Dissipation / Power Factor and Capacitance Measurement on Rotating Machines using the CPC100/TD1 and CR500

Transcription

1 Application Note Dissipation / Power Factor and Capacitance Measurement on Rotating Machines using the CPC100/TD1 and CR500 Author s Fabian Öttl Stefan Böhler Martin Anglhuber Fabian.oettl@omicronenergy.com Stefan.boehler@omicronenergy.com Martin.Anglhuber@omicronenergy.com Date Related OMICRON Product CPC 100, CP TD1, CP CR500 Application Area Rotating Machines Version v1.0 Document ID ANP_16002_ENU Abstract This application note shows how to perform Dissipation Factor (Power Factor) and Capacity Measurements on rotating machines using the CPC 100 in combination with the CP TD1 and the CP CR500. OMICRON Page 1 of 32

2 Content 1 Using this document Operator qualifications and safety standards Safety measures Related documents Dielectric measurement on rotating machines Dissipation factor theory Dissipation factor vs. power factor Preparing the dissipation factor test CPC Editor and Test Card overview Choosing the right parameter in the TanDelta-PF test card Measurement Setup Compensation Choosing the right compensation Steps to perform the compensation Omicron TD1 High-Voltage Source Test Card OMICRON Excel compensation tool Preparation of the test object Buildup of the test circuit Overview: Dissipation factor measurement step by step Assessment of the measurement Influence of the end potential grading Practical example Fingerprint measurement Phase comparison Parameters of the dissipation factor measurement Use cases Capacitance measurement Limitation of the measurement Please use this note only in combination with the related product manual which contains several important safety instructions. The user is responsible for every application that makes use of an OMICRON product. OMICRON electronics GmbH including all international branch offices is henceforth referred to as OMICRON. OMICRON All rights reserved. This application note is a publication of OMICRON. All rights including translation reserved. Reproduction of any kind, for example, photocopying, microfilming, optical character recognition and/or storage in electronic data processing systems, requires the explicit consent of OMICRON. Reprinting, wholly or in part, is not permitted. The product information, specifications, and technical data embodied in this application note represent the technical status at the time of writing and are subject to change without prior notice. We have done our best to ensure that the information given in this application note is useful, accurate and entirely reliable. However, OMICRON does not assume responsibility for any inaccuracies which may be present. OMICRON translates this application note from the source language English into a number of other languages. Any translation of this document is done for local requirements, and in the event of a dispute between the English and a non- English version, the English version of this note shall govern. OMICRON 2015 Page 2 of 32

3 1 Using this document This application guide provides detailed information on how to measure and asses a dissipation (power) factor measurement on a high-voltage rotating machine using the OMICRON CPC 100 in combination with the CP TD1 and the CP CR500. Please refer to national and international safety regulations relevant to working with high voltage (HV). The regulation EN ("The Erection and Operation of Electrical Test Equipment") as well as all the applicable regulations for accident prevention in the country and at the site of operation has to be fulfilled. 1.1 Operator qualifications and safety standards Working on HV devices is extremely dangerous. Measurements must be carried out only by qualified, skilled and authorized personnel. Before starting to work, clearly establish the responsibilities. Personnel receiving training, instructions, directions, or education on the measurement setup must be under constant supervision of an experienced operator while working with the equipment. The measurement must comply with the relevant national and international safety standards listed below: EN (VDE 0104) "Erection and Operation of Electrical Equipment" EN (VDE 0105 Part 100) "Operation of Electrical Installations" IEEE 510 "Recommended Practices for Safety in High-Voltage and High-Power Testing" LAPG NASA "Electrical Safety" Moreover, additional relevant laws and internal safety standards have to be followed. 1.2 Safety measures Before starting a measurement, read the safety rules in the CPC 100, CP TD1 and CP CR500 Reference Manual and observe the application specific safety instructions in this application note when performing measurements to protect yourself from high-voltage hazards. 1.3 Related documents Title CPC 100 Reference Manual CP TD1 Reference Manual CP CR500 Reference Manual Application Guide: Capacitance and Dissipation Factor Measurement with CPC CP TD1 Description Contains information on how to use the CPC 100 test system and relevant safety instructions. Contains information on how to use the CP TD1 test system and relevant safety instructions. Contains information on how to use the CP CR500 test system and relevant safety instructions Contains general information about dissipation factor (power factor) measurement on different assets. OMICRON 2015 Page 3 of 32

4 2 Dielectric measurement on rotating machines The following document applies to all kind of vacuum impregnated or resin-rich taped single bars, coils or complete windings. Dielectric tests of rotating machines can help to identify potential insulation problems before defects or damage occurs. The dielectric measurement is performed with high AC voltage at nominal frequency in order to have a close to realistic behavior of the insulation system. The test is non-destructive for healthy insulation systems and is performed during a standstill of the machine. 2.1 Dissipation factor theory In an ideal capacitor without any dielectric losses, the insulation current is exactly 90 leading according to the applied voltage. For a real insulation with dielectric losses, this angle is less than 90. The insulation of a rotating electrical machine can be modelled by a loss-free capacitance with a parallel ohmic resistance representing the losses in the insulation system (Figure 1). The latter ones are caused by: Surface currents Leakage currents Partial discharges Polarization losses Figure 1: Parallel equivalent circuit diagram and vector diagram of a rotating machine insulation system. Applying a voltage to the parallel components causes a current IC to flow through the capacitance as well as through the resistivity IR. The overall current I therefore has a resistive and a capacitive component. A causal relation between losses and the resistive part can be assumed, as the higher the losses are, the higher the resistive current will be. The angle δ = 90 - φ is caused by the resistive part of the overall current and is proportional to the losses, which leads to the definition of the tan(δ) expressed in Figure 1 as an indication of the overall condition of the insulation system of an electrical machine: 1 tan (δ) = ω C P R P OMICRON 2015 Page 4 of 32

5 2.2 Dissipation factor vs. power factor Another parameter commonly used in North America when speaking of dielectric tests on rotating machines is the so-called power factor. In this case, the definition is different from the dissipation factor. Assuming again that the vector diagram in Figure 1 shows power factor (PF) as the quotient of the resistive part of the current IR to the overall current I: PF = I R I = cos φ tan δ PF = 1 + tan² δ If the dissipation factor (tanδ) is very small typically less than 10 %, which can be presumed as given when measuring healthy electrical machine insulation, the dissipation factor and the power factor differ in a negligible amount and can be assumed to have the same value. Table 1 gives an idea about this statement. Table 1: Comparison between correlating values of dielectric power factor cos(φ) and dielectric loss factor tan(δ) and their difference (IEC 60034_27_3). Switching between the two forms can be done in the CPC software very easy, which is explained in Figure 2. For reasons of simplicity and to avoid repetition, only the dissipation factor is mentioned in the following chapter. As the difference between dissipation factor and power factor is minimal. When testing rotating machines, all of the statements in the following chapter apply to the dissipation factor as well as to the power factor. OMICRON 2015 Page 5 of 32

6 Figure 2: Switching between the two forms of parameter in a dielectric measurement and overview of the OMICRON rotating machine template according to the CPC Start page. OMICRON 2015 Page 6 of 32

7 3 Preparing the dissipation factor test The preparation of the test can be easily done in the office with the CPC start page. OMICRON offers a template for rotating machine testing, which can be adapted and extended to your individual needs. Figure 2 shows this template, which is described in this chapter in more detail. 3.1 CPC Editor and Test Card overview The generator template can be found in the CPC start page (Figure 3). Figure 3: CPC start page with a path to the generator template By opening the template file, the CPC editor appears with the test card combination shown in Figure 4. The default template consists of two comment test cards and eight dissipation factor test cards. The build-up is shown according the test procedure and requires a capacitance measurement and a check of the resonant circuit after the nameplate information of the generator. The reason for these two measurements are explained in chapter 4. After performing the measurement, the Excel file loader template automatically adopts the measurement data from the measurement file for a fast and easy report. Both tools the test template with the test cards and the excel file loader can be adapted by the user according to their needs. For a detailed description of the CPC start page, please refer to the CPC 100 reference manual. OMICRON 2015 Page 7 of 32

8 Figure 4: View of the generator template in the CPC Editor Choosing the right parameter in the TanDelta-PF test card The highest voltage during the test is defined in national or international standards, or it is given by the internal test procedures according to the operator needs. OMICRON suggests to measure the dissipation factor in steps of 10 % of the maximum voltage defined for the machine. The resulting ramp is driven upwards (i.e. from 10 % to 100 % of the defined maximum voltage) and downwards (from 100 % to i.e. 10 % of the maximum voltage). The reason for driving the ramp in both directions is a plus on information out of the results (more details are explained in chapter 5 ). Steps of 20 % of the test voltage are also common. The only difference to the OMICRON suggestion is a lower resolution of the partial discharge measurement, which is potentially performed in parallel. The voltage steps are defined in the Auto Test Point List in the TanDelta-PF test card of the CPC. If doing the measurement in steps up AND downwards, it is not possible to choose exactly the same voltage for both ramps as no second auto test point with the same value can be set in the device. In this case, simply choose a voltage which differs by 1 V (in the example in Figure V instead of 6000 V) from the chosen value. Figure 5: Left side: Auto Test Points slightly differ in the upwards and downwards ramp; Right side: Number of measurements for an averaging result OMICRON 2015 Page 8 of 32

9 If the measurement frequency is the same as the default frequency in the device (the default frequency can be checked by pressing the options button on the device and go to the device setup tap), then the CPC uses a noise suppression algorithm, which performs a measurement on two different frequencies then on the rated frequency. The result is an interpolation of the two measurements. As for rotating machines, the signal-to-noise level is much higher, so the noise suppression algorithm is not needed. Additionally, the compensation (see chapter 4.1 ) is optimized for one frequency. Therefore, please set the default frequency in the device to a frequency different than the test frequency. A workaround could be to test at Hz if 50Hz is the rated frequency, or Hz with 60Hz as rated frequency. With the number of averaging marked in Figure 5, the number of measurements is chosen to determine a single result. In the example shown in Figure 5 above six measurements are averaged to one result and are performed at each voltage step. When a PD measurement is performed simultaneously, the duration of one voltage step can be varied by using the averaging number. On the CPC display (not in the CPC editor), the approximate duration of the step is displayed when changing the value. The measurement of the dissipation factor has to be performed in GSTg-A+B mode in order to guard the current from the CP CR500 parallel path (see also Figure 16). The same applies to the capacitance measurement. The resonance frequency check is performed in GST mode in order to measure the whole measurement circuit. 4 Measurement Setup To ensure a lightweight system with a minimum need of real power, the OMICRON approach is to use a parallel resonant system together with the test capacitance. This means that the test object s large capacitance (represented by CTest) causes a huge apparent power when applying high voltage, but this is compensated with parallel resonators (CR 500). The remaining required power is caused by losses and the rest of the apparent power which is not ideally compensated. This is supplied by the CPC 100 in combination with the CP TD1. To get a better idea of the concept please refer to the following chapter on Compensation. OMICRON 2015 Page 9 of 32

10 4.1 Compensation The picture a) in Figure 6 shows the principle of a parallel resonance system; where the parallel resonators and the test capacitance are considered as ideal, causing a minimum of resulting test current Itotal as a vectorial sum of ITest and IComp. a) b) Figure 6: Equivalent circuit diagram of the parallel resonance circuit with vector diagram Picture a) in Figure 6 represents an ideal parallel resonance circuit, where the theoretical capacity and inductivity are free of losses. Picture b) in Figure 6 represents the vector diagram with real lossy components. The compensation of the reactive part is not done in a perfect way, and the total current Itotal has a real part as well. The real part results from the losses of the test object itself (resulting in the loss angle δ in the picture), from losses in the cables to the test object and from the losses of the compensation reactor. Nevertheless Itotal is usually much smaller than the needed test current ITest at a certain voltage and can be supplied by the CPC 100 in combination of the CP TD Choosing the right compensation For a better understanding about which parallel resonators (CR 500) are used for a certain capacitive load, an example of a real measurement on a 56 MVA machine is presented. The hydro generator to test has the following data: 10 kv 352 nf capacitance phase to ground 50 Hz The nominal voltage is considered as test voltage for a single phase measurement. OMICRON 2015 Page 10 of 32

11 Figure 7: Tan(delta) testing without compensation reactors Without compensation, the source has to provide the full capacitive current ITest, including the whole active part ITa (losses of the test object) as well as reactive part ITr. This case is displayed in Figure 7, where the reactive part of the test current as usual is much higher than the active part. As mentioned above, the overall maximum apparent power, which the CP TD1 has to provide, can be calculated as follws: S = U Test I Test = U 2 jωc Test = 11,1kVA By establishing a parallel resonant circuit, the capacitive reactive part ITr can be compensated by the CP CR 500. Considering the ideal circuit in Figure 6, the system can be assumed in resonance if the following condition is fulfilled: 1 = 2 π f L 2 π f C f C L test frequency capacitance of test object compensator reactance The calculation of the compensation is then: L = 1 4 π 2 f 2 C f C L test frequency capacitance of test object compensator reactance In our example, a parallel reactor of LComp = 28,78 H would be required for a perfect compensation displayed in Figure 8. The resulting current ITotal is the remaining sum of the active part from the lossy inductance as well as from the lossy capacitance. OMICRON 2015 Page 11 of 32

12 Figure 8: Parallel resonant circuit with vector diagram As OMICRON offers 80 H and 40 H reactors, the best compensation can be achieved by connecting one 40 H reactor and an 80 H reactor in parallel, which results in a reactance of 26,6 H. Figure 9: Compensation with OMICRON reactors The current ITotal is now slightly higher than in Figure 8 and will be in the range of 100 ma depending to the losses in the test object and the measurement circuit in general. Nevertheless, a perfect compensation is not required as the CP TD1 can provide the required current of 100 ma without problems. The loss angle δ is not influenced as the current through the compensation coils is guarded. OMICRON 2015 Page 12 of 32

13 In general, an inversely proportional relation between the test capacitance and the compensating reactance is given (i.e. a larger capacitance requires a smaller compensating inductance please see the formulas above). Smaller values for the compensating inductance can be achieved by connecting multiple CP CR500 reactors in parallel. The following diagram shows the types and quantities of CR 500 that are required to compensate a certain capacitance. It also shows the current ITotal, which is not compensated and has to be supplied by the CPC 100 and the CP TD1. The test voltage is assumed as 12 kv, the maximum output current of the CP TD1 is 300 ma. Figure 10: Example of different compensation connections by a given test capacitance in nf Without any compensation, the CPC 100 in combination with the CP TD1 is able to supply an 80 nf capacitive load. If the test object has a capacitance of e.g. 120nF, a compensating reactor of 80 H needs to be connected in parallel to reach the 12 kv. Compensating reactors have to be used to ensure the compensation of capacitances larger than 80 nf and also for testing on nominal frequency. A combined 40 H / 80 H can be used to compensate for capacities up to about 450 µf. For larger capacitances, further 40 H reactors have to be connected in parallel. Due to the insulation level of the LV output of the CP CR 500, connecting in serial of the coils is not permitted, and will cause a damage of the device! Choosing the best measurement configuration, respectively the best compensation, is described in the following chapter OMICRON 2015 Page 13 of 32

14 4.1.2 Steps to perform the compensation There are two possible ways to calculate the best compensation -- either you can do it with the CPC100 or on your computer by using the OMICRON calculation tool. Both alternatives require the capacitance of the test object against ground. If this value is not given by a former measurement or a factory acceptance test, you have to measure the capacitance on site using the following setup without compensation: Figure 11: Measurement setup for capacitance measurement The OMICRON test template for rotating machines in the CPC Start page provides the C-MEAS test card for measuring the capacitance of the test object. Measurement of the capacitance on the test object After calculating and connecting the necessary compensation reactors, the resonance frequency is measured to ensure that the compensation has been done correctly. Figure 12: OMICRON rotating machines test card overview Performing a tan-delta test for each phase (in this case additional test-cards are in the template for measuring the dissipation factor over a frequency range) OMICRON 2015 Page 14 of 32

15 Please note: The reference frequency test card is a tool for double-checking if the compensation is performed correctly. The measurements, which identify the dissipation factor of the single phases, can be performed at rated frequency. After getting the value for compensation, the compensation calculation can be done either: Manually (please refer to the chapter ) By using the OMICRON high-voltage test card By using the OMICRON excel compensation tool Omicron TD1 High-Voltage Source Test Card Using the TD1 High-Voltage Source Test Card is one possibility to measure the test object capacitance, to calculate the correct compensation and to measure the resonant frequency is. The TD1 High-Voltage Source Test Card is designed for using the CP TD1 as a high-voltage source, but is not able to measure the dissipation factor. However, the feature to measure the capacitance, calculate the compensation and to measure the resulting resonant frequency is available. Figure 13: High-Voltage Source Test Card; Detail Following the description of the points in Figure 13: 1. Insert a test voltage: The test voltage should be lower than the rated voltage of the machine. Please keep in mind that no compensation is done when measuring the capacitance, therefore a voltage level between 500 V and 2000 V is suggested. 2. Check the capacitance of test object and calculate the compensation: The calculation of the compensation is done automatically. After the definition of the available CP CR500 coils, the device displays the best measurement setup. 3. Second step: Build up the test setup. By pressing the options button, the screen in Figure 14 will appear. In the TD1 High-Voltage Source Test Card, it is possible to enter the available amount and type of CP CR500 to calculate the best suitable compensation in the test setup. Connect them according to the suggestion of the test card. 4. Double check if the compensation is performed correctly by measuring the resonance frequency. The corresponding wiring diagram is displayed by pressing the Wiring button. Please note: Even if the resonance frequency is not equal to the rated frequency, the measurement can be performed at nominal frequency, as the CPC 100 in combination with the TD1 is able to supply the missing reactive power. OMICRON 2015 Page 15 of 32

16 Available amount and type of CR500 Suggested amount and type of CR500 to use and calculated resonant frequency Figure 14: View of the high-voltage test card by pressing the options button OMICRON Excel compensation tool The calculation tool for the calculation of the necessary compensation for a dedicated test object can be downloaded in the customer area of the OMICRON homepage. This tool calculates the required current for all possible combinations of all coils currently available for the test. The current is the total needed current from the TD1 and includes: The remaining current from the compensation if the resonant frequency is not the test frequency The active part of the current as result of the dissipation factor (estimated value, based on OMICRON experience) The active part of the current from the losses of the CR500 coils, as an average value over a certain test sample amount. Figure 15: Input Form - All yellow cells are required input values provided by the use OMICRON 2015 Page 16 of 32

17 Following is a short explanation of the different input values: Test frequency: Frequency at which the test is performed (usually rated frequency). Highest test voltage: The highest voltage reached during the test. Test capacitance: Capacitance of the test object including additional capacitances (e.g. coupling capacitors). Actual inductance required for perfect compensation: Theoretical value of inductance required for compensation exactly to the test frequency (test frequency = resonant frequency). Available coils: Enter the amount of available coils for the measurement (Please note: the amount of CR500 amount of coils, 1 x CR500 = 2 coils). Coils for best compensation: Best combination of coils (with lowest current) for compensation at the entered test frequency. Actual inductance (with suggested coils): Total inductance as a result of the suggested coils. After you fill in the input form, click on the button Optimize compensation and all the calculations will be executed. The output is similar to the one in the high-voltage source test card explained in chapter Preparation of the test object On rotating machines, usually phase-to-ground of the stator insulation is measured. As a comparison of the different phases is very helpful for the analysis, it is important to separate the phases from each other (e.g. by opening the star point). Any other electrical components, which are directly connected to the phases like cables etc., should also be disconnected, as they would otherwise influence the measurement. If this is not possible, it has to be considered that these components will be included in the measurement as well. 4.3 Buildup of the test circuit Before Starting Make sure that the CPC 100, the CP TD1 and the CP CR500 are properly grounded and the generator is disconnected from the line. Always touch the leads and terminals with a grounding rod first. For your own safety: Be sure to ALWAYS follow the following five safety rules: 1. Disconnect mains 2. Prevent reconnection 3. Test for absence of harmful voltages 4. Ground and short circuit 5. Cover or close off nearby live parts OMICRON 2015 Page 17 of 32

18 The buildup of the test circuit is displayed in Figure 16, where some components are marked with numbers. The following list gives an overview about the different components: 1) Grounding cable 2) CPC 100 booster cable 3) Serial cable from CPC 100 to CP TD1 4) Safety connection to the CP CR500 5) Guard connection to the CP CR500 6) High-voltage cable from CR500 to the test object 7) CP TD1 high-voltage cable Figure 16: Setup of the test circuit with numbered items Following is a description about the items listed above: 1) Grounding Cable: Before proceeding with the test setup or switching on the CPC 100, please ensure that all devices are grounded properly. 2) CPC 100 booster cable: The power connection from the CPC 100 to the voltage booster CP TD1 3) Serial cable from CPC 100 to CP TD1 Ensures the communication between the CPC 100 and the CP TD1 4) Safety connection to the CP CR500 The safety connection to the CP CR500 implements the resonator in the safety system of the CPC 100. If there is an issue with overcurrent (i.e. due to improper test setup), the CP CR500 will stop the test automatically. The connection is done with the special accessory cable displayed in Figure 17 on the left side, where the serial plug replaces the CPC 100 safety dongle, and the special plug terminares the SAFETY A input (Figure 17, right side). The safety dongle also terminates the SAFETY B plug. OMICRON 2015 Page 18 of 32

19 By using more than one CP CR500, the connection between the different CP CR500 units is performed with the special accessory cable. The SAFETY B input of the last CP CR500 is terminated by the safety dongle. Figure 17: Connection cable for implementing the CP CR500 into the CPC 100 safety concept 5) Guard connection to the CP CR500: The guard connection to the CP CR500 ensures that the resonator itself is not influencing the measurement results. The guard connection is performed with the CP TD1 measurement cable connecting special plug into the banana socket (See Figure 17 on the left side). 6) High-voltage cable from CR500 to the test object: The connection from the CP CR 500 to the test object is performed with the red cables in Figure 18 by using the clamp which provides enough sockets for 4 coils and the CD TD1 high voltage cable. Figure 18: Connection cables from the CP CR500 to the test object and clamp for connection to the test object. The CP CR500 cables are non-shielded cables and therefore they can be the starting point of partial discharge activities, when touching ground potential. Especially when performing a tan(delta) measurement and a partial discharge (PD) measurement in parallel, this factor influeces the PD measurement. The picture in Figure 19 on the right side shows a PD-free test setup for a parallel measurement of the tan(delta) and PD activity of the machine. Figure 19 OMICRON 2015 Page 19 of 32

20 1) High-voltage supply for combined dissipation factor and partial discharge measurement 2) Connections to the terminals are kept at a safe distance with Nomex paper from any ground potetial 3) Plastic material tube to ensure that the cables are not touching the floor Figure 19: Measurement set-up at a generator site for a combined dissipation factor and partial discharge measurement (note that the red cables are not touching ground) The picture on the left side in Figure 19 shows a fully-connected CP CR500 with the special plug of the safety cable, the measurement cable from the CP TD1 connected to the banana socket of the cable (blue banana plug) and the safety dongle terminating the safety B plug. In this case, both coils of the CP CR500 are used. The housing is grounded via the grounding cable and is connected at the grounding screw. 7) Blocking Impedance BLI1: A passive filter device when measuring tan(delta) and partial discharge in parallel, which allows the partial discharge measurement at the same point where the injection is taking place. The BLI1 blocks the high frequency disturbances coming from the source. 8) CP TD1 high voltage cable The connection to the CP TD1 is a shielded cable where the shield current is guarded internally. Important for not influencing the measurement results is the section to the right of the connection to the test object, which has a stress coating in order to ensure a smooth potential grading from the high voltage to the cable shield. Under high voltage, this section should not touch any grounded parts and respectively should have a proper distance to any grounded part according to the high voltage output. An example is shown in Figure 20, where the first section of the cable from the blocking impedance BLI 1 is not touching any ground component. OMICRON 2015 Page 20 of 32

21 Figure 20: High-voltage connection to the BLI 1 for a combined dissipation factor and PD measurement with coupling capacitor and the MPD system OMICRON 2015 Page 21 of 32

22 4.4 Overview: Dissipation factor measurement step by step Preparation Prepare the test in the office Capacitance of DUT Measure the capacitance if the value is not given by another source Due to the symmetric design, all three phases have similar values Compensation Connect the required amount of CP CR500 coils to establish the parallel resonance circuit Perform the Test Double check of the right compensation by measuring the resonant frequency Perform the dissipation factor mesaurement Analyzing of the results Analyze the test results by exporting the data to the Excel file loader OMICRON 2015 Page 22 of 32

23 5 Assessment of the measurement Rotating machines are small serial products or non-serial products with wide spread material characteristics. An overall limit regarding a dissipation factor does not exist, even in consultation with the manufacturer. Nevertheless, the dissipation factor measurement gives a reliable statement regarding the overall state of the rotating machine by comparing fingerprint measurements using the same conditions. These are explained in the following chapters. 5.1 Influence of the end potential grading The End Potential Grading (EPG) is a semi conductive area in the end winding of every medium voltage machine. It can be found under various names in the literature: semi conductive grading, end winding stress grading, or the mentioned EPG are only a few names, which all indicate the same area, as shown in Figure 21. Figure 21: Position of the End Potential Grading (EPG) at the single bar The semi conductive paint or tape causes a defined reduction of the high-voltage potential from the soldering contacts to the grounded laminated core. Even if the insulation is applied at the end winding area, this area is on high-voltage potential due to the capacitive coupling. The slot portion by definition has ground potential, therefore a boundary surface with high potential gradients is created at the end winding area. Without the EPG, high electrical field gradients would appear at the laminated core, causing high discharges as displayed in Figure 23. Figure 22: Simulated field strength without EPG and with EPG OMICRON 2015 Page 23 of 32

24 Figure 22 shows the theoretical simulation of the EPG area, which smoothens the electric potential at the boundary surface from high-voltage to ground. Figure 23: Different test probes without EPG (top); EPG on the left side (middle); and with EPG (bottom); (Weidner, 2008) Silicon carbide (SiC) is the primary component and it is responsible for the behavior of the end potential grading area during the dissipation factor test. It is well known in the abrasives industry and is applied as a filler in a varnish or a tape component at the bar. The electrical behavior can be assumed as micro-varistor. Its conduction mechanism is displayed in Figure 24. Figure 24: Schematic illustration of the influence of different conduction mechanisms of a Schottky barrier After reaching the breakdown voltage VBD, the Schottky barrier acts as a conductor. For the dissipation factor measurement this results in a voltage-dependent conducting behavior in the endwinding zone. It acts as a bypass to the ground wall insulation. Depending on the material properties, the avalanche point is reached sooner or later, and makes the dissipation factor measurement a fingerprint measurement as stated above. This leads to the equivalent circuit diagram shown in Figure 25. OMICRON 2015 Page 24 of 32

25 Figure 25: Equivalent circuit diagram of the end potential grading Another influence of the EPG behavior can be seen in the capacitance measurement. As with higher voltage, the SiC-layer becomes more and more conductive. Depending on the voltage, it extends the OCP layer. Therefore, a rise in capacity value is normal during a measurement starting from 0,1 Un up to 1,0 Un (Un is the highest test voltage). Also PD-activity (partial discharge activity) influences the capacitance value by changing the dielectric properties of the voids in the insulation. The effect of capacitance increase can be observed in every machine, no matter if they were in operation over years (most likely high PD-activity) or new machines (most likely low PD-activity). Figure 26: Current paths during a tan(delta) measurement EPG currents are mentioned as Surface Currents OMICRON 2015 Page 25 of 32

26 5.1.1 Practical example To point out the influence of the end potential grading, one limb of a 6kV coil was modified by scratching an isolating gap in the OCP. This allows a measurement with the possibility of guarding the surface currents and measure the pure dielectric parameters of the ground wall insulation. The principle is explained in Figure 27, where the simulated slot section is representing the measuring electrode and the guard electrodes were used as Input A or B in the CP TD1. Figure 27: Modified limb of the coil according to IEEE 286/2000, 2001 The results are represented in Figure 28, where on the left side the measurement setup without guard electrodes and on the right side the measurement with guard electrodes are displayed. Figure 28: Results of the different measurement setups For the measurement without guarding rings, a much higher increase of the tan(delta) can be observed, as the EPG is relatively conductive and responsible for an almost linear increase. The initial value for both diagrams is almost equal, since the EPG is not that conductive at 600V. On the right side of Figure 28, the tan(delta) value is almost equal until 3kV and starts increasing slowly up to 2% compared to 3.75% in the measurement setup without guard rings. OMICRON 2015 Page 26 of 32

27 5.2 Fingerprint measurement As the dissipation factor measurement is a fingerprint measurement and heavily influenced by the end winding surface currents (as described above), the condition of the end winding area should be the same for a reproducible measurement. Therefore it is suggested to clean this part of the winding to ensure a clean and dry surface and to give reproducible conditions for every measurement without any additional current paths. IMPORTANT: Prior to cleaning the end winding area, please look for any traces of PD (e.g. white powder) and document it for the parallel dissipation factor and PD-measurement (Figure 29). The traces provide an indication of the PD activity in the end winding and also the possible failure mechanism. Figure 29: Partial discharge traces in the end winding area ( white powder ) The comparison of different measurements over time under the same circumstances (i.e. end winding surface, temperature, humidity) offers the best possibility to assess the condition of the winding Phase comparison If there is no fingerprint measurement available, a phase comparison can be a proper equivalent solution for a reliable statement regarding the phase condition. Due to the symmetric buildup of generator stator windings, the results of different phases vary only slightly. OMICRON 2015 Page 27 of 32

28 Figure 30: Example of a phase comparison The measurement represented in Figure 30, shows an example of a phase comparison of a healthy winding. A comparison between the phases shows a similar increase of tan(delta) between each steps and the absolute value of the dissipation factor, as well as the hysteresis of the ramp up and downwards. A trend analysis with a measurement performed four years earlier showed that the values have shown consistent behavior over the interval. 5.3 Parameters of the dissipation factor measurement Four parameters are essential for analyzing the results of a dissipation factor measurement: Dissipation factor at low voltages (first voltage step) as a fingerprint value Delta tan(delta) in each interval to identify the highest tan(delta) increase Characteristic tan(delta) tip-up between different pre-defined voltage steps (e.g. 0.6Un and 0.2Un) Shape of the curve of the tan(delta) ramp By using the OMICRON approach with the decreasing ramp, a fifth parameter, the PD hysteresis, can be included in the analysis. This gives a statement about the PD activity in the winding (keywords: inception and extinction voltage). A higher tan(delta) hysteresis is an indicator of higher PD-activity in the ground wall insulation. Generally an assessment can be done by comparing the absolute tan(delta) values or using the graph of dissipation factor over voltage Use cases The following examples of different measurements should give an idea of the diversity of possible dissipation factor graphs. The measurements are from different high-voltage machines, representing different stages of deterioration. A PD-measurement was performed in parallel. OMICRON 2015 Page 28 of 32

29 Figure 31: Tan(delta) measurement example (1) The example in Figure 31 shows a relatively constant dissipation factor value at lower voltages. At 4 kv, the EPG becomes more conductive. Approximately at the same voltage, PD activity also starts. This was confirmed by the parallel PD measurement. The rise of tan(delta) is approximately 2 % starting from 1 % at 1 kv up to 3 % at 12 kv. Figure 32: tan(delta) measurement example (2) The example in Figure 32 shows a generator from 1968 with low PD activity (no PD hysteresis until 6kV) and a relatively non-conductive EPG up to 4kV. In general the tan(delta) value is quite low for an entire winding measurement and increases from 0.6 % up to 0.85 %. OMICRON 2015 Page 29 of 32

30 Figure 33: tan(delta) measurement example (3) Figure 33 shows results from a brand new generator. The machine shows no PD activity but has a very conductive EPG. Therefore, the tan(delta) values are rising up to close to 4 %. Another parameter increasing the dissipation factor is the fact that the machine is a very slow hydro generator. This means a high-end winding surface area in comparison with the slot portion Capacitance measurement The capacitance measurement is done in parallel to the tan(delta) measurement and acts as a fingerprint measurement, as in the case with the dissipation factor. As mentioned above in the EPG description, the capacitance will increase due to the behavior of the semi conductive area. Figure 34 shows two different capacitance measurements with a maximum increment of ~3% and ~4%. a) b) Figure 34: Capacitance measurement of two different generators with an increment of the capacitance value of ~3% and ~4% OMICRON 2015 Page 30 of 32

31 Figure 34 a) is the capacitance measurement from the dissipation factor measurement in Figure 33 and Figure 34 b) is correlating with the measurement in Figure Limitation of the measurement The measured dissipation factor is an integrative value over the whole winding. Therefore, the value does not represent the most deteriorated part of the winding, as a large dissipation factor value can be due to voids distributed all over the winding or a few single heavily aged bars or coils. The latter case is potentially more severe regarding the longevity of the winding. In such cases, where high dissipation factor values or high tip-up factor values can be observed, further investigations like visual inspections or PD measurements should be executed. OMICRON 2015 Page 31 of 32

32 OMICRON is an international company serving the electrical power industry with innovative testing and diagnostic solutions. The application of OMICRON products provides users with the highest level of confidence in the condition assessment of primary and secondary equipment on their systems. Services offered in the area of consulting, commissioning, testing, diagnosis, and training make the product range complete. Customers in more than 140 countries rely on the company's ability to supply leading edge technology of excellent quality. Broad application knowledge and extraordinary customer support provided by offices in North America, Europe, South and East Asia, and the Middle East, together with a worldwide network of distributors and representatives, make the company a market leader in its sector.