factor is sensitive to the entire winding.
Off-line PD testing evaluates various factors that influence the performance of electrical insulation associated with sparking discharges near the terminal region of the stator winding. Testing should be performed in accordance with the requirements of IEEE 1434. The test plan is shown in Figure 3.
Figure 3: Partial Discharge
Measurement of the partial discharge quantities QIEC, as well as extinction voltage and inception voltage are required using coupling capacitors installed at both the line end and neutral end of the winding for simultaneous measurements. Each is calibrated individually and the attenuation through the winding can also be measured. Measurements of leakage currents from grounded phases using HFCT will assist with analysis. Testing should be performed in a single-phase configuration, and when possible in a three-phase configuration. The test set-up is shown in Figure 4.
Figure 4: Typical PD Test Setup for Single Phase testing
As the test voltage does not approach the hipot level, it is preferred to connect the test voltage supply to the neutral terminal of the winding under test, without shorting the line and neutral terminals together.
The coupling capacitor should be connected to the line and neutral terminals. This will allow the stator core to filter out any influence from the power supply and allow the most accurate PD measurements at the line terminals. In some rare cases the test set-up may need to be reversed to allow more detailed investigation of the neutral end of the winding. Refer to Figure 5 for configuration of the test equipment.
Figure 5: Typical PD Test Setup using an Omicron System
The connections between the coupling capacitors and the winding under test should be kept as short as possible and high-voltage leads should be of sufficient diameter to not produce PD during the test.
Partial discharge extinction voltage (PDEV): The highest voltage at which partial discharges (PDs) above some stated magnitude (which may define the limit of permissible background noise) no longer occur as the applied voltage is gradually decreased from above the inception voltage
Partial discharge inception voltage (PDIV): The lowest voltage at which continuous partial discharges (PDs) above some stated magnitude (which may define the limit of permissible background noise) occur as the applied voltage is increased.
Largest repeatedly occurring PD magnitude (QIEC): The largest magnitude recorded which has a pulse repetition rate of 10 pulses per second, which can be directly inferred from a pulse magnitude distribution measured in the frequency range of 250kHZ±150kHZ.
PD tests have origins as a rewind criterion and have evolved into a diagnostic or condition assessment tool, as well as, a maintenance tool. My preference for performing an off-line partial discharge test was using capacitive couplers located at the line end and neutral end of the machine. The remaining two phases were grounded, and the ground lead was monitored with a HFCT.
The model for a generator is very complicated, with impedance mismatches between the end windings and slot sections which result in many reflections. Cross coupling of partial discharge through the end arm regions bypasses the core section. Poorly designed circuit ring taps or ring busses can also be a source of partial discharge when the line and neutral rings are in close proximity.
Some may argue that it is not possible to calibrate for apparent charge due to the complex impedance but using the same calibration procedure and instrumentation during subsequent testing will ensure repeatable results.
While in service the winding has voltage that increases from zero at the neutral connection to rated voltage at the line connection. Off-line testing will subject the entire length of the winding to the test voltage. Off-line partial discharge measurement does not have the same voltage distribution within the winding that would occur while a generator is in service. It will identify weakness in the entire insulation system that may not be detectable with an on-line system. Since these two types of tests are done under different voltage distributions, the results are complimentary and can assist with analysis of problems when compared.
During testing I have been able to identify partial discharge sources located both at the line end of the winding and at the neutral end of the winding. The localization of line end portions can be aided through physical indications of white powder deposits or use of a device which amplified light in the ultra-violet spectrum generated by corona (Corona Finder). White powder was typically limited to the last 10% of the winding closest to the tine ring bus during normal operation, while corona could be observed along the length of the winding during testing. Neutral end problems were typically buried in locations where the winding was in close proximity to windings at line potential. Typically, the first crossover between the high voltage end and the neutral end would show evidence of partial discharge activity, while all crossovers would typical be visible in the ultra-violet spectrum. The end arms of windings would typically show problems at the first coil group; however, all closely spaced groups would show visible corona during off-line testing. Evidence would also be visible in the ring bus where spacing was not maintained.
Comparison of the single-phase and three-phase measurements will show how much PD will be taking place in the end arm regions. Design problems which result in PD from inadequate kickback between the front and back coils, from crossovers in the end arm region, inadequate spacing between end arms between slots, and inadequate spacing in the ring bus will be detectable. The three-phase configuration will result in a reduction PD resulting from phase to phase stress and only Corona and slot PD. I have successfully proven that PD measured was only within the end arm region using this methodology.
The HFCT would show how much partial discharge was taking place between windings. This could be in phase to phase locations already discussed, or through cross-coupling of ‘fast’ PD events.
Measurements within a hydrogen atmosphere will result in measurements much lower than those that would be obtained in air for the same winding.
Humidity has a significant impact on the level of corona that can take place. High humidity environments will result in much lower measurements than would occur in very low humidity environments. Seasonal changes should be considered in the evaluation and it is preferred to perform the testing under the same conditions.
During testing, high background noise can be a problem. Arc Welders and grinding activity are examples of noise sources. Some can be eliminated by installing an RF filter in the power supply; however, some can only be eliminated by performing testing while other activities are not being performed.
For a maintenance flavor, I would suggest investigating how RTV affects PD performance.
Corona probe testing evaluates various factors that influence the performance of electrical insulation within the slot section of a stator core associated with sparking type discharges which could result from insulation delamination, insulation voids, and degradation of the semi-conductive treatment. Testing is performed in accordance with the requirements of IEEE 1434.
When performing the TVA probe test with the rotor in place, there needs to be enough room in the air gap to allow inserting the test probe. The test probe needs to maintain contact with the core. The operator must maintain limits of approach from the ring bus, end windings, and group jumpers. Rotor mounted fan blades or the generator cover H frame can restrict access with a test probe in cases were the rotor is not removed.
Single phase testing is required. During testing of each phase, all slots containing a stator coil in either the top or bottom of the slot are scanned. The slots are typically divided into two or more sections based on test ergonomics.
It is recommended that an AC overvoltage proof test in accordance with IEEE 62.2 section 7.1.5 of at least 1.10 to 1.25 times the rated line-to-neutral voltage of the machine has been performed prior to performing a TVA probe test. This overvoltage test provides some assurance of the integrity of the winding being tested. Care must be taken to ensure that this test does not result in conditioning which reduces the test values.
Hengfeng are exported all over the world and different industries with quality first. Our belief is to provide our customers with more and better high value-added products. Let's create a better future together.
The recommended target voltage for the TVA probe test is the rated line to ground voltage Vln. If relatively high PD readings are obtained, measurements may be made at lower voltages to determine the expected PD activity at operating voltage of specific coils/bars within the winding. If consistently high TVA readings are obtained, then reduce the test voltage to a suitable level to obtain sensible readings
The ferrite core used in the test assembly is very important to achieving correct measurements. Some instruments have interchangeable probes to allow testing with the rotor removed or with the rotor in place. Each probe is carefully designed to ensure measurements remain consistent.
Positioning of the probe will affect test results. Small changes in the gap between the probe and core can have profound changes to measurement. Typically, the ferrite core is insulated with a material having a known thickness to ensure consistent readings when the probe is in contact with the core. Wedges which are proud of the iron surface present challenges because proper contact of the probe may not be possible.
Measurements are very localized within the slot section due to how PD signals propagate within the slot section. Measurements can change as you move from core tooth to core tooth. The entire slot needs to be scanned as a result. My preference is to record the core with at least 2 equal portions. For Turbo-Generators, ergonomics will require it to be broken into more sections that are about one meter each. One person is required to position and move the probe at a slow and steady pace, while the second person observes the instrument and records maximum readings for each measurement section. Locations of measurement anomalies should be marked with an ‘x’ for visual examination at the conclusion of the test.
A field verification that the instrument is functioning is required as no other field check is performed. The instrument should also have a valid calibration certificate. My own preference is to inject a known pulse into a ‘U’ shaped ferrite core that is held in contact with the TVA probe. Each scale on the instrument could then be checked to determine that it’s reading conforms with the expected quantities.
High voltage gloves may be required when holding the TVA test equipment. Evaluation of approach limits with the end winding, and ring bus should be considered. Maintain limits of approach as prescribed by CSA Z462 with the stator end windings
1. Raise the voltage at the rate in accordance with the recommendations of IEEE Standard 4 up to the prescribed test voltage.
Observe a minimum 2-minute conditioning period prior to beginning testing. Note: Pre-conditioning of the stator winding for a period of up to 1 hour before making any measurements may be required, especially in the case of thermoplastic insulation systems.
2. Carryout a scan of the total length of each slot in turn.
Note: Any smell of ozone and if considered excessive and test should be stopped, and appropriate safety equipment should be used.
3. If any high TVA readings are obtained, note and mark the position.
4. Reduce test voltage to zero, turn off high voltage power supply and apply ground to stator winding.
5. Repeat steps 1 through 4 for each phase of the winding as required.
Just as with PD testing, there is also extensive coupling between windings during the TVA probe test. Adjacent slots which contain only windings from grounded phases will test positive using the TVA probe and reference to the winding diagram during analysis is essential.
Once the probe slides past the end of the core, a step change in measured values will also occur because of all the PD activity is now being measured in the end arm region.
Some increase in PD signal can also be observed at the very end of the core due to the grading\Semi-Con interface.
My preference for analysis was to create a spreadsheet with the winding map, showing the phase to ground voltage of each leg of the winding. It would also show the voltage difference between the front and back coils when in normal operation. This mapping would then generate a test sheet which indicated which slots needed to be tested as each phase was energized individually. This information would allow those performing the test to quickly identify measurements which were of greatest concern and perform a more detailed visual inspection of the area. Typically, high phase to phase voltages and high phase to ground voltages are of interest. Sometimes the pattern of X’s left behind can show a geometrical relationship.
My preference for presentation of the information was to provide readings in a couple of graphical formats.
Normally, there should be some increase in readings as the voltage increased, with the 10% of the winding at the line end having the most deterioration. Abnormal results could be associated with the maximum voltage difference between the front and back bars, those connected to the line voltage, or due to geometrical considerations. These would need to be analyzed to determine what was the root cause is.
IEEE 492 and IEEE 1434 do provide some additional guidance.
AC testing results in the highest stress levels of the insulation system within the slot section of the stator core. Acceptance test levels for new machines are 2E+1, while typical maintenance levels are 1.25-1.50E (E = Rated line to line voltage). 1.1E is end of life.
Withstand testing provides some assurance that the winding insulation has a minimum level of electrical strength required to successfully operate for the expected design life or certain period of time. Testing is performed on each phase in accordance with the requirements of IEEE 62.2 on each phase for one minute, with the remaining phases grounded. Partial discharge measurements are recorded at the withstand voltage level during the test.
Monitoring of partial discharge during higher than normal AC testing allows maintenance staff to understand potential weakness in the design of the machine and take corrective action before permanent deterioration of the insulation system occurs. Once the deterioration does occur, it is irreversible. As the insulation ages, partial discharge will occur at lower and lower voltages. Measurements at operating voltage will indicate on-line performance problems
The AC test will typically use a resonant test set to provide 60 Hz power for testing the stator windings; ergo the requirement for knowing winding capacitance. With the size of generators (e.g. capacitance) growing larger and larger, it is becoming increasingly expensive to provide test sets which can test an entire phase at one time. Luckily, the well-informed designer is able to take advantage of the machine design and split the insulated portion of the winding into two or four segments at the isolation points for the line and neutral connection. This allows testing of these larger machines with existing 500 KVA power supplies during acceptance testing, and much smaller test sets for maintenance testing. I have observed some large machines can be 2µF. An RTS capable of energize 0.5µF at 20kV isn’t much larger than an oil barrel and can be powered by a wall plug of 120V at 15A, while a supply capable of energizing 2.1µF at 25kV or 0.65µF at 45kV needs 120A at 208V and is more like a Mini-Cooper.
Acceptance test level for new machines is 2E+1, while typical maintenance levels are 1.25E to 1.50E (E = Rated line to line voltage).
1. When performing acceptance level testing of 2E+1, a spark gap should be used. Set the sphere spacing for a voltage 10% higher than the selected test voltage prior to final connection to the machine. Usually this can be verified with the test set before connection to the generator.
2. Raise the test voltage slowly from zero to the selected voltage. It should be raised sufficiently slowly to permit accurate reading of the measuring instrument, but not so slowly as to cause unnecessarily prolonged stress on the test object at the test voltage. These requirements are met in general if the rate of rise above 75% of the estimated final test voltage is about 2% of the test voltage per second.
3. Once the selected voltage has been reached, the applied voltage is held on the test object for a period of one minute. Corona activity may occur during the time of holding the applied voltage.
4. Once the selected voltage is held for one minute, the applied voltage should be reduced to zero over a period of five to ten seconds.
Traditionally this test is either a pass or fail test. When performed with a series resonant test set, trending of energizing current with respect to time may reveal additional information.
Analysis of PD patterns recorded during testing will give an indication of potential problems. This may be related to activity in the slot section, inadequate spacing in the end arm region, corona, inadequate clearance in the ring bus or end arm region, or several other possible reasons.
Refer to off-line PD testing for more information.
DC testing results in the highest stress levels on the insulation system in the end arm region. Typical maintenance levels are 1.25-1.50E (E = Rated line to line voltage). Testing provides some assurance that the winding insulation has a minimum level of electrical strength required for continued operation. Testing is performed in accordance with the requirements of IEEE 95 on each phase using a 1 to 2 KV/min ramping rate, with the remaining phases grounded.
Typical maintenance levels are 1.25E1.7 to 1.50E1.7 (E = Rated line to line voltage).
The line ends of the stator windings under test and the neutral ends of the windings under test must be connected to each other whenever practical. This will minimize possible damaging surges in the event of failure or flashover during testing. Where a connection between line and neutral is very difficult, the test may be made by connecting the supply to only one end of the winding. Examples of stator test connections are shown in the figure below.
During three (3) phase testing: The line end connectors should be connected together and the star point should be intact.
During single (1) phase testing: The line end and neutral end conductors of the phases not under test must be solidly grounded.
1. Raise the test voltage slowly from zero to the selected voltage. Voltage ramp rates are typically 1 to 2 kV per minute. The initial insulation resistance test performed at 5 kV can be used to estimate current requirements at the target test voltage.
2. Once the selected voltage has been reached, the applied voltage is held on the test object for a period of one minute.
3. Upon completion of the test the test set output supply voltage control should be slowly reduced to zero over a period of 10 seconds. It is not advisable to cause a step change in the applied test voltage. It must be recognized that eliminating the test source will not remove the stored charge from the winding. The voltage should be allowed to decay to half the final applied voltage through the series resistance of the test set. The winding should then be discharged to ground through the discharge resistor of the test set or resistive grounding stick (minimum 1kΩ-2kΩ per kV maximum test voltage). Once the winding capacitance has been fully discharged, the winding should be solidly grounded.
4. The ground connection should be kept in place until the absorbed charge is completely dissipated. This may take several hours. Grounding which may be left unattended after completion of testing should be appropriately tagged with the purpose and importance of the grounding.
I would suggest reviewing “30 years of Ramp DC testing Hydrogenerator Stator Windings test results and interpretations”, Eric Eastment – USBR and Bert Milano – Consultant, 2009 Doble Client Conference.
The DC test voltage is not sufficient to stress the slot section of a stator winding to the same extent as an AC test because it is limited by the end arm region of the winding. In normal operation the surface of an end arm winding is left floating while the grading section is designed to gradually drain away this surface charge. Surface potentials are controlled by capacitance and can be ¾ of copper voltage during normal operation. However, with a contaminated machine, the end arm surface becomes a better conductor and extends the ground plane further out into the less consolidated insulation system located here. This increases stress across the insulation by resistively grading the surface voltage. Experimentation demonstrated the slot section could safety withstand voltages that would rupture the end arm region. The end arm region is also more prone to mechanical damage and cracking. Contamination in these flaws sometimes results in a failure during DC testing.
The response, and the total leakage current at test voltage are helpful in analyzing the condition of the insulation system and can be trended over time.
Winding resistance testing can detect the presence of any high resistance joints, corrosion, or fractures of the conductors. The DC winding resistance of each phase is measured using a Kelvin bridge in accordance with the requirements of IEEE 62.2.
Each phase can be tested individually when the neutral is separated. Otherwise, the phase resistance can be calculated by measuring the phase to phase resistance.
Testing is sensitive to changes in temperature.
Figure 7: Kelvin Bridge Test Connections
1. Record winding temperature using embedded temperature sensors.
2. Connect the current leads and the voltage sensing leads directly to the machine terminals.
3. Perform test measurement using DC current
4. Reverse leads
5. Repeat test.
6. Repeat testing for all 3 connections.
Pay careful attention to aspects of the winding design which could include differences in conductor length, brazed connections and bolted connections.
The stator Slot Wedge Tightness of the machine will be checked, and areas of concern would be noted. Testing is performed in accordance with the requirements of IEEE 62.2
(Future)
Visual Inspection of the stator winding assembly in accordance with the requirements of IEEE 62.2.
(Future)
The condition of the interlaminar resistance between stator punchings of the machine core is evaluated. Testing is performed in accordance with the requirements of IEEE 62.2 using a digital ElCID test set and findings are recorded.
(Future)
Corona camera or ultrasonic testing is used to identify locations of PD with the end arm region of a stator winding. A visual scan of the end arm region of both ends of the stator is performed using the daylight corona camera and locations of partial discharge activity are noted. An audio scan can localize areas of partial discharge.
For the moment I would suggest looking at “Observations on the Detection and Possible Causes of Diffuse Optical Emissions on Stator Coils Subjected to AC Voltage” by Scott Cotner, Hydroelectric Design Center(Future), USACE, 2017 EIC, and “IEC TS 62478 - High voltage test techniques – Measurement of partial discharges by electromagnetic and acoustic methods”.
If you are looking for more details, kindly visit Tan Delta Diagnostic system.