SlideShare a Scribd company logo

High Voltage On-Site Testing with Partial Discharge Measurement (Cigre 502)

AHMED MOHAMED HEGAB, profile picture
AHMED MOHAMED HEGAB

The document details procedures and guidelines for high-voltage on-site testing with partial discharge (PD) measurement, emphasizing its role in assessing the insulation condition of high-voltage apparatus throughout their lifecycle. It discusses the importance of combining withstand voltage tests with PD measurements to identify insulation defects and improve diagnostic accuracy. The content also outlines various testing methods and conditions required for effective PD measurement, aiming to enhance quality assurance and support the maintenance of high-voltage systems.

Engineering
1 of 70
Downloaded 58 times
1
2
3
4
5
6
Most read
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Most read
21
22
23
24
25
26
27
28
29
30
31
32
Most read
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
502 High-Voltage On-Site Testing with Partial Discharge Measurement Working Group D1.33 June 2012
High-Voltage On-Site Testing with Partial Discharge Measurement Working Group D1.33, Task Force 05  Members R. Pietsch (Convenor: 2007-2010) (DE), W. Hauschild (Convenor: 2004-2007) (DE), J. Blackett (AU), R. Bodega (NL), A. Claudi (DE), B. Finlay (AU), M. Gamlin (CH), E. Gockenbach (DE), E. Gulski (NL), R.A. Jongen (NL), T. Leibfried (DE), E. Lemke (DE), S. Meijer (NL), P. Mohaupt (AT), Michael Muhr (AT), F. Petzold (DE), E. Pultrum (NL), G. Rizzi (I), T. Strehl (DE), C. Sumereder (AT), S. Tenbohlen (DE), P. Werle (DE) Copyright © 2012 “Ownership of a CIGRE publication, whether in paper form or on electronic support only infers right of use for personal purposes. Are prohibited, except if explicitly agreed by CIGRE, total or partial reproduction of the publication for use other than personal and transfer to a third party; hence circulation on any intranet or other company network is forbidden. As an exception, CIGRE Collective Members only are allowed to reproduce the publication. ”. Disclaimer notice “CIGRE gives no warranty or assurance about the contents of this publication, nor does it accept any responsibility, as to the accuracy or exhaustiveness of the information. All implied warranties and conditions are excluded to the maximum extent permitted by law”. ISBN : 978-2-85873-194-7
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 3 - High-Voltage On-Site Testing with Partial Discharge Measurement Content 1.  Introduction ................................................................................................................. 5  1.1  Purpose of HV Dielectric On‐Site Tests with PD Measurement .................................. 5  1.2  HV Test Procedures with PD Measurement ............................................................... 6    2.  High Voltage Sources and Accessories for On‐Site Applications .................................... 8  2.1  General Requirements .............................................................................................. 9  2.2  Continuous Alternating Voltage (AC) ...................................................................... 10  2.3  Damped Alternating Voltage (DAC) ........................................................................ 13  2.4  Very‐Low Frequency (VLF) ...................................................................................... 15  2.5  HV Filter, Coupling Capacitor, Connections and Grounding for Conventional PD              Measurement ......................................................................................................... 16    3.  On‐site PD Measurements ......................................................................................... 17  3.1  Conventional Electric PD Measurements ................................................................ 17  3.2  Non‐conventional Electromagnetic PD Detection ................................................... 19  3.3  Noise Reduction ...................................................................................................... 22  3.4  Acoustic PD Detection ............................................................................................ 24  3.5  Important Aspects for PD Evaluation ...................................................................... 25  3.5.1  PD Identification ................................................................................................. 25  3.5.2  PD Localization ................................................................................................... 27    4  Preconditions for On‐Site Testing Including PD Measurement .................................... 28    5  Examples of Test and Measuring Techniques for Apparatus and Systems .................. 29  5.1  Gas‐Insulated Systems (GIS/GIL) ............................................................................. 29  5.1.1  HV Source, PD Measurement and Details of Test Object ..................................... 29  5.1.2  Example 1: On‐Site Test by ACRL (Tuneable Reactor Circuit) Test System and                    Acoustic and UHF PD Measurement Technique ................................................... 32  5.1.3  Example 2: On‐Site Test by ACRF Test System (Tuneable Frequency Circuit)and                  High Frequency PD Measurement Technique ....................................................... 33    5.2  Cable Systems ......................................................................................................... 34  5.2.1  HV Sources and PD measurement ....................................................................... 34 
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 4 - 5.2.2  Required Details of Test Object ........................................................................... 37  5.2.3  Example 1: HV Cable Systems Tested by ACRF Test System and                   Conventional PD Measurement Technique .......................................................... 38  5.2.4  Example 2: MV/HV PILC Cable System Tested by DAC Test System and                   Conventional PD Measurement Technique .......................................................... 40  5.2.5  Example 3: Extruded EHV Cable System Tested by ACRF Test System and                   Non‐Conventional PD Measurement ................................................................... 42  5.2.6  Example 4: MV Cable System Tested by VLF Test system and                   Conventional PD Measurement Technique .......................................................... 44  5.2.7  Example 5: UHF PD Measurement at Cable Accessories in Service....................… 44    5.3  Rotating Machines ................................................................................................. 46  5.3.1  Pre‐Conditions .................................................................................................... 46  5.3.2  Example 1: On‐Site Tests by ACTC and ACRL Test Systems and                   Conventional PD Measurement Technique .......................................................... 47  5.3.3  Example 2: On‐Site Tests by ACRF Test System and                   Conventional PD Measurement Techniques ........................................................ 48  5.3.4  Example 3: On‐Site Tests by DAC Test System and                   Conventional PD Measurement Technique .......................................................... 50    5.4  Power Transformers ............................................................................................... 51  5.4.1  General ............................................................................................................... 51  5.4.2  Preconditions of Test Object ................................................................................ 52  5.4.3  Example 1: On‐Site Test by Motor Generator Set and                   Conventional PD Measurement Technique .......................................................... 54  5.4.4  Example 2: On‐site test by Frequency Converter and                   Conventional PD Measurement Technique .......................................................... 55  5.4.5  Example 3: UHF PD Measurement and Acoustic PD Localization in Service and                   for On‐Site Tests .................................................................................................. 57  5.4.6  Example 4: Synchronous Multi‐Terminal PD Measurement for On‐Site Testing ... 60    6.  Conclusion.................................................................................................................. 62    7.  References ................................................................................................................. 63 
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 5 - 1. Introduction 1.1 Purpose of HV Dielectric On-Site Tests with PD Measurement During the life cycle of high voltage (HV) apparatus or systems (Figure 1.1), many tests and measurements are performed to characterize the insulation condition. The results of these tests and measurements should be compiled in a “life data record”, which supplies information on trends of diagnostic indicator values. The HV on-site test with partial discharge (PD) measurement has an intermediate position between routine tests and in-service monitoring measurements (on-line or off- line): Figure 1.1: Data sampling and recording during the life cycle of HV apparatus or systems In addition to type and routine testing, HV on-site testing is an important part of quality assurance. On- site tests are applied:  as a part of commissioning of equipment on-site to demonstrate that the transport from manufacture to site and erection on-site have not caused any critical defects, which lower the dielectric withstand voltage of the insulation below coordination withstand level Ucw [1.7];  after repair on-site to demonstrate that the equipment has been successfully repaired;  for diagnostic purposes in the framework of the asset management program by providing reference values of diagnostic indicators (e.g. partial discharge values and dielectric parameters) for later trending of test results. On-site tests are normally applied after a warning is received from the condition monitoring program. Furthermore it has to be considered that depending on the particular condition of the HV component:  the expectations from testing depend on the test procedure as selected for the particular component;  some degradation and breakdown risks may be related to the level and the duration of withstand voltage stresses as applied during on-site testing;  voltage withstand testing at the selected voltage level gives go/no-go information for this test voltage level and requires confirmation by methods such as PD measurement, of possible field strength effects on the insulation system.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 6 - HV dielectric withstand testing plays an important role in quality assurance and consequently in contractual matters between manufacturer and user of HV equipment. On-site testing, diagnostics and monitoring should ensure the reliability and remaining life time of an insulating system, and support corresponding decisions of the user (in accordance with their policy related to insurance). The different aims of HV testing on the one hand and insulation diagnostics/monitoring on the other may be seen as having quite different testing philosophies (Table 1.1). But the common aim of the measurement whether for HV testing or condition monitoring is the behaviour of the dielectric in the electric field. Therefore the tests and measurements in the factory, on-site and in-service should be harmonized. This offers the possibility of improvement of on-site test procedures supported by PD measurement. PD measurements play an important role in both quality testing and monitoring for condition assessment. As opposed to diagnostics such as dielectric measurements (tan , εr and dielectric response parameters) which are characteristics of the whole dielectric volume, partial discharges are related to local insulation defects. These defects in the insulation (“weak-points”) will result in PD activity prior to breakdown under certain stresses (e.g. mechanical, chemical, thermal or electrical stresses). In the case of the new insulation, critical defects are the result of assembling failures which can be found by routine testing consisting of the HV withstand test including PD measurement. In case of in service HV equipment insulation (which has been tested successfully and which operates for years), a critical defect might be caused by high electrical, thermal or mechanical stresses or by the “aging” of the insulation itself. This means the partial discharges are symptoms and sometimes also the results of this “integral” process in the volume of the insulation, which causes – over a more or less long period – the mentioned “weak point”. In other words, partial discharges indicate the non-reversible destruction of solid or mixed insulation. This is important for both the detection of production faults (quality assurance test) and of weak points/defects caused by aging during service (diagnostic test). Consequently all HV withstand tests on-site should be combined with PD measurement. Table 1.1: Common characteristics of HV withstand testing and insulation monitoring / diagnostics 1.2 HV Test Procedures with PD Measurement During the HV dielectric withstand test, the insulation is stressed with the specified test voltage (ut) for a specified time duration (tt), see Figure 1.2. This test is successful if the insulation withstands the applied stress. If a disruptive discharge occurs, the insulation fails the withstand test, which means the result delivers a clear answer which does not require any further interpretation. Therefore the withstand test is a direct test, which is intended to destroy a defective insulation whereas healthy insulation should pass the test. Characteristics HV withstand testing monitoring / diagnostics Condition of test object New or repaired aged in service Assessment of development (type test) production (routine test) assembling (on-site test) condition of insulation (remaining life-time) Method withstand voltage test, partly complemented by PD measurement dielectric measurements, PD measurement Aim quality assurance and insulation coordination classification e.g.: - safe for service - conditionally safe - unsafe for service Sometimes characterized as “destructive” (not negative, if insulation with defect is destroyed) “non-destructive” (not positive, if the result gives no clear indication)
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 7 - During an indirect test other parameters, such as the apparent PD is measured. The insulation passes the test when the measured parameter value remains below a predefined limit during the test. The limit value is derived from laboratory and field experience, computer simulation based on physical models or simply from an agreement between the purchaser and manufacturer. The certainty of the interpretation of the on-site test results is very much dependent on how the limit value is originally determined. On-site voltage levels and the test duration should also be taken into account. In most cases, the certainty of the interpretation from indirect tests is very much lower than those of direct tests. Additionally it is necessary to consider the life time consumption of a withstand test. Each electrical stress consumes life time, depending on parameters such as voltage level, test duration and test frequency. A healthy insulation has a higher withstand voltage than insulation with critical defects. A withstand voltage test must be designed in such a way that the life time consumption of healthy insulation is negligible whereas the voltage is high enough to cause breakdown in defective insulation (Figure 1.2). Figure 1.2: Life time characteristic of a solid insulation Figure 1.3: Test cycle for withstand test with PD and test voltage stresses (ut, tt) measurement (schematically); optionally at withstand test level For non-self-restoring insulation it should be noted that insulation with defects may pass the withstand test with the test stress triggering defects which in time lead to a failure in service. To check the presence of insulation defects, the withstand test should be combined with a sensitive PD measurement (Figure 1.3). The voltage should be increased to the highest test voltage value in steps and then decreased in identical steps. The voltage step levels and the time of voltage application on each step can differ when applied to power cables, GIS or power transformers. The PD magnitude at each step (ut) should not exceed a defined limit and the PD inception and extinction voltages should be above the maximum operation voltage. It is recommended that the PD measurement be performed on all voltage steps prior to, during and following the withstand test levels as this may provide an indication of possible damage of the insulation. The combination of withstand testing and PD measurement is strongly recommended. The test voltage sequence as given in Figure 1.3 has been a recommended procedure for quality testing during routine tests for many years (this method cannot be applied to all HV equipment with non-self-restoring insulation however). It is expected that the combination of withstand testing and PD measurement will become mandatory for both routine and on-site tests in future. Additionally, the condition-assessment of insulation should take into account different methods of periodic inspection or monitoring. As a general principle periodic inspections or monitoring should deliver a warning which might be followed by a diagnostic off-line test. This test should apply the same procedure of withstand testing and PD measurement as described above (Figure 1.3). The PD values of that test shall be compared with the reference values (“fingerprints”) of the initial on-site test. ut defective insulation voltagelg ub healthy tb lg time tt voltage ur u1 u2 PD01 PD02 PD12PD11 PD22PD21 PDt ut service withstand test timett
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 8 - Figure 1.4: Examples for HV tests with PD measurement (a) 400 kV GIS by ACRF (frequency tuned resonant circuit) test system of modular reactors via bushing (b) 400 kV GIS by ACRF test system with directly flanged reactor (c) 150 kV HV cable system by DAC test system (damped alternating voltages) (d) 400 kV/433 MVA power transformer by static frequency converter There are many different types of HV apparatus and systems. To be able to test them all requires a wide variety of well adapted mobile HV test facilities [1.1] including PD measuring equipment [1.2, 1.3]. Figure 1.4 shows examples of a selection of the HV test facilities outlining the type of test object, the purpose of the on-site test, and methods of handling. A main aspect should be that the test voltage should represent stresses in service. Related aspects are given in chapter 4 and 5. Clear guidelines for the requirements of the test voltages to be applied are given in IEC 60060-1 and 60060-2 [1.4, 1.5] for routine testing and in IEC 60060-3 [1.6] for on-site testing. For installations which have been in service, lower voltages and/or shorter durations may be used. These test values should be negotiated, taking into account the age, environment, history of breakdowns, and the purpose of carrying out this test, as there is not a general principle for the value of the on-site test voltage levels for all type of HV apparatus. The test voltage levels required for insulation coordination [1.7] are only related to new insulation, and any similar coordination for in-service insulation is not currently available. Furthermore, it is important to coordinate the quality assurance tests (routine tests in factory, commissioning tests on-site, and tests after repair) and the diagnostic tests (off-line or on-line). The results of measurements of partial discharges [1.8], dissipation factor and other quantities of quality tests are used as the references (“fingerprints”) for later diagnostic tests and measurements. Records of repeated measurements may be used to show trends (e.g. PD magnitudes, pulse rates, etc.) which present the most important information for condition assessment. All this enables the establishment of the “life cycle record” (Figure1.1) for important HV equipment as the reference for condition based maintenance (CBM). 2. High Voltage Sources and Accessories for On-Site Applications In 2000 CIGRE WG 33.03/TF 04 published an ELECTRA Report on requirements for HV withstand tests on-site [1.1]. This paper describes the different types of test voltages, their generation as well as the principles for their selection. This chapter repeats the main ideas of the 2000 report and complements them with the developments of recent years. a d c b
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 9 - 2.1 General Requirements The requirements for stationary and mobile HV test systems are different (Table 2.1). In the past, the selection of mobile on-site test systems has been determined by compactness and transportability with the result, that for example HVDC testing has been used for HVAC equipment which contradicts to the above-mentioned principles of insulation coordination assessment [1.7] and HV testing [1.4, 1.5]. Today’s mobile withstand testing systems supply test voltages representative of stresses in service. On-site testing as defined by IEC 60060-3 [1.6] with its larger tolerances now allows a wide application of newer mobile test systems such as frequency-tuned resonant test systems (ACRF) for withstand and PD testing, damped alternating voltage (DAC) sources for diagnostic PD measurement and very low frequency (VLF) test as used for medium voltage cables and generators. A survey of the present application of different test voltages for on-site testing of HV apparatus/systems is shown in Table 2.2. PD measurement is not considered for DC and impulse voltages here. Only continuous AC voltage, damped AC voltage and very-low frequency voltage will be considered in more detail. Fig 2.1 shows the influence of these voltages on PD pattern. Whereas AC voltage (2 min) generates a very clear pattern (a) and DAC voltage shows similarities (b, c, d), VLF voltage (3min) did not generate enough pulses per time interval (e) and requires longer test duration for the generation of a sufficient pattern [2.1]. Table 2.1: Requirements for stationary and mobile HV test systems Figure 2.1: Typical phase resolved PD patterns for AC (a), DAC (b, c, d) and VLF (e) voltages (same defect) [2.18] requirement stationary mobile application development and quality testing in test fields quality and diagnostic testing on-site test voltages IEC 60071-1, IEC 60060-1, apparatus standards comparable to stresses in service, IEC 60060-3, apparatus standards compactness transportability usually enough space in HV labs very important for transportation weight-to-test power ratio usually not of high priority very important mechanical robustness usually no problem very important for frequent transportation and assembling handling control often automatic testing; connection to LAN easy to handle, robust
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 10 - Table 2.2: Test voltages used on-site for various HV apparatus/systems (modified according to [1.1]) equipment to be tested on-site cables GIS GIL instru- ment trans- formers power transformers rotating machines arres- ters oil-paper cables extruded cables MV HV on-site test voltage shape for details see 5.2 5.2 5.2 5.1 - 5.4 5.3 - direct voltage (DC) W W 1) W 1 very low frequency voltage (VLF) W W, DM, PD W alternating voltage (AC) by ACTC W, PD W, PD W, PD, DM W, PD, DM W by ACRL W, DM, PD W, PD, DM W, PD W, PD W, PD W, PD, DM W, PD, DM W by ACRF W, DM, PD W, PD, DM W, PD W, PD W, PD W, PD, DM W, PD, DM W damped alternating voltage (DAC) PD, DM 2) PD, DM 2) PD, DM PD, DM impulse voltage lightning (LI, OLI) W W switching (SI) W W Abbreviations: W - withstand voltage test (e.g. 60 s with AC) PD - voltage test with PD measurement DM - voltage test with dielectric measurement (mainly tan δ, no withstand test) ACTC - transformer circuit for AC voltage generation ACRL - inductance-tuned resonant circuit for AC voltage generation ACRF - frequency-tuned resonant circuit for AC voltage generation MV - medium voltage HV - high-voltage 1) applied in the past and no longer recommended 2) mainly for PD diagnostics 2.2 Continuous Alternating Voltage (AC) Continuous AC voltage corresponds to the in-service or operating voltage of HV apparatus/systems and is therefore the most important voltage for withstand and PD testing. Some main characteristics of the different principles of AC test voltage generation (Figure 2.2) are summarized in Table 2.3 and described in the following. It is considered that the most test objects (Table 2.2) are a capacitive load. Compensated transformer circuits (ACTC – Figure 2.2a) use the traditional HV transformer and a LV or MV regulator. Because of the capacitive test object Cc, an inductive compensation LC is applied on the primary side. TH is the test transformer inductance. The circuit is completed by a coupling capacitor CH, a PD measuring impedance Z and a blocking impedance LS. The weight-to-power ratio (Table 2.3), the large dimensions and the high power consumption of this equipment restrict the use of ACTC test systems to on-site applications of relatively low voltage and low test power demand (e.g. MV switchgear and MV rotating machines of relatively low power). Inductance-tuned resonant circuits (ACRL – Figure 2.2b) are resonant circuits consisting of the capacitive load Cc and a reactor of variable inductance LH tuned in such a way that its natural frequency is equal to that of the power supply (50 Hz or 60 Hz). This is done by an adjustable gap between the fixed and the moveable part of the magnetic core of the reactor. The ACRL system has a limited load range of about Cmax/Cmin = 20. To generate a voltage without the test object connected, a basic load such as a voltage divider or coupling capacitor must be provided. The ACRL series resonant circuits, as described above, have a much better weight to power ratio than ACTC systems (Table 2.3) but they are increasingly being replaced by ACRF systems (see below). The ACRL system is still applied when a certain test frequency is mandatory. In such cases the ACRF system operates often as a “parallel” resonant system. Then, the tuneable reactor also acts as an auto-transformer, the test voltage is fixed by the regulator and the test current is optimally compensated. The parallel mode is helpful e.g. for on-site testing of large generators with fluctuating losses due to PD phenomena which would disturb the voltage stability in the series mode [2.2].
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 11 - a) b) Figure 2.2: Circuits for AC test voltage generation c) Frequency-tuned resonant circuits (ACRF – Figure 2.2c) operate with a fixed reactor LH. Resonance is realised when the power loss of the oscillating circuit is supplied at its natural frequency via a frequency converter. ACRF systems have a much wider load range, a better quality factor, lower power consumption, a lower weight (Table 2.3) and a much lower price than ACRL systems. As the frequency depends on the capacitance of the test object a frequency range must be defined which satisfactorily represents power frequency. Based on experimental investigations [2.3] for extruded cables a range 20 Hz to 300 Hz is standardized [2.4, 2.5], for GIS the range 10 Hz to 300 Hz is used [2.6] and the horizontal standard IEC 60060-3 [1.6] defines even 10 Hz to 500 Hz. For all these frequencies, it is presumed that they generate a stress of the insulation based on a capacitive voltage distribution (as for 50/60 Hz). The advantages of ACRF systems have been proven in practice. GIS systems have been tested for more than 20 years and large scale tests on cable systems have been made for more than 10 years [2.7, 2.8]. The ACRF systems of different design are well adapted to the test objects (Figure 1.4) being robust and reliable. They are available up to 1000 kV and for more than 100 MVA (50 Hz equivalent power). TR LC TH LS Z PD CC AC ACTC Compensated transformer CH Z ACRL Resonant test system with tuned inductance TR LH TE CH Z PD CC ACLS FC TE LH CH Z PD CC AC AC f LS ACRF Resonant test system with variable frequency
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 12 - Table 2.3: Comparison of circuits for AC test voltages on-site transformer circuits 1) (ACTC) tuneable reactor circuits (ACRL) tuneable frequency circuits (ACRF) frequency power frequency 2) 50 Hz or 60 Hz power frequency 2) 50 Hz or 60 Hz 20 Hz to 300 Hz quality factor q = Ptest / Psupply up to 5 40 to 100 70 to 200 power supply single phase (or two phases) single phase (or two phases) three phases weight to power ratio for testing medium voltage components > 10 kg / kVA 2 to 10 kg / kVA 3) 1,0 to 2,0 kg / kVA 3) HV cables, GIL not applicable 2 to 7,5 kg / kVA 3) 0,8 to 1,5 kg / kVA 3) GIS, HV components > 10 kg / kVA 2 to 8 kg / kVA 3) 0,6 to 1,0 kg / kVA 3) harmonics requires caution < 1 % < 1 % generation of PD noise no no IGBT switching pulses (can be suppressed) mechanical characteristic moving parts in the regulator moving parts in the regulator and in the tuneable reactor no moving parts transportation weight and volume Total weight and volume are in close relation to the weight to power ratio 1) Including compensation reactors 2) With additional frequency converter for other frequencies 3) The range depends on design parameters (duty cycle, power, materials, etc.). Typical duty cycles for oil-insulated sources used for medium voltage components are 30 min ON, 30 min OFF 6 to 12 times per day at rated voltage and current. Typical duty cycles for test on HV cables and GIL are 60 min ON, 60 min OFF, 3 to 6 times per day. SF6- insulated sources for tests on GIS can be used at rated parameters for 15 min per day (e.g. 3 times, each of 5 min). Motor-generator (M/G)-sets and recently static frequency converters are used for induced voltage on-site testing of power transformers [2.9, 2.10, and 2.11]. The HV AC test voltage is generated by the power transformer under test itself. Static frequency converters are smaller and lighter than M/G-sets and deliver an AC voltage of a wide frequency range (e.g. 15 Hz to 200 Hz) [2.12]). They can be combined with ACRF test systems (for applied voltage tests) to form very compact mobile transformer test systems [2.11, 2.12] (Figure 1.4 e and 2.3). The frequency converter and the test circuit must be designed in such a way that the wave shape of the voltage is sinusoidal (total harmonic distortion THD < 5 %). The PD noise level qN must be sufficiently low (qN < 20 … 100 pC) to be able to measure critical PD in power transformers. Figure 2.3: Mobile transformer test system for induced and applied voltage tests based on a static frequency converter Frequency Converter Compensation unit Step-up transformer Control room HV- FilterHV-test lead Power supply cable HV connection cable
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 13 - 2.3 Damped Alternating Voltage (DAC) The damped alternating voltage (DAC) system supplies an oscillating switching impulse voltage (OSI) of quite low damping [1.1, 1.6, 2.13 - 2.18]. The test circuit (Figure 2.4a, b) consists basically of a direct voltage source, which charges the cable capacitance, and a suitable inductance to maintain an oscillating circuit with a frequency range between 20 and 1000 Hz (Figure 2.4d). To obtain PD occurrence conditions (e.g. PD inception, PD magnitude) similar to 50 (60) Hz AC voltage stresses, the DAC frequency has to be below approx. 500 Hz [2.1]. After the DC voltage ramp has reached the ignition voltage of a thyristor switch or the disruptive voltage of a spark gap, the capacitance is discharged via the inductance. As in resonant test systems, the oscillating frequency depends on the capacitive load (Figure 2.4d). The DAC system frequency approximates the natural frequency of the circuit, see Table 2.4. During the charging process, the test object is stressed with an increasing unipolar voltage. The charging time depends on the maximum available load current of the power supply (Iload, max), the test voltage (Utest) and the capacitance CC. An example for the charging times for various cable capacitances and test voltages are also shown in Figure 2.4c when for example a maximum load current of 8 mA of the applied voltage supply is assumed. The DAC voltage is never higher than the charging DC voltage and due to a shorter duration and decaying characteristic it is not comparable to withstand voltage testing with continuous AC. The quality factor QC of the oscillating circuit, which is responsible for the attenuation of the AC voltage wave, can be expressed by: QC = (L / (Cc • RA 2 ). RA is the equivalent circuit resistance. The maximum capacitance Cmax which can be tested using the DAC system, can be calculated by Cmax = (Ip / Vmax)2 • L, where Vmax is the maximum applied test voltage and Ip is the permissible AC current in the oscillating circuit, (Figure 2.4f). DAC system voltages are applied for diagnostic PD and dielectric loss measurement on power cables [2.13-2.15] (see chapter 5.2.4) and for the diagnostics of the stator insulation of rotating machines [2.16]. Major advantages of the DAC systems are their light weight and low input power requirements.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 14 - 0.1 1 10 100 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 C[uF] t[s] 150kV 120kV 90kV 60kV 30kV (a) (b) (c) (d) (e) (f) Figure 2.4: On-site generation of DAC voltages: (a) Schematic structure of a DAC system (b) Schematic view of a DAC measuring circuit (c) DAC charging time as function of power cable test capacitance (with max. load current of 8 mA) (d) Dependence of the frequency of damped AC test voltage on the power cable test capacitance (e) Damped sinusoidal AC voltage wave (f) Max. power cable capacitance in function of the applied damped AC voltage for an e.g. max. current of 300 A Table 2.4: Examples of the typical damped AC voltage frequencies in [Hz] for different lengths of two typical 150 kV power cables Length [km] XLPE (C = 154 pF/m) [Hz] Oil filled (C = 373 pF/m) [Hz] 0,25 300 194 0,5 213 137 1 151 97 2 107 69 4 76 49 8 53 34 16 38 24 20 34 22 0 20 40 60 80 100 120 140 160 180 0 1 2 3 4 5 6 7 8 9 10 11 12 13 C[uF ] f[ Hz ] da m pe d A C fre qu en cy [H ] HV power cable capacitance [ μF] 0 20 40 60 80 100 120 140 160 180 0 1 2 3 4 5 6 7 8 9 10 11 12 13 C[uF ] f[ Hz ] 0 20 40 60 80 100 120 140 160 180 0 1 2 3 4 5 6 7 8 9 10 11 12 13 C[uF ] f[ Hz ] dam ped AC freq uenc y [Hz] HV power cable capacitance [ μF] 1 10 100 1000 10000 10 30 50 70 90 110 130 150 Damped AC Voltage [kV] C[uF] 300A
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 15 - 2.4 Very-Low Frequency (VLF) VLF testing, primarily of frequencies of approximately 0,1 Hz but in some cases (high capacitive load) down to 0,01 Hz has been an accepted method for voltage withstand testing for all types of medium voltage cables for several years (CENELEC HD 620 S1 and IEEE 400.2) [2.17]. Figure 2.5: VLF voltage generation by controlled DC charging and discharging circuit [2.17] There are several principles of VLF voltage generation utilising inductors with diode chains or power electronic switches, Figure 2.5 illustrates cosine-square wave voltages, where the polarity reversal follows a cosine function and Figure 2.6 gives an example of electronic invertors used for sine-wave voltage generation [2.19]. Voltage levels up to 200 kV are available. This type of on-site equipment has the benefit of low weight, ease of transportation, low power consumption and low cost. One has to consider that VLF causes a voltage cycle of 10 s (or even 100 s) which leads to a resistive voltage distribution in the insulation, longer test duration and a higher test voltage than with power frequency. Further consequences must be considered in PD and tan  measurements [2.20-2.26]. Figure 2.7 shows that the PD pattern at 0,1 Hz may have some similarities with that of 50 Hz test voltage if the test duration at 0,1 Hz is correspondingly extended. Figure 2.6: VLF voltages a) cosine-square wave, b) sine wave a b
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 16 - Figure 2.7: Phase resolved PD patterns for 50 Hz (top) and 0,1 Hz (bottom) of an artificial defect in a cable joint of a 110 kV cable. The defect was in the semiconducting layer [2.26] 2.5 HV Filter, Coupling Capacitor, Connections and Grounding for Conventional PD Measurement The signal-to-noise ratio required for successful PD on-site measurements is dependent upon the HV circuit consisting of HV source, HV connections, test object and ground connection, e.g. Figure 2.2. The circuit acts as an antenna for radiated noise signals, such as those from external corona in the substation or from broadcasting stations. Therefore the spatial dimensions of the circuit should be as small as possible and itself PD free. A second source of noise is related to noise signals conducted from the power source or from power- electronic components to the test object (e.g. static frequency converters or power electronic switches). Efficient damping of this conducted noise may be achieved by the use of HV C-L filters consisting of a filter capacitor (CF; e.g. the bushing of a transformer or a voltage divider) and a blocking impedance (Ls > 15 mH up to 100 mH). The filter is connected to the coupling capacitor (CH, see Figure 2.2). As a rule of thumb a single well designed C-L filter can reduce the noise level up to one tenth (1/10), two such filters in series (“” filter) up to 1/20. The HV filter and the coupling capacitor should be arranged as near as possible to the test object. If the connection between the voltage source and the HV filter is by a shielded cable and not by an “open” air-insulated tube or wire, the “antenna effect” can be disregarded and the cable will act as an efficient filter capacitor CF. The coupling capacitor CH (Figure 2.2) must be adapted to the test object. The traditional rule CH ≈ (0.1 … 1) CC (capacitance of test object) cannot be generally recommended in this case. In a noisy surrounding a lower CH/CC may actually result in a better signal-to-noise ratio. The HV circuit includes the ground connection which should be as short as possible, of low impedance and bypass free. Because many test objects are grounded on-site directly, the grounding of the voltage source and its ground connection to the test object are a matter of optimisation of the grounding conditions.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 17 - 3. On-site PD Measurements When performing on-site PD tests in compliance with the relevant IEC standards the signal-to-noise (S/N) ratio becomes decisive. This is because the background noise level may range between several tens and a few hundreds of pC. In some cases, however, peaks of more than 1000 pC occur. Therefore the apparent charge can be measured sensitively only in “noise-free” surroundings, such as well shielded test areas using additional HV and LV noise filters or special electronic features for noise rejection [1.2]. To overcome the poor signal-to-noise (S/N) ratio under on-site conditions, conventional PD measurements are increasingly substituted by so-called non-conventional methods, as shown in Figure 3.1. In this context it should be noted, that an application guide on PD measurements in compliance with IEC 60270 [1.2] is available as Technical Brochure No. 366 [3.1]. Later, another Technical Brochure has been prepared by CIGRE WG D1.33, which deals with the so-called non-conventional PD detection methods [3.2]. Additionally a new IEC guide is currently under consideration to provide recommendations and specifications of the electromagnetic and acoustic PD detection methods [3.3]. Therefore, this brochure will only briefly consider the fundamentals of both conventional and non- conventional PD detection techniques but cover the practical experiences gained from on-site PD tests in more detail. Figure 3.1: Survey on PD measurement methods applied for HV apparatus under on-site condition 3.1 Conventional Electric PD Measurements Conventional partial discharge measurements complying with the relevant IEC standards [1.2, 1.8] have been proven over years as an indispensable tool for quality assurance testing of HV apparatus. This technique is based on the evaluation of the apparent charge which is only a small fraction of that charge created at the PD site, due to the condition that the stray capacitance Ca between the PD source and the test object terminals is much larger than the virtual capacitance Cb of the PD source itself [3.1]. Because the ratio Ca/Cb is not known at all, and the apparent charge depends additionally on the size and the location of PD defects, the PD severity of HV apparatus cannot be assessed by evaluating the apparent charge alone. in time New HV components GIS Transformers Cables Service aged Conventional methods (IEC 60270) Non- conventional methods measurement in frequency domain domain Electro - magnetic detection Acoustical detection Detection of optical occurrences Chemical compounds analysis measurement
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 18 - As a consequence, knowledge rules for PD diagnosis in service have been established in the past based on practical experiences in the laboratory and on-site [3.4-3.6]. PD instruments intended for apparent charge measurements are generally equipped with a band-pass filter amplifier connected in series with a peak detector [1.2, 3.1]. Without going into detail her, it can be stated that the magnitude of the output pulse trains is a measure of the apparent charge of the input PD pulse sequences, due to the so-called quasi-integration performance of the band-pass filter. The condition for the quasi-integration is generally satisfied as long as the band-pass filter selects the measuring frequency in a range where the density of the frequency spectrum of the PD signal, appearing across the terminals of the test object, is nearly constant [3.7]. This is generally satisfied by limiting the measuring frequency below 500 kHz [1.2, 3.1]. To evaluate the apparent charge in terms of pC, the scale factor of the PD instrument must be determined by a specified calibration procedure. This is based on a simulation of the internal charge transfer by an injection of artificial PD pulses between the terminals of the test object [1.2, 3.1, 3.8- 3.12]. Therefore, the apparent charge of a PD pulse is defined in IEC 60270 [1.2] as “that charge which, if injected within a very short time between the terminals of the test object in a specified test circuit, would give the same reading on the measuring instrument as the PD current pulse itself.” In practice the magnitude of the injected calibrating charge q0 should exceed twice the background noise level. Moreover, q0 should be tuned to a limit which is close to the apparent charge as specified by the relevant apparatus standards or as set by an agreement. In practice it may also be useful to tune q0 to the level of the apparent charge to be expected for the particular HV apparatus under test. The advantage of PD measurement in compliance with IEC 60270 [1.2] is that the specified calibration procedure ensures reproducible and comparable PD test results, even when the tests are performed in different HV laboratories using PD measuring facilities of different manufacturers. The main drawback of conventional PD measurements is, however, that the S/N ratio is strongly reduced by the limitation of the measuring frequency below 500 kHz [see 1.2, 3.1]. Therefore, the applicability of PD measurements in compliance with IEC 60270 [1.2] may strongly be restricted under noisy on-site condition. Currently the classical analogue PD signal processing method is being replaced more and more by the advanced digital signal processing. This technique offers not only a more informative visualization of phase-resolved PD patterns (see Figure 3.2) but also a deeper statistical analysis of the very complex PD data with respect to the identification and classification of PD events (see example reported in [3.13-3.17]). Moreover, the digital technique can be used for the localization of PD sites in power cables [3.18, 3.19] and provides powerful features for de-noising of PD signals [3.20-3.28]. Figure 3.2: Example of phase-resolved PD pattern [3.15]
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 19 - 3.2 Non-conventional Electromagnetic PD Detection Electromagnetic PD detection has been established over the past three decades as a valuable tool for on-site diagnostics of HV apparatus [3.29-3.62] due to the excellent S/N ratio, which becomes more enhanced at higher measurement frequencies. PD pulses due to discharges in gaseous dielectrics, such as SF6 and SF6-gas mixtures, are characterized by a rise time down to the sub-nanosecond range. Therefore the frequency spectrum covers a range up to about 3 GHz. In coaxial structures (e.g. GIS, GIL) such fast transients release electromagnetic waves traveling not only in the basic transverse electromagnetic mode (TEM) but also in higher order modes (TM, TE) [3.59]. The TEM waves which are weakly absorbed propagate through the whole plant. Above a critical frequency, however, the PD signals spread as TM and TE waves which are not coupled to conductors. These propagating fast transients can be detected by antennas (field sensors). PD pulses due to discharges in dielectric imperfections, such as voids and cracks in polymeric insulation, are characterized by rise times in the nanosecond range. Therefore, the resulting frequency spectrum may cover several hundreds of MHz and even more. This high frequency spectrum is well detectable by means of inductive and capacitive sensors. Due to the wide range of the frequency spectrum of PD transients, electromagnetic PD detection methods cover the frequency ranges of HF (3 MHz-30 MHz), VHF (30 MHz-300 MHz) and UHF (300 MHz-3000 MHz). Measuring frequencies in the UHF and VHF range are usually applied to GIS/GIL [3.29-3.31, 3.34-3.39] and power transformers [3.40-3.46]. The VHF and HF range is most suitable for rotating machines [3.47-3.51] as well as for power cable accessories [3.32, 3.33, 3.52- 3.55, 3.61, 3.62], see Table 3.1. Table 3.1: Most suitable frequency band for on-site / on-line PD detection in different components A survey on the structure and the significant steps of electromagnetic PD detection methods is schematically presented in Figure 3.3. Figure 3.3: Survey on the structure and steps of the electromagnetic PD detection Non- conventional methods Electro- magnetic detection HF/VHF UHF Sensors EM signal transmission Performance and sensitivity check On-site test Signal processing Sensitivity check PD measurement Cables Transformers GIS Generators HF + + - + VHF + + + + UHF - + + - Legend: + indicates suitable - indicates not suitable
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 20 - Figure 3.4: Survey on common PD measurement procedures in the VHF and UHF range [3.33] The PD detection technique in the UHF and VHF range is mostly performed in the frequency domain, using either the zero-span or the full-spectra mode (see Figure 3.4). UHF and VHF PD diagnostics performed in the time domain, however, may also offer different advantages as reported, for instance, in [3.32, 3.33, and 3.61]. Without going in details it can be stated that the magnitude of the captured PD pulses and therefore the S/N ratio is enhanced as the measuring frequency increases. It should be noted that as the measuring frequency increases the detected signal magnitude is affected by attenuation and dispersion phenomena, mainly due to the skin effect in conductors and to polarization losses in dielectrics. The magnitude of propagating PD transients depends, for instance on [3.59]:  the distance between PD site and PD sensor  the material properties of the conductors and the insulation of the test object  the geometrical size / dimensions of the test object  the characteristic propagation modes, for instance in GIS/GIL: 1-2 dB/km (TEM, TE-mode); 4 dB/km (TM-mode) The signal magnitude can also be affected by reflections at discontinuities in the propagation path which lead to resonances and standing waves. Therefore, reproducible PD test results cannot be expected under these conditions, even with low scattering of the apparent charge magnitudes. Consequently, in these circumstances the electromagnetic PD detection method cannot be calibrated in terms of pC. As an alternative, a performance check as well as a sensitivity check is recommended as reported, for instance, in the references [3.2, 3.3, 3.56-3.62]. A performance check includes the verification of the functionality of the whole signal transmission path consisting of the PD sensor, the measuring cable, and the PD measuring system. The sensitivity check is intended to verify the PD threshold level of the complete measuring arrangement in terms of pC. This is best determined by comparative PD studies using both the conventional and non-conventional circuit. FREQUENCY DOMAIN MEASUREMENT TIME DOMAIN MEASUREMENT PRINCIPLE Zero-span mode Full-spectra mode Ultra-wide-band mode MEASURING FREQUENCY SPECTRUM PD SIGNATURES PROCEDURE test voltage PD signal test voltage PD signal Frequency sweep centre frequency PD signal
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 21 - The individual steps which are required to carry out the performance and sensitivity checks are summarized in Figure 3.5. These procedures have been proven for GIS / GIL, power transformers and cable accessories, where the specific steps for each type of HV apparatus may differ slightly. As a practical example, the steps required for the sensitivity check for power cable accessories are listed in Table 3.2. Figure 3.5: General steps for the performance / sensitivity check Table 3.2: Steps for the sensitivity check for cable systems Laboratory set-up Single port Dual port PD source with known magnitude Injection of artificial pulses Sensitivity Check of sensor arrangement Sensitivity check Performance check On-site Injection of artificial pulses in same sensor Injection of artificial pulses in other sensor Transformers Cables, GIS Transformers 1. Laboratory set-up 2. On-site 1.1 Installation of an artificial PD defect (e.g. PD magnitude below 100pC) on a full size combination of power cable part and cable accessories 2.1 Performance check (sensitivity check) of the whole system: cable accessory, sensor, PD detection system. Injection of artificial voltage pulses according 1.5 1.2 Installation of given type of HF/VHF/UHF sensors on the cable accessories and parallel connection of a PD detection circuit (IEC 60270) 2.2 Tuning of the PD detection system to HF/VHF/UHF frequency range with best signal/noise ratio 1.3 HV energizing of the full size setup and PD detection using both systems according to IEC 60270 and IEC 62478 2.3 Estimation of the PD detection sensitivity of the whole system in terms of pC 1.4 For PD defect from (1.1) ratio estimation of the lowest [pC] reading and [μV] for the HF/VHF/UHF circuits 1.5 Search for an artificial voltage pulse with similar frequency/amplitude characteristics as the PD defect from (1.1) and (1.3)
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 22 - 3.3 Noise Reduction Background noise can be coupled into the PD measuring circuit in different ways. Generally it can be distinguished between conducted and radiated noise. Conducted noise:  PD transients originating from poorly designed HV test circuit components, such as discharges from protrusions on HV shielding electrodes (Figure 3.6a) and discharges between floating parts (Figure 3.6b).  Periodical or stochastic occurring HF signals caused by switching operations in the mains and power electronic components and maintenance work in the measurement surroundings being conducted into the test source see Figure 3.6c.  Continuous radio frequency (RF) signals received from radio broadcasting via the grounding system and/or the feeding facility of the PD test circuit. Radiated noise:  Continuous RF waves radiated from radio broadcasting and TV stations, as well as noises from cell phones, see Figure 3.6d.  Periodically and stochastically pulse-type interferences from rectifiers, frequency converters and switches.  PD transients radiated from nearby HV components especially in substations. When using conventional PD measurement circuits in accordance with IEC 60270 [1.2], conducted noises can be reduced or even suppressed by suitable filter circuits installed in power electronic components such as frequency converters, in the power supply or between the HV test facility and the test object. The latter is usually designed as a -filter equipped with blocking impedances in series and low inductive capacitors in parallel (see 2.5). a) Discharges from protrusions of HV shielding electrodes b) Discharges between floating metallic parts c) Maintenance work (drill) d) Signals form a cell phone call Figure 3.6: 3D-graphs of characteristic noise signatures [3.32] (x-phase angle, y-pulse number, z-apparent charge)
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 23 - Such noise filters are well proven in HV test laboratories and could in principle also be applied to on- site testing. However, with respect to radiated noise, the application of HV filters for on-site PD testing can be questionable because such filters act like antennas for radiated noise. Therefore on-site test set-ups should be designed with very compact filters with low longitudinal capacitance and the HV and grounding connections should be of low impedance. Even with these characteristics however, it is recommended that LV and HV filters and internal filters in power electronic components be recommended for on-site PD tests. The reduction of background noise picked up by conventional PD circuits according to IEC 60270 [1.2] requires a well-adapted coupling capacitor connected in series with the measuring impedance. Both should be located as close as possible to the test object. If for example the test voltage is supplied by a cable some tens of metres long from the HVAC source then the blocking impedance should be connected adjacent to the coupling capacitor. This is because the measuring cable can also act as a filter capacitor (see 3.3). Depending on the time characteristic of noise signals, appropriate counter measures are also recommended [3.16], such as:  Continuous wave (CW) noise, e.g. radio, TV, cellular phones. Counter measure: band pass filter with adjustable centre frequency and bandwidth tuned to “proper” spectral areas (see Figure 3.7) and/or use of adaptive notch filters to remove only certain narrow-band spectral interferences for improved signal-to-noise ratios.  Periodic pulse-type interference with constant phase angle (rectifier, converter): Counter measure: as this is easy to identify due to the constant phase position and usual small derivation in amplitude (deterministic behaviour), window gating is the preferred method.  Periodic pulse-type interference with shifting phase angle (frequency converter): Counter measure: as this is easily identified by rotating phase position, dynamic gating is recommended.  Pulse-type interference with stochastic phase angle (e.g. switching): Counter measure: the stochastic behaviour in phase and amplitude and rare (random) occurrence makes an appropriate choice of measurement time interval the best method.  PD-type interference (e.g. corona): Counter measure: de-noising based on pulse waveform analysis or multi-channel PD measurements is the preferred method because of the stochastic behaviour in phase and amplitude of this noise. Digital PD measurement and data evaluation, in combination with analogue methods, enables an effective rejection of electromagnetic interferences. An actual comparison of digital de-noising techniques for partial discharge measurements, including FFT, low pass filtering, short-time FFT, frequency-domain adaptive filtering, matched filter, notch filter, wavelet de-noising and others is given in [3.28]. Optimum results were obtained by combining notch filter and wavelet de-noising. In this context it should be noted that not all digital filter algorithms are suitable for real-time applications because of their algorithmic structure or demand for processing power. In many practical on-site cases, a combination of analog and digital filtering techniques delivered excellent results. For a limited number of dominant narrow-band interferences (e.g. radio), analog notch filters can help to reduce the input amplitude for the A/D converter, resulting in an improved dynamic range for the PD signals of interest. Very promising de-noising results can also be obtained by pulse waveform analysis [3.20-3.28]. This is because interference pulses and PD pulses often show different waveforms, which can, in principle, be analysed in the frequency domain or in the time domain. Pulse waveform analysis extracts a set of pulse features, which include, in the time domain, pulse rise and fall time, and pulse width and the characteristic frequency spectrum.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 24 - Another promising method for eliminating electromagnetic interference and for the separation of different PD sources is multi-terminal PD measurement. Within a short time window PD signals of the three phases of the test object (e.g. at three different bushings of a power transformer) are measured. The amplitudes of the signals are added as vectors in a “star diagram”. The three axes of this star diagram correspond to the three PD measuring points. Each PD source has a characteristic ratio of the amplitudes and will build up a cluster of events in the star diagram. By selecting only the signals which fulfill the condition of the cluster, the common "Phase Related Diagram" shows only the pattern of this specific PD source. With this principle, PD from different sources and noise pulses can be separated [3.17, 3.28]. For more details see 5.4.6. As mentioned previously, the application of VHF and UHF PD measurements improves significantly the S/N ratio, especially if this technique is combined with advanced de-noising features, as presented in [3.20-3.28]. As a result of such tools, one gets a lower background noise level in the non-energized HV circuit. However, this may increase after the HV test facility is switched on. In [3.2] a summary of well proven noise rejection methods is given for the electromagnetic PD detection technique, in particular for GIS and GIL. Similar procedures are applicable to other high- voltage assets, such as power transformers and cable accessories. Although these examples refer only to the UHF/VHF technique, similar noise rejection tools can also be followed for other non- conventional methods, such as the acoustic PD detection technique. 3.4 Acoustic PD Detection The acoustic detection of partial discharges uses the fact that an acoustic signal (e.g. a mechanical vibration) is emitted as a result of the pressure build-up by PD pulses in the insulation medium. The signals can be picked up by means of piezoelectric transducers, fibre optic sensors, accelerometers, condenser microphones and sound resonance sensors. Because of the short duration of the PD pulses the resulting compressed wave contains a frequency spectrum considerably higher than the audible sound band. The frequency spectrum generally used ranges from 10 kHz and 300 kHz. The acoustic wave propagation from the PD source to the acoustic sensor (transducer) is strongly influenced by the geometry of the test object as well as by the medium (oil, gas, steel). Consequently, different wave types with different propagation velocities appear and reflections and refractions at boundaries lead to changes of the sound propagation, resulting into damping, absorption and scattering of the signal [3.2, 3.42, 3.43, 3.45, 3.63-3.66, 5.26]. Using acoustic PD detection methods an attempt is made to locate the PD source. Ultrasonic PD location has been proven for partial discharges of high energy. The field of application of this method spans all HV apparatus discussed in this paper (e.g. see chapter 5.1.1, 5.1.2 and 5.4.5). To reach an optimum sensitivity the complete acoustic-mechanical system must be thoroughly understood (e.g. the influence of the sensor size). Comparative measurements have revealed that for GIS for example, an equivalent sensitivity as low as 5 pC could be achieved under on-site condition. Figure 3.7: Probability of spectral interference [3.16]
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 25 - 3.5 Important Aspects for PD Evaluation To support the interpretation of PD measurements to recognize aging processes in the insulation of HV components, links between specific insulation characteristic and the information provided by particular measured PD quantities are necessary. As a result, systematic knowledge rules need to be formulated to support insulation condition assessment (see Figure 3.8). In particular, to identify measurable and derivable quantities for particular types of HV components, several aspects have to be taken into account (e.g. typical insulation defects, external factors like propagation effects, and disturbances or cross-talking). In addition to the detection of partial discharge activity, other important aspects of PD measurements include the identification of the discharge source and type of defect and the location of the defect to allow condition assessment of the insulation. Figure 3.8: General aspects of creating PD knowledge rules [3.6] 3.5.1 PD Identification The use of partial discharge measurements for the condition assessment of HV components depends on the method chosen for PD detection and on the derived quantities to evaluate the measured information. A combination of both aspects results in PD diagnosis, which generates information suitable for making PD knowledge rules. Generally, in a HV on-site test including PD detection, the PD magnitude, phase-resolved patterns and location of the discharges are usually determined as a function of the applied voltage. The discharge detection can be obtained with a discharge detection system performing in accordance with IEC 60270. Figure 3.9 shows a scheme of how to obtain the typical PD quantities. Partial discharges cause sequences of current impulses in the leads to the object under test. This can happen several times during a power frequency cycle. As a result, groups of recurrent discharges, which occur during the positive half and the negative half of the voltage cycle, can be found. If a sample has several discharging sites, more discharges will occur within the same time intervals. Different quantities have been introduced over the years to describe the characteristics of a discharge.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 26 - These quantities can be divided into three main groups:  basic PD quantities (according to IEC 60270)  derived PD quantities  phase, amplitude-related derived quantities (PD pattern) It is well known that the sequences of PD pulses contain random events, which are characterized by considerable variation in magnitude and phase angle. These significant PD parameters are also strongly influenced by the level of the applied AC test voltage. Therefore, informative diagnosis of insulation defects can only be achieved successfully if several PD quantities are measured and evaluated simultaneously (e.g. charge, phase and number of occurrence (, q, n)). While the actual ageing processes generated by the interaction of partial discharges are still not well understood, considerable information has been gathered over the years on insulation type, defect type and the effect of load conditions etc. As the HV components age, the number of faults will increase and the power supply quality can deteriorate. Important symptoms of this degradation process can be related to discharging local insulation imperfections or defects, which may occur in the particular HV component (Figure 3.10a and Figure 3.10b). Figure 3.9: Principle of digital PD data processing to obtain basic and derived PD quantities Figure 3.10: a) Generation process for PD knowledge rules
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 27 - Figure 3.10: b) Relationship between service condition specific factors, defect detection, knowledge rules and condition assessment to support Condition Based Maintenance. With regard to a particular type of HV component, the combination of information about characteristics observed for each PD quantity can be used to support the interpretation of measuring data. In particular, to evaluate the effectiveness of PD measurement for different components, the following structure can be used:  Insulation defects: short description of general insulation problems (no particular construction problems) possible in this type of HV component;  PD processes and PD signals: short description of HV component specific aspects important for PD measurement;  Representative results: typical ‘’acceptable’’ and ‘’not acceptable’’ examples of using PD quantities to detect insulation degradation;  PD knowledge rules: examples of combining PD quantities to determine certain insulation condition of a HV component;  Conclusion: benefits/limitations of using PD diagnosis for this type of HV component. 3.5.2 PD Localization To estimate the criticality of PD, fault-type identification [3.22] has to be complemented by PD site localization. In expanded geometries, like long cables or GIS, or large components like power transformers, the PD identification without localization would be useless. The method most used to calculate the location of a partial discharge site is based on the difference in arrival times of PD pulses at one or several sensors. Different calculation methods are used for different HV components. Methods based on the attenuation of PD signals are also used. Further details will be found in the corresponding chapters. Regarding PD location calculation the following can be concluded:  time-domain-reflectometry is the most important method, where the arrival time of original and reflected PD signals are measured (cables, GIS, transformer); PD signal attenuation, however, may limit the applicability of this method.  with several sensors the difference of the time-of-arrival of either acoustic, electric or acoustic/electric signals can be used for evaluation; in long cable systems, GPS synchronization of PD measurements is often necessary [3.68];  other techniques that are being used, or are under investigation, are based on the attenuation effects that occur during the progress of EM-waves as they travel through the components, see e.g. [3.69].
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 28 - 4 Preconditions for On-Site Testing Including PD Measurement On-site tests including PD measurement can be applied to different components and for different purposes. In particular, they can be part of commissioning tests after erection, for condition assessment (diagnostic or monitoring) or to check the effectiveness of repair operations. Moreover the systems used (concerning voltage stress amplitude and shape, PD measuring systems characteristics etc.) are very different and depend on the component to be tested, their insulation system or sometimes on the preferences of the suppliers and users. The main reference documents, such as International Standards (IEC, IEEE), typically only specify the voltage shape for testing [1.6]. Generally no indication is given about details and preconditions of the on-site test or the testing procedure and algorithms for the evaluation of the results. The technical details required for on-site testing are given in the following list. The list is subdivided into three parts referring to the characteristics of the component under test and its installation, the test requirements and the characteristics of the test system. Requirements for component under test and its installation  Ratings and main physical characteristics, including insulation system, specified in details in the next paragraphs;  Type of installation (e.g. indoor or outdoor);  Results of type and routine testing including PD magnitude, capacitance and loss factor at nominal voltage and frequency;  Date of installation (age) and average service load;  Characteristics of the power network and grounding system;  Air clearances requirements, available space for the HV test system;  Atmospheric (temperature, air pressure and humidity) and geographical (height above sea level, pollution class) information. Test requirements  Aim of the tests (e.g.: monitoring, commissioning after installation, commissioning after major maintenance, diagnostics etc.);  Availability of previous test data or history of the test object;  Test voltage shape and frequency requirements;  Restrictions for the test object regarding a possible, immediate failure during test;  Test procedure: voltage level(s), duration of application or number of impulses (in case of DAC) for each voltage levels, highest voltage (for both PD measurement and withstand voltage if required), PD inception and extinction voltage level;  Required PD measuring method and maximum accepted PD magnitude and max. background noise level;  Possible required evaluation of the results: PD location, PD phase resolved pattern, defect recognition and severity of defects (if present). Requirements for HV test and PD measuring system  Type and ratings of the test system;  Supply requirements (voltage, frequency, power and their tolerances);  Weight, size and environmental restrictions for transportation and operation;  Grounding requirements;  Type, dimensions and length required for the HV connection to the component under test;  Characteristics of PD measuring systems (sensitivity, bandwidth, conformity with IEC Standard 60270);  Additional PD evaluation features of the measuring systems (filter to enhance the S/N-ratio, possible location of PD source, automatic defect recognition etc.).
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 29 - 5 Examples of Test and Measuring Techniques for Apparatus and Systems This chapter describes the different HV test sources and test procedures required to perform on-site PD tests on various types of HV apparatus including gas-insulated switchgears (GIS), power cables, rotating machines and power transformers. 5.1 Gas-Insulated Systems (GIS/GIL) 5.1.1 HV Source, PD Measurement and Details of Test Object On-site testing of GIS is an important step during commissioning of a new GIS/GIL or after modification and maintenance tasks. This test assures a correct erection or repair of a GIS/GIL and verifies that there are no defects within the GIS/GIL that could lead to a major failure during operation. Various methods can be performed depending on rated voltage level and GIS type and design. Experiences of on-site testing with different waveforms and procedures have shown which method is suitable to detect major defects within a GIS/GIL [3.57]. HV Sources and Test Procedures It has been common practice for nearly two decades to carry out AC high voltage tests by resonant test systems with variable frequencies (ACRF). IEC 62271-2003 [2.6] defines a test frequency range between 10 Hz and 300 Hz. Typically test frequencies above 80 Hz are used if voltage transformers are connected to the GIS. The results of sensitive partial discharge measurements have proved the success of this method. Typical defects that cause PD such as protrusions, free moving particles, floating parts or cracks in spacers, can be produced during transportation and erection. Due to the different physical behaviour of the defects under the different voltage stresses (HV AC stress, lightning impulse or switching surges) it is impossible to find all defects by withstand testing of only one waveform. Table 5.1-1 gives an overview of the effectiveness of the various applied HV voltage waveforms to detect the different defects [3.57]. It can be seen that the application of PD measurement enables the detection of all kind of serious defects during AC voltage testing. The standards [2.6] define no single technique or the methodology which must be used for on-site testing. This still depends on an agreement between the manufacturer and customer as to which test procedure (including PD measurement or not) and which combination of test voltages (AC, LI, OLI or OSI) is to be applied. The following main PD measuring methods are in used to detect partial discharges in GIS:  Conventional measuring systems based on the IEC 60270 recommendations;  UHF measuring systems using narrowband or wideband filter detection in a frequency range up to several GHz;  Acoustic measurements, mainly for PD location, using externally mounted acoustic sensors which detect the acoustic signals emitted by a PD source. Neither the UHF nor the acoustic PD measurement techniques can be calibrated in accordance with IEC 60270. This means, that the output signal of these measurement techniques cannot be directly associated with the apparent charge. This is not really a drawback; however, the main purpose of PD measurement on-site is the PD detection itself (yes/no condition) and the identification of the PD sources. Therefore the signal does not need to be calibrated. To compare the sensitivity of different UHF sensors and acoustic sensors a CIGRE brochure has been published which describes in detail the sensitivity check of such sensors [3.58].
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 30 - Table 5.1-1: Relative effectiveness of on-site test on GIS defects /3.57/ Defect High AC LI SI Low AC with PD High AC with PD High AC with SI High AC with LI High AC with PD and LI Sharp protrusions fixed on live parts + - + + Round protrusions fixed on live parts (assembly faults) - + + + + + + Particles on spacers -/+ + - -/+ -/+ + + Cracks in spacers -/+ - - - + -/+ -/+ + Free particles + - + + + + + Parts floating - + + - - + Left foreign bodies + + - - + + + + - less effective + effective UHF-Sensors Specially designed capacitive sensors (antennas) are suitable for GIS PD measurements in the UHF range (e.g. frequencies above 300 MHz and up to 3 GHz) and have been well proven in the past for testing GIS on-site. In principal two types of UHF PD detection systems can be used, the narrow band technique with a bandwidth of around 5 MHz and the wide band technique using a bandwidth of up to 2 GHz. The UHF technique, which needs adequate electrical sensors, is very sensitive and can also locate the PD source with a sufficient spatial resolution. Different kinds of sensor are used (Figure 5.1-1):  Fixed UHF Sensors (disc sensor, cone sensor)  Mobile UHF-window sensor  Field grading electrodes. While in a new plant the UHF sensors can be installed simply (disc sensor arranged in a assembly flange, Figure 5.1-2a), only portable window sensors can be used in an existing plant (Figure 5.1-2b). In addition, field grading electrodes can also serve for the detection of the PD signals with less sensitivity, however. A comparison of the first two kinds of sensors shows that with an appropriately sized window, the sensitivity of conventional UHF sensors can be achieved. With very small flange diameters however, the UHF measurement is not successful [3.36]. Field grading electrodes (used to make the electric field distribution more uniform inside barrier insulators) show a sensitivity corresponding to those of the window sensors [3.36, 3.37, and 3.38].
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 31 - Figure 5.1-1: Different types of PD sensors/couplers: a) disc sensor, b) and d) field grading electrodes, c) external plate sensor [3.36] Figure 5.1-2a: Sectional view and mounting of an UHF PD disk sensor Figure 5.1-2b: Sectional view and mounting of an UHF PD window coupler PD Localization Different procedures and methods for the localization of the PD source can be applied. Sectionalizing and electrical time-of-flight measurements are the most practical procedures and are typical for the UHF PD measurement method. These reflectometry techniques are explained in general in the CIGRE Brochure 297 [3.19] and are also applicable for acoustic measurements. The very fast electric pulse emitted by a PD source of rise time below 1 ns, propagates in all directions along the GIS bus duct. It arrives at couplers which might be located on both sides of the PD source. By using the “time-off-flight” technique the time difference between the two wave fronts arriving at the couplers can provide information on the location of the PD source (Fig 5.1-3). As the time difference is of the order of tens of nanoseconds a fast digital oscilloscope is required for the measurements [3.39]. a ) b ) c ) d )
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 32 - The electric waves attenuate when propagating through the GIS because of damping and resonance phenomena. This can also be applied to locate the PD site if the influence of different GIS components on signal attenuation is known. Figure 5.1-3: Principle of PD localization with two sensors and the time-of-flight method [3.39] Required Details of Test Object The following information about GIS/GIL has to be considered in addition to chapter 4, in order to be able to perform HV tests in combination with sensitive PD measurements (UHF- or acoustic methods):  Type / layout of installation: number of busbars, cable, transformer connections  Ratings  Capacitance of bay, components at nominal voltage and frequency  Date of installation (age) and average service load  Characteristics of the power network and grounding system  Air clearances requirements if HV test voltage is supplied via bushing to GIS/GIL  Restrictions for the test object regarding a possible, immediate failure during test  Required PD measuring method and maximum accepted PD magnitude and max. background noise level. 5.1.2 Example 1: On-Site Test by ACRL Test System (Tuneable Reactor Circuit) and Acoustic and UHF PD Measurement Technique To perform HV on-site testing on two GIS substations, rated voltage 145 kV and 300 kV, a resonant test system with variable inductance (ACRL), was used. Both acoustic and UHF PD measurements were made. Details are given in [5.1]. An acoustic signal was detected using an acoustic emission sensor with a resonant frequency of 32 kHz, placed on the external surface of the GIS enclosure. As the acoustic signal coming from the defects is subjected to strong attenuation along the GIS, particularly when there are spacers between the PD source and the measuring probe, the sensor was positioned in different places on the GIS. At least one measuring point was placed in each compartment. Acoustic measurements were performed at U=0 kV (background level) and at U=191 kV (1,1 x operating voltage). The acoustic signal coming from the sensor was amplified and then sent to an oscilloscope and to PD detector where time-domain analyses were performed. The tests were made with a single-phase test circuit. Each phase conductor was connected in turn to the voltage supply, the conductors of the other phases being earthed together with the enclosure. For the UHF measurements, pre-installed UHF-sensors were used to pick up the PD signals [5.1]. The combination of both PD measurement techniques is a very powerful tool to detect, locate and interpret the recorded PD event and to distinguish it from signals not originating from harmful defects.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 33 - During the on-site tests various signals could be detected, analysed and located. In addition loose electric shields, free moving particles of size 2 mm to 6 mm, were detected and localized. Figure 5.1- 4a shows the signal of free moving particles at power frequency. Figure 5.1-4b shows the corresponding phase resolved PD pattern. After opening the GIS, numerous metallic (dust-like) particles were found and removed. Figure 5.1-4a: Acoustic signal of harmful free particles [5.1] Figure: 5.1-4b Phase resolved PD pattern of free moving particles 5.1.3 Example 2: On-Site Test by ACRF Test System (Tuneable Frequency Circuit) and High Frequency PD Measurement Technique It is now common practice to test GIS on-site with ACRF test systems. Two primary methods are available. Metal-enclosed, SF6-insulated HV reactors, which can be directly flanged to the GIS (Figure 1.4b), achieve a high PD sensitivity with low background noise. They require the lowest space demand and no additional safety requirements. Maximum test voltage is about 740 kV. Oil-insulated modular reactors on the other hand (Figure 1.4a) cannot be directly flanged to the GIS. The HV has to be supplied via a bushing to the GIS. The inductance of the reactors can be well adapted to the test object capacitance by series and parallel connection of reactor modules. Test voltages in the range from 50 kV to 800 kV can be realized. A typical test sequence, consisting of a conditioning phase and the withstand voltage test for 1 min. is given in Figure 5.1-5. Testing frequencies above 80 Hz to 90 Hz allow testing of the GIS including the voltage transformers. This avoids a special disconnector for the voltage transformer or, if not available, further disassembly of the GIS and the addition of other testing components. This additional assembly work can't be tested in the same way as the GIS as this becomes a weak point in the commissioning. If an addition to an existing GIS without a separating switch for the voltage transformer is provided, higher test frequencies allow the commissioning test of any additions with the existing GIS without re-moving the existing voltage transformer. Results achieved during HV testing at a 220 kV GIS substation, equipped with 144 UHF sensors of two types, are presented in Figure 5.1-6a,b. The substation consisted of eleven bays with ten feeders and a double busbar configuration with one bypass coupling. A variable frequency resonant test system was set up on-site. The HV was supplied via a bushing to the GIS. The sensitivity of the UHF sensors was checked before HV test according to the CIGRE sensitivity check described in [3.58]. Applications of wide and narrow band UHF method were applied to the GIS and the detected PD signals resulted in their origin being localized to within +/- 3 m. By means of acoustic measurements the PD location could be localised to +/- 0,5 m. The signature of the measured PD pattern indicated a free moving particle, Figure 5.1-6a, and b. The compartment was opened and two moving particle with a length of 2 mm were found [5.2].
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 34 - Figure 5.1-5: Typical test procedure on-site, consisting of a conditioning phase (Ur / √3) up to rated voltage Ur and the withstand test voltage for 1 min (0.8 x rated short-duration power-frequency withstand voltage Ud) Figure 5.1-6a: Spectrum of free moving particle Figure 5.1-6b: PD pattern of free mowing particle, integrated over 100 s in zero span modus, centre frequency is 637 MHz [5.2] 5.2 Cable Systems 5.2.1 HV Sources and PD measurement Two principal methods can be considered for the detection of on-site partial discharges in external energized power cables, (see Figure 5.2-1):  (A) Non-standard “unconventional” method according to IEC 62478  (B) Standard “conventional” detection method according to IEC 60270 [1.2, 1.8, 2.13] Table 5.2-1: PD detection during on-site testing TEST TYPE DESCRIPTION (A) AC voltage test [1.6] PD measurement (unconventional) not yet standardized a) frequency of detected signal within the radio Frequency (RF) range (e.g. up to 500 MHz), PD measurement in [μV] or [mV] b) calibration in [pC] not possible c) sensitivity check recommended d) detection of PD is mainly focused on cable accessories (e.g. joints and terminations) e) satisfactory signal / noise ratio is of importance (B) AC voltage test [1.6], [2.13] PD measurement in accordance to [1.2], [1.8] a) frequency of the detected signal in the kHz range, PD magnitude measurement in [pC] b) detection of PD on the whole cable system insulation c) localization of PD in cable accessories d) measurement is quite sensitive to external noise.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 35 - On-site after-laying testing of new installed (or repaired) EHV cable systems utilizes non-conventional PD measurement today using VHF or UHF frequency ranges, with focus on the accessories (e.g. joints and terminations). The external noise and disturbances can be suppressed down to a certain level and the eventual PD activity can be detected after the sensitivity checks have been done for the specific configuration of the cable accessories (e.g. joint, termination, the VHF/UHF coupler type and the signal processing device, see Figure 5.2-1). Figure 5.2-1: Principles of off-line PD detection methods for power cables; off-line standardized detection applicable for both routine testing and on-site testing. Using conventional PD detection, discharging activity can be detected in both the cable insulation and the cable accessories. The sensitivity of time-domain-reflectometry (TDR) can be estimated on the basis of the frequency characteristic of the cable impedance, Figure 5.2-2 [3.19]. When establishing the required sensitivity and maximum acceptable background noise level for the on-site PD measurement, the following aspects should be taken into account.  Type of cable system (e.g. type of cable and the type of accessories). Polymeric cable system are generally quite sensitive to partial discharges, whereas fluid-fill or mass-impregnated systems are usually less sensitive to PD. Regarding accessories, fluid-filled accessories can usually withstand higher PD activity than fluid-free accessories and for a longer time before failure occurs.  The operational electric stress level of the cable systems. In solid dielectrics the severity of PD activity increases quickly with increasing operational stress. As a consequence, HV and particularly EHV cable systems, which are designed for operating at relatively high electric fields, are as a rule much more sensitive to PD’s than MV systems, which operate at lower electric stress.  Test voltage level reached during the tests. When cable systems are being tested at voltages higher than the nominal voltage, the actual electric stress in the system is higher than the nominal stress. Consequently, the severity of the PDs increases with the test voltage. PD activity which may originate outside the test object such as corona discharges originating at the HV connection between the voltage source and the test object also increases with the test (over)voltage; for instance. As an example, severe defects in oil filled MV cables (PILC) or defects in accessories of MV XLPE cables may induce PDs in the range of hundreds of pCs without leading to an immediate failure. For those situations, a background noise level in the order of 100 pC can be accepted for the PD test on-site. On the other hand, for (E)HV polymeric extruded cables systems, PDs of a few tens of pCs may lead to failure within a relatively short time. Consequently, for this case it is important to maintain on-site background noise level as low as possible (10 pC or lower). Option B
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 36 - Figure 5.2-2: PD detection sensitivity versus the cable length calculated on the basis of a reference value of 10 pC achievable for a power cable of 1 km length [3.19] These examples also show the complexity of establishing general threshold values for PD levels during on-site measurements on cable systems. The threshold values used today are based on experience, long term observation of measured cable systems and following laboratory investigations of examined parts of the cable system such as joints. For these reasons the combination of HV withstand testing and PD measurement should be taken into consideration (see 1.2). In addition to the apparent charge, the evaluation of the phase-resolved PD patterns as well as the shape of the individual PD pulses can be very helpful for an assessment of the PD severity. For more information refer to chapter 3.3. In particular it is important to be able to discriminate between internal PD activity (e.g. PDs originated inside the test object) and external PD activity (e.g. PDs originated outside the test object). Sensitive PD measurements according to IEC 60270 can only be applied for a cable length up to few km. For longer cable lengths the PD threshold level is increased substantially due to attenuation and dispersion of propagating PD pulses (see Figure 5.2-2). PD Localization Depending on the cable system length, techniques based on time-domain-reflectometry (TDR) commonly used in laboratory testing are also applicable for off-line measurements on-site. For longer lines, with several accessories, approximate location estimates can be derived by looking at the magnitude and frequency content of the pulses, since, owing to attenuation, both quantities will be higher in proximity of the source [3.19, 3.69]. Synchronized measurements of arrival times can be used to derive a more exact location, by using GPS, fibre optics, or reference injected pulses, see e.g. [3.68]. Techniques based on time-domain-reflectometry (TDR) commonly used in laboratory testing are also applicable for off-line measurements on-site. Signal reduction analysis is based on the fact that PD pulses propagate through the power cable in different modes [3.19]. As the attenuation of the amplitude of the signal is dependent on the frequency of the propagating signal this effect can be used to roughly estimate the origin of the PD signals. 1 10 100 1000 0 1 2 3 4 5 6 7 8 9 10 11 Power cable length [km] DetectablePDpulsecharge[pC]
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 37 - 5.2.2 Required Details of Test Object The setting up of optimised testing procedures relies on an understanding of the goal of the on-site test and the cable characteristics. The following check list is reported for on-site tests including PD measurement. Cable, Joint, Termination and Lay-down Characteristics  Rated voltage;  Type of insulation of cables and accessories e.g. XLPE, ethylene-propylene rubber (EPR), mass-impregnated, Fluid-fill, etc.;  Single core or multi-core cable (for MV cable), propagation velocity of PD signals;  Cable length (capacitance, loss factor at rated voltage and frequency);  Type of installation (e.g. indoor or outdoor);  Type of earthing (e.g. solid bonding, single point bonding or cross-bonding);  Year of installation, average service load and other relevant historical data (e.g. previous failure, sheath defects);  Number of joints and their distances from the measuring point. Moreover number and position of cross bonding joints (whenever used) where it is usually possible to install unconventional PD sensors;  Type of termination (e.g. outdoor termination, metal-enclosed termination, etc.);  Availability and characteristics of the power supply (e.g. network earthing: insulated neutral or earthed network);  Clearances, position of terminations and presence of external structures that may interfere with the test. Test Requirement  Aim of the tests (e.g.: commissioning after installation, commissioning after major maintenance, diagnostics, monitoring etc.);  Voltage shape and frequency requirement (see chapter 2);  Test procedure: voltage level(s), duration of application or number of impulses (in case of DAC) for each voltage levels, highest voltage (for both PD measurement and withstand voltage if required);  Location of defected points (e.g. cables insulation, joint or termination for later risk analysis);  Required PD measuring method;  Required maximum accepted PD magnitude, or PD inception and extinction voltages, or alternatively definition of criticality degree of the possible defected points and indication of the consequent actions, (e.g. repetition of the test within a given time or fix the defected point), related to the position of the defect (cable, joint, termination and their insulation materials);  Calibration of the measuring system or sensitivity check (in case of unconventional PD measurement). Test System and Testing Conditions  Supply requirements (voltage, frequency and power and their tolerances);  Grounding requirements;  Weight, size and environmental (temperature, humidity, air pressure and pollution class) restrictions for transport and operation;  Type, and dimensions of connection from the generator to the cable;  Characteristics of PD measuring systems (sensitivity, bandwidth, conformity with IEC 60270);  Additional PD evaluation features of the measuring systems (filter to enhance the signal-to- noise ratio, possible location of PD spot, automatic defect recognition etc.);  Possible presence of sources of electromagnetic disturbances in the vicinity of the test objects (e.g. power electronic devices, antennas, etc.);  Requirements regarding propagation velocity (e.g. at mixed cables);  Necessity and availability of PD pin pointing equipment.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 38 - 5.2.3 Example 1: HV Cable Systems Tested by ACRF Test System and Conventional PD Measurement Technique An ACRF (see 2.2) test system including PD measurement facility based on a spectrum analyser was used for commissioning of cable systems [5.3]. This system allows the PD measurement to be performed without internal PD sensors being integrated into the cable accessories. The external sensor is based on a coupling capacitor or a high frequency current transformer using the principle of IEC 60270 [1.2]. Over four years, more than 60 PD commissioning tests have been performed in conjunction with withstand tests. In this example the discharge activity ranged from as low as 15 pC to 250 pC. The noise level ranged from 5 pC up to 70 pC. On average a noise level of 20 pC was experienced. The variation in this noise level was determined by local circumstances. Figure 5.2-3 shows a measurement on a HV cable resulting in a noise level of approximately 10 pC. The four spikes (1) to (4) in this figure are caused by the frequency converter of the ACRF test system. The measurement did not show any partial discharge activity in the cable circuit. Unfortunately, some PD measurements could not be performed because of extremely high noise levels, ranging up to 200 pC. These high levels were experienced in a substation environment with live equipment in the vicinity and under rainy conditions. On three occasions evidence of PD activity in the cable circuit was found. PD activity outside the cable circuit has also been measured including PD patterns caused by floating parts close to high voltage (Figure 5.2-4) and discharges generated by an insulating rod as a result of wear (Figure 5.2-5). The rod was used to support the tube connecting the reactor of the ACRF test system to the cable under test. After replacing the rod the disturbance disappeared. Even corona activity inside a connected GIS was measured. This PD activity was found to be from the GIS cable termination which was not correctly equipped with a temporary corona shield. After installing such a shield, the discharges have disappeared. Finally the situation on-site may influence the sensitivity of the test. It is essential, for example, for the ACRF test system to be positioned as close to the termination as possible. In order to minimize disturbances, the leads must be connected to the terminations in short straight lines. It is especially important to consider fences, screens and other earthed objects when connecting the test object to the voltage source (see Figure 5.2-6). Although the clearances under such circumstances are sufficient from a voltage point of view, disturbing influences for PD measurements cannot always be excluded. Figure 5.2-3: Example of a PD-free measurement, noise level approximately 10 pC (1) (3)(2) (4)
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 39 - Figure 5.2-4: Example of discharges from a floating electrode Figure 5.2-5: Example of discharges from an insulating rod used for support of the HV connection Figure 5.2-6: Example of a situation requiring some improvisation for the HV connection; a) HV source, b) HV electrode, c) connection to test object via bushing Time [s] a b c
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 40 - 5.2.4 Example 2: MV/HV PILC Cable System Tested by DAC Test System and Conventional PD Measurement Technique This chapter is related to the practical experience of on-site PD measurements on MV and HV power cables using damped AC (DAC) test voltages with a frequency between 20 Hz and 500 Hz [5.4 - 5.7]. For a 10 kV PILC (paper insulated, lead covered) cable system, the PD parameters (PD inception voltage (PDIV), PD extinction voltage (PDEV), PD magnitude at PDIV and at voltages up to 2,0 U0 were measured to determine the severity of the PD defects. In addition to the PD parameter and mapping of PD occurrences the ground noise level has also been documented to determine the test conditions (Table 5.2-2). The result of a typical PD measurement is given in Figure 5.2-7. Typical patterns from PD in an oil- filled joint can be distinguished, with some expertise, from PD in voids (Figure 5.2-7b) and PD between paper layers in a dry area of PILC cables (Figure 5.2-7a). Stressing the cable insulation with DAC voltages and applying TDR to PD signals in HF range (up to 50 MHz) provides the PD site location in cable insulation. As a result the PD mapping of a cable provides an indication of the location and statistics of PD occurrences (Figure 5.2-8). Table 5.2-2: Example of typical PD and test parameters of a three phase 10 kV PILC cable (length: 1950 m) With knowledge of the type of components in a circuit combined with the mapped PD activities, the operational history and importance of the circuit, an evaluation of PD defects and recommendations for maintenance or replacement can be determined. In general HV power cables (50 kV to 150 kV) should be PD free. This should also be the case during after-laying testing of new or repaired HV cables, or during condition assessment testing of service aged power cables (Figure 1.4c). Figure 5.2-7: Typical phase resolved PD patterns as observed at damped AC stresses of: a) an insulation defect in oil filled cable system b) an insulation defect in epoxy insulation of joint or dry area of PILC insulation
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 41 - Performing test voltages also up to 1,73 U0 (where the maximum is advisable for new or recently installed cables, lower levels should be used for aged cables) is important for several reasons:  To evaluate if there are insulation defects with a PD inception voltage (PDIV) > U0. Such defects may initiate insulation failure in the case of temporary AC over-voltages.  To conclude following successful after-laying test that there are no PD detected up to 1,73 U0 and that during the service operation the power cable insulation is free of discharging defects.  That the on-site test has not initiated any discharging processes in the insulation. This information is important to confirm the non-destructive character of the diagnostic on-site test itself. An example of a 55 year-old oil-filled cable section which is not PD-free is shown in Figures 5.2-9 to 5.2.-11. In particular as the PDIV was lower than U0 it was important to localize the PD activity in the cable. It follows from the PD mappings (Figure 5.2-10) that all discharges as observed in phase L1 are localized in cable joints. Also it was observed that by increasing the test voltage up to 1,2 U0 the PD magnitudes did not significantly change. Additional measurements of dielectric losses (tan ) as a function of the test voltage confirmed that phase L1 had higher values of dielectric losses compared to phases L2 and L3 (Figure 5.2-11). Information of the presence of PD activity in one of the phases of the 150 kV cable section and the differences in dielectric losses between the three phases were valuable inputs for further decisions about the maintenance and the operation of this particular HV cable section. Figure 5.2-8: Example of a PD mapping of mixed PILC and XLPE cable circuit Figure 5.2-9: Example of phase-resolved PD patterns as observed during on-site testing of a 150 kV, oil- filled, 55 year-old, and 6,2 km long HV cable section: a) at 0,7 U0: background noise is < 10 pC, the cable section is PD-free b) at 1,0 U0: PD activity of 40 pC starting at 0,9 U0 has been observed in phase L1 c) at 1,2 U0: PD activity of 90 pC starting at 0,9 U0 has been observed in phase L1
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 42 - Figure 5.2-10: Examples of PD mappings as made during on-site testing of a 150 kV, oil-filled, 6,2 km long,and 55 year-old HV cable section: a) up to 1 U0: PD activity between 20 pC and 105 pC has been observed in four joints (phase L1) b) up to 1.2 U0: as compared to 1 U0 no significant change of PD behavior Figure 5.2-11: Example of dielectric losses (tan ) measurement between 0,3 U0 and U0 on a 150 kV, oil filled, 6,2 km long and 55 years old HV cable section. 5.2.5 Example 3: Extruded EHV Cable System Tested by ACRF Test System and Non-Conventional PD Measurement After the installation of the 380 kV cable connection beneath the waterways “Caland Canal” and “Nieuwe Waterweg” in The Netherlands [5.6], an AC voltage test with non-conventional PD measurements were performed. The 380 kV cable connection consists of two 2200 m long XLPE cable circuits. Each circuit is divided into two sections by means of joints and it is terminated with composite outdoor sealing ends. Each phase of the two circuits was tested for 1 hour at 374 KV AC voltage (phase to ground), according to the IEC 62067 [5.8]. The voltage was applied by means of an ACRF system (see paragraph 2.1). The AC voltage frequency during the test was 43 Hz. During the AC voltage test, partial discharge measurements were made at the cable accessories. For this purpose three autonomous PD measuring units were installed at the two sealing ends and at the joint of each cable phase. A measuring unit consists of a VHF/UHF PD sensor integrated in the cable joint, a pre-amplifier and a PC-controlled spectrum analyzer [5.9, 5.10]. In addition, a trigger unit was adopted to synchronize the AC voltage and the PD signal. Finally, a wireless network was used for the communication between the three stand-alone measuring systems. In Figure 5.2-12 the PD detection system is represented. Before starting the PD measurement, the sensitivity check was performed as described in [5.10]. This was done in order to correlate the measured voltage signal in μV to a PD signal in pC. Using a fast impulse generator which was calibrated in the laboratory, artificial PD pulses were injected into each cable termination. 0 10 20 30 40 50 60 40 60 80 100 115 V[kV] value*10^-4 L1 L2 L3
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 43 - Figure 5.2-12: VHF/UHF PD detection system as used during the on-site PD measurements performed on the 380 kV cable accessories [5.9] To achieve optimal performances of the measuring system, the spectrum of each PD detection unit was tuned to maximize the signal-to-noise ratio. For the specific spectrum range used for the PD detection, the noise level was lower 10 pC. However, due to the fact that the terminations themselves act as large antennas, some external noise activity was observed. In order to separate these external disturbances from potential partial discharge signals, the phase-resolved PD noise patterns were displayed in real time and stored during the PD measurements. In this way it was possible to establish that all accessories employed in the cable connection were PD-free, at voltage of 374 kV. Some of the results of the PD measurements are shown in Figure 5.2-13. Thanks to the real-time phase-resolved PD plotting, random disturbances could be discriminated, as shown Figure 5.2-13, at termination 2. PC Termination 1 Joint Termination 2 Figure 5.2-13. Measuring results simultaneously obtained at the three cable accessories during the acceptance test of one cable: the PD magnitude versus time for one hour testing and examples of phase-resolved patterns obtained during several cycles of the test [5.10]. Displayed PD pulses are caused by external noise. sensor spectrum analyser pre-amplifier
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 44 - 5.2.6 Example 4: MV Cable System Tested by VLF Test system and Conventional PD Measurement Technique PD-measurement and fault localization of PIL- and XLPE-insulated cable systems is also performed by VLF voltage with conventional PD measurement. Testing equipment is compact and light-weight with low power consumption and can be stored in a small van, (Figure 5.2-14). When performing PD measurements under VLF test voltages one has to consider that the electrical field may be controlled more by the resistance of the dielectric, which is different from the capacitive field control under service condition. Hence, the PD occurrence may change substantially under VLF test voltages. Figure 5.2-14: VLF System on-site test arrangement Because of the low frequency and the low voltage gradient of the VLF voltage against the 50/60 Hz operating voltage, it may take considerable time to get a sufficient number of single PD pulses for PD detection and localization (see 2.4 and Figure 2.7). This time factor does not influence PD detection and location however. On the other hand PD activities are activated at comparatively low voltage levels [5.10]. As the general PD activity is much higher in PIL-cable systems than in XLPE-cables, PD recognition is less complicated. Experience also shows that the time required for data handling is much longer than the time taken to make the measurements. Furthermore well trained operators are necessary. In most cases the PD source is not located within the cable insulation itself, but in joints and terminations. Therefore it is very important to identify the exact location of the PD source. Experiences with VLF systems gained in laboratory and on-site show that the localization accuracy is in the range of several meters As for all PD measurements, the measured levels by different type of test equipment cannot be compared against each other, while using identical equipment test and applying identical test procedures leads to comparable results. 5.2.7 Example 5: UHF PD Measurement at Cable Accessories in Service Statistics show that most failures in the insulation of cable systems occur in accessories, such as joints and terminations. Critical PD defects in cable accessories can sensitively be detected by the UHF method, as is well reported in [5.13], [3.54]. Due to the strong attenuation of the ultra-high frequency spectrum, only PD signals adjacent to the PD source are captured. Therefore the UHF method can be characterized as spatially selective, which ensures a high signal-to-noise ratio. Consequently, the coupling sensors should be placed as close as possible to the test object.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 45 - PD activity detected within accessories is a clear signal for replacement. Consequently, those accessories that are about to fail should be replaced immediately, thereby reducing the risk of cable system failures. In the following a practical example will be presented which refers to on-site on-line UHF PD diagnostics of MV cable terminations (the described coupling sensors are also useful for off-line UHF PD measurements with an external test voltage source). To screen the termination effectively against external electromagnetic interferences a barrel sleeve is clamped around the plug-in connector (size 4, up to 72 kV) as shown in Figure 5.2-15. Inside of the sleeve there are three 5 cm long monopole antennas, each placed 1-2 cm above the cable surface and oriented at a tangent to the cable cross-section. Antennas are shifted by 120° from each other to embrace the whole circumference of the test object. The measured signal from the antenna via is fed to the digital oscilloscope via a 40 dB pre-amplifier (in a frequency range from 1 MHz to 1000 MHz). The calibration of the UHF method in terms of the amount of charge is impossible. However comparative PD measurements were performed on an identical test set-up assembled in laboratory, using both the UHF method and the conventional method according to IEC 60270 [1.2]. This was done to establish the maximum sensitivity of the UHF method. Raising the applied voltage controlled the magnitude of PD. A sensitivity, corresponding to 4 pC, was detectable by the UHF method. PD-measurements in an unshielded laboratory were performed on several 45 kV cable connectors. Connectors with simulated PD emitted fast electromagnetic pulses whereas measurements on connectors without PD showed only slight background noise. So the detection of UHF PD signals constitutes a strong sign for internal PD activity. Figure 5.2-16 shows such a typical fast pulse signal picked up on a connector with PD. The frequency spectrum of this pulse is plotted in Figure 5.2-17. Figure 5.2-16: Typical fast pulse emitted by some connectors Figure 5.2-17: Frequency spectrum of PD-pulse, acquired with a bandwidth of 3 GHz There are several high frequency components that indicate presence of PD activity in addition to the broadcast and GSM frequency spikes. The failure of this connector a few weeks after the measurement confirmed this diagnosis. Range of amplification Reproducible frequency components of the measured fast impulse Range of amplification Reproducible frequency components of the measured fast impulse 1 2 Figure 5.2-15: Test set-up consisting of (1) plug-in connector and (2) metallic barrel sleeve with antennas
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 46 - 5.3 Rotating Machines For more than 50 years insulation systems of rotating machines (generators and motors) have been studied by applying technical diagnostic tests. Over the years there have been different standards for the testing of winding insulation: including national recommendations, international standards and several routine tests specified by customers. To ensure reliable operation and to increase the lifetime of rotating machines, on-site diagnostic tests have gained more importance over the last few years. 5.3.1 Pre-Conditions The following test methods are usually applied for on-site diagnostics:  Visual inspection  Insulation resistance (Charging and Discharging Currents: PDC, DRA Analysis)  Dissipation factor (tan  and capacitance as a function of test voltage and frequency  Partial discharge as a function over test voltage This chapter focuses on HV testing in conjunction with PD measurement. The standards for PD measurement of rotating machines are not at the same level as on-site PD measurement for other equipment such as GIS or cables. One reason for this is the complex frequency dependent signal propagation function of PD pulses within the extended stator winding. Recent publication dealing with PD measurements and the characterization of PD patterns include IEEE Std. 1434-2000 [5.14] and IEC 60034-27 [5.15]. The practice of PD measurement is different in various countries. In USA broad band measurement systems (up to several tens of MHz) have been common whereas in Europe small band measurement systems (e.g. 10 kHz bandwidth) have also been applied for detailed failure characterization and detection. For a risk evaluation of the insulation system condition, the PD magnitude is measured in the power station with the machine off-line. To be able to perform the testing on-site, adequate mobile high voltage sources and sensitive PD measurement equipment are needed. Due to the environmental effects, the background noise level can be high compared to laboratory conditions. Fortunately the PD magnitude of a new respectively healthy winding insulation is 1000 pC and more, depending on the insulation technology. In addition tan measurement can be performed. In this case environmental influence plays less of a role, but the temperature of the windings must be observed to get comparable results. Restrictions of PD measurements Due to the strong dependency of the PD signal propagation within a rotating machine, these machines cannot be treated as lumped devices in circuit diagrams. As a major consequence the PD magnitude itself has no clear meaning and gives no direct hint about the harmfulness to the machine of a PD source. Instead a trend analysis is more important and a comparison of results gained from machines of the same type, ratings and age allows a much better and more secure evaluation of the PD measurement. This is also one reason that PD magnitudes for rotating machines are widely given in mV instead of pC, as this gives users the erroneous belief that the reading are absolute and that for example a 10 nC discharge indicates the same amount of damage regardless of winding characteristics and machine design. Test system requirements The rated voltages of generators lie in the medium voltage range. Voltages of 6,3 kV or 10,5 kV are often used. It is necessary that the AC test voltage can be regulated in steps of 10% or 20% of the rated voltage. The generators have the dielectric property of a high capacitive load, which demands medium voltage test systems with high power in the range of some MVA or the capability to compensate the reactive load. Therefore ACRL and ACRF test systems are recommended. An AC test system based on transformers (ACT) can only be used for machines of lower power and low test voltages (< 20 kV). The ACRL or the ACRF circuit can be successfully used for testing rotating machines on-site. Two examples of realisations are given below.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 47 - 5.3.2 Example 1: On-Site Tests by ACTC and ACRL Test Systems and Conventional PD Measurement Technique The resonance test system with variable inductance (ACRL) shown in Figure 5.3-1 has a rated power of 350 kVA and two taps of 50 kV and 30 kV respectively. The whole test system is mounted into a 3 m container, together with control room and switching cubicle. As an example, the stators of a 42 MVA/10,5 kV and a 20 MVA/6,3 kV generator were tested with this mobile system. The HV test was done together with a PD and dissipation factor measurement. The capacitances of the generators to be tested were about 780 nF and 691 nF respectively. The maximum applied test voltages were 12 kV and 7,6 kV at 50 Hz. Figure 5.3-1: Mobile HV test system on-site in a pumped-storage power station In Figure 5.3-2 the results of the dissipation factor measurements are shown. Generator 1 shows an almost linear behaviour of tan  with increasing test voltage and generator 2 a characteristic bend at the PD inception voltage. In Figure 5.3-3 the test results of PD measurements are displayed as a function of applied test voltage. The PD magnitudes were evaluated as QIEC values according to IEC 60270 (2001). Figure 5.3-2: Measured dissipation factor tan  as function of test voltage for two generators Dissipation Factor Test Results 5,0 10,0 15,0 20,0 25,0 30,0 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 Test Voltage (50Hz) [V] tan[‰] Generator 1: 42MVA, 10.5 kV Generator 2: 20 MVA, 6.3 kV Ionization bend due to PD inception C. Sumereder, TU GRAZ
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 48 - Figure 5.3-3: Measured partial discharge intensity and PD inception as function of test voltage for two generators 5.3.3 Example 2: On-Site Tests by ACRF Test System and Conventional PD Measurement Techniques The test was performed with an ACRF test systems on stator windings of turbines of ratings of 137 MVA/15,8 kV and 1640 MVA/27 kV. This ACRF system of test voltage generation is not affected by intensive partial discharges. Quality factor of the resonant circuit is smaller than with VPE cable test circuits, but the test power was sufficient, not only for single phase tests with 42 Hz respectively 46 Hz, but also to test the parallel connection of all three phases with approximately 26 Hz. The high voltage test was performed together with partial discharge and tan  measurement, the voltage being increased in steps. Figure 5.3-4: Compact resonant test system (ACRF), 36 kV/10 A for testing stator windings; 1 – reactor, 2a – basic load, 2b – cable to the test specimen, 3 – exciter transformer, 4 – frequency converter & control, 5 – PC and PD measuring system Results of partial discharge measurements made at 50 Hz were compared with results using the mobile ACRF test system (Figure 5.3-4). All measurements were made in accordance with common practice on generator windings with a wide-band partial discharge measuring instrument and with a band-pass filter characteristic (20 kHz to 20 MHz). In order to receive a reproducible partial discharge characteristic, the winding component to be tested was conditioned in each case with rated voltage Un for 5 minutes before the PD measurement was performed. The impulse distributions from the phase- dependent partial discharge patterns (-q-n-pattern) were recorded after each voltage step of 0,1 Un. Partial Discharge Level according to IEC 60270-2001 0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 Test Voltage (50 Hz) [V] PartialDischargeIntensityaccordingQIEC[pC] Generator 2: 20 MVA, 6.3 kV PD inception at 30% rated voltage Generator 1: 420MVA, 10.5 kV PD inception at 60% rated voltage Ionization bend due to PD inception C. Sumereder, TU GRAZ
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 49 - Figure 5.3-7: tan  as function of normalized test voltage of a 16 kV stator winding; parameter is the test frequency; also shown is the calculated value for 50 Hz PD patterns are shown of the 27 kV stator winding of a nuclear power station generator, measured at rated voltage Un = 27 kV and at operating voltage 3/Un = 16,2 kV and a test frequency at 45,7 Hz (Figure 5.3-5 and 5.3-6). These PD patterns are typical for mica based machine insulating systems. The variation of the test frequency did not affect the typical partial discharge behaviour expected in the mica insulation of rotating machines. As PD activity will increase tan  the influence of the test voltage on this diagnostic quantity will be briefly discussed. To allow a consistent interpretation of the test results over decades of the entire machine life span, the test conditions must always be kept comparable. The dissipation factor tan  is by definition proportional to the test frequency. This enables a conversion of the test results to other frequencies. Therefore the actual effects of the test frequency on the dissipation factor were examined in comparative measurements. The voltage-dependent dissipation factor values were therefore determined with an ACRF test system at 26 Hz and 42 Hz at a 16-kV-stator winding. These results are represented in Figure 5.3-7 together with the conventional 50 Hz measurement. The curves at different frequencies are parallel-shifted, where, as expected, the higher test frequency leads to larger tan  values. The conversion of the different test frequencies to the 50 Hz reference value provided larger tan  values however. One possible cause of this anomaly may be the non-linear resistive behaviour of the materials in the glow protection units. The frequency dependant capacitive loss current of these units causes an under-proportional change of dissipation factor at reduced test frequency. The rise of the dissipation factor over the voltage, tan /U, which is an indication of the quality of the impregnation during manufacturing and of the degree of delamination with aging, is to a large extent independent of the test voltage frequency (Figure 5.3-7). Figure 5.3-5: PD pattern measured at Un = 16,2 kV (27 kV/sqrt (3)) at a 1640 MVA generator with an ACRF test system at 45,7 Hz Figure 5.3-6: PD pattern measured at Un = 27 kV at a 1640 MVA generator with an ACRF test system at 45,7 Hz
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 50 - 5.3.4 Example 3: On-Site Tests by DAC Test System and Conventional PD Measurement Technique Using a damped AC system (Figure 5.3-8), the PD behaviour and the level of dielectric losses of four different motors of ages of between 5 and 10 years have been tested on-site (Table 5.3-1). Table 5.3-1: Overview of motors tested on-site. Rated voltage Rated power Machine No 1 6600 V 10750 kW Machine No 2 6600 V 12240 kW Machine No 3 6600 V 7840 kW Machine No 4 3300 V 4580 kW The test procedure was the same for all cases and consisted of a number of tests between the partial discharge inception voltage (PDIV) and 1.7 U0. For data analysis the following information has been evaluated:  The PDIV for phases U, V and W  The PD magnitude at PDIV  The PD magnitude up to 1,7 U0 The diagnostic parameters were collected for a series of machines as shown in Table 5.3-2. In particular the tests have been performed at both the HV and LV side of the stator coil. To explain this process the results obtained for the Machine No. 3 will be discussed below. In Tables 5.3-2 to 5.3-3 the measuring data is shown. In Figure 5.3-9 the PD behaviour as a function of the test voltage is shown. It can be concluded, that in comparison to HV connection of the machine at the LV side (star connection), a significantly higher PD activity has been observed. This evidence suggests that further tests are needed to identify the source and asses its harmfulness. Figure 5.3-8: Example of on-site PD diagnosis and with dielectric loss measurement of stator insulation, using damped AC voltages.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 51 - Figure 5.3-9: PD behavior as function of the test voltage, measured on a stator winding Table 5.3-2: PD magnitudes in [nC]; Table 5.3-3: PD magnitudes in [nC] detected at LV measured at HV side of the stator side of the stator 5.4 Power Transformers 5.4.1 General PD measurement during the acceptance test with induced voltage on power transformers has been used in the manufacturers’ test bays since the seventies. The determination of maximum allowed PD magnitudes during the AC voltage test is strongly dependent on the philosophy of AC voltage withstand testing. Up to the early eighties the test voltage levels continuously increased. Usually the test voltage level was applied for 6000 cycles (e.g. 50 Hz for 2 min.) In the early eighties a new test procedure was introduced with IEC 60076-3 [5.17]. There was no PD magnitude specified for test voltage levels but after that, the test voltage was kept at 1,3 / 3mU or 1,5 / 3mU for a period of 30 minutes (so-called method 2). During this time the PD magnitude was required not to exceed a HV Phase HV Phases Voltage L1 L2 L3 L1,L2,L3 1,3 kV 1,3 30 0.03 30 2,6 kV 25 68 9.6 25 3,9 kV 65 70 20 80 4,5 kV 65 80 25 100 6,6 kV 70 90 55 110 HV Phase HV Phases Voltage L1 L2 L3 L1,L2,L3 1,3 kV - - - - 2,6 kV 23 11 10 13 3,9 kV 25 25 12 15 4,5 kV 25 25 14 17 6,6 kV 25 35 15 25
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 52 - certain limit. In 2000 the philosophy of having a short duration test (ACSD) at high voltage levels and a long duration test (ACLD) at lower levels was consequently introduced in the new edition of IEC 60076-3 [5.17]. The idea behind this test is to test the design of the transformer using a high voltage level (ACSD) and to check the quality of manufacturing by the long duration test (ACLD). Thereby, maximum PD magnitudes have been defined for the ACLD as well as for the ACSD test procedure. Recently, the on-site PD measurement gained more and more importance. This is mainly due to the lower cost of on-site repair of power transformers as an alternative to the transportation of the transformer back to the factory for repair. However, on-site repair of power transformers requires not only a certain manufacturing competence on-site but also the capability for on-site testing [2.10, 2.11, 5.18, 5.19]. Therefore, recent developments have been aimed into two directions, firstly PD measuring techniques including PD localization for on-site application have been developed and secondly, frequency converters have been developed as a new type of voltage source which are much easier to handle than motor-generator sets. 5.4.2 Preconditions of Test Object Various preparations have to be done on-site before a high-voltage test on a transformer can be performed. First of all it is essential to have the final acceptance testing protocol of the transformer determined, defining the test levels and the standards (usually IEC or IEEE) to which the tests have to be performed. Based on the final acceptance protocol and the defined test levels it will be possible to estimate the power is needed to perform the tests allowing the size of the test system itself to be adapted in advance. This requires the knowledge of the active and reactive power requirements during the induced voltage test. However, usually this data has never been recorded during the induced voltage test since the apparent power in the manufacturer’s transformer test bay is sufficient for the induced voltage test on all transformers. The situation is completely different on-site. The following aspects have to be taken into consideration:  Power supply of the test voltage source (motor-generator set or frequency converter) On-site the power for testing of transformers must be taken out of the power network or from a mobile diesel-generator set. The required power can be limited with respect to reactive power.  Compensation Sometimes on-site measurements are far away from civilisation. Therefore, if the compensation of reactive power is not sufficient, testing is not possible and the test voltage may not be able to be achieved. Organizing additional compensation within a short time period can be very difficult and depends on the test location. Therefore, it is extremely helpful to have a good estimate of the power consumption of the induced voltage test. Due to the limited power supply often available for on-site testing, the frequency and compensation can be adjusted in a way that the power consumption is minimised because the transformer is an inductance at lower frequencies and capacitance at higher frequencies (Figure 5.4-1) [5.15, 5.20, 5.21]. At the frequency of self-compensation the reactive power consumption by the test object is theoretically zero and practically extremely low. If this effect is used, it must be considered that the output power of the test voltage source may also depend on the frequency and on the relation between reactive and total power, expressed by the power factor cos  (Figure 5.4-2). Therefore, knowledge of the dependence of the capacitive and inductive behaviour on the frequency would be valuable. For older transformers this information is often not available. Experience of the test engineer and flexibility of the test equipment are therefore prerequisite for successful testing. Furthermore it is necessary to have information on environmental conditions on-site to be able to determine the location of the test system on-site, and the location of the power supply for the test source. One should also inspect the site to determine whether power equipment like overhead lines, surge arresters, etc. have to be modified or removed during the HV-tests in order to reach the required
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 53 - distances between components at high voltage and ground potential. It is also necessary to check the climatic conditions expected during the test so that these can be taken into consideration. Figure 5.4-1: Test power demand of a 100 MVA power transformer as function of test frequency 10 100 1.000 10.000 0 50 100 150 200 250 Frequency in Hz TestingpowerinkWorkVA S P P active power [kW] S overall power [kVA] „self compensation“ capacitive loadinductive load 10 100 1.000 10.000 0 50 100 150 200 250 Frequency in Hz TestingpowerinkWorkVA S P P active power [kW] S overall power [kVA] „self compensation“ capacitive loadinductive load 40 60 80 100 120 140 160 180 200 0 50 100 150 200 250 300 350 400 450 500 550 2-phase capacitive load 2-phase inductive load 3-phase capacitive load 3-phase inductive load           )=-0.2 )=-0.8 )=1.0 )=0.8 )=0.2 )=-0.2 )=-0.8 )=1.0 )=0.8 )=0.2 cos( cos( cos( cos( cos( cos( cos( cos( cos( cos( PowerLimit(kVA) Test frequency (Hz) Figure 5.4-2: Available output power of a 500 kV frequency converter depending on frequency and power factor cos .
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 54 - Following the above general preparations it is necessary to perform a number of preliminary on-site tests to assess whether the condition of the transformer will allow high voltage testing. For this condition assessment the following tests are necessary:  insulation resistance  winding resistance  transformation ratio  dielectric dissipation factor of transformer and bushings  frequency domain spectroscopy in order to determine the moisture content inside the solid insulation and  frequency response analysis in order to detect winding distortions. Finally the insulation liquid parameters have to be tested in order to ensure the oil has adequate breakdown voltage and low moisture content. If all these tests are passed, the set-up for the HV-testing can be arranged including shields on the HV-bushings to avoid corona signals and short circuits on the current transformers. Usually the induced voltage test with PD measurement is combined with other HVAC tests. The following AC test-sequence is recommendable unless otherwise agreed:  determination of no-load losses  applied voltage test LV/MV/HV  induced voltage test including partial discharge measurements. 5.4.3 Example 1: On-Site Test by Motor Generator Set and Conventional PD Measurement Technique The low-voltage side of the transformer under test is fed by the M/G set via a matching transformer. The induced voltage test can be performed as a single-phase or a three-phase test depending on the power transformer and the available motor-generator set (M/G-set) [5.22, 5.23]. On-site also three- phase power transformers can be tested with a single-phase M/G set (Figure 5.4-3 a), but a three- phase test (Figure 5.4-3 b) is recommended if a related M/G set is available. The power supply of the M/G set must be able to provide the required active power. The generator must be able to provide the active as well as the reactive power required by the test object. Therefore, compensation of the reactive power has to be taken into account. Usually the compensation does not need to be very exact if the generator can also provide a certain amount of reactive power. For conventional PD measurements in the test bay (according to IEC 60270 [1.2], see also chapter 3.1), the transformer bushings are usually used as coupling capacitors. Some years ago so- called “radiation noise detectors” (selective signal measurement) were frequently in use – even in manufacturers’ test bays. However, special multi-channel PD measuring systems are now in service. The preparation of the PD measurement starts with the calibration procedure. During the test, PD magnitudes at different test voltages are recorded by the PD measuring system. A synchronised multichannel PD measuring system is another measurement technique for three phase power transformers to localize the PD source. It offers the possibility of improving the S/N ratio and the localization of the PD source [5.24]. On-site PD measurement can be difficult because of the presence of external disturbances. One method to distinguish internal from external PD, is to correlate signals from AE sensors or piezoelectric transducers (mounted on the tank wall) together with the electrical PD measurement. If the microphones show signals the PD is generated inside the transformer with high probability. Otherwise, it is an external PD. However, the electrically measured PD signals contain both, the internal and the external partial discharges. Therefore, the sources of external corona discharges should be removed as far as possible. An extremely useful tool is an UV camera which is able to detect the ultra violet (UV) emission of external partial discharges caused by corona discharges, e.g. at the bushings. In principle, acoustic corona discharge detection would be also possible.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 55 - Figure 5.4-3: Circuit diagram of the induced voltage test arrangement (without compensation) a. single phase test b. three phase test Figure 5.4-4: On-site test of a 740 MVA/420 kV/27 kV transformer using a M/G set on PD measurement according to IEC 60270 However, an exact location of corona discharges is usually difficult using only acoustic detectors. During an induced voltage test of a 740 MVA power transformer (420/27 kV) (Figure 5.4-4) pollutions on the insulator of one of the bushings resulted in corona discharges and in a measured PD magnitude of about 1000 pC. Cleaning of the entire bushing is extremely time-consuming. In that case the UV camera was useful in order to detect all corona discharge sources within a reasonable time span of a few hours. 5.4.4 Example 2: On-site test by Frequency Converter and Conventional PD Measurement Technique The use of a static frequency converter system for high voltage testing of power transformers [5.19- 5.21] allows a simple and rapid set-up on-site, because the whole HV test system can be installed into one container as shown in Figure 2.3 [5.20]. This allows high mobility because the system can be transported easily by road or by ship. Mobile systems of power ratings of up to 500 kVA have been in use for some years. That power is sufficient for most transformers. Recently systems with a power in a range of 1 MVA have become available, removing the restrictions in testing even the largest GSU-transformers. As static frequency converter are operating with semi-conductors like IGBTs (insulated gate bipolar transistors), voltage transients are produced leading to an increasing PD measurement noise level in certain frequency ranges. Therefore special PD measurement methods have been developed [5.25] to address this issue. Figure 5.4-5 shows a mobile test system based on a frequency converter when testing a 400 MVA GSU transformer. The system is provided for applied voltage testing (a, b), induced voltage testing (c, d) and loss measurements. The applied voltage test is similar to GIS testing (see chapter 5.1), whereas the induced voltage test requires the three-phase configuration shown in Figure 5.4-6 (as single line diagram). The three-phase output of the converter is adjusted to the required input voltage of the transformer under test by a matching transformer with compensation on both sides. With respect to noise suppression a medium-voltage filter is applied and the filter capacitor is also used as a voltage divider. The three bushings of the transformer under test are used for synchronous PD measurement by a multi-channel PD measuring system. The whole test is computer controlled.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 56 - (a) (b) (c) (d) Figure 5.4-5: Mobile test system based on a frequency converter when testing a 400 MVA GSU transformer. (a) and (b) show the applied voltage test and (c) and (d) show the induced voltage test The low-voltage side of the 400 MVA GSU transformer under test is connected to the step-up transformer by three medium-voltage cables. The test frequency adjusted by the self-compensation (see chapter 5.4.2) is 140 Hz, the test voltage 80% of that for routine tests and the step-test procedure according to Figure 1.3 is applied with a duration of 43 s at the highest level. As can be seen from Figure 5.4-7, a sensitivity ranging below 15 pC is achievable, although the sensitivity is different from case to case depending on surrounding noise influences. The figure shows the PD behaviour of one phase during the highest test voltage, where no PD signals could be seen resulting in the test being passed. Figure 5.4-6: Block diagram of an induced voltage test based on a frequency converter (single line diagram of a three-phase arrangement)
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 57 - Before the HV test can be performed the PD measurement equipment has to be installed and calibrated. Figure 5.4-8 shows the spectrum of the noise level with and without the frequency converter operating and the calibration signal. Based on a comparison of these curves, the mid- frequency is chosen in such a way that the highest sensitivity is reached, but as low a frequency as possible (at least in a range of a few 100 kHz). If this frequency is found, which in this case was around 1 MHz, the used spectrum analyzer operates for the measurement in “zero span” mode at the selected mid-frequency and the whole system operates as a tuneable filter. Using this frequency-tuned resonant test system, transformers of up to 500 MVA and 400 kV have been successfully tested in Europe and South East Asia without having major problems in transportation. Figure 5.4-7: Noise below 15 pC at induced withstand voltage Figure 5.4-8: PD noise frequency spectrum and calibration level 5.4.5 Example 3: UHF PD Measurement and Acoustic PD Localization in Service and for On-Site Tests Partial discharges under oil emit acoustic and also electromagnetic waves with frequencies up to the UHF range (300 MHz to 3000 MHz). The electromagnetic waves are detectable with UHF probes. As a result of the shielding characteristics of the transformer tank against external electromagnetic waves, normally signals can clearly be distinguished between internal and external PDs. The UHF probes can be installed at an oil flange during on-site testing or even during operation. The probes withstand all mechanical and thermal stresses in the transformer and can be used as a permanently installed measuring device. A calibration comparable to the measurement according to the IEC60270 [1.2] is impossible for UHF sensors. Therefore it is necessary to examine the installed probes for its function with a performance check [3.62]. In Figure 5.4-9 an UHF signal measured at a 200 MVA single-phase transformer and its corresponding spectrum is shown. a) b) Figure 5.4-9: a) UHF PD signal on a 200 MVA single-phase transformer; b) corresponding UHF PD spectrum 0.00 0.05 0.10 0.15 0.20 -0.03 -0.02 -0.01 0.00 0.01 0.02 0.03 0.04 amplitude(V) time (us) online UHF signal 0 250 500 750 1000 1250 1500 0.0 0.1 0.2 0.3 0.4 0.5 0.6 spectrum of online UHF signal frequency (MHz) amplitude(Vs)
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 58 - Identification and filtering of on-site disturbances During on-site measurements sometimes UHF noise signals are detectable even though the sensor is inside the tank of the transformer. Processing of these measured signals can show known narrowband disturbances. Disturbances around 0,5 GHz are caused by the digital video broadcasting service, at nearly 0,9 GHz and 1,8 GHz are the mobile phone disturbances and at about 2,1 GHz UMTS signals are detectable. Because these disturbances have a small bandwidth, a narrowband filter can be applied easily. Narrow-band measurement systems with adaptive filtering are recommended. With narrow-band systems disturbances can be suppressed simply by measuring “next to them” in an unaffected frequency range. Another powerful tool for field measurements is phase resolved UHF PD measurement. Internal PDs produce phase stable UHF signals and phase-resolved PD patterns (PRPD) can identify interference and disturbances because of their non-phase-stable occurrence, especially if the test frequency is different from power frequency. Localization of PD by combining UHF and acoustic methods Two main tasks are encountered concerning PD measurements. The first is to provide evidence of PD e.g. a “yes/no” decision. The second and most important is the determination of the PD location (localization). The on-site PD detection sensitivity may be hampered in the conventional case whereas the non-conventional methods such as electromagnetic (UHF) and acoustic, do generally not suffer from external disturbances. Application of the two unconventional methods can provide an advantageous combination for optimised detection and localization. Other methods for the localization of the PD based on the acoustic signal arrival time include three different approaches for the system of non-linear observation equations. Depending on whether mixed-acoustic (e.g. electric or electromagnetic triggering) or all-acoustic (acoustic triggering) measurements are used, the equations have three (space coordinates (x, y, z) of the PD) or four unknowns (an additionally unknown temporal origin). A new approach to acoustic signal processing works with so called pseudo-times, allowing the usage of robust direct GPS (Global Positioning System) solvers instead of the previously used iterative algorithms. In the presence of inevitable measuring errors, sensitivity limits or wrongly assumed acoustic propagation velocities much more stable results were featured by the direct GPS solver. Another important part of the localization procedure is a correct objective arrival time determination. Here good experiences have been gathered with signal-energy based criteria [5.26]. Increasing the acoustic signal-to-noise ratio with an averaging process (continuous evaluation of the mean value) has long been known and used. To be successful a stable trigger is required, signal and noise should be uncorrelated and the noise should be white. In terms of PD measurements and the goal to de-noise acoustic signals (e.g. to quantify their arrival times hidden in the noise) one needs to have a physical signal related to the PD with a significantly higher sensitivity than the acoustic one. Here, electromagnetic UHF PD signals have proven their applicability. During the averaging process, the noise contained in the acoustic signal tends towards its statistic mean value, which is zero if white noise is assumed. The acoustic signal itself is superimposed constructively and the presence of an acoustic signal with a stable relation to the UHF trigger can be verified with high sensitivity. The theoretically maximum signal-to-noise-ratio gain is N x 0,5, where N is the number of superposition. For the example of Figure 5.4-10, within a single acoustic PD impulse of a 132 pC (electric equivalent PD), there is no clearly observable information above the noise level in the acoustic signal. Using the same set-up with UHF-triggered averaged acoustic signal with 500 superimpositions of PD impulses of maximums of only 9 pC, a clear acoustic impulse can be seen, which allows the determination of the travelling time of the impulse. So, a successive localization is possible [5.27]. Figure 5.4-10 shows this comparison of a single acoustic impulse of a 132 pC with no clearly observable information and 500 super-impositions of maximum 9 pC where a clear impulse is visible (same experimental arrangement, same sensor position).
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 59 - Figure 5.4-10: Comparison of a 132 pC single acoustic impulse and 500 superpositions of maximum 9 pC Application on Power Transformers: all-acoustic on-site/on-line PD localization in a 200 MVA, 380/220 kV-single-phase transformer Over a period of several months all-acoustic on-site measurements were performed on a 200 MVA single-phase transformer. Its gas-in-oil diagnosis indicated partial discharge. During an offline applied voltage test an electric PD measurement revealed PD magnitudes up to 600 pC and an autonomous acoustic measurement recorded an impulse on four sensors simultaneously. The oil-temperature was about 26°C (which corresponds to 1387 m/s sound velocity in oil) according to information from the operating company. Figure 5.4-11 pictures the top view of the housing of the transformer with the sensor positions and the average of the calculated PD location. Figure 5.4-11: Top view of the housing of the 200 MVA transformer with attached acoustic sensors and the calculated PD location inside the transformer In Table 5.4-1 results of location are summarized. The result of “measurement 1” was determined iteratively because the direct solver generated a complex solution due to strongly erroneous time information. For the remaining two measurements the direct solver was used.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 60 - Tab. 5.4-1: Acoustic PD measurements over a period of several months with changing sensor positions (Figure 5.4-11) calculated PD-origin x [m] y [m] z [m] measurement 1 (offline a, b, c, d) 1,40 3,12 2,27 measurement 2 (online c, d, e, f, g) 1,25 3,19 2,23 measurement 3 (online a, c, d, e, f, g) 1,27 3,22 2,19 5.4.6 Example 4: Synchronous Multi-Terminal PD Measurement for On-Site Testing The interpretation of phase resolved (PRPD) patterns from on-site transformer measurements might be difficult because of the external disturbances and the cross-talk of the transformer windings. Evaluation of synchronous multi-terminal PD measurements establishes a straight forward approach to remove external disturbances and, furthermore, to distinguish between multiple PD and noise sources [3.16, 3.17, 3.28]. The synchronous multi-terminal PD measurement is based on the standard measurement circuit of IEC 60270. For PD measurement on power transformers a three phase measuring system is used, in which all three phases are measured synchronously. Therefore one PD-impulse is measured at all three phases. The measuring result is illustrated in a STAR diagram [5.24, 3.67]. The STAR diagram is a two-dimensional plot with a 120° phase shift of the three phase axis. Figure 5.4-12 (right) shows the impulse signals on all phases. In this example the PD source is located on phase L1 and the PD signals of phase L2 and L3 occur because of the cross-talk of the windings. An addition of the signal amplitude vectors of a single PD activity (value in pC) of the three phases builds a point in the diagram. In this example the point is close to L1 and therefore the PD source is located in phase L1. Assuming that each individual PD fault leads to a unique cluster, each cluster represents one specific fault location within the transformer. External disturbances, like corona or noise, can be measured with the multi-terminal IEC circuit on all three phases. Measured external impulses often possess similarities of the shape and amplitude. Consequently the vector addition of those impulses leads to clusters located next to the point of origin. Therefore internal PD and external impulse sources can be differentiated from each other. Retransformations of clusters of the STAR diagram into a PRPD pattern are possible. Therefore, a cluster can be extracted from the STAR diagram and can be illustrated in the well-known PRPD pattern. The selected cluster and the retransformation of a single PD sources enables pattern recognition. The results of an onsite, offline measurement carried out on a 110/220 kV, 210 MVA grid-coupling transformer by means of a motor generator set showed PD activity at operating voltage level in the transformer [5.27]. The measurements were done on the 110 kV and the 220 kV side. For demonstrating purpose a copper wire was fastened to the conductor of the bushing on phase L3,110 (Figure 5.4-13) representing an external corona with a considerable low amplitude of 30 pC. Figure 5.4-14 shows the PRPD pattern of an internal PD on phase L2,110 of the transformer and some disturbances from the corona on phase 3. Because the PD level of the corona is quite small the coupling from phase L3,110 to L2,110 is not very significant, but can be depicted in the rectangle of Figure 5.4-14. For phase L1,110 and L3,110 internal PD sources could not be detected and no cross-coupling from the active phase L2,110 was visible in the PRPD pattern. The STAR diagram in Figure 5.4-15 confirms the results of the two PD sources. One cluster is close to phase L2,110 (rectangle) which is the internal measured PD and the external corona discharge on L3,110 (ellipse).
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 61 - Figure 5.4-12: STAR diagram evaluation of a PD impulse signal by means of vector addition of a three phase measurement Figure 5.4-13: Copper wire (here indicated in black) as artificial defect on phase L3,110 with bushing and connector between conductor and coupling capacitor Figure 5.4-14: PRPD-pattern on phase L2,110 disturbed by corona on phase L3,110 (marked rectangle) Figure 5.4-15: STAR diagram of the PD activity (rectangle showing internal PD and ellipse showing corona discharge at attached wire) Figure 5.4-16: PRPD-pattern on phase L2,110 after retransformation without corona disturbances The retransformation of all partial discharges on phase 2 (rectangular cluster) is shown in Figure 5.4-16. Now the pattern does not include the disturbances from the external corona (see rectangle in Figure 5.4-16). In the case of a stronger external PD this effect would be more pronounced. Thus PRPD patterns can be de-noised by means of the STAR diagram and retransformed into specific clusters. Furthermore the PRPD pattern of just one partial discharge can be generated out of overlapping PD patterns in order to simplify pattern recognition [5.24].
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 62 - 6. Conclusion Based on practical examples this brochure deals with the state of the art of high voltage (HV) on-site testing in combination with partial discharge (PD) measurements. The aim of this brochure is to give background information and helpful hints in performing such tests on various types of apparatus. In addition to type and routine testing in the factory, HV on-site testing is an important part of quality assurance. On-site tests are applied  as a part of commissioning of equipment on-site to demonstrate that the transport from the factory to site and erection on-site have not caused any new and critical defects which lower the dielectric withstand voltage of the insulation below the withstand level of the insulation coordination determined values;  after on-site repair to demonstrate that the equipment has been successfully repaired;  for diagnostic purposes in the framework of asset management by providing reference values of diagnostic indicators (e.g. partial discharge values and dielectric parameters) for later tests. Warnings from the PD monitoring system are a critical trigger in the condition-based maintenance program. HV withstand testing plays an important role in quality assurance of HV apparatus. However, additional measurements of the dielectric properties, such as significant PD quantities and the loss factor tan δ may enhance the reliability and remaining life time of electrical insulation. The common physical background of the measured phenomena is the behaviour of dielectrics in the electric field. Therefore the tests and measurements in the factory, on-site and in-service should be harmonized. This includes the possibility of improvement of on-site test procedures supported by partial discharge measurement. HV withstand voltage testing (e.g. up to 2,5 U0) of defect-free/non aged insulation does not have a destructive influence on the service life of the tested component. Due to life consumption of installed HV equipment, however, lower AC test voltages are recommended for on-site testing. AC voltage test levels higher than 1,0 U0 of defective/aged insulation may have a destructive influence on the service life of the component, even if no breakdown has occurred. Therefore a combination of AC voltage on- site testing with PD measurement is recommended. The PD measurement can provide an indication of whether the tested component is in comparable condition following the withstand test as it was before the test. In this context PD measurements play an important role in both quality testing and monitoring for condition assessment. In contrast to most dielectric measurements (tan , r, dielectric response parameters) which are characteristics of the whole dielectric volume, partial discharges are “weak point phenomena”. In the case of new insulation, the critical defects/weak points are the result of assembly failure which can be found by a routine test consisting of the HV withstand test combined with PD measurement. In case of service aged HV equipment insulation (which has been tested successfully and which operates for years) a critical defect may be caused by high electrical, thermal or mechanical stresses and by the “aging” of the insulation itself. This means that the partial discharges can be symptoms and/or the results of the aging process which causes – over a more or less long period – the above mentioned “weak point”. As there are several methods of on-site test voltage generation which can be applied to withstand testing and on-site testing, this brochure presents and discusses the various methods and the commonly used HV sources and PD measurement techniques for on-site testing of HV insulation. Typical methods applied to the on-site testing of HV apparatus, such as gas-insulated systems (GIS), cable systems, rotating machines and power transformers, are presented in this brochure. The benefits and the limitations of on-site HV testing combined with sensitive PD measurement for this apparatus is also discussed.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 63 - 7. References [1.1] W. Hauschild et al.: CIGRE WG 33.03/TF 04: Proposed requirements for HV withstand tests on-site. Electra No. 195, pp. 13–21, 2000 [1.2] IEC 60270: 2000: HV test techniques – Partial discharge measurements [1.3] M. Muhr et al.: Sensors and sensing for non-conventional PD detection. Report D1-102, CIGRE Paris, 2006 [1.4] IEC 60060-1: 1989-11: HV test techniques – Part 1: General definitions and test requirements [1.5] IEC 60060-2: 1994: HV test techniques – Part 2: Measuring Systems [1.6] IEC 60060-3: 2006-02: HV test techniques – Part 3: Definitions and requirements for on-site testing [1.7] IEC 60071-1: 1993-12: Insulation coordination. Part 1: Definitions, principles and rules [1.8] IEC 885-3: 1988: High-voltage test techniques – Test methods for partial discharge measurement of extruded power cables [2.1] E. Gulski, P., Cichecki, F.J. Wester, J.J, Smit, R. Bodega, T.J.W.H. Hermans, P.P. Seitz, B. Quak, F. de Vries: On-site testing and PD diagnosis of high voltage power cables. IEEE transactions on dielectrics and electrical insulation 15(6), 1691-1700, 2008, (ISSN 1070- 9878) [2.2] H. Michitsch, J. Spiegelberg, W. Golbig: HV tests of large rotating machines for commissioning and for diagnosis on the installation site (in German). ew (Elektrizitätswirtschaft) 104,17/18, pp. 62-70, 2005 [2.3] E. Gockenbach, W. Hauschild: The selection of the frequency range for HV on-site testing of extruded insulation cable systems. IEEE Electrical Insulation Magazine, Vol. 16, No. 6, pp. 11-16, 2000 [2.4] IEC 60840: 2004: Power cables with extended insulation for rated voltages 30 kV up to 150 kV – Test methods and requirements [2.5] IEC 62067: 2001: Power cables with extruded insulation for rated voltages 150 kV up to 500 kV – Test methods and requirements [2.6] IEC 62271: 2003: High-voltage switchgear and controlgear, Part 203: Gas-insulated metal- enclosed switchgear for rated voltages above 52 kV [2.7] W. Hauschild, P. Coors et al.: The technique of AC on-site testing of HV cables by frequency-tuned resonant test systems. CIGRE Conf., report 33-304, 2002 [2.8] W. Hauschild, S. Schierig; P. Coors: Resonant test systems for HV testing of super-long cables and gas-insulated transmission lines. 14 th ISH Beijing, Paper J-02 (p. 558), 2005 [2.9] Th. Aschwanden, A. Schenk et al.: On-site repair and refurbishment and HV tests of large power transformers in the transmission grid. CIGRE Report A2-203, 2004 [2.10] J. C. Mendes, R. A. Marcondes, J. Nakamura: On-site tests on HV power transformers. CIGRE Report A2-211, Paris 2004 [2.11] P. Werle: On-site test of power transformers. HIGHVOLT Kolloquium ’07, Paper 4.5, 2007 [2.12] A. Winter, P. Coors, U. Stephan: A mobile transformer test system based on a static frequency converter. 15 th ISH Ljubljana, Paper T10-T732, 2007
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 64 - [2.13] IEEE 400.3-2006: Guide for partial discharge testing of shielded power cable systems in a field environment [2.14] CIGRE WG 33.03: Experiences in partial discharge detection of distribution power cable systems. Electra, p. 34-43, 2003 [2.15] E. Gulski, F. J. Wester, J. J. Smit, P. N. Seitz, M. Turner: Advanced PD diagnosis of MV power cables using oscillating wave test system. IEEE El Magazine March/April, Vol. 16 No. 2, 2000 [2.16] E. Gulski, J. J. Smit, H. van Breen, F. de Vries, P. P. Seitz, F. Petzold: Partial discharges diagnosis of stator insulation using damped ac voltages. Proceedings International Conference on Condition Monitoring and Diagnosis 2-5 April 2006, Changwon, Korea, 2006 [2.17] IEEE 400-2001: IEEE Guide for field testing and evaluation of the insulation of shielded power cable systems [2.18] F.J. Wester: Condition assessment of power cables using PD diagnosis at damped AC voltages. ISBN 90-8559-019-1, PhD thesis TU Delft, 2004 [2.19] W. Boone, C. C. Damschild et al.: VLF HV generators for testing cables after laying. 5 th ISH Braunschweig, paper 62-04, 1987 [2.20] G. C. Damstra, W. Zhang: Series resonant inverter with triggered vacuum gaps. ISDEIV Conference Moscow, Vol. 2259, pp. 270-272, 1994 [2.21] H. Emanuel et al.: A new high voltage dielectric test system for insulation diagnosis and partial discharge measurement. NORD-IS, 1996 [2.22] C. Forssén, H. Edin: Modelling partial discharges in a cavity at different applied frequencies. CEIDP, 2007 [2.23] G. Voigt, P. Mohaupt: Partial discharge measurements on service aged medium voltage cables at different frequencies. JICABLE, 2003 [2.24] M. Kuschel, R. Plath, I. Stepputat, W. Kalkner: Diagnostic techniques for service- aged XLPE-insulated medium voltage cables. REE, 1996 [2.25] R. Plath, W. Kalkner, I. Krage: Comparison of diagnostic systems for service-aged PE/XLPE-insulated medium voltage cables. Elektrizitätswirtschaft, Heft 20, 1997 (in German) [2.26] K. Rethmeier, P. Mohaupt, S. Seifert, V. Bergmann, W. Kalkner: Sinusoidal 0.1 Hz VLF test voltage for diagnostic PD measurements of high voltage cable systems. 15 th ISH Ljubljana, T7-448 [3.1] E. Lemke, S. Berlijn, E. Gulski, M. Muhr, E. Pultrum, T. Strehl, W. Hauschild, J. Rickmann, G. Rizzi: Guide for partial discharge measurements in compliance to IEC 60270. CIGRE WG D1.33, Technical Brochure 366, ELECTRA No. 241, pp. 61-67, 2008 [3.2] E. Gulski et al.: Guidelines for un-conventional PD measurements. CIGRE WG D1.33 (TF 03), Technical Brochure 444, 2010 [3.3] IEC 64278: High voltage test techniques – Measurement of partial discharges by electromagnetic and acoustic methods (under preparation) [3.4] CIGRE WG 21-03: Significance of discharge detection. ELECTRA 11, pp. 53–60, 1969 [3.5] CIGRE WG 21-03: Recognition of discharges. ELECTRA 11, pp. 61–98, 1969
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 65 - [3.6] E. Gulski, D. Allan, T.R. Blackburn, A. Contin, E. Gockenbach, E. Lemke, T. Lundgaard, T. Mizutani, G.C. Montanari, M. Muhr, T. Okamoto, B.T. Phung, H. Sedding, H.J. van Breen, F.J. Wester: Knowledge rules for partial discharge diagnosis in service. CIGRE JWG-TF 15.11 / 33.03.02, Technical Brochure No. 226, 2003 [3.7] K. Schon: Konzept der Teilentladungsmessung bei Teilentladungsprüfungen. ETZ Archiv Vol. 8, No. 9, pp. 319-324, 1986 [3.8] E. Lemke, P. Schmiegel, H. Elze, D. Russwurm: Experience in the calibration technique for PD calibrators. ERA Conference on High Voltage Measurements and Calibration. Milan, Italy, 1966 [3.9] W. Lukas, K. Schon, E. Lemke, H. Elze: Comparison of two techniques for calibrating PD calibrators. 10 th ISH Montreal, 1997 [3.10] E. Lemke: Moderne TE-Mess-Systeme und ihre Kalibrierung. HIGHVOLT Kolloquium ’97 Dresden, paper 7.7, pp 95-106, 1997 [3.11] C. Cherbaucich, G. Rizzi, R. Gobbo, G. Pesavento, O. Saracco, A. Sardi: Partial discharge calibrators: Practical experience in their calibration. ERA Conference on High Voltage Measurements and Calibration, London, U.K., 1998 [3.12] E. Lemke, E. Gockenbach, W. Kalkner: Messtechnik für die Diagnose elektrischer Betriebsmittel. ETG-Fachtagung „Diagnostik elektrischer Betriebsmittel“, Berlin, 2002 [3.13] E. Gulski: Computer-aided recognition of Partial Discharges using statistical tools. Ph.D. Thesis, TU Delft, 1991 [3.14] A. Claudi et al: Instrumentation and measurement for in-service monitoring of HV insulation. CIGRE TF D1.02.08, Technical Brochure No. 286, 2005 [3.15] K.D. Plath, R. Plath, H. Emanuel, W. Kalkner: Synchrone dreiphasige Teilentladungsmessungen an Leistungstransformatoren vor Ort und im Labor. ETG-Fachtagung „Diagnostik elektrischer Betriebsmittel“, Berlin, paper 2.1, 2002 [3.16] R. Plath: Digital measurements and analysis of partial discharges – drawbacks and opportunities. HIGHVOLT Kolloquium ’07, Dresden, paper 2.1, pp. 63-71, 2007 [3.17] W. Koltunowicz, R. Plath: Synchronous multi-channel PD measurements. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 15, No.6, pp. 1715-1723, 2008 [3.18] E. Lemke, T. Strehl, M. Boltze: Advanced diagnostic tool for PD fault location in power cables using the CDA technology. 13 th International Symposium on High Voltage Engineering, Bangalore, India, paper 6-72, 2001 [3.19] E. Lemke, E. Gulski, W. Hauschild, R. Malewski, P. Mohaupt, M. Muhr, J. Rickmann, T. Strehl, F.J. Wester: Practical aspects of the detection of partial discharges in power cables. CIGRE WG D1.33.TF 05, Technical Brochure 297, ELECTRA No. 232, pp. 63-70, 2006 [3.20] M. Hoof and R. Patsch: Analyzing Partial Discharge Pulse Sequences – A New Approach to Investigate Degradation Phenomena. Conference Record of the IEEE International Symposium on Electrical Insulation, Pittsburgh, PA, USA, pp. 327-331, 1994 [3.21] L. Satish, B. Nazneen: Wavelet-based denoising of partial discharge signals buried in excessive noise and interference. IEEE Transactions on Electrical Insulation, Vol. 10, pp. 354-367, 2003 [3.22] A. Cavallini, A. Contin, G.C. Montanari, F. Puletti: Advanced PD interference in on-field measurements, part I: Noise rejection. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 10, No. 2, pp 216-224, 2003 and A. Cavallini, A. Contin, G.C. Montanari, F. Puletti: Advanced PD interference in on-field
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 66 - measurements, part II: Identification of defects in solid insulation systems. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 10, No. 3, pp 528-538, June 2003 [3.23] S. Happe, H.-G. Kranz, W. Krause: Advanced Suppression of Stochastic Pulse Shaped Partial Discharge Disturbances. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 12, No. 6, pp. 265-275, 2005 [3.24] N. Hayakawa, Y. Sugimori, H. Okubo: Partial discharge current pulse waveform analysis (PD-CPWA) based on PD mechanisms in solid and gas media. Proceedings of International Symposium on Electrical Insulating Materials, Kitakyushu, Japan, pp. 812-815, 2005 [3.25] S. Sriram, S. Nitin, K.M.U. Prabhu, M. J. Bastiaans: Signal denoising techniques for partial discharge measurements. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 12, No. 6, pp. 1182-1191, 2005 [3.26] J.-T. Kim, J.-H. Kim, J.-Y. Koo: Noise discriminations in measuring partial discharges using pulse wave shape analysis. CMD 2006, Changwon, Korea, 2006 [3.27] T.-K. Abdel-Galil, Ayman H. El-Hag, A.M. Gaouda, M.M.A. Salama, R. Bartnikas: De-noising of partial discharge signal using Eigen-decomposition technique. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 15, No.6, pp. 1657-1662, 2008 [3.28] K. Rethmeier, M. Krüger, A. Kraetge, R. Plath, W. Koltunowicz, A. Obralic, W. Kalkner: Experiences in on-site partial discharge measurements and prospects for PD monitoring. CMD Beijing, 2008 [3.29] S.A. Boggs, G.L. Ford, R. C. Madge: Coupling devices for the detection of partial discharges in gas-insulated switchgear. IEEE Transactions on Power Apparatus and Systems, Vol. 10, pp. 3969-3973, 1981 [3.30] S.A. Boggs, G.C. Stone: Fundamental limitations in the measurement of corona and partial discharges. IEEE Transactions on Electrical Insulation, Vol. 17 pp. 143-150, 1982 [3.31] Y. Takahashi: Diagnostic methods for gas-insulated substations. IEEE Transactions on Electrical Insulation, Vol. 21, No. 6, pp. 1037-1043, 1986 [3.32] E. Lemke, H. Elze, W. Weissenberg: Experience in PD diagnostics of HV cable terminations in service using an ultra-wideband PD probing in the real-time mode. 13 th International Symposium on High Voltage Engineering, Delft, The Netherlands, paper 404, 2003 [3.33] E. Lemke, T. Strehl, W. Weissenberg: Practical experience in on-site PD diagnosis tests of HV cable accessories in service. IEEE International Symposium on Electrical Insulation, Toronto, Canada, pp. 498-501, 2006 [3.34] B. Krampe, C. Neumann: Betriebsverhalten mit der UHF-TE-Messung an gasisolierten metallgekapselten Schaltanlagen. ETG Fachbericht 87, VDE Verlag, Berlin und Offenbach, 2002 [3.35] R. Kurrer, K. Feser: TE-Messung an SF6-Anlagen im UHF-Bereich. ETG Fachbericht 56, VDE Verlag, Berlin und Offenbach, 1995 [3.36] O. Rudolph, A. Breuer, P. Erning: Vorort-Diagnose von gasisolierten Schaltanlagen: Teilentladungsmessung. IEH-Prüftechnik, Symposium, Karlsruhe, 2000 [3.37] U. Riechert A. Đurđić, M. Kudoke, M. Stanek: Monitoring und Diagnose von gasisolierten Schaltanlagen - Entwicklungstrends und Einsatzmöglichkeiten. ETG Fachtagung „Diagnostik elektrischer Betriebsmittel“, Kassel, 2006, ETG-Fachbericht 104, 2006, S. 161- 166, VDE Verlag GmbH, Berlin und Offenbach, ISBN 978-3-8007-2977-7 [3.38] S. Neuhold, Th. Heizmann, R. Bräunlich; D. Koechli, U. Riechert, Ch. Dehne: Experiences with UHF PD detection in GIS using external capacitive sensors on windows and disk-
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 67 - insulators. 15th International Symposium on High Voltage Engineering, Ljubljana, Slovenia, paper T7-480, 2007 [3.39] S. Hoek, M. Bornowski, S. Tenbohlen, T. Strehl, U. Riechert: Partial discharge detection and localization in gas-insulated switchgears. Stuttgarter Hochspannungssymposium, Conference Record, pp. 211-218, 2008, ISBN 978-3-00-024098-0 [3.40] W.R. Rutgers, Y.H. Fu: UHF PD-detection in a power transformer. 10th ISH, Montreal, Canada, 1997 [3.41] M.D. Judd, G.P. Cleary, S. Meijer: Testing UHF partial discharge detection on a laboratory based power transformer. Conf. Proc. 13th ISH, Delft, Netherlands, 2003 [3.42] S. M. Markalous, S. Tenbohlen, K. Feser: Improvement of acoustic detection and localization accuracy by sensitive electromagnetic PD measurements under oil in the UHF range. 14th ISH, Beijing, China, 2005 [3.43] S. M. Markalous: Detection and location of partial discharges in power transformers using acoustic and electromagnetic signals. Ph.D. Thesis, University of Stuttgart, Germany, 2006 [3.44] S. Tenbohlen, D. Denissov, S.M. Hoek, S.M. Markalous: Partial discharge measurement in the ultra high frequency (UHF) range. IEEE Transactions on Dielectrics and Electrical Insulation (ITDEIS), Vol. 15, No.6, pp. 1544-1552, 2008 [3.45] S. Markalous, S. Tenbohlen, K. Feser: Detection and location of partial discharges in power transformers using acoustic and electromagnetic signals. IEEE Transactions on Dielectrics and Electrical Insulation (ITDEIS), Vol. 15, No.6, pp. 1576-1583, 2008 [3.46] J. Lopez-Roldan, T. Tang, M. Gaskin: Optimisation of a sensor for onsite detection of partial discharges in power transformers by the UHF method. IEEE Transactions on Dielectrics and Electrical Insulation (ITDEIS), Vol. 15, No.6, pp. 1634-1639, 2008 [3.47] G.C. Stone: Partial discharge signal measurements to assess rotating machine insulation condition: A survey. International Symposium on Electrical Insulation, Montreal, 1996 [3.48] E. Binder, A. Draxler, H. Egger, A. Hummer, H.R. Fuchs, H. Koklek, F. Müller, M. Drpic, M. Hoof, R. Käfer, S. Lanz: Entwicklungen und Nachweisuntersuchungen von Diagnosemethoden für Wasserkraftgeneratoren. E&I Vol. 117, No. 12, 2000 [3.49] T. Strehl, E. Lemke, H. Elze: On-line PD measurement, diagnostic tools and monitoring strategy for generators and power transformers. 12 th ISH, Bangalore, India, 2001 [3.50] M. Hoof: TE-On-Line-Monitoring an rotierenden Maschinen, Moderne Verfahren, Möglichkeiten und Grenzen der Zustandsbewertung. HIGHVOLT Kolloquium, pp. 225-232, 2003 [3.51] C. Hudon, M. Belec, A. Contin, A. Cavallini, D.N. Nguyen, G.C. Montanari, M.Conti: Evolution in automatic phase resolved partial discharge pattern recognition for machine diagnosis. 13th ISH Delft, paper O.14.06, 2003 [3.52] D. Pommerenke, I. Krage, W. Kalkner, E. Lemke, P. Schmiegel: On-site PD measurement on high voltage cable accessories using integrated sensors. 9th ISH Graz, Austria, paper 56.08, 1995 [3.53] D. Denissov, R. Grund, T. Klein, W. Köhler, S. Tenbohlen: UHF partial discharge diagnosis of plug-in cable terminations. 7th International Conference on Insulated Power Cables, Jicable’07, Paris, France, paper 0133, 2007 [3.54] D. Denissov, W. Köhler, S. Tenbohlen, R. Grund, T. Klein: Simulation of UHF PD wave propagation in plug-in cable terminations. IEEE International Symposium on Electrical Insulation (ISEI), Vancouver, Canada, 2008, pp. 547-550, 2008
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 68 - [3.55] T. Klein, D. Denissow, W. Köhler, S. Tenbohlen: Application of the UHF method for on-line PD diagnosis of cable terminations. CIGRE Session, Paris, France, paper B1-203, 2008. [3.56] H.D. Schlemper, R. Kurrer, K. Feser: Sensitivity for on-site partial discharge detection in GIS. 8th International Symposium on High Voltage Engineering, Yokohama, Japan, Vol. 3, pp. 157-160, 1993 [3.57] CIGRE JWG 33/23.12: Insulation co-ordination of GIS: Return of experience, on-site tests and diagnostic techniques. ELECTRA, Nr. 176, 1998 [3.58] CIGRE TF 15/33.03.05: PD detection system for GIS: Sensitivity verification for the UHF method and the acoustic method. ELECTRA No. 183, 1999 [3.59] R. Feger: Sensoren und Verfahren der UHF Teilentladungs-Messtechnik an gasisolierten Schaltanlagen. Ph.D., D 93, Universität Stuttgart, 2001 [3.60] S. Meijer, P.D. Agoris, J.J. Smit: UHF PD sensitivity check on power transformers. 14 th International Symposium on High Voltage Engineering, Beijing, China, 2005 [3.61] E. Lemke, T. Strehl, S. Markalous, W. Weissenberg: Ultra-wide-band PD diagnostics of power cable terminations in service. IEEE Transactions on Dielectrics and Electrical Insulation (ITDEIS), Vol. 15, No.6, pp. 1570-1575, 2008 [3.62] S. Coenen, S. Tenbohlen, S. Markalous, T. Strehl: Sensitivity of UHF PD measurements in power transformers. IEEE Transactions on Dielectrics and Electrical Insulation (ITDEIS), Vol. 15, No.6, pp. 1553-1558, 2008 [3.63] E. Howells and E.T. Norton: Detection of PDs in transformes using acoustic emission techniques. IEEE Trans, Vol. PAS-97, No. 5, pp. 1538-1549, 1978 [3.64] CIGRE WG 21-05: Partial discharge detection by means of acoustic detection. 1993 [3.65] L. E. Lungard, M. Runde, B. Skyberg: Acoustic Diagnosis of gas insulated substations; a theoretical and experimental basis. IEEE PD, Vo.5, No. 4, pp. 1751-1759, 1990 [3.66] A. -D. Schlemper, K. Feser, H. Blaum, P. Hirchesch: Sensitivity of acoustic PD detection in GIS; Laboratory experiments and on-site experience. Conf. Rec. of IEEE Symp. On El. Ins., pp. 99-102, 1996 [3.67] A. Pfeffer, S. Coenen, S. Tenbohlen, T. Strehl, S. Markalous: Onsite Experiences with multi- terminal IEC PD measurements and UHF PD measurement. 16 th ISH, Cape Town, 2009 [3.68] A. Cavallini, G. C. Montanari, G. Pasini, A. Salsi: Improving PD source localization techniques in cable systems. International Symposium on Electrical Insulating Materials, 2008 (ISEIM 2008), pp. 175-178, Tokyo, Japan, Sept. 2008 [3.69] A. Cavallini, G. C. Montanari and F. Puletti: A novel method to locate PD in polymeric cable systems based on amplitude-frequency (AF) map. IEEE Trans. on Dielec. Elec. Insul., Vol. 14, No. pp. 726-734, 2007 [5.1] M.L. Philip, Y.S Goh, W.H. Leong, M. Darus A. Sethuraman, S.A. Fuad, W. Koltunowicz: Partial Discharge Test of GIS Installation in Malaysia: Techniques and Experiences. CIGRE Conf., report 15-304, 2000 [5.2] G.J. Behrmann, S. Neuhold, R. Pietsch: Results of UHF measurements in a 220 kV Substation during on-site commissioning tests. Conf.Proc. 10th ISH, Montreal, Canada, 1997 [5.3] E. Pultrum, R.J.B. Gruntjes, S. A. M. Verhoeven: On-site partial discharge measurement and commissioning test: the experience. CIGRE 2006, paper B1-108.
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 69 - [5.4] F. Petzold, M. Beigert, E. Gulski: Experience with Off-line PD diagnosis on MV cables knowledge rules for asset decisions. Proceedings Jicable '07. 7th International Conference on Insulated Power Cables. France, Versailles, June 2007, paper C7.2.5 [5.5] E. Gulski, J.J. Smit, P. Cichecki, P.P. Seitz, B. Quak, F. de Vries, F. Petzold: Insulation diagnosis of HV power cables. Proceedings Jicable '07. 7th International Conference on Insulated Power Cables, France, Versailles, June 2007, paper C7.1.2 [5.6] P.P. Seitz, B. Quak, E. Gulski, J.J. Smit, P. Cichecki, F. de Vries, F. Petzold: Novel method for on-site testing and diagnosis of transmission cables up to 250kV. Proceedings Jicable '07. 7th International Conference on Insulated Power Cables, France, Versailles, June 2007, paper C7.2.10 [5.7] C.G.A. Koreman et al.: Development of a new 380 kV double circuit XLPE insulated cable system in The Netherlands. Cigré 2006, B1-107 [5.8] IEC 62067: 2006-03: Power cables with extruded insulation and their accessories for rated voltages above 150 kV (Um = 170 kV) up to 500 kV (Um = 550 kV) - Test methods and requirements [5.9] S. Meijer, et. al.: Advanced partial discharge measuring system for simultaneous testing of cable Accessories. CMD conference proceedings, Changwon, Korea, 2006. [5.10] S. Meijer, et. al.: Simultaneous condition assessment of accessories of power cables using a wireless VHF/UHF PD detection system. CMD conference proceedings, Changwon, Korea, 2006 [5.11] Elektrizitätswirtschaft EW JG 106, Heft 4, page 20-26, 2007 (in German) [5.12] Elektrizitätswirtschaft EW JG 106, Heft 24, page 42-53, 2007 (in German) [5.13] D. Denissov, R. Grund, T. Klein, W. Köhler, S. Tenbohlen: UHF partial discharge diagnosis of plug-in cable terminations. 7 th International Conference on Insulated Power Cables, Jicable’07, June 24-28, 2007, Paris, France, paper No. 0133 [5.14] IEEE 1434-2000: Trial-Use Guide to the Measurement of Partial Discharges in Rotating Machinery [5.15] IEC 60034-27: 2006-12: Rotating electrical machines - Part 27: Off-line partial discharge measurements on the stator winding insulation of rotating electrical machines [5.16] H.J. van Breen: Condition assessment of stator insulation using partial discharge diagnostics. ISBN 90-9020128-9, PhD Thesis TU Delft, The Netherlands, 2005 [5.17] IEC 60076-3: 2000: Power transformers - Part 3: Insulation levels, dielectric tests and external clearances in air [5.18] T. Leibfried, U. Thieß: Electrical testing and diagnostic measurements in the life cycle of power transformers. HIGHVOLT-Kolloquium ‘07, pp. 115-121, 2007 [5.19] P. Werle et al.: Repair and HV testing of power transformers on site. (in German) ETG-Fachtagung “Diagnostik Elektrischer Betriebsmittel” Kassel, 2006 [5.20] W. Hauschild, A. Thiede, T. Leibfried, F. Martin: Static frequency converters for HV testing of power transformers (in German). Stuttgarter Hochspannungssymposium, 2006 [5.21] A. Winter, P. Coors, U. Stephan: A mobile transformer test system based on static frequency converters. HIGHVOLT-Kolloquium ‘07, pp. 137-142, 2007 [5.22] T. Aschwanden, A. Schenk et al.: On-site repair and refurbishment and HV tests of large power transformers in the transmission grid. CIGRE Session Paris, paper A2-203, 2004
High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 70 - [5.23] J. C. Mendes, R. A. Marcondes, J. Westberg: On-site repair of power transformers. CIGRE Session Paris, paper 12-114, 2002 [5.24] W. Koltunowicz, R. Plath: Synchronous Multi-channel PD Measurements. IEEE Transactions on Dielectrics and Electrical Insulation (IEEE TDEI), Vol. 15, No.6, pp. 1715- 1723, 2008 [5.25] P. Werle: On-site test of power transformers. Omicron Symposium. Austria 2006 [5.26] S. Markalous, S. Tenbohlen, K. Feser: New robust non-iterative algorithms for acoustic PD- localization in oil/paper-insulated transformers. 14 th ISH Symposium Beijing, paper No. G- 039, 2005 [5.27] S. Tenbohlen, S. Hoek, D. Denissov, R. Huber, U. Riechert, S. Markalous, T. Strehl, T. Klein: Electromagnetic (UHF) PD diagnosis of GIS, cable accessories and oil-paper insulated power transformers for improved PD detection and localization. CIGRE Session Paris, paper No. D1-104, 2006

More Related Content

PDF
Guidelines for Unconventional Partial Discharge Measurement (CIGRE 444)
AHMED MOHAMED HEGAB
 
PPTX
Ppt on protection of power transformers
siddharam kantoli
 
PPTX
Top 10 Interview Questions and Answers
Learn By Watch
 
PDF
IEEE Guide for Field Testing of Shielded Power Cable (IEEE 400.2 2013)
AHMED MOHAMED HEGAB
 
PDF
Cable fault location methods (Megger)
AHMED MOHAMED HEGAB
 
PPTX
Transformer maintenance
Molla Morshad
 
PPTX
Partial discharge testing
Debojyoti Mukherjee
 
PDF
TRANSFORMER VECTOR GROUP.pdf
Amala Joseph
 
Guidelines for Unconventional Partial Discharge Measurement (CIGRE 444)
AHMED MOHAMED HEGAB
 
Ppt on protection of power transformers
siddharam kantoli
 
Top 10 Interview Questions and Answers
Learn By Watch
 
IEEE Guide for Field Testing of Shielded Power Cable (IEEE 400.2 2013)
AHMED MOHAMED HEGAB
 
Cable fault location methods (Megger)
AHMED MOHAMED HEGAB
 
Transformer maintenance
Molla Morshad
 
Partial discharge testing
Debojyoti Mukherjee
 
TRANSFORMER VECTOR GROUP.pdf
Amala Joseph
 

What's hot (20)

PDF
Principles of Cable Sizing
Leonardo ENERGY
 
PPTX
Earthing in electrical network
BHAGWAN PRASAD
 
PDF
NEETRAC (Chapter 8: Partial Discharge for HV and EHV Cable Systems)
AHMED MOHAMED HEGAB
 
PDF
Testing of Transformers.
Geetesh Verma
 
PPTX
Tap Changer
Pankaj Singh
 
PDF
Sheath Fault Location On Power Cables - Sheath Testing, Sheath Fault Location...
Thorne & Derrick International
 
PPT
Installation, Testing and Troubleshooting of Transformers
Living Online
 
PDF
Principles of Cable Selection
Leonardo ENERGY
 
PPTX
Current Transformer.pptx
MadhusmitaMohanty29
 
PPTX
presentation on substation layout and BUS bar arrangement.
Hemendra Kumar Rajput
 
PDF
Power transformer testing and commissioning guidelines r
Ashish Patel
 
PDF
Design of substation (with Transformer Design)
SayanSarkar55
 
PPTX
Switchgear and protection 1
Md Irshad Ahmad
 
PPTX
Cable Sizing Calculations
Siddhartha mudgal
 
PPT
Instrument transformer CT & PT
Chandan Singh
 
PPTX
Transformer Designing
Praveen Sharma
 
PPTX
SUBSTATION OVERVIEW - A presentation on substation
raiamitesh8
 
PDF
Substation Maintenance
karikalan123
 
PDF
Bill of Material of 132/33 KV 15 MVA Pooling Substation (15-07-2019)
Gensol Engineering Limited
 
PPTX
High voltage testing of electrical transformer
PraveenKumar5880
 
Principles of Cable Sizing
Leonardo ENERGY
 
Earthing in electrical network
BHAGWAN PRASAD
 
NEETRAC (Chapter 8: Partial Discharge for HV and EHV Cable Systems)
AHMED MOHAMED HEGAB
 
Testing of Transformers.
Geetesh Verma
 
Tap Changer
Pankaj Singh
 
Sheath Fault Location On Power Cables - Sheath Testing, Sheath Fault Location...
Thorne & Derrick International
 
Installation, Testing and Troubleshooting of Transformers
Living Online
 
Principles of Cable Selection
Leonardo ENERGY
 
Current Transformer.pptx
MadhusmitaMohanty29
 
presentation on substation layout and BUS bar arrangement.
Hemendra Kumar Rajput
 
Power transformer testing and commissioning guidelines r
Ashish Patel
 
Design of substation (with Transformer Design)
SayanSarkar55
 
Switchgear and protection 1
Md Irshad Ahmad
 
Cable Sizing Calculations
Siddhartha mudgal
 
Instrument transformer CT & PT
Chandan Singh
 
Transformer Designing
Praveen Sharma
 
SUBSTATION OVERVIEW - A presentation on substation
raiamitesh8
 
Substation Maintenance
karikalan123
 
Bill of Material of 132/33 KV 15 MVA Pooling Substation (15-07-2019)
Gensol Engineering Limited
 
High voltage testing of electrical transformer
PraveenKumar5880
 
Ad

Similar to High Voltage On-Site Testing with Partial Discharge Measurement (Cigre 502) (20)

PDF
SPE 171517_Estimating Probability of Failure _2014_Final
Katrina Carter-Journet
 
PDF
MIL-STD-810H.pdf
ssusera393ca
 
PDF
Making a Viable Electron Device into a Reliable Product: Brief Review
Crimsonpublishers-Electronics
 
PPTX
Reliability prediction of electronic components
PRANAY GUPTA
 
PDF
Early Life Failure Rate Calculation Procedure for Semiconductor Components
Hilaire (Ananda) Perera P.Eng.
 
PDF
6 structural analysis of pipeline spans oti 93 613
Dither Gutiérrez
 
PPTX
A Statistical Approach to Optimize Parameters for Electrodeposition of Indium...
Arkansas State University
 
PDF
Norma certificadora militar MIL-STD-810H.pdf
Revista Sociedade Militar
 
DOC
Ut procedure general
naiduvs
 
PDF
iMovR ZipDesk ANSI/BIFMA X5.5 Test Report
iMovR
 
PPTX
Sensors Field monitoring application in engineering
SomilYadav5
 
PPTX
The Role of Process Characterization in Process Validation
David Goodrich
 
PDF
Endurance Testing of Aircraft Electro-Hydraulic Actuator Using LabVIEW
Waqas Tariq
 
DOCX
Engineering Science WorksheetsTENSILE TESTER KIT ES6TE.docx
SALU18
 
PPTX
QUALITY ASSURANCE TESTS OF REMOTE AFTERLOADING BRACHYTHERAPY UNIT 1.pptx
TaushifulHoque
 
PDF
ZMPCZM016000.13.03 Certificate of compliance
painezeeman
 
PDF
DNV GL Scorecard_final version
SANJAY KUMAR
 
PDF
Methods to determine the overall health and condition of large power transfor...
Alexander Decker
 
PDF
11.methods to determine the overall health and condition of large power trans...
Alexander Decker
 
PDF
TENSILE TEST REPORT
musadoto
 
SPE 171517_Estimating Probability of Failure _2014_Final
Katrina Carter-Journet
 
MIL-STD-810H.pdf
ssusera393ca
 
Making a Viable Electron Device into a Reliable Product: Brief Review
Crimsonpublishers-Electronics
 
Reliability prediction of electronic components
PRANAY GUPTA
 
Early Life Failure Rate Calculation Procedure for Semiconductor Components
Hilaire (Ananda) Perera P.Eng.
 
6 structural analysis of pipeline spans oti 93 613
Dither Gutiérrez
 
A Statistical Approach to Optimize Parameters for Electrodeposition of Indium...
Arkansas State University
 
Norma certificadora militar MIL-STD-810H.pdf
Revista Sociedade Militar
 
Ut procedure general
naiduvs
 
iMovR ZipDesk ANSI/BIFMA X5.5 Test Report
iMovR
 
Sensors Field monitoring application in engineering
SomilYadav5
 
The Role of Process Characterization in Process Validation
David Goodrich
 
Endurance Testing of Aircraft Electro-Hydraulic Actuator Using LabVIEW
Waqas Tariq
 
Engineering Science WorksheetsTENSILE TESTER KIT ES6TE.docx
SALU18
 
QUALITY ASSURANCE TESTS OF REMOTE AFTERLOADING BRACHYTHERAPY UNIT 1.pptx
TaushifulHoque
 
ZMPCZM016000.13.03 Certificate of compliance
painezeeman
 
DNV GL Scorecard_final version
SANJAY KUMAR
 
Methods to determine the overall health and condition of large power transfor...
Alexander Decker
 
11.methods to determine the overall health and condition of large power trans...
Alexander Decker
 
TENSILE TEST REPORT
musadoto
 
Ad

More from AHMED MOHAMED HEGAB (20)

PDF
Testing of Transformer (AREVA)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 14: Glossary)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 13: Benefits of Diagnostics)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 12: Other Diagnostic Techniques)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 11: Metallic Shield Assessment)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 10: Monitored Withstand Techniques)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 9: Simple Dielectric Withstand)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 7: Medium Voltage Cable System Partial Discharge) )
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 6: Dissipation Factor)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 5: Time Domain Reflectomertry)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 4: How to Start)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 3: HV and EHV Cable System Aging and Testing Issues)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 2: MV Cable System Issues)
AHMED MOHAMED HEGAB
 
PDF
NEETRAC (Chapter 1: introduction)
AHMED MOHAMED HEGAB
 
PDF
IEEE Guide for the Application and Interpretation of FRA for oil Transformer
AHMED MOHAMED HEGAB
 
PDF
Tan Delta on MV Cables (Megger)
AHMED MOHAMED HEGAB
 
PDF
Guide for Transformer Maintenance (Cigre-445)
AHMED MOHAMED HEGAB
 
PDF
Introduction to power quality (Megger)
AHMED MOHAMED HEGAB
 
PDF
Electrical Classic Control (Basics)
AHMED MOHAMED HEGAB
 
PDF
Reading of electrical drawing
AHMED MOHAMED HEGAB
 
Testing of Transformer (AREVA)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 14: Glossary)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 13: Benefits of Diagnostics)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 12: Other Diagnostic Techniques)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 11: Metallic Shield Assessment)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 10: Monitored Withstand Techniques)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 9: Simple Dielectric Withstand)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 7: Medium Voltage Cable System Partial Discharge) )
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 6: Dissipation Factor)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 5: Time Domain Reflectomertry)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 4: How to Start)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 3: HV and EHV Cable System Aging and Testing Issues)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 2: MV Cable System Issues)
AHMED MOHAMED HEGAB
 
NEETRAC (Chapter 1: introduction)
AHMED MOHAMED HEGAB
 
IEEE Guide for the Application and Interpretation of FRA for oil Transformer
AHMED MOHAMED HEGAB
 
Tan Delta on MV Cables (Megger)
AHMED MOHAMED HEGAB
 
Guide for Transformer Maintenance (Cigre-445)
AHMED MOHAMED HEGAB
 
Introduction to power quality (Megger)
AHMED MOHAMED HEGAB
 
Electrical Classic Control (Basics)
AHMED MOHAMED HEGAB
 
Reading of electrical drawing
AHMED MOHAMED HEGAB
 

Recently uploaded (20)

PDF
III.4.1.2_The_Space_Environment.p pdffdf
sittiporn2
 
PPTX
Current and future trends in Computer Vision.pptx
Subramanyam Natarajan
 
PDF
The CXO Playbook 2025 – Future-Ready Strategies for C-Suite Leaders Cerebrai...
Cerebraix Technologies
 
PDF
PREDICTION OF DIABETES FROM ELECTRONIC HEALTH RECORDS
ijaia
 
PDF
Enhancing Cyber Defense Against Zero-Day Attacks using Ensemble Neural Networks
IJCNCJournal
 
PPTX
MET 305 2019 SCHEME MODULE 2 COMPLETE.pptx
VinayB68
 
PDF
composite construction of structures.pdf
AinieButt1
 
PPTX
additive manufacturing of ss316l using mig welding
premcdr7799
 
PPTX
Safety Seminar civil to be ensured for safe working.
DILJEETKUMAR8
 
PDF
PPT on Performance Review to get promotions
prashant593122
 
PPTX
Foundation to blockchain - A guide to Blockchain Tech
Davies Chacko
 
PDF
keyrequirementskkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
mojeeburahmanwardak
 
PPTX
FINAL REVIEW FOR COPD DIANOSIS FOR PULMONARY DISEASE.pptx
rubeshbme
 
PDF
Embodied AI: Ushering in the Next Era of Intelligent Systems
Yu Huang
 
PPT
Mechanical Engineering MATERIALS Selection
arsalanahmad705384
 
PPTX
Fundamentals of safety and accident prevention -final (1).pptx
Palani kumar
 
PDF
737-MAX_SRG.pdf student reference guides
ssusere2119a1
 
PDF
TFEC-4-2020-Design-Guide-for-Timber-Roof-Trusses.pdf
AinieButt1
 
PPTX
UNIT-1 - COAL BASED THERMAL POWER PLANTS
Chandra Kumar S
 
PDF
BMEC211 - INTRODUCTION TO MECHATRONICS-1.pdf
ryankakungu
 
III.4.1.2_The_Space_Environment.p pdffdf
sittiporn2
 
Current and future trends in Computer Vision.pptx
Subramanyam Natarajan
 
The CXO Playbook 2025 – Future-Ready Strategies for C-Suite Leaders Cerebrai...
Cerebraix Technologies
 
PREDICTION OF DIABETES FROM ELECTRONIC HEALTH RECORDS
ijaia
 
Enhancing Cyber Defense Against Zero-Day Attacks using Ensemble Neural Networks
IJCNCJournal
 
MET 305 2019 SCHEME MODULE 2 COMPLETE.pptx
VinayB68
 
composite construction of structures.pdf
AinieButt1
 
additive manufacturing of ss316l using mig welding
premcdr7799
 
Safety Seminar civil to be ensured for safe working.
DILJEETKUMAR8
 
PPT on Performance Review to get promotions
prashant593122
 
Foundation to blockchain - A guide to Blockchain Tech
Davies Chacko
 
keyrequirementskkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
mojeeburahmanwardak
 
FINAL REVIEW FOR COPD DIANOSIS FOR PULMONARY DISEASE.pptx
rubeshbme
 
Embodied AI: Ushering in the Next Era of Intelligent Systems
Yu Huang
 
Mechanical Engineering MATERIALS Selection
arsalanahmad705384
 
Fundamentals of safety and accident prevention -final (1).pptx
Palani kumar
 
737-MAX_SRG.pdf student reference guides
ssusere2119a1
 
TFEC-4-2020-Design-Guide-for-Timber-Roof-Trusses.pdf
AinieButt1
 
UNIT-1 - COAL BASED THERMAL POWER PLANTS
Chandra Kumar S
 
BMEC211 - INTRODUCTION TO MECHATRONICS-1.pdf
ryankakungu
 

High Voltage On-Site Testing with Partial Discharge Measurement (Cigre 502)

  • 1. 502 High-Voltage On-Site Testing with Partial Discharge Measurement Working Group D1.33 June 2012
  • 2. High-Voltage On-Site Testing with Partial Discharge Measurement Working Group D1.33, Task Force 05  Members R. Pietsch (Convenor: 2007-2010) (DE), W. Hauschild (Convenor: 2004-2007) (DE), J. Blackett (AU), R. Bodega (NL), A. Claudi (DE), B. Finlay (AU), M. Gamlin (CH), E. Gockenbach (DE), E. Gulski (NL), R.A. Jongen (NL), T. Leibfried (DE), E. Lemke (DE), S. Meijer (NL), P. Mohaupt (AT), Michael Muhr (AT), F. Petzold (DE), E. Pultrum (NL), G. Rizzi (I), T. Strehl (DE), C. Sumereder (AT), S. Tenbohlen (DE), P. Werle (DE) Copyright © 2012 “Ownership of a CIGRE publication, whether in paper form or on electronic support only infers right of use for personal purposes. Are prohibited, except if explicitly agreed by CIGRE, total or partial reproduction of the publication for use other than personal and transfer to a third party; hence circulation on any intranet or other company network is forbidden. As an exception, CIGRE Collective Members only are allowed to reproduce the publication. ”. Disclaimer notice “CIGRE gives no warranty or assurance about the contents of this publication, nor does it accept any responsibility, as to the accuracy or exhaustiveness of the information. All implied warranties and conditions are excluded to the maximum extent permitted by law”. ISBN : 978-2-85873-194-7
  • 3. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 3 - High-Voltage On-Site Testing with Partial Discharge Measurement Content 1.  Introduction ................................................................................................................. 5  1.1  Purpose of HV Dielectric On‐Site Tests with PD Measurement .................................. 5  1.2  HV Test Procedures with PD Measurement ............................................................... 6    2.  High Voltage Sources and Accessories for On‐Site Applications .................................... 8  2.1  General Requirements .............................................................................................. 9  2.2  Continuous Alternating Voltage (AC) ...................................................................... 10  2.3  Damped Alternating Voltage (DAC) ........................................................................ 13  2.4  Very‐Low Frequency (VLF) ...................................................................................... 15  2.5  HV Filter, Coupling Capacitor, Connections and Grounding for Conventional PD              Measurement ......................................................................................................... 16    3.  On‐site PD Measurements ......................................................................................... 17  3.1  Conventional Electric PD Measurements ................................................................ 17  3.2  Non‐conventional Electromagnetic PD Detection ................................................... 19  3.3  Noise Reduction ...................................................................................................... 22  3.4  Acoustic PD Detection ............................................................................................ 24  3.5  Important Aspects for PD Evaluation ...................................................................... 25  3.5.1  PD Identification ................................................................................................. 25  3.5.2  PD Localization ................................................................................................... 27    4  Preconditions for On‐Site Testing Including PD Measurement .................................... 28    5  Examples of Test and Measuring Techniques for Apparatus and Systems .................. 29  5.1  Gas‐Insulated Systems (GIS/GIL) ............................................................................. 29  5.1.1  HV Source, PD Measurement and Details of Test Object ..................................... 29  5.1.2  Example 1: On‐Site Test by ACRL (Tuneable Reactor Circuit) Test System and                    Acoustic and UHF PD Measurement Technique ................................................... 32  5.1.3  Example 2: On‐Site Test by ACRF Test System (Tuneable Frequency Circuit)and                  High Frequency PD Measurement Technique ....................................................... 33    5.2  Cable Systems ......................................................................................................... 34  5.2.1  HV Sources and PD measurement ....................................................................... 34 
  • 4. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 4 - 5.2.2  Required Details of Test Object ........................................................................... 37  5.2.3  Example 1: HV Cable Systems Tested by ACRF Test System and                   Conventional PD Measurement Technique .......................................................... 38  5.2.4  Example 2: MV/HV PILC Cable System Tested by DAC Test System and                   Conventional PD Measurement Technique .......................................................... 40  5.2.5  Example 3: Extruded EHV Cable System Tested by ACRF Test System and                   Non‐Conventional PD Measurement ................................................................... 42  5.2.6  Example 4: MV Cable System Tested by VLF Test system and                   Conventional PD Measurement Technique .......................................................... 44  5.2.7  Example 5: UHF PD Measurement at Cable Accessories in Service....................… 44    5.3  Rotating Machines ................................................................................................. 46  5.3.1  Pre‐Conditions .................................................................................................... 46  5.3.2  Example 1: On‐Site Tests by ACTC and ACRL Test Systems and                   Conventional PD Measurement Technique .......................................................... 47  5.3.3  Example 2: On‐Site Tests by ACRF Test System and                   Conventional PD Measurement Techniques ........................................................ 48  5.3.4  Example 3: On‐Site Tests by DAC Test System and                   Conventional PD Measurement Technique .......................................................... 50    5.4  Power Transformers ............................................................................................... 51  5.4.1  General ............................................................................................................... 51  5.4.2  Preconditions of Test Object ................................................................................ 52  5.4.3  Example 1: On‐Site Test by Motor Generator Set and                   Conventional PD Measurement Technique .......................................................... 54  5.4.4  Example 2: On‐site test by Frequency Converter and                   Conventional PD Measurement Technique .......................................................... 55  5.4.5  Example 3: UHF PD Measurement and Acoustic PD Localization in Service and                   for On‐Site Tests .................................................................................................. 57  5.4.6  Example 4: Synchronous Multi‐Terminal PD Measurement for On‐Site Testing ... 60    6.  Conclusion.................................................................................................................. 62    7.  References ................................................................................................................. 63 
  • 5. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 5 - 1. Introduction 1.1 Purpose of HV Dielectric On-Site Tests with PD Measurement During the life cycle of high voltage (HV) apparatus or systems (Figure 1.1), many tests and measurements are performed to characterize the insulation condition. The results of these tests and measurements should be compiled in a “life data record”, which supplies information on trends of diagnostic indicator values. The HV on-site test with partial discharge (PD) measurement has an intermediate position between routine tests and in-service monitoring measurements (on-line or off- line): Figure 1.1: Data sampling and recording during the life cycle of HV apparatus or systems In addition to type and routine testing, HV on-site testing is an important part of quality assurance. On- site tests are applied:  as a part of commissioning of equipment on-site to demonstrate that the transport from manufacture to site and erection on-site have not caused any critical defects, which lower the dielectric withstand voltage of the insulation below coordination withstand level Ucw [1.7];  after repair on-site to demonstrate that the equipment has been successfully repaired;  for diagnostic purposes in the framework of the asset management program by providing reference values of diagnostic indicators (e.g. partial discharge values and dielectric parameters) for later trending of test results. On-site tests are normally applied after a warning is received from the condition monitoring program. Furthermore it has to be considered that depending on the particular condition of the HV component:  the expectations from testing depend on the test procedure as selected for the particular component;  some degradation and breakdown risks may be related to the level and the duration of withstand voltage stresses as applied during on-site testing;  voltage withstand testing at the selected voltage level gives go/no-go information for this test voltage level and requires confirmation by methods such as PD measurement, of possible field strength effects on the insulation system.
  • 6. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 6 - HV dielectric withstand testing plays an important role in quality assurance and consequently in contractual matters between manufacturer and user of HV equipment. On-site testing, diagnostics and monitoring should ensure the reliability and remaining life time of an insulating system, and support corresponding decisions of the user (in accordance with their policy related to insurance). The different aims of HV testing on the one hand and insulation diagnostics/monitoring on the other may be seen as having quite different testing philosophies (Table 1.1). But the common aim of the measurement whether for HV testing or condition monitoring is the behaviour of the dielectric in the electric field. Therefore the tests and measurements in the factory, on-site and in-service should be harmonized. This offers the possibility of improvement of on-site test procedures supported by PD measurement. PD measurements play an important role in both quality testing and monitoring for condition assessment. As opposed to diagnostics such as dielectric measurements (tan , εr and dielectric response parameters) which are characteristics of the whole dielectric volume, partial discharges are related to local insulation defects. These defects in the insulation (“weak-points”) will result in PD activity prior to breakdown under certain stresses (e.g. mechanical, chemical, thermal or electrical stresses). In the case of the new insulation, critical defects are the result of assembling failures which can be found by routine testing consisting of the HV withstand test including PD measurement. In case of in service HV equipment insulation (which has been tested successfully and which operates for years), a critical defect might be caused by high electrical, thermal or mechanical stresses or by the “aging” of the insulation itself. This means the partial discharges are symptoms and sometimes also the results of this “integral” process in the volume of the insulation, which causes – over a more or less long period – the mentioned “weak point”. In other words, partial discharges indicate the non-reversible destruction of solid or mixed insulation. This is important for both the detection of production faults (quality assurance test) and of weak points/defects caused by aging during service (diagnostic test). Consequently all HV withstand tests on-site should be combined with PD measurement. Table 1.1: Common characteristics of HV withstand testing and insulation monitoring / diagnostics 1.2 HV Test Procedures with PD Measurement During the HV dielectric withstand test, the insulation is stressed with the specified test voltage (ut) for a specified time duration (tt), see Figure 1.2. This test is successful if the insulation withstands the applied stress. If a disruptive discharge occurs, the insulation fails the withstand test, which means the result delivers a clear answer which does not require any further interpretation. Therefore the withstand test is a direct test, which is intended to destroy a defective insulation whereas healthy insulation should pass the test. Characteristics HV withstand testing monitoring / diagnostics Condition of test object New or repaired aged in service Assessment of development (type test) production (routine test) assembling (on-site test) condition of insulation (remaining life-time) Method withstand voltage test, partly complemented by PD measurement dielectric measurements, PD measurement Aim quality assurance and insulation coordination classification e.g.: - safe for service - conditionally safe - unsafe for service Sometimes characterized as “destructive” (not negative, if insulation with defect is destroyed) “non-destructive” (not positive, if the result gives no clear indication)
  • 7. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 7 - During an indirect test other parameters, such as the apparent PD is measured. The insulation passes the test when the measured parameter value remains below a predefined limit during the test. The limit value is derived from laboratory and field experience, computer simulation based on physical models or simply from an agreement between the purchaser and manufacturer. The certainty of the interpretation of the on-site test results is very much dependent on how the limit value is originally determined. On-site voltage levels and the test duration should also be taken into account. In most cases, the certainty of the interpretation from indirect tests is very much lower than those of direct tests. Additionally it is necessary to consider the life time consumption of a withstand test. Each electrical stress consumes life time, depending on parameters such as voltage level, test duration and test frequency. A healthy insulation has a higher withstand voltage than insulation with critical defects. A withstand voltage test must be designed in such a way that the life time consumption of healthy insulation is negligible whereas the voltage is high enough to cause breakdown in defective insulation (Figure 1.2). Figure 1.2: Life time characteristic of a solid insulation Figure 1.3: Test cycle for withstand test with PD and test voltage stresses (ut, tt) measurement (schematically); optionally at withstand test level For non-self-restoring insulation it should be noted that insulation with defects may pass the withstand test with the test stress triggering defects which in time lead to a failure in service. To check the presence of insulation defects, the withstand test should be combined with a sensitive PD measurement (Figure 1.3). The voltage should be increased to the highest test voltage value in steps and then decreased in identical steps. The voltage step levels and the time of voltage application on each step can differ when applied to power cables, GIS or power transformers. The PD magnitude at each step (ut) should not exceed a defined limit and the PD inception and extinction voltages should be above the maximum operation voltage. It is recommended that the PD measurement be performed on all voltage steps prior to, during and following the withstand test levels as this may provide an indication of possible damage of the insulation. The combination of withstand testing and PD measurement is strongly recommended. The test voltage sequence as given in Figure 1.3 has been a recommended procedure for quality testing during routine tests for many years (this method cannot be applied to all HV equipment with non-self-restoring insulation however). It is expected that the combination of withstand testing and PD measurement will become mandatory for both routine and on-site tests in future. Additionally, the condition-assessment of insulation should take into account different methods of periodic inspection or monitoring. As a general principle periodic inspections or monitoring should deliver a warning which might be followed by a diagnostic off-line test. This test should apply the same procedure of withstand testing and PD measurement as described above (Figure 1.3). The PD values of that test shall be compared with the reference values (“fingerprints”) of the initial on-site test. ut defective insulation voltagelg ub healthy tb lg time tt voltage ur u1 u2 PD01 PD02 PD12PD11 PD22PD21 PDt ut service withstand test timett
  • 8. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 8 - Figure 1.4: Examples for HV tests with PD measurement (a) 400 kV GIS by ACRF (frequency tuned resonant circuit) test system of modular reactors via bushing (b) 400 kV GIS by ACRF test system with directly flanged reactor (c) 150 kV HV cable system by DAC test system (damped alternating voltages) (d) 400 kV/433 MVA power transformer by static frequency converter There are many different types of HV apparatus and systems. To be able to test them all requires a wide variety of well adapted mobile HV test facilities [1.1] including PD measuring equipment [1.2, 1.3]. Figure 1.4 shows examples of a selection of the HV test facilities outlining the type of test object, the purpose of the on-site test, and methods of handling. A main aspect should be that the test voltage should represent stresses in service. Related aspects are given in chapter 4 and 5. Clear guidelines for the requirements of the test voltages to be applied are given in IEC 60060-1 and 60060-2 [1.4, 1.5] for routine testing and in IEC 60060-3 [1.6] for on-site testing. For installations which have been in service, lower voltages and/or shorter durations may be used. These test values should be negotiated, taking into account the age, environment, history of breakdowns, and the purpose of carrying out this test, as there is not a general principle for the value of the on-site test voltage levels for all type of HV apparatus. The test voltage levels required for insulation coordination [1.7] are only related to new insulation, and any similar coordination for in-service insulation is not currently available. Furthermore, it is important to coordinate the quality assurance tests (routine tests in factory, commissioning tests on-site, and tests after repair) and the diagnostic tests (off-line or on-line). The results of measurements of partial discharges [1.8], dissipation factor and other quantities of quality tests are used as the references (“fingerprints”) for later diagnostic tests and measurements. Records of repeated measurements may be used to show trends (e.g. PD magnitudes, pulse rates, etc.) which present the most important information for condition assessment. All this enables the establishment of the “life cycle record” (Figure1.1) for important HV equipment as the reference for condition based maintenance (CBM). 2. High Voltage Sources and Accessories for On-Site Applications In 2000 CIGRE WG 33.03/TF 04 published an ELECTRA Report on requirements for HV withstand tests on-site [1.1]. This paper describes the different types of test voltages, their generation as well as the principles for their selection. This chapter repeats the main ideas of the 2000 report and complements them with the developments of recent years. a d c b
  • 9. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 9 - 2.1 General Requirements The requirements for stationary and mobile HV test systems are different (Table 2.1). In the past, the selection of mobile on-site test systems has been determined by compactness and transportability with the result, that for example HVDC testing has been used for HVAC equipment which contradicts to the above-mentioned principles of insulation coordination assessment [1.7] and HV testing [1.4, 1.5]. Today’s mobile withstand testing systems supply test voltages representative of stresses in service. On-site testing as defined by IEC 60060-3 [1.6] with its larger tolerances now allows a wide application of newer mobile test systems such as frequency-tuned resonant test systems (ACRF) for withstand and PD testing, damped alternating voltage (DAC) sources for diagnostic PD measurement and very low frequency (VLF) test as used for medium voltage cables and generators. A survey of the present application of different test voltages for on-site testing of HV apparatus/systems is shown in Table 2.2. PD measurement is not considered for DC and impulse voltages here. Only continuous AC voltage, damped AC voltage and very-low frequency voltage will be considered in more detail. Fig 2.1 shows the influence of these voltages on PD pattern. Whereas AC voltage (2 min) generates a very clear pattern (a) and DAC voltage shows similarities (b, c, d), VLF voltage (3min) did not generate enough pulses per time interval (e) and requires longer test duration for the generation of a sufficient pattern [2.1]. Table 2.1: Requirements for stationary and mobile HV test systems Figure 2.1: Typical phase resolved PD patterns for AC (a), DAC (b, c, d) and VLF (e) voltages (same defect) [2.18] requirement stationary mobile application development and quality testing in test fields quality and diagnostic testing on-site test voltages IEC 60071-1, IEC 60060-1, apparatus standards comparable to stresses in service, IEC 60060-3, apparatus standards compactness transportability usually enough space in HV labs very important for transportation weight-to-test power ratio usually not of high priority very important mechanical robustness usually no problem very important for frequent transportation and assembling handling control often automatic testing; connection to LAN easy to handle, robust
  • 10. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 10 - Table 2.2: Test voltages used on-site for various HV apparatus/systems (modified according to [1.1]) equipment to be tested on-site cables GIS GIL instru- ment trans- formers power transformers rotating machines arres- ters oil-paper cables extruded cables MV HV on-site test voltage shape for details see 5.2 5.2 5.2 5.1 - 5.4 5.3 - direct voltage (DC) W W 1) W 1 very low frequency voltage (VLF) W W, DM, PD W alternating voltage (AC) by ACTC W, PD W, PD W, PD, DM W, PD, DM W by ACRL W, DM, PD W, PD, DM W, PD W, PD W, PD W, PD, DM W, PD, DM W by ACRF W, DM, PD W, PD, DM W, PD W, PD W, PD W, PD, DM W, PD, DM W damped alternating voltage (DAC) PD, DM 2) PD, DM 2) PD, DM PD, DM impulse voltage lightning (LI, OLI) W W switching (SI) W W Abbreviations: W - withstand voltage test (e.g. 60 s with AC) PD - voltage test with PD measurement DM - voltage test with dielectric measurement (mainly tan δ, no withstand test) ACTC - transformer circuit for AC voltage generation ACRL - inductance-tuned resonant circuit for AC voltage generation ACRF - frequency-tuned resonant circuit for AC voltage generation MV - medium voltage HV - high-voltage 1) applied in the past and no longer recommended 2) mainly for PD diagnostics 2.2 Continuous Alternating Voltage (AC) Continuous AC voltage corresponds to the in-service or operating voltage of HV apparatus/systems and is therefore the most important voltage for withstand and PD testing. Some main characteristics of the different principles of AC test voltage generation (Figure 2.2) are summarized in Table 2.3 and described in the following. It is considered that the most test objects (Table 2.2) are a capacitive load. Compensated transformer circuits (ACTC – Figure 2.2a) use the traditional HV transformer and a LV or MV regulator. Because of the capacitive test object Cc, an inductive compensation LC is applied on the primary side. TH is the test transformer inductance. The circuit is completed by a coupling capacitor CH, a PD measuring impedance Z and a blocking impedance LS. The weight-to-power ratio (Table 2.3), the large dimensions and the high power consumption of this equipment restrict the use of ACTC test systems to on-site applications of relatively low voltage and low test power demand (e.g. MV switchgear and MV rotating machines of relatively low power). Inductance-tuned resonant circuits (ACRL – Figure 2.2b) are resonant circuits consisting of the capacitive load Cc and a reactor of variable inductance LH tuned in such a way that its natural frequency is equal to that of the power supply (50 Hz or 60 Hz). This is done by an adjustable gap between the fixed and the moveable part of the magnetic core of the reactor. The ACRL system has a limited load range of about Cmax/Cmin = 20. To generate a voltage without the test object connected, a basic load such as a voltage divider or coupling capacitor must be provided. The ACRL series resonant circuits, as described above, have a much better weight to power ratio than ACTC systems (Table 2.3) but they are increasingly being replaced by ACRF systems (see below). The ACRL system is still applied when a certain test frequency is mandatory. In such cases the ACRF system operates often as a “parallel” resonant system. Then, the tuneable reactor also acts as an auto-transformer, the test voltage is fixed by the regulator and the test current is optimally compensated. The parallel mode is helpful e.g. for on-site testing of large generators with fluctuating losses due to PD phenomena which would disturb the voltage stability in the series mode [2.2].
  • 11. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 11 - a) b) Figure 2.2: Circuits for AC test voltage generation c) Frequency-tuned resonant circuits (ACRF – Figure 2.2c) operate with a fixed reactor LH. Resonance is realised when the power loss of the oscillating circuit is supplied at its natural frequency via a frequency converter. ACRF systems have a much wider load range, a better quality factor, lower power consumption, a lower weight (Table 2.3) and a much lower price than ACRL systems. As the frequency depends on the capacitance of the test object a frequency range must be defined which satisfactorily represents power frequency. Based on experimental investigations [2.3] for extruded cables a range 20 Hz to 300 Hz is standardized [2.4, 2.5], for GIS the range 10 Hz to 300 Hz is used [2.6] and the horizontal standard IEC 60060-3 [1.6] defines even 10 Hz to 500 Hz. For all these frequencies, it is presumed that they generate a stress of the insulation based on a capacitive voltage distribution (as for 50/60 Hz). The advantages of ACRF systems have been proven in practice. GIS systems have been tested for more than 20 years and large scale tests on cable systems have been made for more than 10 years [2.7, 2.8]. The ACRF systems of different design are well adapted to the test objects (Figure 1.4) being robust and reliable. They are available up to 1000 kV and for more than 100 MVA (50 Hz equivalent power). TR LC TH LS Z PD CC AC ACTC Compensated transformer CH Z ACRL Resonant test system with tuned inductance TR LH TE CH Z PD CC ACLS FC TE LH CH Z PD CC AC AC f LS ACRF Resonant test system with variable frequency
  • 12. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 12 - Table 2.3: Comparison of circuits for AC test voltages on-site transformer circuits 1) (ACTC) tuneable reactor circuits (ACRL) tuneable frequency circuits (ACRF) frequency power frequency 2) 50 Hz or 60 Hz power frequency 2) 50 Hz or 60 Hz 20 Hz to 300 Hz quality factor q = Ptest / Psupply up to 5 40 to 100 70 to 200 power supply single phase (or two phases) single phase (or two phases) three phases weight to power ratio for testing medium voltage components > 10 kg / kVA 2 to 10 kg / kVA 3) 1,0 to 2,0 kg / kVA 3) HV cables, GIL not applicable 2 to 7,5 kg / kVA 3) 0,8 to 1,5 kg / kVA 3) GIS, HV components > 10 kg / kVA 2 to 8 kg / kVA 3) 0,6 to 1,0 kg / kVA 3) harmonics requires caution < 1 % < 1 % generation of PD noise no no IGBT switching pulses (can be suppressed) mechanical characteristic moving parts in the regulator moving parts in the regulator and in the tuneable reactor no moving parts transportation weight and volume Total weight and volume are in close relation to the weight to power ratio 1) Including compensation reactors 2) With additional frequency converter for other frequencies 3) The range depends on design parameters (duty cycle, power, materials, etc.). Typical duty cycles for oil-insulated sources used for medium voltage components are 30 min ON, 30 min OFF 6 to 12 times per day at rated voltage and current. Typical duty cycles for test on HV cables and GIL are 60 min ON, 60 min OFF, 3 to 6 times per day. SF6- insulated sources for tests on GIS can be used at rated parameters for 15 min per day (e.g. 3 times, each of 5 min). Motor-generator (M/G)-sets and recently static frequency converters are used for induced voltage on-site testing of power transformers [2.9, 2.10, and 2.11]. The HV AC test voltage is generated by the power transformer under test itself. Static frequency converters are smaller and lighter than M/G-sets and deliver an AC voltage of a wide frequency range (e.g. 15 Hz to 200 Hz) [2.12]). They can be combined with ACRF test systems (for applied voltage tests) to form very compact mobile transformer test systems [2.11, 2.12] (Figure 1.4 e and 2.3). The frequency converter and the test circuit must be designed in such a way that the wave shape of the voltage is sinusoidal (total harmonic distortion THD < 5 %). The PD noise level qN must be sufficiently low (qN < 20 … 100 pC) to be able to measure critical PD in power transformers. Figure 2.3: Mobile transformer test system for induced and applied voltage tests based on a static frequency converter Frequency Converter Compensation unit Step-up transformer Control room HV- FilterHV-test lead Power supply cable HV connection cable
  • 13. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 13 - 2.3 Damped Alternating Voltage (DAC) The damped alternating voltage (DAC) system supplies an oscillating switching impulse voltage (OSI) of quite low damping [1.1, 1.6, 2.13 - 2.18]. The test circuit (Figure 2.4a, b) consists basically of a direct voltage source, which charges the cable capacitance, and a suitable inductance to maintain an oscillating circuit with a frequency range between 20 and 1000 Hz (Figure 2.4d). To obtain PD occurrence conditions (e.g. PD inception, PD magnitude) similar to 50 (60) Hz AC voltage stresses, the DAC frequency has to be below approx. 500 Hz [2.1]. After the DC voltage ramp has reached the ignition voltage of a thyristor switch or the disruptive voltage of a spark gap, the capacitance is discharged via the inductance. As in resonant test systems, the oscillating frequency depends on the capacitive load (Figure 2.4d). The DAC system frequency approximates the natural frequency of the circuit, see Table 2.4. During the charging process, the test object is stressed with an increasing unipolar voltage. The charging time depends on the maximum available load current of the power supply (Iload, max), the test voltage (Utest) and the capacitance CC. An example for the charging times for various cable capacitances and test voltages are also shown in Figure 2.4c when for example a maximum load current of 8 mA of the applied voltage supply is assumed. The DAC voltage is never higher than the charging DC voltage and due to a shorter duration and decaying characteristic it is not comparable to withstand voltage testing with continuous AC. The quality factor QC of the oscillating circuit, which is responsible for the attenuation of the AC voltage wave, can be expressed by: QC = (L / (Cc • RA 2 ). RA is the equivalent circuit resistance. The maximum capacitance Cmax which can be tested using the DAC system, can be calculated by Cmax = (Ip / Vmax)2 • L, where Vmax is the maximum applied test voltage and Ip is the permissible AC current in the oscillating circuit, (Figure 2.4f). DAC system voltages are applied for diagnostic PD and dielectric loss measurement on power cables [2.13-2.15] (see chapter 5.2.4) and for the diagnostics of the stator insulation of rotating machines [2.16]. Major advantages of the DAC systems are their light weight and low input power requirements.
  • 14. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 14 - 0.1 1 10 100 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 C[uF] t[s] 150kV 120kV 90kV 60kV 30kV (a) (b) (c) (d) (e) (f) Figure 2.4: On-site generation of DAC voltages: (a) Schematic structure of a DAC system (b) Schematic view of a DAC measuring circuit (c) DAC charging time as function of power cable test capacitance (with max. load current of 8 mA) (d) Dependence of the frequency of damped AC test voltage on the power cable test capacitance (e) Damped sinusoidal AC voltage wave (f) Max. power cable capacitance in function of the applied damped AC voltage for an e.g. max. current of 300 A Table 2.4: Examples of the typical damped AC voltage frequencies in [Hz] for different lengths of two typical 150 kV power cables Length [km] XLPE (C = 154 pF/m) [Hz] Oil filled (C = 373 pF/m) [Hz] 0,25 300 194 0,5 213 137 1 151 97 2 107 69 4 76 49 8 53 34 16 38 24 20 34 22 0 20 40 60 80 100 120 140 160 180 0 1 2 3 4 5 6 7 8 9 10 11 12 13 C[uF ] f[ Hz ] da m pe d A C fre qu en cy [H ] HV power cable capacitance [ μF] 0 20 40 60 80 100 120 140 160 180 0 1 2 3 4 5 6 7 8 9 10 11 12 13 C[uF ] f[ Hz ] 0 20 40 60 80 100 120 140 160 180 0 1 2 3 4 5 6 7 8 9 10 11 12 13 C[uF ] f[ Hz ] dam ped AC freq uenc y [Hz] HV power cable capacitance [ μF] 1 10 100 1000 10000 10 30 50 70 90 110 130 150 Damped AC Voltage [kV] C[uF] 300A
  • 15. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 15 - 2.4 Very-Low Frequency (VLF) VLF testing, primarily of frequencies of approximately 0,1 Hz but in some cases (high capacitive load) down to 0,01 Hz has been an accepted method for voltage withstand testing for all types of medium voltage cables for several years (CENELEC HD 620 S1 and IEEE 400.2) [2.17]. Figure 2.5: VLF voltage generation by controlled DC charging and discharging circuit [2.17] There are several principles of VLF voltage generation utilising inductors with diode chains or power electronic switches, Figure 2.5 illustrates cosine-square wave voltages, where the polarity reversal follows a cosine function and Figure 2.6 gives an example of electronic invertors used for sine-wave voltage generation [2.19]. Voltage levels up to 200 kV are available. This type of on-site equipment has the benefit of low weight, ease of transportation, low power consumption and low cost. One has to consider that VLF causes a voltage cycle of 10 s (or even 100 s) which leads to a resistive voltage distribution in the insulation, longer test duration and a higher test voltage than with power frequency. Further consequences must be considered in PD and tan  measurements [2.20-2.26]. Figure 2.7 shows that the PD pattern at 0,1 Hz may have some similarities with that of 50 Hz test voltage if the test duration at 0,1 Hz is correspondingly extended. Figure 2.6: VLF voltages a) cosine-square wave, b) sine wave a b
  • 16. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 16 - Figure 2.7: Phase resolved PD patterns for 50 Hz (top) and 0,1 Hz (bottom) of an artificial defect in a cable joint of a 110 kV cable. The defect was in the semiconducting layer [2.26] 2.5 HV Filter, Coupling Capacitor, Connections and Grounding for Conventional PD Measurement The signal-to-noise ratio required for successful PD on-site measurements is dependent upon the HV circuit consisting of HV source, HV connections, test object and ground connection, e.g. Figure 2.2. The circuit acts as an antenna for radiated noise signals, such as those from external corona in the substation or from broadcasting stations. Therefore the spatial dimensions of the circuit should be as small as possible and itself PD free. A second source of noise is related to noise signals conducted from the power source or from power- electronic components to the test object (e.g. static frequency converters or power electronic switches). Efficient damping of this conducted noise may be achieved by the use of HV C-L filters consisting of a filter capacitor (CF; e.g. the bushing of a transformer or a voltage divider) and a blocking impedance (Ls > 15 mH up to 100 mH). The filter is connected to the coupling capacitor (CH, see Figure 2.2). As a rule of thumb a single well designed C-L filter can reduce the noise level up to one tenth (1/10), two such filters in series (“” filter) up to 1/20. The HV filter and the coupling capacitor should be arranged as near as possible to the test object. If the connection between the voltage source and the HV filter is by a shielded cable and not by an “open” air-insulated tube or wire, the “antenna effect” can be disregarded and the cable will act as an efficient filter capacitor CF. The coupling capacitor CH (Figure 2.2) must be adapted to the test object. The traditional rule CH ≈ (0.1 … 1) CC (capacitance of test object) cannot be generally recommended in this case. In a noisy surrounding a lower CH/CC may actually result in a better signal-to-noise ratio. The HV circuit includes the ground connection which should be as short as possible, of low impedance and bypass free. Because many test objects are grounded on-site directly, the grounding of the voltage source and its ground connection to the test object are a matter of optimisation of the grounding conditions.
  • 17. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 17 - 3. On-site PD Measurements When performing on-site PD tests in compliance with the relevant IEC standards the signal-to-noise (S/N) ratio becomes decisive. This is because the background noise level may range between several tens and a few hundreds of pC. In some cases, however, peaks of more than 1000 pC occur. Therefore the apparent charge can be measured sensitively only in “noise-free” surroundings, such as well shielded test areas using additional HV and LV noise filters or special electronic features for noise rejection [1.2]. To overcome the poor signal-to-noise (S/N) ratio under on-site conditions, conventional PD measurements are increasingly substituted by so-called non-conventional methods, as shown in Figure 3.1. In this context it should be noted, that an application guide on PD measurements in compliance with IEC 60270 [1.2] is available as Technical Brochure No. 366 [3.1]. Later, another Technical Brochure has been prepared by CIGRE WG D1.33, which deals with the so-called non-conventional PD detection methods [3.2]. Additionally a new IEC guide is currently under consideration to provide recommendations and specifications of the electromagnetic and acoustic PD detection methods [3.3]. Therefore, this brochure will only briefly consider the fundamentals of both conventional and non- conventional PD detection techniques but cover the practical experiences gained from on-site PD tests in more detail. Figure 3.1: Survey on PD measurement methods applied for HV apparatus under on-site condition 3.1 Conventional Electric PD Measurements Conventional partial discharge measurements complying with the relevant IEC standards [1.2, 1.8] have been proven over years as an indispensable tool for quality assurance testing of HV apparatus. This technique is based on the evaluation of the apparent charge which is only a small fraction of that charge created at the PD site, due to the condition that the stray capacitance Ca between the PD source and the test object terminals is much larger than the virtual capacitance Cb of the PD source itself [3.1]. Because the ratio Ca/Cb is not known at all, and the apparent charge depends additionally on the size and the location of PD defects, the PD severity of HV apparatus cannot be assessed by evaluating the apparent charge alone. in time New HV components GIS Transformers Cables Service aged Conventional methods (IEC 60270) Non- conventional methods measurement in frequency domain domain Electro - magnetic detection Acoustical detection Detection of optical occurrences Chemical compounds analysis measurement
  • 18. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 18 - As a consequence, knowledge rules for PD diagnosis in service have been established in the past based on practical experiences in the laboratory and on-site [3.4-3.6]. PD instruments intended for apparent charge measurements are generally equipped with a band-pass filter amplifier connected in series with a peak detector [1.2, 3.1]. Without going into detail her, it can be stated that the magnitude of the output pulse trains is a measure of the apparent charge of the input PD pulse sequences, due to the so-called quasi-integration performance of the band-pass filter. The condition for the quasi-integration is generally satisfied as long as the band-pass filter selects the measuring frequency in a range where the density of the frequency spectrum of the PD signal, appearing across the terminals of the test object, is nearly constant [3.7]. This is generally satisfied by limiting the measuring frequency below 500 kHz [1.2, 3.1]. To evaluate the apparent charge in terms of pC, the scale factor of the PD instrument must be determined by a specified calibration procedure. This is based on a simulation of the internal charge transfer by an injection of artificial PD pulses between the terminals of the test object [1.2, 3.1, 3.8- 3.12]. Therefore, the apparent charge of a PD pulse is defined in IEC 60270 [1.2] as “that charge which, if injected within a very short time between the terminals of the test object in a specified test circuit, would give the same reading on the measuring instrument as the PD current pulse itself.” In practice the magnitude of the injected calibrating charge q0 should exceed twice the background noise level. Moreover, q0 should be tuned to a limit which is close to the apparent charge as specified by the relevant apparatus standards or as set by an agreement. In practice it may also be useful to tune q0 to the level of the apparent charge to be expected for the particular HV apparatus under test. The advantage of PD measurement in compliance with IEC 60270 [1.2] is that the specified calibration procedure ensures reproducible and comparable PD test results, even when the tests are performed in different HV laboratories using PD measuring facilities of different manufacturers. The main drawback of conventional PD measurements is, however, that the S/N ratio is strongly reduced by the limitation of the measuring frequency below 500 kHz [see 1.2, 3.1]. Therefore, the applicability of PD measurements in compliance with IEC 60270 [1.2] may strongly be restricted under noisy on-site condition. Currently the classical analogue PD signal processing method is being replaced more and more by the advanced digital signal processing. This technique offers not only a more informative visualization of phase-resolved PD patterns (see Figure 3.2) but also a deeper statistical analysis of the very complex PD data with respect to the identification and classification of PD events (see example reported in [3.13-3.17]). Moreover, the digital technique can be used for the localization of PD sites in power cables [3.18, 3.19] and provides powerful features for de-noising of PD signals [3.20-3.28]. Figure 3.2: Example of phase-resolved PD pattern [3.15]
  • 19. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 19 - 3.2 Non-conventional Electromagnetic PD Detection Electromagnetic PD detection has been established over the past three decades as a valuable tool for on-site diagnostics of HV apparatus [3.29-3.62] due to the excellent S/N ratio, which becomes more enhanced at higher measurement frequencies. PD pulses due to discharges in gaseous dielectrics, such as SF6 and SF6-gas mixtures, are characterized by a rise time down to the sub-nanosecond range. Therefore the frequency spectrum covers a range up to about 3 GHz. In coaxial structures (e.g. GIS, GIL) such fast transients release electromagnetic waves traveling not only in the basic transverse electromagnetic mode (TEM) but also in higher order modes (TM, TE) [3.59]. The TEM waves which are weakly absorbed propagate through the whole plant. Above a critical frequency, however, the PD signals spread as TM and TE waves which are not coupled to conductors. These propagating fast transients can be detected by antennas (field sensors). PD pulses due to discharges in dielectric imperfections, such as voids and cracks in polymeric insulation, are characterized by rise times in the nanosecond range. Therefore, the resulting frequency spectrum may cover several hundreds of MHz and even more. This high frequency spectrum is well detectable by means of inductive and capacitive sensors. Due to the wide range of the frequency spectrum of PD transients, electromagnetic PD detection methods cover the frequency ranges of HF (3 MHz-30 MHz), VHF (30 MHz-300 MHz) and UHF (300 MHz-3000 MHz). Measuring frequencies in the UHF and VHF range are usually applied to GIS/GIL [3.29-3.31, 3.34-3.39] and power transformers [3.40-3.46]. The VHF and HF range is most suitable for rotating machines [3.47-3.51] as well as for power cable accessories [3.32, 3.33, 3.52- 3.55, 3.61, 3.62], see Table 3.1. Table 3.1: Most suitable frequency band for on-site / on-line PD detection in different components A survey on the structure and the significant steps of electromagnetic PD detection methods is schematically presented in Figure 3.3. Figure 3.3: Survey on the structure and steps of the electromagnetic PD detection Non- conventional methods Electro- magnetic detection HF/VHF UHF Sensors EM signal transmission Performance and sensitivity check On-site test Signal processing Sensitivity check PD measurement Cables Transformers GIS Generators HF + + - + VHF + + + + UHF - + + - Legend: + indicates suitable - indicates not suitable
  • 20. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 20 - Figure 3.4: Survey on common PD measurement procedures in the VHF and UHF range [3.33] The PD detection technique in the UHF and VHF range is mostly performed in the frequency domain, using either the zero-span or the full-spectra mode (see Figure 3.4). UHF and VHF PD diagnostics performed in the time domain, however, may also offer different advantages as reported, for instance, in [3.32, 3.33, and 3.61]. Without going in details it can be stated that the magnitude of the captured PD pulses and therefore the S/N ratio is enhanced as the measuring frequency increases. It should be noted that as the measuring frequency increases the detected signal magnitude is affected by attenuation and dispersion phenomena, mainly due to the skin effect in conductors and to polarization losses in dielectrics. The magnitude of propagating PD transients depends, for instance on [3.59]:  the distance between PD site and PD sensor  the material properties of the conductors and the insulation of the test object  the geometrical size / dimensions of the test object  the characteristic propagation modes, for instance in GIS/GIL: 1-2 dB/km (TEM, TE-mode); 4 dB/km (TM-mode) The signal magnitude can also be affected by reflections at discontinuities in the propagation path which lead to resonances and standing waves. Therefore, reproducible PD test results cannot be expected under these conditions, even with low scattering of the apparent charge magnitudes. Consequently, in these circumstances the electromagnetic PD detection method cannot be calibrated in terms of pC. As an alternative, a performance check as well as a sensitivity check is recommended as reported, for instance, in the references [3.2, 3.3, 3.56-3.62]. A performance check includes the verification of the functionality of the whole signal transmission path consisting of the PD sensor, the measuring cable, and the PD measuring system. The sensitivity check is intended to verify the PD threshold level of the complete measuring arrangement in terms of pC. This is best determined by comparative PD studies using both the conventional and non-conventional circuit. FREQUENCY DOMAIN MEASUREMENT TIME DOMAIN MEASUREMENT PRINCIPLE Zero-span mode Full-spectra mode Ultra-wide-band mode MEASURING FREQUENCY SPECTRUM PD SIGNATURES PROCEDURE test voltage PD signal test voltage PD signal Frequency sweep centre frequency PD signal
  • 21. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 21 - The individual steps which are required to carry out the performance and sensitivity checks are summarized in Figure 3.5. These procedures have been proven for GIS / GIL, power transformers and cable accessories, where the specific steps for each type of HV apparatus may differ slightly. As a practical example, the steps required for the sensitivity check for power cable accessories are listed in Table 3.2. Figure 3.5: General steps for the performance / sensitivity check Table 3.2: Steps for the sensitivity check for cable systems Laboratory set-up Single port Dual port PD source with known magnitude Injection of artificial pulses Sensitivity Check of sensor arrangement Sensitivity check Performance check On-site Injection of artificial pulses in same sensor Injection of artificial pulses in other sensor Transformers Cables, GIS Transformers 1. Laboratory set-up 2. On-site 1.1 Installation of an artificial PD defect (e.g. PD magnitude below 100pC) on a full size combination of power cable part and cable accessories 2.1 Performance check (sensitivity check) of the whole system: cable accessory, sensor, PD detection system. Injection of artificial voltage pulses according 1.5 1.2 Installation of given type of HF/VHF/UHF sensors on the cable accessories and parallel connection of a PD detection circuit (IEC 60270) 2.2 Tuning of the PD detection system to HF/VHF/UHF frequency range with best signal/noise ratio 1.3 HV energizing of the full size setup and PD detection using both systems according to IEC 60270 and IEC 62478 2.3 Estimation of the PD detection sensitivity of the whole system in terms of pC 1.4 For PD defect from (1.1) ratio estimation of the lowest [pC] reading and [μV] for the HF/VHF/UHF circuits 1.5 Search for an artificial voltage pulse with similar frequency/amplitude characteristics as the PD defect from (1.1) and (1.3)
  • 22. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 22 - 3.3 Noise Reduction Background noise can be coupled into the PD measuring circuit in different ways. Generally it can be distinguished between conducted and radiated noise. Conducted noise:  PD transients originating from poorly designed HV test circuit components, such as discharges from protrusions on HV shielding electrodes (Figure 3.6a) and discharges between floating parts (Figure 3.6b).  Periodical or stochastic occurring HF signals caused by switching operations in the mains and power electronic components and maintenance work in the measurement surroundings being conducted into the test source see Figure 3.6c.  Continuous radio frequency (RF) signals received from radio broadcasting via the grounding system and/or the feeding facility of the PD test circuit. Radiated noise:  Continuous RF waves radiated from radio broadcasting and TV stations, as well as noises from cell phones, see Figure 3.6d.  Periodically and stochastically pulse-type interferences from rectifiers, frequency converters and switches.  PD transients radiated from nearby HV components especially in substations. When using conventional PD measurement circuits in accordance with IEC 60270 [1.2], conducted noises can be reduced or even suppressed by suitable filter circuits installed in power electronic components such as frequency converters, in the power supply or between the HV test facility and the test object. The latter is usually designed as a -filter equipped with blocking impedances in series and low inductive capacitors in parallel (see 2.5). a) Discharges from protrusions of HV shielding electrodes b) Discharges between floating metallic parts c) Maintenance work (drill) d) Signals form a cell phone call Figure 3.6: 3D-graphs of characteristic noise signatures [3.32] (x-phase angle, y-pulse number, z-apparent charge)
  • 23. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 23 - Such noise filters are well proven in HV test laboratories and could in principle also be applied to on- site testing. However, with respect to radiated noise, the application of HV filters for on-site PD testing can be questionable because such filters act like antennas for radiated noise. Therefore on-site test set-ups should be designed with very compact filters with low longitudinal capacitance and the HV and grounding connections should be of low impedance. Even with these characteristics however, it is recommended that LV and HV filters and internal filters in power electronic components be recommended for on-site PD tests. The reduction of background noise picked up by conventional PD circuits according to IEC 60270 [1.2] requires a well-adapted coupling capacitor connected in series with the measuring impedance. Both should be located as close as possible to the test object. If for example the test voltage is supplied by a cable some tens of metres long from the HVAC source then the blocking impedance should be connected adjacent to the coupling capacitor. This is because the measuring cable can also act as a filter capacitor (see 3.3). Depending on the time characteristic of noise signals, appropriate counter measures are also recommended [3.16], such as:  Continuous wave (CW) noise, e.g. radio, TV, cellular phones. Counter measure: band pass filter with adjustable centre frequency and bandwidth tuned to “proper” spectral areas (see Figure 3.7) and/or use of adaptive notch filters to remove only certain narrow-band spectral interferences for improved signal-to-noise ratios.  Periodic pulse-type interference with constant phase angle (rectifier, converter): Counter measure: as this is easy to identify due to the constant phase position and usual small derivation in amplitude (deterministic behaviour), window gating is the preferred method.  Periodic pulse-type interference with shifting phase angle (frequency converter): Counter measure: as this is easily identified by rotating phase position, dynamic gating is recommended.  Pulse-type interference with stochastic phase angle (e.g. switching): Counter measure: the stochastic behaviour in phase and amplitude and rare (random) occurrence makes an appropriate choice of measurement time interval the best method.  PD-type interference (e.g. corona): Counter measure: de-noising based on pulse waveform analysis or multi-channel PD measurements is the preferred method because of the stochastic behaviour in phase and amplitude of this noise. Digital PD measurement and data evaluation, in combination with analogue methods, enables an effective rejection of electromagnetic interferences. An actual comparison of digital de-noising techniques for partial discharge measurements, including FFT, low pass filtering, short-time FFT, frequency-domain adaptive filtering, matched filter, notch filter, wavelet de-noising and others is given in [3.28]. Optimum results were obtained by combining notch filter and wavelet de-noising. In this context it should be noted that not all digital filter algorithms are suitable for real-time applications because of their algorithmic structure or demand for processing power. In many practical on-site cases, a combination of analog and digital filtering techniques delivered excellent results. For a limited number of dominant narrow-band interferences (e.g. radio), analog notch filters can help to reduce the input amplitude for the A/D converter, resulting in an improved dynamic range for the PD signals of interest. Very promising de-noising results can also be obtained by pulse waveform analysis [3.20-3.28]. This is because interference pulses and PD pulses often show different waveforms, which can, in principle, be analysed in the frequency domain or in the time domain. Pulse waveform analysis extracts a set of pulse features, which include, in the time domain, pulse rise and fall time, and pulse width and the characteristic frequency spectrum.
  • 24. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 24 - Another promising method for eliminating electromagnetic interference and for the separation of different PD sources is multi-terminal PD measurement. Within a short time window PD signals of the three phases of the test object (e.g. at three different bushings of a power transformer) are measured. The amplitudes of the signals are added as vectors in a “star diagram”. The three axes of this star diagram correspond to the three PD measuring points. Each PD source has a characteristic ratio of the amplitudes and will build up a cluster of events in the star diagram. By selecting only the signals which fulfill the condition of the cluster, the common "Phase Related Diagram" shows only the pattern of this specific PD source. With this principle, PD from different sources and noise pulses can be separated [3.17, 3.28]. For more details see 5.4.6. As mentioned previously, the application of VHF and UHF PD measurements improves significantly the S/N ratio, especially if this technique is combined with advanced de-noising features, as presented in [3.20-3.28]. As a result of such tools, one gets a lower background noise level in the non-energized HV circuit. However, this may increase after the HV test facility is switched on. In [3.2] a summary of well proven noise rejection methods is given for the electromagnetic PD detection technique, in particular for GIS and GIL. Similar procedures are applicable to other high- voltage assets, such as power transformers and cable accessories. Although these examples refer only to the UHF/VHF technique, similar noise rejection tools can also be followed for other non- conventional methods, such as the acoustic PD detection technique. 3.4 Acoustic PD Detection The acoustic detection of partial discharges uses the fact that an acoustic signal (e.g. a mechanical vibration) is emitted as a result of the pressure build-up by PD pulses in the insulation medium. The signals can be picked up by means of piezoelectric transducers, fibre optic sensors, accelerometers, condenser microphones and sound resonance sensors. Because of the short duration of the PD pulses the resulting compressed wave contains a frequency spectrum considerably higher than the audible sound band. The frequency spectrum generally used ranges from 10 kHz and 300 kHz. The acoustic wave propagation from the PD source to the acoustic sensor (transducer) is strongly influenced by the geometry of the test object as well as by the medium (oil, gas, steel). Consequently, different wave types with different propagation velocities appear and reflections and refractions at boundaries lead to changes of the sound propagation, resulting into damping, absorption and scattering of the signal [3.2, 3.42, 3.43, 3.45, 3.63-3.66, 5.26]. Using acoustic PD detection methods an attempt is made to locate the PD source. Ultrasonic PD location has been proven for partial discharges of high energy. The field of application of this method spans all HV apparatus discussed in this paper (e.g. see chapter 5.1.1, 5.1.2 and 5.4.5). To reach an optimum sensitivity the complete acoustic-mechanical system must be thoroughly understood (e.g. the influence of the sensor size). Comparative measurements have revealed that for GIS for example, an equivalent sensitivity as low as 5 pC could be achieved under on-site condition. Figure 3.7: Probability of spectral interference [3.16]
  • 25. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 25 - 3.5 Important Aspects for PD Evaluation To support the interpretation of PD measurements to recognize aging processes in the insulation of HV components, links between specific insulation characteristic and the information provided by particular measured PD quantities are necessary. As a result, systematic knowledge rules need to be formulated to support insulation condition assessment (see Figure 3.8). In particular, to identify measurable and derivable quantities for particular types of HV components, several aspects have to be taken into account (e.g. typical insulation defects, external factors like propagation effects, and disturbances or cross-talking). In addition to the detection of partial discharge activity, other important aspects of PD measurements include the identification of the discharge source and type of defect and the location of the defect to allow condition assessment of the insulation. Figure 3.8: General aspects of creating PD knowledge rules [3.6] 3.5.1 PD Identification The use of partial discharge measurements for the condition assessment of HV components depends on the method chosen for PD detection and on the derived quantities to evaluate the measured information. A combination of both aspects results in PD diagnosis, which generates information suitable for making PD knowledge rules. Generally, in a HV on-site test including PD detection, the PD magnitude, phase-resolved patterns and location of the discharges are usually determined as a function of the applied voltage. The discharge detection can be obtained with a discharge detection system performing in accordance with IEC 60270. Figure 3.9 shows a scheme of how to obtain the typical PD quantities. Partial discharges cause sequences of current impulses in the leads to the object under test. This can happen several times during a power frequency cycle. As a result, groups of recurrent discharges, which occur during the positive half and the negative half of the voltage cycle, can be found. If a sample has several discharging sites, more discharges will occur within the same time intervals. Different quantities have been introduced over the years to describe the characteristics of a discharge.
  • 26. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 26 - These quantities can be divided into three main groups:  basic PD quantities (according to IEC 60270)  derived PD quantities  phase, amplitude-related derived quantities (PD pattern) It is well known that the sequences of PD pulses contain random events, which are characterized by considerable variation in magnitude and phase angle. These significant PD parameters are also strongly influenced by the level of the applied AC test voltage. Therefore, informative diagnosis of insulation defects can only be achieved successfully if several PD quantities are measured and evaluated simultaneously (e.g. charge, phase and number of occurrence (, q, n)). While the actual ageing processes generated by the interaction of partial discharges are still not well understood, considerable information has been gathered over the years on insulation type, defect type and the effect of load conditions etc. As the HV components age, the number of faults will increase and the power supply quality can deteriorate. Important symptoms of this degradation process can be related to discharging local insulation imperfections or defects, which may occur in the particular HV component (Figure 3.10a and Figure 3.10b). Figure 3.9: Principle of digital PD data processing to obtain basic and derived PD quantities Figure 3.10: a) Generation process for PD knowledge rules
  • 27. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 27 - Figure 3.10: b) Relationship between service condition specific factors, defect detection, knowledge rules and condition assessment to support Condition Based Maintenance. With regard to a particular type of HV component, the combination of information about characteristics observed for each PD quantity can be used to support the interpretation of measuring data. In particular, to evaluate the effectiveness of PD measurement for different components, the following structure can be used:  Insulation defects: short description of general insulation problems (no particular construction problems) possible in this type of HV component;  PD processes and PD signals: short description of HV component specific aspects important for PD measurement;  Representative results: typical ‘’acceptable’’ and ‘’not acceptable’’ examples of using PD quantities to detect insulation degradation;  PD knowledge rules: examples of combining PD quantities to determine certain insulation condition of a HV component;  Conclusion: benefits/limitations of using PD diagnosis for this type of HV component. 3.5.2 PD Localization To estimate the criticality of PD, fault-type identification [3.22] has to be complemented by PD site localization. In expanded geometries, like long cables or GIS, or large components like power transformers, the PD identification without localization would be useless. The method most used to calculate the location of a partial discharge site is based on the difference in arrival times of PD pulses at one or several sensors. Different calculation methods are used for different HV components. Methods based on the attenuation of PD signals are also used. Further details will be found in the corresponding chapters. Regarding PD location calculation the following can be concluded:  time-domain-reflectometry is the most important method, where the arrival time of original and reflected PD signals are measured (cables, GIS, transformer); PD signal attenuation, however, may limit the applicability of this method.  with several sensors the difference of the time-of-arrival of either acoustic, electric or acoustic/electric signals can be used for evaluation; in long cable systems, GPS synchronization of PD measurements is often necessary [3.68];  other techniques that are being used, or are under investigation, are based on the attenuation effects that occur during the progress of EM-waves as they travel through the components, see e.g. [3.69].
  • 28. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 28 - 4 Preconditions for On-Site Testing Including PD Measurement On-site tests including PD measurement can be applied to different components and for different purposes. In particular, they can be part of commissioning tests after erection, for condition assessment (diagnostic or monitoring) or to check the effectiveness of repair operations. Moreover the systems used (concerning voltage stress amplitude and shape, PD measuring systems characteristics etc.) are very different and depend on the component to be tested, their insulation system or sometimes on the preferences of the suppliers and users. The main reference documents, such as International Standards (IEC, IEEE), typically only specify the voltage shape for testing [1.6]. Generally no indication is given about details and preconditions of the on-site test or the testing procedure and algorithms for the evaluation of the results. The technical details required for on-site testing are given in the following list. The list is subdivided into three parts referring to the characteristics of the component under test and its installation, the test requirements and the characteristics of the test system. Requirements for component under test and its installation  Ratings and main physical characteristics, including insulation system, specified in details in the next paragraphs;  Type of installation (e.g. indoor or outdoor);  Results of type and routine testing including PD magnitude, capacitance and loss factor at nominal voltage and frequency;  Date of installation (age) and average service load;  Characteristics of the power network and grounding system;  Air clearances requirements, available space for the HV test system;  Atmospheric (temperature, air pressure and humidity) and geographical (height above sea level, pollution class) information. Test requirements  Aim of the tests (e.g.: monitoring, commissioning after installation, commissioning after major maintenance, diagnostics etc.);  Availability of previous test data or history of the test object;  Test voltage shape and frequency requirements;  Restrictions for the test object regarding a possible, immediate failure during test;  Test procedure: voltage level(s), duration of application or number of impulses (in case of DAC) for each voltage levels, highest voltage (for both PD measurement and withstand voltage if required), PD inception and extinction voltage level;  Required PD measuring method and maximum accepted PD magnitude and max. background noise level;  Possible required evaluation of the results: PD location, PD phase resolved pattern, defect recognition and severity of defects (if present). Requirements for HV test and PD measuring system  Type and ratings of the test system;  Supply requirements (voltage, frequency, power and their tolerances);  Weight, size and environmental restrictions for transportation and operation;  Grounding requirements;  Type, dimensions and length required for the HV connection to the component under test;  Characteristics of PD measuring systems (sensitivity, bandwidth, conformity with IEC Standard 60270);  Additional PD evaluation features of the measuring systems (filter to enhance the S/N-ratio, possible location of PD source, automatic defect recognition etc.).
  • 29. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 29 - 5 Examples of Test and Measuring Techniques for Apparatus and Systems This chapter describes the different HV test sources and test procedures required to perform on-site PD tests on various types of HV apparatus including gas-insulated switchgears (GIS), power cables, rotating machines and power transformers. 5.1 Gas-Insulated Systems (GIS/GIL) 5.1.1 HV Source, PD Measurement and Details of Test Object On-site testing of GIS is an important step during commissioning of a new GIS/GIL or after modification and maintenance tasks. This test assures a correct erection or repair of a GIS/GIL and verifies that there are no defects within the GIS/GIL that could lead to a major failure during operation. Various methods can be performed depending on rated voltage level and GIS type and design. Experiences of on-site testing with different waveforms and procedures have shown which method is suitable to detect major defects within a GIS/GIL [3.57]. HV Sources and Test Procedures It has been common practice for nearly two decades to carry out AC high voltage tests by resonant test systems with variable frequencies (ACRF). IEC 62271-2003 [2.6] defines a test frequency range between 10 Hz and 300 Hz. Typically test frequencies above 80 Hz are used if voltage transformers are connected to the GIS. The results of sensitive partial discharge measurements have proved the success of this method. Typical defects that cause PD such as protrusions, free moving particles, floating parts or cracks in spacers, can be produced during transportation and erection. Due to the different physical behaviour of the defects under the different voltage stresses (HV AC stress, lightning impulse or switching surges) it is impossible to find all defects by withstand testing of only one waveform. Table 5.1-1 gives an overview of the effectiveness of the various applied HV voltage waveforms to detect the different defects [3.57]. It can be seen that the application of PD measurement enables the detection of all kind of serious defects during AC voltage testing. The standards [2.6] define no single technique or the methodology which must be used for on-site testing. This still depends on an agreement between the manufacturer and customer as to which test procedure (including PD measurement or not) and which combination of test voltages (AC, LI, OLI or OSI) is to be applied. The following main PD measuring methods are in used to detect partial discharges in GIS:  Conventional measuring systems based on the IEC 60270 recommendations;  UHF measuring systems using narrowband or wideband filter detection in a frequency range up to several GHz;  Acoustic measurements, mainly for PD location, using externally mounted acoustic sensors which detect the acoustic signals emitted by a PD source. Neither the UHF nor the acoustic PD measurement techniques can be calibrated in accordance with IEC 60270. This means, that the output signal of these measurement techniques cannot be directly associated with the apparent charge. This is not really a drawback; however, the main purpose of PD measurement on-site is the PD detection itself (yes/no condition) and the identification of the PD sources. Therefore the signal does not need to be calibrated. To compare the sensitivity of different UHF sensors and acoustic sensors a CIGRE brochure has been published which describes in detail the sensitivity check of such sensors [3.58].
  • 30. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 30 - Table 5.1-1: Relative effectiveness of on-site test on GIS defects /3.57/ Defect High AC LI SI Low AC with PD High AC with PD High AC with SI High AC with LI High AC with PD and LI Sharp protrusions fixed on live parts + - + + Round protrusions fixed on live parts (assembly faults) - + + + + + + Particles on spacers -/+ + - -/+ -/+ + + Cracks in spacers -/+ - - - + -/+ -/+ + Free particles + - + + + + + Parts floating - + + - - + Left foreign bodies + + - - + + + + - less effective + effective UHF-Sensors Specially designed capacitive sensors (antennas) are suitable for GIS PD measurements in the UHF range (e.g. frequencies above 300 MHz and up to 3 GHz) and have been well proven in the past for testing GIS on-site. In principal two types of UHF PD detection systems can be used, the narrow band technique with a bandwidth of around 5 MHz and the wide band technique using a bandwidth of up to 2 GHz. The UHF technique, which needs adequate electrical sensors, is very sensitive and can also locate the PD source with a sufficient spatial resolution. Different kinds of sensor are used (Figure 5.1-1):  Fixed UHF Sensors (disc sensor, cone sensor)  Mobile UHF-window sensor  Field grading electrodes. While in a new plant the UHF sensors can be installed simply (disc sensor arranged in a assembly flange, Figure 5.1-2a), only portable window sensors can be used in an existing plant (Figure 5.1-2b). In addition, field grading electrodes can also serve for the detection of the PD signals with less sensitivity, however. A comparison of the first two kinds of sensors shows that with an appropriately sized window, the sensitivity of conventional UHF sensors can be achieved. With very small flange diameters however, the UHF measurement is not successful [3.36]. Field grading electrodes (used to make the electric field distribution more uniform inside barrier insulators) show a sensitivity corresponding to those of the window sensors [3.36, 3.37, and 3.38].
  • 31. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 31 - Figure 5.1-1: Different types of PD sensors/couplers: a) disc sensor, b) and d) field grading electrodes, c) external plate sensor [3.36] Figure 5.1-2a: Sectional view and mounting of an UHF PD disk sensor Figure 5.1-2b: Sectional view and mounting of an UHF PD window coupler PD Localization Different procedures and methods for the localization of the PD source can be applied. Sectionalizing and electrical time-of-flight measurements are the most practical procedures and are typical for the UHF PD measurement method. These reflectometry techniques are explained in general in the CIGRE Brochure 297 [3.19] and are also applicable for acoustic measurements. The very fast electric pulse emitted by a PD source of rise time below 1 ns, propagates in all directions along the GIS bus duct. It arrives at couplers which might be located on both sides of the PD source. By using the “time-off-flight” technique the time difference between the two wave fronts arriving at the couplers can provide information on the location of the PD source (Fig 5.1-3). As the time difference is of the order of tens of nanoseconds a fast digital oscilloscope is required for the measurements [3.39]. a ) b ) c ) d )
  • 32. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 32 - The electric waves attenuate when propagating through the GIS because of damping and resonance phenomena. This can also be applied to locate the PD site if the influence of different GIS components on signal attenuation is known. Figure 5.1-3: Principle of PD localization with two sensors and the time-of-flight method [3.39] Required Details of Test Object The following information about GIS/GIL has to be considered in addition to chapter 4, in order to be able to perform HV tests in combination with sensitive PD measurements (UHF- or acoustic methods):  Type / layout of installation: number of busbars, cable, transformer connections  Ratings  Capacitance of bay, components at nominal voltage and frequency  Date of installation (age) and average service load  Characteristics of the power network and grounding system  Air clearances requirements if HV test voltage is supplied via bushing to GIS/GIL  Restrictions for the test object regarding a possible, immediate failure during test  Required PD measuring method and maximum accepted PD magnitude and max. background noise level. 5.1.2 Example 1: On-Site Test by ACRL Test System (Tuneable Reactor Circuit) and Acoustic and UHF PD Measurement Technique To perform HV on-site testing on two GIS substations, rated voltage 145 kV and 300 kV, a resonant test system with variable inductance (ACRL), was used. Both acoustic and UHF PD measurements were made. Details are given in [5.1]. An acoustic signal was detected using an acoustic emission sensor with a resonant frequency of 32 kHz, placed on the external surface of the GIS enclosure. As the acoustic signal coming from the defects is subjected to strong attenuation along the GIS, particularly when there are spacers between the PD source and the measuring probe, the sensor was positioned in different places on the GIS. At least one measuring point was placed in each compartment. Acoustic measurements were performed at U=0 kV (background level) and at U=191 kV (1,1 x operating voltage). The acoustic signal coming from the sensor was amplified and then sent to an oscilloscope and to PD detector where time-domain analyses were performed. The tests were made with a single-phase test circuit. Each phase conductor was connected in turn to the voltage supply, the conductors of the other phases being earthed together with the enclosure. For the UHF measurements, pre-installed UHF-sensors were used to pick up the PD signals [5.1]. The combination of both PD measurement techniques is a very powerful tool to detect, locate and interpret the recorded PD event and to distinguish it from signals not originating from harmful defects.
  • 33. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 33 - During the on-site tests various signals could be detected, analysed and located. In addition loose electric shields, free moving particles of size 2 mm to 6 mm, were detected and localized. Figure 5.1- 4a shows the signal of free moving particles at power frequency. Figure 5.1-4b shows the corresponding phase resolved PD pattern. After opening the GIS, numerous metallic (dust-like) particles were found and removed. Figure 5.1-4a: Acoustic signal of harmful free particles [5.1] Figure: 5.1-4b Phase resolved PD pattern of free moving particles 5.1.3 Example 2: On-Site Test by ACRF Test System (Tuneable Frequency Circuit) and High Frequency PD Measurement Technique It is now common practice to test GIS on-site with ACRF test systems. Two primary methods are available. Metal-enclosed, SF6-insulated HV reactors, which can be directly flanged to the GIS (Figure 1.4b), achieve a high PD sensitivity with low background noise. They require the lowest space demand and no additional safety requirements. Maximum test voltage is about 740 kV. Oil-insulated modular reactors on the other hand (Figure 1.4a) cannot be directly flanged to the GIS. The HV has to be supplied via a bushing to the GIS. The inductance of the reactors can be well adapted to the test object capacitance by series and parallel connection of reactor modules. Test voltages in the range from 50 kV to 800 kV can be realized. A typical test sequence, consisting of a conditioning phase and the withstand voltage test for 1 min. is given in Figure 5.1-5. Testing frequencies above 80 Hz to 90 Hz allow testing of the GIS including the voltage transformers. This avoids a special disconnector for the voltage transformer or, if not available, further disassembly of the GIS and the addition of other testing components. This additional assembly work can't be tested in the same way as the GIS as this becomes a weak point in the commissioning. If an addition to an existing GIS without a separating switch for the voltage transformer is provided, higher test frequencies allow the commissioning test of any additions with the existing GIS without re-moving the existing voltage transformer. Results achieved during HV testing at a 220 kV GIS substation, equipped with 144 UHF sensors of two types, are presented in Figure 5.1-6a,b. The substation consisted of eleven bays with ten feeders and a double busbar configuration with one bypass coupling. A variable frequency resonant test system was set up on-site. The HV was supplied via a bushing to the GIS. The sensitivity of the UHF sensors was checked before HV test according to the CIGRE sensitivity check described in [3.58]. Applications of wide and narrow band UHF method were applied to the GIS and the detected PD signals resulted in their origin being localized to within +/- 3 m. By means of acoustic measurements the PD location could be localised to +/- 0,5 m. The signature of the measured PD pattern indicated a free moving particle, Figure 5.1-6a, and b. The compartment was opened and two moving particle with a length of 2 mm were found [5.2].
  • 34. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 34 - Figure 5.1-5: Typical test procedure on-site, consisting of a conditioning phase (Ur / √3) up to rated voltage Ur and the withstand test voltage for 1 min (0.8 x rated short-duration power-frequency withstand voltage Ud) Figure 5.1-6a: Spectrum of free moving particle Figure 5.1-6b: PD pattern of free mowing particle, integrated over 100 s in zero span modus, centre frequency is 637 MHz [5.2] 5.2 Cable Systems 5.2.1 HV Sources and PD measurement Two principal methods can be considered for the detection of on-site partial discharges in external energized power cables, (see Figure 5.2-1):  (A) Non-standard “unconventional” method according to IEC 62478  (B) Standard “conventional” detection method according to IEC 60270 [1.2, 1.8, 2.13] Table 5.2-1: PD detection during on-site testing TEST TYPE DESCRIPTION (A) AC voltage test [1.6] PD measurement (unconventional) not yet standardized a) frequency of detected signal within the radio Frequency (RF) range (e.g. up to 500 MHz), PD measurement in [μV] or [mV] b) calibration in [pC] not possible c) sensitivity check recommended d) detection of PD is mainly focused on cable accessories (e.g. joints and terminations) e) satisfactory signal / noise ratio is of importance (B) AC voltage test [1.6], [2.13] PD measurement in accordance to [1.2], [1.8] a) frequency of the detected signal in the kHz range, PD magnitude measurement in [pC] b) detection of PD on the whole cable system insulation c) localization of PD in cable accessories d) measurement is quite sensitive to external noise.
  • 35. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 35 - On-site after-laying testing of new installed (or repaired) EHV cable systems utilizes non-conventional PD measurement today using VHF or UHF frequency ranges, with focus on the accessories (e.g. joints and terminations). The external noise and disturbances can be suppressed down to a certain level and the eventual PD activity can be detected after the sensitivity checks have been done for the specific configuration of the cable accessories (e.g. joint, termination, the VHF/UHF coupler type and the signal processing device, see Figure 5.2-1). Figure 5.2-1: Principles of off-line PD detection methods for power cables; off-line standardized detection applicable for both routine testing and on-site testing. Using conventional PD detection, discharging activity can be detected in both the cable insulation and the cable accessories. The sensitivity of time-domain-reflectometry (TDR) can be estimated on the basis of the frequency characteristic of the cable impedance, Figure 5.2-2 [3.19]. When establishing the required sensitivity and maximum acceptable background noise level for the on-site PD measurement, the following aspects should be taken into account.  Type of cable system (e.g. type of cable and the type of accessories). Polymeric cable system are generally quite sensitive to partial discharges, whereas fluid-fill or mass-impregnated systems are usually less sensitive to PD. Regarding accessories, fluid-filled accessories can usually withstand higher PD activity than fluid-free accessories and for a longer time before failure occurs.  The operational electric stress level of the cable systems. In solid dielectrics the severity of PD activity increases quickly with increasing operational stress. As a consequence, HV and particularly EHV cable systems, which are designed for operating at relatively high electric fields, are as a rule much more sensitive to PD’s than MV systems, which operate at lower electric stress.  Test voltage level reached during the tests. When cable systems are being tested at voltages higher than the nominal voltage, the actual electric stress in the system is higher than the nominal stress. Consequently, the severity of the PDs increases with the test voltage. PD activity which may originate outside the test object such as corona discharges originating at the HV connection between the voltage source and the test object also increases with the test (over)voltage; for instance. As an example, severe defects in oil filled MV cables (PILC) or defects in accessories of MV XLPE cables may induce PDs in the range of hundreds of pCs without leading to an immediate failure. For those situations, a background noise level in the order of 100 pC can be accepted for the PD test on-site. On the other hand, for (E)HV polymeric extruded cables systems, PDs of a few tens of pCs may lead to failure within a relatively short time. Consequently, for this case it is important to maintain on-site background noise level as low as possible (10 pC or lower). Option B
  • 36. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 36 - Figure 5.2-2: PD detection sensitivity versus the cable length calculated on the basis of a reference value of 10 pC achievable for a power cable of 1 km length [3.19] These examples also show the complexity of establishing general threshold values for PD levels during on-site measurements on cable systems. The threshold values used today are based on experience, long term observation of measured cable systems and following laboratory investigations of examined parts of the cable system such as joints. For these reasons the combination of HV withstand testing and PD measurement should be taken into consideration (see 1.2). In addition to the apparent charge, the evaluation of the phase-resolved PD patterns as well as the shape of the individual PD pulses can be very helpful for an assessment of the PD severity. For more information refer to chapter 3.3. In particular it is important to be able to discriminate between internal PD activity (e.g. PDs originated inside the test object) and external PD activity (e.g. PDs originated outside the test object). Sensitive PD measurements according to IEC 60270 can only be applied for a cable length up to few km. For longer cable lengths the PD threshold level is increased substantially due to attenuation and dispersion of propagating PD pulses (see Figure 5.2-2). PD Localization Depending on the cable system length, techniques based on time-domain-reflectometry (TDR) commonly used in laboratory testing are also applicable for off-line measurements on-site. For longer lines, with several accessories, approximate location estimates can be derived by looking at the magnitude and frequency content of the pulses, since, owing to attenuation, both quantities will be higher in proximity of the source [3.19, 3.69]. Synchronized measurements of arrival times can be used to derive a more exact location, by using GPS, fibre optics, or reference injected pulses, see e.g. [3.68]. Techniques based on time-domain-reflectometry (TDR) commonly used in laboratory testing are also applicable for off-line measurements on-site. Signal reduction analysis is based on the fact that PD pulses propagate through the power cable in different modes [3.19]. As the attenuation of the amplitude of the signal is dependent on the frequency of the propagating signal this effect can be used to roughly estimate the origin of the PD signals. 1 10 100 1000 0 1 2 3 4 5 6 7 8 9 10 11 Power cable length [km] DetectablePDpulsecharge[pC]
  • 37. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 37 - 5.2.2 Required Details of Test Object The setting up of optimised testing procedures relies on an understanding of the goal of the on-site test and the cable characteristics. The following check list is reported for on-site tests including PD measurement. Cable, Joint, Termination and Lay-down Characteristics  Rated voltage;  Type of insulation of cables and accessories e.g. XLPE, ethylene-propylene rubber (EPR), mass-impregnated, Fluid-fill, etc.;  Single core or multi-core cable (for MV cable), propagation velocity of PD signals;  Cable length (capacitance, loss factor at rated voltage and frequency);  Type of installation (e.g. indoor or outdoor);  Type of earthing (e.g. solid bonding, single point bonding or cross-bonding);  Year of installation, average service load and other relevant historical data (e.g. previous failure, sheath defects);  Number of joints and their distances from the measuring point. Moreover number and position of cross bonding joints (whenever used) where it is usually possible to install unconventional PD sensors;  Type of termination (e.g. outdoor termination, metal-enclosed termination, etc.);  Availability and characteristics of the power supply (e.g. network earthing: insulated neutral or earthed network);  Clearances, position of terminations and presence of external structures that may interfere with the test. Test Requirement  Aim of the tests (e.g.: commissioning after installation, commissioning after major maintenance, diagnostics, monitoring etc.);  Voltage shape and frequency requirement (see chapter 2);  Test procedure: voltage level(s), duration of application or number of impulses (in case of DAC) for each voltage levels, highest voltage (for both PD measurement and withstand voltage if required);  Location of defected points (e.g. cables insulation, joint or termination for later risk analysis);  Required PD measuring method;  Required maximum accepted PD magnitude, or PD inception and extinction voltages, or alternatively definition of criticality degree of the possible defected points and indication of the consequent actions, (e.g. repetition of the test within a given time or fix the defected point), related to the position of the defect (cable, joint, termination and their insulation materials);  Calibration of the measuring system or sensitivity check (in case of unconventional PD measurement). Test System and Testing Conditions  Supply requirements (voltage, frequency and power and their tolerances);  Grounding requirements;  Weight, size and environmental (temperature, humidity, air pressure and pollution class) restrictions for transport and operation;  Type, and dimensions of connection from the generator to the cable;  Characteristics of PD measuring systems (sensitivity, bandwidth, conformity with IEC 60270);  Additional PD evaluation features of the measuring systems (filter to enhance the signal-to- noise ratio, possible location of PD spot, automatic defect recognition etc.);  Possible presence of sources of electromagnetic disturbances in the vicinity of the test objects (e.g. power electronic devices, antennas, etc.);  Requirements regarding propagation velocity (e.g. at mixed cables);  Necessity and availability of PD pin pointing equipment.
  • 38. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 38 - 5.2.3 Example 1: HV Cable Systems Tested by ACRF Test System and Conventional PD Measurement Technique An ACRF (see 2.2) test system including PD measurement facility based on a spectrum analyser was used for commissioning of cable systems [5.3]. This system allows the PD measurement to be performed without internal PD sensors being integrated into the cable accessories. The external sensor is based on a coupling capacitor or a high frequency current transformer using the principle of IEC 60270 [1.2]. Over four years, more than 60 PD commissioning tests have been performed in conjunction with withstand tests. In this example the discharge activity ranged from as low as 15 pC to 250 pC. The noise level ranged from 5 pC up to 70 pC. On average a noise level of 20 pC was experienced. The variation in this noise level was determined by local circumstances. Figure 5.2-3 shows a measurement on a HV cable resulting in a noise level of approximately 10 pC. The four spikes (1) to (4) in this figure are caused by the frequency converter of the ACRF test system. The measurement did not show any partial discharge activity in the cable circuit. Unfortunately, some PD measurements could not be performed because of extremely high noise levels, ranging up to 200 pC. These high levels were experienced in a substation environment with live equipment in the vicinity and under rainy conditions. On three occasions evidence of PD activity in the cable circuit was found. PD activity outside the cable circuit has also been measured including PD patterns caused by floating parts close to high voltage (Figure 5.2-4) and discharges generated by an insulating rod as a result of wear (Figure 5.2-5). The rod was used to support the tube connecting the reactor of the ACRF test system to the cable under test. After replacing the rod the disturbance disappeared. Even corona activity inside a connected GIS was measured. This PD activity was found to be from the GIS cable termination which was not correctly equipped with a temporary corona shield. After installing such a shield, the discharges have disappeared. Finally the situation on-site may influence the sensitivity of the test. It is essential, for example, for the ACRF test system to be positioned as close to the termination as possible. In order to minimize disturbances, the leads must be connected to the terminations in short straight lines. It is especially important to consider fences, screens and other earthed objects when connecting the test object to the voltage source (see Figure 5.2-6). Although the clearances under such circumstances are sufficient from a voltage point of view, disturbing influences for PD measurements cannot always be excluded. Figure 5.2-3: Example of a PD-free measurement, noise level approximately 10 pC (1) (3)(2) (4)
  • 39. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 39 - Figure 5.2-4: Example of discharges from a floating electrode Figure 5.2-5: Example of discharges from an insulating rod used for support of the HV connection Figure 5.2-6: Example of a situation requiring some improvisation for the HV connection; a) HV source, b) HV electrode, c) connection to test object via bushing Time [s] a b c
  • 40. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 40 - 5.2.4 Example 2: MV/HV PILC Cable System Tested by DAC Test System and Conventional PD Measurement Technique This chapter is related to the practical experience of on-site PD measurements on MV and HV power cables using damped AC (DAC) test voltages with a frequency between 20 Hz and 500 Hz [5.4 - 5.7]. For a 10 kV PILC (paper insulated, lead covered) cable system, the PD parameters (PD inception voltage (PDIV), PD extinction voltage (PDEV), PD magnitude at PDIV and at voltages up to 2,0 U0 were measured to determine the severity of the PD defects. In addition to the PD parameter and mapping of PD occurrences the ground noise level has also been documented to determine the test conditions (Table 5.2-2). The result of a typical PD measurement is given in Figure 5.2-7. Typical patterns from PD in an oil- filled joint can be distinguished, with some expertise, from PD in voids (Figure 5.2-7b) and PD between paper layers in a dry area of PILC cables (Figure 5.2-7a). Stressing the cable insulation with DAC voltages and applying TDR to PD signals in HF range (up to 50 MHz) provides the PD site location in cable insulation. As a result the PD mapping of a cable provides an indication of the location and statistics of PD occurrences (Figure 5.2-8). Table 5.2-2: Example of typical PD and test parameters of a three phase 10 kV PILC cable (length: 1950 m) With knowledge of the type of components in a circuit combined with the mapped PD activities, the operational history and importance of the circuit, an evaluation of PD defects and recommendations for maintenance or replacement can be determined. In general HV power cables (50 kV to 150 kV) should be PD free. This should also be the case during after-laying testing of new or repaired HV cables, or during condition assessment testing of service aged power cables (Figure 1.4c). Figure 5.2-7: Typical phase resolved PD patterns as observed at damped AC stresses of: a) an insulation defect in oil filled cable system b) an insulation defect in epoxy insulation of joint or dry area of PILC insulation
  • 41. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 41 - Performing test voltages also up to 1,73 U0 (where the maximum is advisable for new or recently installed cables, lower levels should be used for aged cables) is important for several reasons:  To evaluate if there are insulation defects with a PD inception voltage (PDIV) > U0. Such defects may initiate insulation failure in the case of temporary AC over-voltages.  To conclude following successful after-laying test that there are no PD detected up to 1,73 U0 and that during the service operation the power cable insulation is free of discharging defects.  That the on-site test has not initiated any discharging processes in the insulation. This information is important to confirm the non-destructive character of the diagnostic on-site test itself. An example of a 55 year-old oil-filled cable section which is not PD-free is shown in Figures 5.2-9 to 5.2.-11. In particular as the PDIV was lower than U0 it was important to localize the PD activity in the cable. It follows from the PD mappings (Figure 5.2-10) that all discharges as observed in phase L1 are localized in cable joints. Also it was observed that by increasing the test voltage up to 1,2 U0 the PD magnitudes did not significantly change. Additional measurements of dielectric losses (tan ) as a function of the test voltage confirmed that phase L1 had higher values of dielectric losses compared to phases L2 and L3 (Figure 5.2-11). Information of the presence of PD activity in one of the phases of the 150 kV cable section and the differences in dielectric losses between the three phases were valuable inputs for further decisions about the maintenance and the operation of this particular HV cable section. Figure 5.2-8: Example of a PD mapping of mixed PILC and XLPE cable circuit Figure 5.2-9: Example of phase-resolved PD patterns as observed during on-site testing of a 150 kV, oil- filled, 55 year-old, and 6,2 km long HV cable section: a) at 0,7 U0: background noise is < 10 pC, the cable section is PD-free b) at 1,0 U0: PD activity of 40 pC starting at 0,9 U0 has been observed in phase L1 c) at 1,2 U0: PD activity of 90 pC starting at 0,9 U0 has been observed in phase L1
  • 42. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 42 - Figure 5.2-10: Examples of PD mappings as made during on-site testing of a 150 kV, oil-filled, 6,2 km long,and 55 year-old HV cable section: a) up to 1 U0: PD activity between 20 pC and 105 pC has been observed in four joints (phase L1) b) up to 1.2 U0: as compared to 1 U0 no significant change of PD behavior Figure 5.2-11: Example of dielectric losses (tan ) measurement between 0,3 U0 and U0 on a 150 kV, oil filled, 6,2 km long and 55 years old HV cable section. 5.2.5 Example 3: Extruded EHV Cable System Tested by ACRF Test System and Non-Conventional PD Measurement After the installation of the 380 kV cable connection beneath the waterways “Caland Canal” and “Nieuwe Waterweg” in The Netherlands [5.6], an AC voltage test with non-conventional PD measurements were performed. The 380 kV cable connection consists of two 2200 m long XLPE cable circuits. Each circuit is divided into two sections by means of joints and it is terminated with composite outdoor sealing ends. Each phase of the two circuits was tested for 1 hour at 374 KV AC voltage (phase to ground), according to the IEC 62067 [5.8]. The voltage was applied by means of an ACRF system (see paragraph 2.1). The AC voltage frequency during the test was 43 Hz. During the AC voltage test, partial discharge measurements were made at the cable accessories. For this purpose three autonomous PD measuring units were installed at the two sealing ends and at the joint of each cable phase. A measuring unit consists of a VHF/UHF PD sensor integrated in the cable joint, a pre-amplifier and a PC-controlled spectrum analyzer [5.9, 5.10]. In addition, a trigger unit was adopted to synchronize the AC voltage and the PD signal. Finally, a wireless network was used for the communication between the three stand-alone measuring systems. In Figure 5.2-12 the PD detection system is represented. Before starting the PD measurement, the sensitivity check was performed as described in [5.10]. This was done in order to correlate the measured voltage signal in μV to a PD signal in pC. Using a fast impulse generator which was calibrated in the laboratory, artificial PD pulses were injected into each cable termination. 0 10 20 30 40 50 60 40 60 80 100 115 V[kV] value*10^-4 L1 L2 L3
  • 43. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 43 - Figure 5.2-12: VHF/UHF PD detection system as used during the on-site PD measurements performed on the 380 kV cable accessories [5.9] To achieve optimal performances of the measuring system, the spectrum of each PD detection unit was tuned to maximize the signal-to-noise ratio. For the specific spectrum range used for the PD detection, the noise level was lower 10 pC. However, due to the fact that the terminations themselves act as large antennas, some external noise activity was observed. In order to separate these external disturbances from potential partial discharge signals, the phase-resolved PD noise patterns were displayed in real time and stored during the PD measurements. In this way it was possible to establish that all accessories employed in the cable connection were PD-free, at voltage of 374 kV. Some of the results of the PD measurements are shown in Figure 5.2-13. Thanks to the real-time phase-resolved PD plotting, random disturbances could be discriminated, as shown Figure 5.2-13, at termination 2. PC Termination 1 Joint Termination 2 Figure 5.2-13. Measuring results simultaneously obtained at the three cable accessories during the acceptance test of one cable: the PD magnitude versus time for one hour testing and examples of phase-resolved patterns obtained during several cycles of the test [5.10]. Displayed PD pulses are caused by external noise. sensor spectrum analyser pre-amplifier
  • 44. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 44 - 5.2.6 Example 4: MV Cable System Tested by VLF Test system and Conventional PD Measurement Technique PD-measurement and fault localization of PIL- and XLPE-insulated cable systems is also performed by VLF voltage with conventional PD measurement. Testing equipment is compact and light-weight with low power consumption and can be stored in a small van, (Figure 5.2-14). When performing PD measurements under VLF test voltages one has to consider that the electrical field may be controlled more by the resistance of the dielectric, which is different from the capacitive field control under service condition. Hence, the PD occurrence may change substantially under VLF test voltages. Figure 5.2-14: VLF System on-site test arrangement Because of the low frequency and the low voltage gradient of the VLF voltage against the 50/60 Hz operating voltage, it may take considerable time to get a sufficient number of single PD pulses for PD detection and localization (see 2.4 and Figure 2.7). This time factor does not influence PD detection and location however. On the other hand PD activities are activated at comparatively low voltage levels [5.10]. As the general PD activity is much higher in PIL-cable systems than in XLPE-cables, PD recognition is less complicated. Experience also shows that the time required for data handling is much longer than the time taken to make the measurements. Furthermore well trained operators are necessary. In most cases the PD source is not located within the cable insulation itself, but in joints and terminations. Therefore it is very important to identify the exact location of the PD source. Experiences with VLF systems gained in laboratory and on-site show that the localization accuracy is in the range of several meters As for all PD measurements, the measured levels by different type of test equipment cannot be compared against each other, while using identical equipment test and applying identical test procedures leads to comparable results. 5.2.7 Example 5: UHF PD Measurement at Cable Accessories in Service Statistics show that most failures in the insulation of cable systems occur in accessories, such as joints and terminations. Critical PD defects in cable accessories can sensitively be detected by the UHF method, as is well reported in [5.13], [3.54]. Due to the strong attenuation of the ultra-high frequency spectrum, only PD signals adjacent to the PD source are captured. Therefore the UHF method can be characterized as spatially selective, which ensures a high signal-to-noise ratio. Consequently, the coupling sensors should be placed as close as possible to the test object.
  • 45. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 45 - PD activity detected within accessories is a clear signal for replacement. Consequently, those accessories that are about to fail should be replaced immediately, thereby reducing the risk of cable system failures. In the following a practical example will be presented which refers to on-site on-line UHF PD diagnostics of MV cable terminations (the described coupling sensors are also useful for off-line UHF PD measurements with an external test voltage source). To screen the termination effectively against external electromagnetic interferences a barrel sleeve is clamped around the plug-in connector (size 4, up to 72 kV) as shown in Figure 5.2-15. Inside of the sleeve there are three 5 cm long monopole antennas, each placed 1-2 cm above the cable surface and oriented at a tangent to the cable cross-section. Antennas are shifted by 120° from each other to embrace the whole circumference of the test object. The measured signal from the antenna via is fed to the digital oscilloscope via a 40 dB pre-amplifier (in a frequency range from 1 MHz to 1000 MHz). The calibration of the UHF method in terms of the amount of charge is impossible. However comparative PD measurements were performed on an identical test set-up assembled in laboratory, using both the UHF method and the conventional method according to IEC 60270 [1.2]. This was done to establish the maximum sensitivity of the UHF method. Raising the applied voltage controlled the magnitude of PD. A sensitivity, corresponding to 4 pC, was detectable by the UHF method. PD-measurements in an unshielded laboratory were performed on several 45 kV cable connectors. Connectors with simulated PD emitted fast electromagnetic pulses whereas measurements on connectors without PD showed only slight background noise. So the detection of UHF PD signals constitutes a strong sign for internal PD activity. Figure 5.2-16 shows such a typical fast pulse signal picked up on a connector with PD. The frequency spectrum of this pulse is plotted in Figure 5.2-17. Figure 5.2-16: Typical fast pulse emitted by some connectors Figure 5.2-17: Frequency spectrum of PD-pulse, acquired with a bandwidth of 3 GHz There are several high frequency components that indicate presence of PD activity in addition to the broadcast and GSM frequency spikes. The failure of this connector a few weeks after the measurement confirmed this diagnosis. Range of amplification Reproducible frequency components of the measured fast impulse Range of amplification Reproducible frequency components of the measured fast impulse 1 2 Figure 5.2-15: Test set-up consisting of (1) plug-in connector and (2) metallic barrel sleeve with antennas
  • 46. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 46 - 5.3 Rotating Machines For more than 50 years insulation systems of rotating machines (generators and motors) have been studied by applying technical diagnostic tests. Over the years there have been different standards for the testing of winding insulation: including national recommendations, international standards and several routine tests specified by customers. To ensure reliable operation and to increase the lifetime of rotating machines, on-site diagnostic tests have gained more importance over the last few years. 5.3.1 Pre-Conditions The following test methods are usually applied for on-site diagnostics:  Visual inspection  Insulation resistance (Charging and Discharging Currents: PDC, DRA Analysis)  Dissipation factor (tan  and capacitance as a function of test voltage and frequency  Partial discharge as a function over test voltage This chapter focuses on HV testing in conjunction with PD measurement. The standards for PD measurement of rotating machines are not at the same level as on-site PD measurement for other equipment such as GIS or cables. One reason for this is the complex frequency dependent signal propagation function of PD pulses within the extended stator winding. Recent publication dealing with PD measurements and the characterization of PD patterns include IEEE Std. 1434-2000 [5.14] and IEC 60034-27 [5.15]. The practice of PD measurement is different in various countries. In USA broad band measurement systems (up to several tens of MHz) have been common whereas in Europe small band measurement systems (e.g. 10 kHz bandwidth) have also been applied for detailed failure characterization and detection. For a risk evaluation of the insulation system condition, the PD magnitude is measured in the power station with the machine off-line. To be able to perform the testing on-site, adequate mobile high voltage sources and sensitive PD measurement equipment are needed. Due to the environmental effects, the background noise level can be high compared to laboratory conditions. Fortunately the PD magnitude of a new respectively healthy winding insulation is 1000 pC and more, depending on the insulation technology. In addition tan measurement can be performed. In this case environmental influence plays less of a role, but the temperature of the windings must be observed to get comparable results. Restrictions of PD measurements Due to the strong dependency of the PD signal propagation within a rotating machine, these machines cannot be treated as lumped devices in circuit diagrams. As a major consequence the PD magnitude itself has no clear meaning and gives no direct hint about the harmfulness to the machine of a PD source. Instead a trend analysis is more important and a comparison of results gained from machines of the same type, ratings and age allows a much better and more secure evaluation of the PD measurement. This is also one reason that PD magnitudes for rotating machines are widely given in mV instead of pC, as this gives users the erroneous belief that the reading are absolute and that for example a 10 nC discharge indicates the same amount of damage regardless of winding characteristics and machine design. Test system requirements The rated voltages of generators lie in the medium voltage range. Voltages of 6,3 kV or 10,5 kV are often used. It is necessary that the AC test voltage can be regulated in steps of 10% or 20% of the rated voltage. The generators have the dielectric property of a high capacitive load, which demands medium voltage test systems with high power in the range of some MVA or the capability to compensate the reactive load. Therefore ACRL and ACRF test systems are recommended. An AC test system based on transformers (ACT) can only be used for machines of lower power and low test voltages (< 20 kV). The ACRL or the ACRF circuit can be successfully used for testing rotating machines on-site. Two examples of realisations are given below.
  • 47. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 47 - 5.3.2 Example 1: On-Site Tests by ACTC and ACRL Test Systems and Conventional PD Measurement Technique The resonance test system with variable inductance (ACRL) shown in Figure 5.3-1 has a rated power of 350 kVA and two taps of 50 kV and 30 kV respectively. The whole test system is mounted into a 3 m container, together with control room and switching cubicle. As an example, the stators of a 42 MVA/10,5 kV and a 20 MVA/6,3 kV generator were tested with this mobile system. The HV test was done together with a PD and dissipation factor measurement. The capacitances of the generators to be tested were about 780 nF and 691 nF respectively. The maximum applied test voltages were 12 kV and 7,6 kV at 50 Hz. Figure 5.3-1: Mobile HV test system on-site in a pumped-storage power station In Figure 5.3-2 the results of the dissipation factor measurements are shown. Generator 1 shows an almost linear behaviour of tan  with increasing test voltage and generator 2 a characteristic bend at the PD inception voltage. In Figure 5.3-3 the test results of PD measurements are displayed as a function of applied test voltage. The PD magnitudes were evaluated as QIEC values according to IEC 60270 (2001). Figure 5.3-2: Measured dissipation factor tan  as function of test voltage for two generators Dissipation Factor Test Results 5,0 10,0 15,0 20,0 25,0 30,0 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 Test Voltage (50Hz) [V] tan[‰] Generator 1: 42MVA, 10.5 kV Generator 2: 20 MVA, 6.3 kV Ionization bend due to PD inception C. Sumereder, TU GRAZ
  • 48. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 48 - Figure 5.3-3: Measured partial discharge intensity and PD inception as function of test voltage for two generators 5.3.3 Example 2: On-Site Tests by ACRF Test System and Conventional PD Measurement Techniques The test was performed with an ACRF test systems on stator windings of turbines of ratings of 137 MVA/15,8 kV and 1640 MVA/27 kV. This ACRF system of test voltage generation is not affected by intensive partial discharges. Quality factor of the resonant circuit is smaller than with VPE cable test circuits, but the test power was sufficient, not only for single phase tests with 42 Hz respectively 46 Hz, but also to test the parallel connection of all three phases with approximately 26 Hz. The high voltage test was performed together with partial discharge and tan  measurement, the voltage being increased in steps. Figure 5.3-4: Compact resonant test system (ACRF), 36 kV/10 A for testing stator windings; 1 – reactor, 2a – basic load, 2b – cable to the test specimen, 3 – exciter transformer, 4 – frequency converter & control, 5 – PC and PD measuring system Results of partial discharge measurements made at 50 Hz were compared with results using the mobile ACRF test system (Figure 5.3-4). All measurements were made in accordance with common practice on generator windings with a wide-band partial discharge measuring instrument and with a band-pass filter characteristic (20 kHz to 20 MHz). In order to receive a reproducible partial discharge characteristic, the winding component to be tested was conditioned in each case with rated voltage Un for 5 minutes before the PD measurement was performed. The impulse distributions from the phase- dependent partial discharge patterns (-q-n-pattern) were recorded after each voltage step of 0,1 Un. Partial Discharge Level according to IEC 60270-2001 0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 Test Voltage (50 Hz) [V] PartialDischargeIntensityaccordingQIEC[pC] Generator 2: 20 MVA, 6.3 kV PD inception at 30% rated voltage Generator 1: 420MVA, 10.5 kV PD inception at 60% rated voltage Ionization bend due to PD inception C. Sumereder, TU GRAZ
  • 49. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 49 - Figure 5.3-7: tan  as function of normalized test voltage of a 16 kV stator winding; parameter is the test frequency; also shown is the calculated value for 50 Hz PD patterns are shown of the 27 kV stator winding of a nuclear power station generator, measured at rated voltage Un = 27 kV and at operating voltage 3/Un = 16,2 kV and a test frequency at 45,7 Hz (Figure 5.3-5 and 5.3-6). These PD patterns are typical for mica based machine insulating systems. The variation of the test frequency did not affect the typical partial discharge behaviour expected in the mica insulation of rotating machines. As PD activity will increase tan  the influence of the test voltage on this diagnostic quantity will be briefly discussed. To allow a consistent interpretation of the test results over decades of the entire machine life span, the test conditions must always be kept comparable. The dissipation factor tan  is by definition proportional to the test frequency. This enables a conversion of the test results to other frequencies. Therefore the actual effects of the test frequency on the dissipation factor were examined in comparative measurements. The voltage-dependent dissipation factor values were therefore determined with an ACRF test system at 26 Hz and 42 Hz at a 16-kV-stator winding. These results are represented in Figure 5.3-7 together with the conventional 50 Hz measurement. The curves at different frequencies are parallel-shifted, where, as expected, the higher test frequency leads to larger tan  values. The conversion of the different test frequencies to the 50 Hz reference value provided larger tan  values however. One possible cause of this anomaly may be the non-linear resistive behaviour of the materials in the glow protection units. The frequency dependant capacitive loss current of these units causes an under-proportional change of dissipation factor at reduced test frequency. The rise of the dissipation factor over the voltage, tan /U, which is an indication of the quality of the impregnation during manufacturing and of the degree of delamination with aging, is to a large extent independent of the test voltage frequency (Figure 5.3-7). Figure 5.3-5: PD pattern measured at Un = 16,2 kV (27 kV/sqrt (3)) at a 1640 MVA generator with an ACRF test system at 45,7 Hz Figure 5.3-6: PD pattern measured at Un = 27 kV at a 1640 MVA generator with an ACRF test system at 45,7 Hz
  • 50. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 50 - 5.3.4 Example 3: On-Site Tests by DAC Test System and Conventional PD Measurement Technique Using a damped AC system (Figure 5.3-8), the PD behaviour and the level of dielectric losses of four different motors of ages of between 5 and 10 years have been tested on-site (Table 5.3-1). Table 5.3-1: Overview of motors tested on-site. Rated voltage Rated power Machine No 1 6600 V 10750 kW Machine No 2 6600 V 12240 kW Machine No 3 6600 V 7840 kW Machine No 4 3300 V 4580 kW The test procedure was the same for all cases and consisted of a number of tests between the partial discharge inception voltage (PDIV) and 1.7 U0. For data analysis the following information has been evaluated:  The PDIV for phases U, V and W  The PD magnitude at PDIV  The PD magnitude up to 1,7 U0 The diagnostic parameters were collected for a series of machines as shown in Table 5.3-2. In particular the tests have been performed at both the HV and LV side of the stator coil. To explain this process the results obtained for the Machine No. 3 will be discussed below. In Tables 5.3-2 to 5.3-3 the measuring data is shown. In Figure 5.3-9 the PD behaviour as a function of the test voltage is shown. It can be concluded, that in comparison to HV connection of the machine at the LV side (star connection), a significantly higher PD activity has been observed. This evidence suggests that further tests are needed to identify the source and asses its harmfulness. Figure 5.3-8: Example of on-site PD diagnosis and with dielectric loss measurement of stator insulation, using damped AC voltages.
  • 51. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 51 - Figure 5.3-9: PD behavior as function of the test voltage, measured on a stator winding Table 5.3-2: PD magnitudes in [nC]; Table 5.3-3: PD magnitudes in [nC] detected at LV measured at HV side of the stator side of the stator 5.4 Power Transformers 5.4.1 General PD measurement during the acceptance test with induced voltage on power transformers has been used in the manufacturers’ test bays since the seventies. The determination of maximum allowed PD magnitudes during the AC voltage test is strongly dependent on the philosophy of AC voltage withstand testing. Up to the early eighties the test voltage levels continuously increased. Usually the test voltage level was applied for 6000 cycles (e.g. 50 Hz for 2 min.) In the early eighties a new test procedure was introduced with IEC 60076-3 [5.17]. There was no PD magnitude specified for test voltage levels but after that, the test voltage was kept at 1,3 / 3mU or 1,5 / 3mU for a period of 30 minutes (so-called method 2). During this time the PD magnitude was required not to exceed a HV Phase HV Phases Voltage L1 L2 L3 L1,L2,L3 1,3 kV 1,3 30 0.03 30 2,6 kV 25 68 9.6 25 3,9 kV 65 70 20 80 4,5 kV 65 80 25 100 6,6 kV 70 90 55 110 HV Phase HV Phases Voltage L1 L2 L3 L1,L2,L3 1,3 kV - - - - 2,6 kV 23 11 10 13 3,9 kV 25 25 12 15 4,5 kV 25 25 14 17 6,6 kV 25 35 15 25
  • 52. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 52 - certain limit. In 2000 the philosophy of having a short duration test (ACSD) at high voltage levels and a long duration test (ACLD) at lower levels was consequently introduced in the new edition of IEC 60076-3 [5.17]. The idea behind this test is to test the design of the transformer using a high voltage level (ACSD) and to check the quality of manufacturing by the long duration test (ACLD). Thereby, maximum PD magnitudes have been defined for the ACLD as well as for the ACSD test procedure. Recently, the on-site PD measurement gained more and more importance. This is mainly due to the lower cost of on-site repair of power transformers as an alternative to the transportation of the transformer back to the factory for repair. However, on-site repair of power transformers requires not only a certain manufacturing competence on-site but also the capability for on-site testing [2.10, 2.11, 5.18, 5.19]. Therefore, recent developments have been aimed into two directions, firstly PD measuring techniques including PD localization for on-site application have been developed and secondly, frequency converters have been developed as a new type of voltage source which are much easier to handle than motor-generator sets. 5.4.2 Preconditions of Test Object Various preparations have to be done on-site before a high-voltage test on a transformer can be performed. First of all it is essential to have the final acceptance testing protocol of the transformer determined, defining the test levels and the standards (usually IEC or IEEE) to which the tests have to be performed. Based on the final acceptance protocol and the defined test levels it will be possible to estimate the power is needed to perform the tests allowing the size of the test system itself to be adapted in advance. This requires the knowledge of the active and reactive power requirements during the induced voltage test. However, usually this data has never been recorded during the induced voltage test since the apparent power in the manufacturer’s transformer test bay is sufficient for the induced voltage test on all transformers. The situation is completely different on-site. The following aspects have to be taken into consideration:  Power supply of the test voltage source (motor-generator set or frequency converter) On-site the power for testing of transformers must be taken out of the power network or from a mobile diesel-generator set. The required power can be limited with respect to reactive power.  Compensation Sometimes on-site measurements are far away from civilisation. Therefore, if the compensation of reactive power is not sufficient, testing is not possible and the test voltage may not be able to be achieved. Organizing additional compensation within a short time period can be very difficult and depends on the test location. Therefore, it is extremely helpful to have a good estimate of the power consumption of the induced voltage test. Due to the limited power supply often available for on-site testing, the frequency and compensation can be adjusted in a way that the power consumption is minimised because the transformer is an inductance at lower frequencies and capacitance at higher frequencies (Figure 5.4-1) [5.15, 5.20, 5.21]. At the frequency of self-compensation the reactive power consumption by the test object is theoretically zero and practically extremely low. If this effect is used, it must be considered that the output power of the test voltage source may also depend on the frequency and on the relation between reactive and total power, expressed by the power factor cos  (Figure 5.4-2). Therefore, knowledge of the dependence of the capacitive and inductive behaviour on the frequency would be valuable. For older transformers this information is often not available. Experience of the test engineer and flexibility of the test equipment are therefore prerequisite for successful testing. Furthermore it is necessary to have information on environmental conditions on-site to be able to determine the location of the test system on-site, and the location of the power supply for the test source. One should also inspect the site to determine whether power equipment like overhead lines, surge arresters, etc. have to be modified or removed during the HV-tests in order to reach the required
  • 53. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 53 - distances between components at high voltage and ground potential. It is also necessary to check the climatic conditions expected during the test so that these can be taken into consideration. Figure 5.4-1: Test power demand of a 100 MVA power transformer as function of test frequency 10 100 1.000 10.000 0 50 100 150 200 250 Frequency in Hz TestingpowerinkWorkVA S P P active power [kW] S overall power [kVA] „self compensation“ capacitive loadinductive load 10 100 1.000 10.000 0 50 100 150 200 250 Frequency in Hz TestingpowerinkWorkVA S P P active power [kW] S overall power [kVA] „self compensation“ capacitive loadinductive load 40 60 80 100 120 140 160 180 200 0 50 100 150 200 250 300 350 400 450 500 550 2-phase capacitive load 2-phase inductive load 3-phase capacitive load 3-phase inductive load           )=-0.2 )=-0.8 )=1.0 )=0.8 )=0.2 )=-0.2 )=-0.8 )=1.0 )=0.8 )=0.2 cos( cos( cos( cos( cos( cos( cos( cos( cos( cos( PowerLimit(kVA) Test frequency (Hz) Figure 5.4-2: Available output power of a 500 kV frequency converter depending on frequency and power factor cos .
  • 54. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 54 - Following the above general preparations it is necessary to perform a number of preliminary on-site tests to assess whether the condition of the transformer will allow high voltage testing. For this condition assessment the following tests are necessary:  insulation resistance  winding resistance  transformation ratio  dielectric dissipation factor of transformer and bushings  frequency domain spectroscopy in order to determine the moisture content inside the solid insulation and  frequency response analysis in order to detect winding distortions. Finally the insulation liquid parameters have to be tested in order to ensure the oil has adequate breakdown voltage and low moisture content. If all these tests are passed, the set-up for the HV-testing can be arranged including shields on the HV-bushings to avoid corona signals and short circuits on the current transformers. Usually the induced voltage test with PD measurement is combined with other HVAC tests. The following AC test-sequence is recommendable unless otherwise agreed:  determination of no-load losses  applied voltage test LV/MV/HV  induced voltage test including partial discharge measurements. 5.4.3 Example 1: On-Site Test by Motor Generator Set and Conventional PD Measurement Technique The low-voltage side of the transformer under test is fed by the M/G set via a matching transformer. The induced voltage test can be performed as a single-phase or a three-phase test depending on the power transformer and the available motor-generator set (M/G-set) [5.22, 5.23]. On-site also three- phase power transformers can be tested with a single-phase M/G set (Figure 5.4-3 a), but a three- phase test (Figure 5.4-3 b) is recommended if a related M/G set is available. The power supply of the M/G set must be able to provide the required active power. The generator must be able to provide the active as well as the reactive power required by the test object. Therefore, compensation of the reactive power has to be taken into account. Usually the compensation does not need to be very exact if the generator can also provide a certain amount of reactive power. For conventional PD measurements in the test bay (according to IEC 60270 [1.2], see also chapter 3.1), the transformer bushings are usually used as coupling capacitors. Some years ago so- called “radiation noise detectors” (selective signal measurement) were frequently in use – even in manufacturers’ test bays. However, special multi-channel PD measuring systems are now in service. The preparation of the PD measurement starts with the calibration procedure. During the test, PD magnitudes at different test voltages are recorded by the PD measuring system. A synchronised multichannel PD measuring system is another measurement technique for three phase power transformers to localize the PD source. It offers the possibility of improving the S/N ratio and the localization of the PD source [5.24]. On-site PD measurement can be difficult because of the presence of external disturbances. One method to distinguish internal from external PD, is to correlate signals from AE sensors or piezoelectric transducers (mounted on the tank wall) together with the electrical PD measurement. If the microphones show signals the PD is generated inside the transformer with high probability. Otherwise, it is an external PD. However, the electrically measured PD signals contain both, the internal and the external partial discharges. Therefore, the sources of external corona discharges should be removed as far as possible. An extremely useful tool is an UV camera which is able to detect the ultra violet (UV) emission of external partial discharges caused by corona discharges, e.g. at the bushings. In principle, acoustic corona discharge detection would be also possible.
  • 55. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 55 - Figure 5.4-3: Circuit diagram of the induced voltage test arrangement (without compensation) a. single phase test b. three phase test Figure 5.4-4: On-site test of a 740 MVA/420 kV/27 kV transformer using a M/G set on PD measurement according to IEC 60270 However, an exact location of corona discharges is usually difficult using only acoustic detectors. During an induced voltage test of a 740 MVA power transformer (420/27 kV) (Figure 5.4-4) pollutions on the insulator of one of the bushings resulted in corona discharges and in a measured PD magnitude of about 1000 pC. Cleaning of the entire bushing is extremely time-consuming. In that case the UV camera was useful in order to detect all corona discharge sources within a reasonable time span of a few hours. 5.4.4 Example 2: On-site test by Frequency Converter and Conventional PD Measurement Technique The use of a static frequency converter system for high voltage testing of power transformers [5.19- 5.21] allows a simple and rapid set-up on-site, because the whole HV test system can be installed into one container as shown in Figure 2.3 [5.20]. This allows high mobility because the system can be transported easily by road or by ship. Mobile systems of power ratings of up to 500 kVA have been in use for some years. That power is sufficient for most transformers. Recently systems with a power in a range of 1 MVA have become available, removing the restrictions in testing even the largest GSU-transformers. As static frequency converter are operating with semi-conductors like IGBTs (insulated gate bipolar transistors), voltage transients are produced leading to an increasing PD measurement noise level in certain frequency ranges. Therefore special PD measurement methods have been developed [5.25] to address this issue. Figure 5.4-5 shows a mobile test system based on a frequency converter when testing a 400 MVA GSU transformer. The system is provided for applied voltage testing (a, b), induced voltage testing (c, d) and loss measurements. The applied voltage test is similar to GIS testing (see chapter 5.1), whereas the induced voltage test requires the three-phase configuration shown in Figure 5.4-6 (as single line diagram). The three-phase output of the converter is adjusted to the required input voltage of the transformer under test by a matching transformer with compensation on both sides. With respect to noise suppression a medium-voltage filter is applied and the filter capacitor is also used as a voltage divider. The three bushings of the transformer under test are used for synchronous PD measurement by a multi-channel PD measuring system. The whole test is computer controlled.
  • 56. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 56 - (a) (b) (c) (d) Figure 5.4-5: Mobile test system based on a frequency converter when testing a 400 MVA GSU transformer. (a) and (b) show the applied voltage test and (c) and (d) show the induced voltage test The low-voltage side of the 400 MVA GSU transformer under test is connected to the step-up transformer by three medium-voltage cables. The test frequency adjusted by the self-compensation (see chapter 5.4.2) is 140 Hz, the test voltage 80% of that for routine tests and the step-test procedure according to Figure 1.3 is applied with a duration of 43 s at the highest level. As can be seen from Figure 5.4-7, a sensitivity ranging below 15 pC is achievable, although the sensitivity is different from case to case depending on surrounding noise influences. The figure shows the PD behaviour of one phase during the highest test voltage, where no PD signals could be seen resulting in the test being passed. Figure 5.4-6: Block diagram of an induced voltage test based on a frequency converter (single line diagram of a three-phase arrangement)
  • 57. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 57 - Before the HV test can be performed the PD measurement equipment has to be installed and calibrated. Figure 5.4-8 shows the spectrum of the noise level with and without the frequency converter operating and the calibration signal. Based on a comparison of these curves, the mid- frequency is chosen in such a way that the highest sensitivity is reached, but as low a frequency as possible (at least in a range of a few 100 kHz). If this frequency is found, which in this case was around 1 MHz, the used spectrum analyzer operates for the measurement in “zero span” mode at the selected mid-frequency and the whole system operates as a tuneable filter. Using this frequency-tuned resonant test system, transformers of up to 500 MVA and 400 kV have been successfully tested in Europe and South East Asia without having major problems in transportation. Figure 5.4-7: Noise below 15 pC at induced withstand voltage Figure 5.4-8: PD noise frequency spectrum and calibration level 5.4.5 Example 3: UHF PD Measurement and Acoustic PD Localization in Service and for On-Site Tests Partial discharges under oil emit acoustic and also electromagnetic waves with frequencies up to the UHF range (300 MHz to 3000 MHz). The electromagnetic waves are detectable with UHF probes. As a result of the shielding characteristics of the transformer tank against external electromagnetic waves, normally signals can clearly be distinguished between internal and external PDs. The UHF probes can be installed at an oil flange during on-site testing or even during operation. The probes withstand all mechanical and thermal stresses in the transformer and can be used as a permanently installed measuring device. A calibration comparable to the measurement according to the IEC60270 [1.2] is impossible for UHF sensors. Therefore it is necessary to examine the installed probes for its function with a performance check [3.62]. In Figure 5.4-9 an UHF signal measured at a 200 MVA single-phase transformer and its corresponding spectrum is shown. a) b) Figure 5.4-9: a) UHF PD signal on a 200 MVA single-phase transformer; b) corresponding UHF PD spectrum 0.00 0.05 0.10 0.15 0.20 -0.03 -0.02 -0.01 0.00 0.01 0.02 0.03 0.04 amplitude(V) time (us) online UHF signal 0 250 500 750 1000 1250 1500 0.0 0.1 0.2 0.3 0.4 0.5 0.6 spectrum of online UHF signal frequency (MHz) amplitude(Vs)
  • 58. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 58 - Identification and filtering of on-site disturbances During on-site measurements sometimes UHF noise signals are detectable even though the sensor is inside the tank of the transformer. Processing of these measured signals can show known narrowband disturbances. Disturbances around 0,5 GHz are caused by the digital video broadcasting service, at nearly 0,9 GHz and 1,8 GHz are the mobile phone disturbances and at about 2,1 GHz UMTS signals are detectable. Because these disturbances have a small bandwidth, a narrowband filter can be applied easily. Narrow-band measurement systems with adaptive filtering are recommended. With narrow-band systems disturbances can be suppressed simply by measuring “next to them” in an unaffected frequency range. Another powerful tool for field measurements is phase resolved UHF PD measurement. Internal PDs produce phase stable UHF signals and phase-resolved PD patterns (PRPD) can identify interference and disturbances because of their non-phase-stable occurrence, especially if the test frequency is different from power frequency. Localization of PD by combining UHF and acoustic methods Two main tasks are encountered concerning PD measurements. The first is to provide evidence of PD e.g. a “yes/no” decision. The second and most important is the determination of the PD location (localization). The on-site PD detection sensitivity may be hampered in the conventional case whereas the non-conventional methods such as electromagnetic (UHF) and acoustic, do generally not suffer from external disturbances. Application of the two unconventional methods can provide an advantageous combination for optimised detection and localization. Other methods for the localization of the PD based on the acoustic signal arrival time include three different approaches for the system of non-linear observation equations. Depending on whether mixed-acoustic (e.g. electric or electromagnetic triggering) or all-acoustic (acoustic triggering) measurements are used, the equations have three (space coordinates (x, y, z) of the PD) or four unknowns (an additionally unknown temporal origin). A new approach to acoustic signal processing works with so called pseudo-times, allowing the usage of robust direct GPS (Global Positioning System) solvers instead of the previously used iterative algorithms. In the presence of inevitable measuring errors, sensitivity limits or wrongly assumed acoustic propagation velocities much more stable results were featured by the direct GPS solver. Another important part of the localization procedure is a correct objective arrival time determination. Here good experiences have been gathered with signal-energy based criteria [5.26]. Increasing the acoustic signal-to-noise ratio with an averaging process (continuous evaluation of the mean value) has long been known and used. To be successful a stable trigger is required, signal and noise should be uncorrelated and the noise should be white. In terms of PD measurements and the goal to de-noise acoustic signals (e.g. to quantify their arrival times hidden in the noise) one needs to have a physical signal related to the PD with a significantly higher sensitivity than the acoustic one. Here, electromagnetic UHF PD signals have proven their applicability. During the averaging process, the noise contained in the acoustic signal tends towards its statistic mean value, which is zero if white noise is assumed. The acoustic signal itself is superimposed constructively and the presence of an acoustic signal with a stable relation to the UHF trigger can be verified with high sensitivity. The theoretically maximum signal-to-noise-ratio gain is N x 0,5, where N is the number of superposition. For the example of Figure 5.4-10, within a single acoustic PD impulse of a 132 pC (electric equivalent PD), there is no clearly observable information above the noise level in the acoustic signal. Using the same set-up with UHF-triggered averaged acoustic signal with 500 superimpositions of PD impulses of maximums of only 9 pC, a clear acoustic impulse can be seen, which allows the determination of the travelling time of the impulse. So, a successive localization is possible [5.27]. Figure 5.4-10 shows this comparison of a single acoustic impulse of a 132 pC with no clearly observable information and 500 super-impositions of maximum 9 pC where a clear impulse is visible (same experimental arrangement, same sensor position).
  • 59. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 59 - Figure 5.4-10: Comparison of a 132 pC single acoustic impulse and 500 superpositions of maximum 9 pC Application on Power Transformers: all-acoustic on-site/on-line PD localization in a 200 MVA, 380/220 kV-single-phase transformer Over a period of several months all-acoustic on-site measurements were performed on a 200 MVA single-phase transformer. Its gas-in-oil diagnosis indicated partial discharge. During an offline applied voltage test an electric PD measurement revealed PD magnitudes up to 600 pC and an autonomous acoustic measurement recorded an impulse on four sensors simultaneously. The oil-temperature was about 26°C (which corresponds to 1387 m/s sound velocity in oil) according to information from the operating company. Figure 5.4-11 pictures the top view of the housing of the transformer with the sensor positions and the average of the calculated PD location. Figure 5.4-11: Top view of the housing of the 200 MVA transformer with attached acoustic sensors and the calculated PD location inside the transformer In Table 5.4-1 results of location are summarized. The result of “measurement 1” was determined iteratively because the direct solver generated a complex solution due to strongly erroneous time information. For the remaining two measurements the direct solver was used.
  • 60. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 60 - Tab. 5.4-1: Acoustic PD measurements over a period of several months with changing sensor positions (Figure 5.4-11) calculated PD-origin x [m] y [m] z [m] measurement 1 (offline a, b, c, d) 1,40 3,12 2,27 measurement 2 (online c, d, e, f, g) 1,25 3,19 2,23 measurement 3 (online a, c, d, e, f, g) 1,27 3,22 2,19 5.4.6 Example 4: Synchronous Multi-Terminal PD Measurement for On-Site Testing The interpretation of phase resolved (PRPD) patterns from on-site transformer measurements might be difficult because of the external disturbances and the cross-talk of the transformer windings. Evaluation of synchronous multi-terminal PD measurements establishes a straight forward approach to remove external disturbances and, furthermore, to distinguish between multiple PD and noise sources [3.16, 3.17, 3.28]. The synchronous multi-terminal PD measurement is based on the standard measurement circuit of IEC 60270. For PD measurement on power transformers a three phase measuring system is used, in which all three phases are measured synchronously. Therefore one PD-impulse is measured at all three phases. The measuring result is illustrated in a STAR diagram [5.24, 3.67]. The STAR diagram is a two-dimensional plot with a 120° phase shift of the three phase axis. Figure 5.4-12 (right) shows the impulse signals on all phases. In this example the PD source is located on phase L1 and the PD signals of phase L2 and L3 occur because of the cross-talk of the windings. An addition of the signal amplitude vectors of a single PD activity (value in pC) of the three phases builds a point in the diagram. In this example the point is close to L1 and therefore the PD source is located in phase L1. Assuming that each individual PD fault leads to a unique cluster, each cluster represents one specific fault location within the transformer. External disturbances, like corona or noise, can be measured with the multi-terminal IEC circuit on all three phases. Measured external impulses often possess similarities of the shape and amplitude. Consequently the vector addition of those impulses leads to clusters located next to the point of origin. Therefore internal PD and external impulse sources can be differentiated from each other. Retransformations of clusters of the STAR diagram into a PRPD pattern are possible. Therefore, a cluster can be extracted from the STAR diagram and can be illustrated in the well-known PRPD pattern. The selected cluster and the retransformation of a single PD sources enables pattern recognition. The results of an onsite, offline measurement carried out on a 110/220 kV, 210 MVA grid-coupling transformer by means of a motor generator set showed PD activity at operating voltage level in the transformer [5.27]. The measurements were done on the 110 kV and the 220 kV side. For demonstrating purpose a copper wire was fastened to the conductor of the bushing on phase L3,110 (Figure 5.4-13) representing an external corona with a considerable low amplitude of 30 pC. Figure 5.4-14 shows the PRPD pattern of an internal PD on phase L2,110 of the transformer and some disturbances from the corona on phase 3. Because the PD level of the corona is quite small the coupling from phase L3,110 to L2,110 is not very significant, but can be depicted in the rectangle of Figure 5.4-14. For phase L1,110 and L3,110 internal PD sources could not be detected and no cross-coupling from the active phase L2,110 was visible in the PRPD pattern. The STAR diagram in Figure 5.4-15 confirms the results of the two PD sources. One cluster is close to phase L2,110 (rectangle) which is the internal measured PD and the external corona discharge on L3,110 (ellipse).
  • 61. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 61 - Figure 5.4-12: STAR diagram evaluation of a PD impulse signal by means of vector addition of a three phase measurement Figure 5.4-13: Copper wire (here indicated in black) as artificial defect on phase L3,110 with bushing and connector between conductor and coupling capacitor Figure 5.4-14: PRPD-pattern on phase L2,110 disturbed by corona on phase L3,110 (marked rectangle) Figure 5.4-15: STAR diagram of the PD activity (rectangle showing internal PD and ellipse showing corona discharge at attached wire) Figure 5.4-16: PRPD-pattern on phase L2,110 after retransformation without corona disturbances The retransformation of all partial discharges on phase 2 (rectangular cluster) is shown in Figure 5.4-16. Now the pattern does not include the disturbances from the external corona (see rectangle in Figure 5.4-16). In the case of a stronger external PD this effect would be more pronounced. Thus PRPD patterns can be de-noised by means of the STAR diagram and retransformed into specific clusters. Furthermore the PRPD pattern of just one partial discharge can be generated out of overlapping PD patterns in order to simplify pattern recognition [5.24].
  • 62. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 62 - 6. Conclusion Based on practical examples this brochure deals with the state of the art of high voltage (HV) on-site testing in combination with partial discharge (PD) measurements. The aim of this brochure is to give background information and helpful hints in performing such tests on various types of apparatus. In addition to type and routine testing in the factory, HV on-site testing is an important part of quality assurance. On-site tests are applied  as a part of commissioning of equipment on-site to demonstrate that the transport from the factory to site and erection on-site have not caused any new and critical defects which lower the dielectric withstand voltage of the insulation below the withstand level of the insulation coordination determined values;  after on-site repair to demonstrate that the equipment has been successfully repaired;  for diagnostic purposes in the framework of asset management by providing reference values of diagnostic indicators (e.g. partial discharge values and dielectric parameters) for later tests. Warnings from the PD monitoring system are a critical trigger in the condition-based maintenance program. HV withstand testing plays an important role in quality assurance of HV apparatus. However, additional measurements of the dielectric properties, such as significant PD quantities and the loss factor tan δ may enhance the reliability and remaining life time of electrical insulation. The common physical background of the measured phenomena is the behaviour of dielectrics in the electric field. Therefore the tests and measurements in the factory, on-site and in-service should be harmonized. This includes the possibility of improvement of on-site test procedures supported by partial discharge measurement. HV withstand voltage testing (e.g. up to 2,5 U0) of defect-free/non aged insulation does not have a destructive influence on the service life of the tested component. Due to life consumption of installed HV equipment, however, lower AC test voltages are recommended for on-site testing. AC voltage test levels higher than 1,0 U0 of defective/aged insulation may have a destructive influence on the service life of the component, even if no breakdown has occurred. Therefore a combination of AC voltage on- site testing with PD measurement is recommended. The PD measurement can provide an indication of whether the tested component is in comparable condition following the withstand test as it was before the test. In this context PD measurements play an important role in both quality testing and monitoring for condition assessment. In contrast to most dielectric measurements (tan , r, dielectric response parameters) which are characteristics of the whole dielectric volume, partial discharges are “weak point phenomena”. In the case of new insulation, the critical defects/weak points are the result of assembly failure which can be found by a routine test consisting of the HV withstand test combined with PD measurement. In case of service aged HV equipment insulation (which has been tested successfully and which operates for years) a critical defect may be caused by high electrical, thermal or mechanical stresses and by the “aging” of the insulation itself. This means that the partial discharges can be symptoms and/or the results of the aging process which causes – over a more or less long period – the above mentioned “weak point”. As there are several methods of on-site test voltage generation which can be applied to withstand testing and on-site testing, this brochure presents and discusses the various methods and the commonly used HV sources and PD measurement techniques for on-site testing of HV insulation. Typical methods applied to the on-site testing of HV apparatus, such as gas-insulated systems (GIS), cable systems, rotating machines and power transformers, are presented in this brochure. The benefits and the limitations of on-site HV testing combined with sensitive PD measurement for this apparatus is also discussed.
  • 63. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 63 - 7. References [1.1] W. Hauschild et al.: CIGRE WG 33.03/TF 04: Proposed requirements for HV withstand tests on-site. Electra No. 195, pp. 13–21, 2000 [1.2] IEC 60270: 2000: HV test techniques – Partial discharge measurements [1.3] M. Muhr et al.: Sensors and sensing for non-conventional PD detection. Report D1-102, CIGRE Paris, 2006 [1.4] IEC 60060-1: 1989-11: HV test techniques – Part 1: General definitions and test requirements [1.5] IEC 60060-2: 1994: HV test techniques – Part 2: Measuring Systems [1.6] IEC 60060-3: 2006-02: HV test techniques – Part 3: Definitions and requirements for on-site testing [1.7] IEC 60071-1: 1993-12: Insulation coordination. Part 1: Definitions, principles and rules [1.8] IEC 885-3: 1988: High-voltage test techniques – Test methods for partial discharge measurement of extruded power cables [2.1] E. Gulski, P., Cichecki, F.J. Wester, J.J, Smit, R. Bodega, T.J.W.H. Hermans, P.P. Seitz, B. Quak, F. de Vries: On-site testing and PD diagnosis of high voltage power cables. IEEE transactions on dielectrics and electrical insulation 15(6), 1691-1700, 2008, (ISSN 1070- 9878) [2.2] H. Michitsch, J. Spiegelberg, W. Golbig: HV tests of large rotating machines for commissioning and for diagnosis on the installation site (in German). ew (Elektrizitätswirtschaft) 104,17/18, pp. 62-70, 2005 [2.3] E. Gockenbach, W. Hauschild: The selection of the frequency range for HV on-site testing of extruded insulation cable systems. IEEE Electrical Insulation Magazine, Vol. 16, No. 6, pp. 11-16, 2000 [2.4] IEC 60840: 2004: Power cables with extended insulation for rated voltages 30 kV up to 150 kV – Test methods and requirements [2.5] IEC 62067: 2001: Power cables with extruded insulation for rated voltages 150 kV up to 500 kV – Test methods and requirements [2.6] IEC 62271: 2003: High-voltage switchgear and controlgear, Part 203: Gas-insulated metal- enclosed switchgear for rated voltages above 52 kV [2.7] W. Hauschild, P. Coors et al.: The technique of AC on-site testing of HV cables by frequency-tuned resonant test systems. CIGRE Conf., report 33-304, 2002 [2.8] W. Hauschild, S. Schierig; P. Coors: Resonant test systems for HV testing of super-long cables and gas-insulated transmission lines. 14 th ISH Beijing, Paper J-02 (p. 558), 2005 [2.9] Th. Aschwanden, A. Schenk et al.: On-site repair and refurbishment and HV tests of large power transformers in the transmission grid. CIGRE Report A2-203, 2004 [2.10] J. C. Mendes, R. A. Marcondes, J. Nakamura: On-site tests on HV power transformers. CIGRE Report A2-211, Paris 2004 [2.11] P. Werle: On-site test of power transformers. HIGHVOLT Kolloquium ’07, Paper 4.5, 2007 [2.12] A. Winter, P. Coors, U. Stephan: A mobile transformer test system based on a static frequency converter. 15 th ISH Ljubljana, Paper T10-T732, 2007
  • 64. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 64 - [2.13] IEEE 400.3-2006: Guide for partial discharge testing of shielded power cable systems in a field environment [2.14] CIGRE WG 33.03: Experiences in partial discharge detection of distribution power cable systems. Electra, p. 34-43, 2003 [2.15] E. Gulski, F. J. Wester, J. J. Smit, P. N. Seitz, M. Turner: Advanced PD diagnosis of MV power cables using oscillating wave test system. IEEE El Magazine March/April, Vol. 16 No. 2, 2000 [2.16] E. Gulski, J. J. Smit, H. van Breen, F. de Vries, P. P. Seitz, F. Petzold: Partial discharges diagnosis of stator insulation using damped ac voltages. Proceedings International Conference on Condition Monitoring and Diagnosis 2-5 April 2006, Changwon, Korea, 2006 [2.17] IEEE 400-2001: IEEE Guide for field testing and evaluation of the insulation of shielded power cable systems [2.18] F.J. Wester: Condition assessment of power cables using PD diagnosis at damped AC voltages. ISBN 90-8559-019-1, PhD thesis TU Delft, 2004 [2.19] W. Boone, C. C. Damschild et al.: VLF HV generators for testing cables after laying. 5 th ISH Braunschweig, paper 62-04, 1987 [2.20] G. C. Damstra, W. Zhang: Series resonant inverter with triggered vacuum gaps. ISDEIV Conference Moscow, Vol. 2259, pp. 270-272, 1994 [2.21] H. Emanuel et al.: A new high voltage dielectric test system for insulation diagnosis and partial discharge measurement. NORD-IS, 1996 [2.22] C. Forssén, H. Edin: Modelling partial discharges in a cavity at different applied frequencies. CEIDP, 2007 [2.23] G. Voigt, P. Mohaupt: Partial discharge measurements on service aged medium voltage cables at different frequencies. JICABLE, 2003 [2.24] M. Kuschel, R. Plath, I. Stepputat, W. Kalkner: Diagnostic techniques for service- aged XLPE-insulated medium voltage cables. REE, 1996 [2.25] R. Plath, W. Kalkner, I. Krage: Comparison of diagnostic systems for service-aged PE/XLPE-insulated medium voltage cables. Elektrizitätswirtschaft, Heft 20, 1997 (in German) [2.26] K. Rethmeier, P. Mohaupt, S. Seifert, V. Bergmann, W. Kalkner: Sinusoidal 0.1 Hz VLF test voltage for diagnostic PD measurements of high voltage cable systems. 15 th ISH Ljubljana, T7-448 [3.1] E. Lemke, S. Berlijn, E. Gulski, M. Muhr, E. Pultrum, T. Strehl, W. Hauschild, J. Rickmann, G. Rizzi: Guide for partial discharge measurements in compliance to IEC 60270. CIGRE WG D1.33, Technical Brochure 366, ELECTRA No. 241, pp. 61-67, 2008 [3.2] E. Gulski et al.: Guidelines for un-conventional PD measurements. CIGRE WG D1.33 (TF 03), Technical Brochure 444, 2010 [3.3] IEC 64278: High voltage test techniques – Measurement of partial discharges by electromagnetic and acoustic methods (under preparation) [3.4] CIGRE WG 21-03: Significance of discharge detection. ELECTRA 11, pp. 53–60, 1969 [3.5] CIGRE WG 21-03: Recognition of discharges. ELECTRA 11, pp. 61–98, 1969
  • 65. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 65 - [3.6] E. Gulski, D. Allan, T.R. Blackburn, A. Contin, E. Gockenbach, E. Lemke, T. Lundgaard, T. Mizutani, G.C. Montanari, M. Muhr, T. Okamoto, B.T. Phung, H. Sedding, H.J. van Breen, F.J. Wester: Knowledge rules for partial discharge diagnosis in service. CIGRE JWG-TF 15.11 / 33.03.02, Technical Brochure No. 226, 2003 [3.7] K. Schon: Konzept der Teilentladungsmessung bei Teilentladungsprüfungen. ETZ Archiv Vol. 8, No. 9, pp. 319-324, 1986 [3.8] E. Lemke, P. Schmiegel, H. Elze, D. Russwurm: Experience in the calibration technique for PD calibrators. ERA Conference on High Voltage Measurements and Calibration. Milan, Italy, 1966 [3.9] W. Lukas, K. Schon, E. Lemke, H. Elze: Comparison of two techniques for calibrating PD calibrators. 10 th ISH Montreal, 1997 [3.10] E. Lemke: Moderne TE-Mess-Systeme und ihre Kalibrierung. HIGHVOLT Kolloquium ’97 Dresden, paper 7.7, pp 95-106, 1997 [3.11] C. Cherbaucich, G. Rizzi, R. Gobbo, G. Pesavento, O. Saracco, A. Sardi: Partial discharge calibrators: Practical experience in their calibration. ERA Conference on High Voltage Measurements and Calibration, London, U.K., 1998 [3.12] E. Lemke, E. Gockenbach, W. Kalkner: Messtechnik für die Diagnose elektrischer Betriebsmittel. ETG-Fachtagung „Diagnostik elektrischer Betriebsmittel“, Berlin, 2002 [3.13] E. Gulski: Computer-aided recognition of Partial Discharges using statistical tools. Ph.D. Thesis, TU Delft, 1991 [3.14] A. Claudi et al: Instrumentation and measurement for in-service monitoring of HV insulation. CIGRE TF D1.02.08, Technical Brochure No. 286, 2005 [3.15] K.D. Plath, R. Plath, H. Emanuel, W. Kalkner: Synchrone dreiphasige Teilentladungsmessungen an Leistungstransformatoren vor Ort und im Labor. ETG-Fachtagung „Diagnostik elektrischer Betriebsmittel“, Berlin, paper 2.1, 2002 [3.16] R. Plath: Digital measurements and analysis of partial discharges – drawbacks and opportunities. HIGHVOLT Kolloquium ’07, Dresden, paper 2.1, pp. 63-71, 2007 [3.17] W. Koltunowicz, R. Plath: Synchronous multi-channel PD measurements. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 15, No.6, pp. 1715-1723, 2008 [3.18] E. Lemke, T. Strehl, M. Boltze: Advanced diagnostic tool for PD fault location in power cables using the CDA technology. 13 th International Symposium on High Voltage Engineering, Bangalore, India, paper 6-72, 2001 [3.19] E. Lemke, E. Gulski, W. Hauschild, R. Malewski, P. Mohaupt, M. Muhr, J. Rickmann, T. Strehl, F.J. Wester: Practical aspects of the detection of partial discharges in power cables. CIGRE WG D1.33.TF 05, Technical Brochure 297, ELECTRA No. 232, pp. 63-70, 2006 [3.20] M. Hoof and R. Patsch: Analyzing Partial Discharge Pulse Sequences – A New Approach to Investigate Degradation Phenomena. Conference Record of the IEEE International Symposium on Electrical Insulation, Pittsburgh, PA, USA, pp. 327-331, 1994 [3.21] L. Satish, B. Nazneen: Wavelet-based denoising of partial discharge signals buried in excessive noise and interference. IEEE Transactions on Electrical Insulation, Vol. 10, pp. 354-367, 2003 [3.22] A. Cavallini, A. Contin, G.C. Montanari, F. Puletti: Advanced PD interference in on-field measurements, part I: Noise rejection. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 10, No. 2, pp 216-224, 2003 and A. Cavallini, A. Contin, G.C. Montanari, F. Puletti: Advanced PD interference in on-field
  • 66. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 66 - measurements, part II: Identification of defects in solid insulation systems. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 10, No. 3, pp 528-538, June 2003 [3.23] S. Happe, H.-G. Kranz, W. Krause: Advanced Suppression of Stochastic Pulse Shaped Partial Discharge Disturbances. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 12, No. 6, pp. 265-275, 2005 [3.24] N. Hayakawa, Y. Sugimori, H. Okubo: Partial discharge current pulse waveform analysis (PD-CPWA) based on PD mechanisms in solid and gas media. Proceedings of International Symposium on Electrical Insulating Materials, Kitakyushu, Japan, pp. 812-815, 2005 [3.25] S. Sriram, S. Nitin, K.M.U. Prabhu, M. J. Bastiaans: Signal denoising techniques for partial discharge measurements. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 12, No. 6, pp. 1182-1191, 2005 [3.26] J.-T. Kim, J.-H. Kim, J.-Y. Koo: Noise discriminations in measuring partial discharges using pulse wave shape analysis. CMD 2006, Changwon, Korea, 2006 [3.27] T.-K. Abdel-Galil, Ayman H. El-Hag, A.M. Gaouda, M.M.A. Salama, R. Bartnikas: De-noising of partial discharge signal using Eigen-decomposition technique. IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 15, No.6, pp. 1657-1662, 2008 [3.28] K. Rethmeier, M. Krüger, A. Kraetge, R. Plath, W. Koltunowicz, A. Obralic, W. Kalkner: Experiences in on-site partial discharge measurements and prospects for PD monitoring. CMD Beijing, 2008 [3.29] S.A. Boggs, G.L. Ford, R. C. Madge: Coupling devices for the detection of partial discharges in gas-insulated switchgear. IEEE Transactions on Power Apparatus and Systems, Vol. 10, pp. 3969-3973, 1981 [3.30] S.A. Boggs, G.C. Stone: Fundamental limitations in the measurement of corona and partial discharges. IEEE Transactions on Electrical Insulation, Vol. 17 pp. 143-150, 1982 [3.31] Y. Takahashi: Diagnostic methods for gas-insulated substations. IEEE Transactions on Electrical Insulation, Vol. 21, No. 6, pp. 1037-1043, 1986 [3.32] E. Lemke, H. Elze, W. Weissenberg: Experience in PD diagnostics of HV cable terminations in service using an ultra-wideband PD probing in the real-time mode. 13 th International Symposium on High Voltage Engineering, Delft, The Netherlands, paper 404, 2003 [3.33] E. Lemke, T. Strehl, W. Weissenberg: Practical experience in on-site PD diagnosis tests of HV cable accessories in service. IEEE International Symposium on Electrical Insulation, Toronto, Canada, pp. 498-501, 2006 [3.34] B. Krampe, C. Neumann: Betriebsverhalten mit der UHF-TE-Messung an gasisolierten metallgekapselten Schaltanlagen. ETG Fachbericht 87, VDE Verlag, Berlin und Offenbach, 2002 [3.35] R. Kurrer, K. Feser: TE-Messung an SF6-Anlagen im UHF-Bereich. ETG Fachbericht 56, VDE Verlag, Berlin und Offenbach, 1995 [3.36] O. Rudolph, A. Breuer, P. Erning: Vorort-Diagnose von gasisolierten Schaltanlagen: Teilentladungsmessung. IEH-Prüftechnik, Symposium, Karlsruhe, 2000 [3.37] U. Riechert A. Đurđić, M. Kudoke, M. Stanek: Monitoring und Diagnose von gasisolierten Schaltanlagen - Entwicklungstrends und Einsatzmöglichkeiten. ETG Fachtagung „Diagnostik elektrischer Betriebsmittel“, Kassel, 2006, ETG-Fachbericht 104, 2006, S. 161- 166, VDE Verlag GmbH, Berlin und Offenbach, ISBN 978-3-8007-2977-7 [3.38] S. Neuhold, Th. Heizmann, R. Bräunlich; D. Koechli, U. Riechert, Ch. Dehne: Experiences with UHF PD detection in GIS using external capacitive sensors on windows and disk-
  • 67. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 67 - insulators. 15th International Symposium on High Voltage Engineering, Ljubljana, Slovenia, paper T7-480, 2007 [3.39] S. Hoek, M. Bornowski, S. Tenbohlen, T. Strehl, U. Riechert: Partial discharge detection and localization in gas-insulated switchgears. Stuttgarter Hochspannungssymposium, Conference Record, pp. 211-218, 2008, ISBN 978-3-00-024098-0 [3.40] W.R. Rutgers, Y.H. Fu: UHF PD-detection in a power transformer. 10th ISH, Montreal, Canada, 1997 [3.41] M.D. Judd, G.P. Cleary, S. Meijer: Testing UHF partial discharge detection on a laboratory based power transformer. Conf. Proc. 13th ISH, Delft, Netherlands, 2003 [3.42] S. M. Markalous, S. Tenbohlen, K. Feser: Improvement of acoustic detection and localization accuracy by sensitive electromagnetic PD measurements under oil in the UHF range. 14th ISH, Beijing, China, 2005 [3.43] S. M. Markalous: Detection and location of partial discharges in power transformers using acoustic and electromagnetic signals. Ph.D. Thesis, University of Stuttgart, Germany, 2006 [3.44] S. Tenbohlen, D. Denissov, S.M. Hoek, S.M. Markalous: Partial discharge measurement in the ultra high frequency (UHF) range. IEEE Transactions on Dielectrics and Electrical Insulation (ITDEIS), Vol. 15, No.6, pp. 1544-1552, 2008 [3.45] S. Markalous, S. Tenbohlen, K. Feser: Detection and location of partial discharges in power transformers using acoustic and electromagnetic signals. IEEE Transactions on Dielectrics and Electrical Insulation (ITDEIS), Vol. 15, No.6, pp. 1576-1583, 2008 [3.46] J. Lopez-Roldan, T. Tang, M. Gaskin: Optimisation of a sensor for onsite detection of partial discharges in power transformers by the UHF method. IEEE Transactions on Dielectrics and Electrical Insulation (ITDEIS), Vol. 15, No.6, pp. 1634-1639, 2008 [3.47] G.C. Stone: Partial discharge signal measurements to assess rotating machine insulation condition: A survey. International Symposium on Electrical Insulation, Montreal, 1996 [3.48] E. Binder, A. Draxler, H. Egger, A. Hummer, H.R. Fuchs, H. Koklek, F. Müller, M. Drpic, M. Hoof, R. Käfer, S. Lanz: Entwicklungen und Nachweisuntersuchungen von Diagnosemethoden für Wasserkraftgeneratoren. E&I Vol. 117, No. 12, 2000 [3.49] T. Strehl, E. Lemke, H. Elze: On-line PD measurement, diagnostic tools and monitoring strategy for generators and power transformers. 12 th ISH, Bangalore, India, 2001 [3.50] M. Hoof: TE-On-Line-Monitoring an rotierenden Maschinen, Moderne Verfahren, Möglichkeiten und Grenzen der Zustandsbewertung. HIGHVOLT Kolloquium, pp. 225-232, 2003 [3.51] C. Hudon, M. Belec, A. Contin, A. Cavallini, D.N. Nguyen, G.C. Montanari, M.Conti: Evolution in automatic phase resolved partial discharge pattern recognition for machine diagnosis. 13th ISH Delft, paper O.14.06, 2003 [3.52] D. Pommerenke, I. Krage, W. Kalkner, E. Lemke, P. Schmiegel: On-site PD measurement on high voltage cable accessories using integrated sensors. 9th ISH Graz, Austria, paper 56.08, 1995 [3.53] D. Denissov, R. Grund, T. Klein, W. Köhler, S. Tenbohlen: UHF partial discharge diagnosis of plug-in cable terminations. 7th International Conference on Insulated Power Cables, Jicable’07, Paris, France, paper 0133, 2007 [3.54] D. Denissov, W. Köhler, S. Tenbohlen, R. Grund, T. Klein: Simulation of UHF PD wave propagation in plug-in cable terminations. IEEE International Symposium on Electrical Insulation (ISEI), Vancouver, Canada, 2008, pp. 547-550, 2008
  • 68. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 68 - [3.55] T. Klein, D. Denissow, W. Köhler, S. Tenbohlen: Application of the UHF method for on-line PD diagnosis of cable terminations. CIGRE Session, Paris, France, paper B1-203, 2008. [3.56] H.D. Schlemper, R. Kurrer, K. Feser: Sensitivity for on-site partial discharge detection in GIS. 8th International Symposium on High Voltage Engineering, Yokohama, Japan, Vol. 3, pp. 157-160, 1993 [3.57] CIGRE JWG 33/23.12: Insulation co-ordination of GIS: Return of experience, on-site tests and diagnostic techniques. ELECTRA, Nr. 176, 1998 [3.58] CIGRE TF 15/33.03.05: PD detection system for GIS: Sensitivity verification for the UHF method and the acoustic method. ELECTRA No. 183, 1999 [3.59] R. Feger: Sensoren und Verfahren der UHF Teilentladungs-Messtechnik an gasisolierten Schaltanlagen. Ph.D., D 93, Universität Stuttgart, 2001 [3.60] S. Meijer, P.D. Agoris, J.J. Smit: UHF PD sensitivity check on power transformers. 14 th International Symposium on High Voltage Engineering, Beijing, China, 2005 [3.61] E. Lemke, T. Strehl, S. Markalous, W. Weissenberg: Ultra-wide-band PD diagnostics of power cable terminations in service. IEEE Transactions on Dielectrics and Electrical Insulation (ITDEIS), Vol. 15, No.6, pp. 1570-1575, 2008 [3.62] S. Coenen, S. Tenbohlen, S. Markalous, T. Strehl: Sensitivity of UHF PD measurements in power transformers. IEEE Transactions on Dielectrics and Electrical Insulation (ITDEIS), Vol. 15, No.6, pp. 1553-1558, 2008 [3.63] E. Howells and E.T. Norton: Detection of PDs in transformes using acoustic emission techniques. IEEE Trans, Vol. PAS-97, No. 5, pp. 1538-1549, 1978 [3.64] CIGRE WG 21-05: Partial discharge detection by means of acoustic detection. 1993 [3.65] L. E. Lungard, M. Runde, B. Skyberg: Acoustic Diagnosis of gas insulated substations; a theoretical and experimental basis. IEEE PD, Vo.5, No. 4, pp. 1751-1759, 1990 [3.66] A. -D. Schlemper, K. Feser, H. Blaum, P. Hirchesch: Sensitivity of acoustic PD detection in GIS; Laboratory experiments and on-site experience. Conf. Rec. of IEEE Symp. On El. Ins., pp. 99-102, 1996 [3.67] A. Pfeffer, S. Coenen, S. Tenbohlen, T. Strehl, S. Markalous: Onsite Experiences with multi- terminal IEC PD measurements and UHF PD measurement. 16 th ISH, Cape Town, 2009 [3.68] A. Cavallini, G. C. Montanari, G. Pasini, A. Salsi: Improving PD source localization techniques in cable systems. International Symposium on Electrical Insulating Materials, 2008 (ISEIM 2008), pp. 175-178, Tokyo, Japan, Sept. 2008 [3.69] A. Cavallini, G. C. Montanari and F. Puletti: A novel method to locate PD in polymeric cable systems based on amplitude-frequency (AF) map. IEEE Trans. on Dielec. Elec. Insul., Vol. 14, No. pp. 726-734, 2007 [5.1] M.L. Philip, Y.S Goh, W.H. Leong, M. Darus A. Sethuraman, S.A. Fuad, W. Koltunowicz: Partial Discharge Test of GIS Installation in Malaysia: Techniques and Experiences. CIGRE Conf., report 15-304, 2000 [5.2] G.J. Behrmann, S. Neuhold, R. Pietsch: Results of UHF measurements in a 220 kV Substation during on-site commissioning tests. Conf.Proc. 10th ISH, Montreal, Canada, 1997 [5.3] E. Pultrum, R.J.B. Gruntjes, S. A. M. Verhoeven: On-site partial discharge measurement and commissioning test: the experience. CIGRE 2006, paper B1-108.
  • 69. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 69 - [5.4] F. Petzold, M. Beigert, E. Gulski: Experience with Off-line PD diagnosis on MV cables knowledge rules for asset decisions. Proceedings Jicable '07. 7th International Conference on Insulated Power Cables. France, Versailles, June 2007, paper C7.2.5 [5.5] E. Gulski, J.J. Smit, P. Cichecki, P.P. Seitz, B. Quak, F. de Vries, F. Petzold: Insulation diagnosis of HV power cables. Proceedings Jicable '07. 7th International Conference on Insulated Power Cables, France, Versailles, June 2007, paper C7.1.2 [5.6] P.P. Seitz, B. Quak, E. Gulski, J.J. Smit, P. Cichecki, F. de Vries, F. Petzold: Novel method for on-site testing and diagnosis of transmission cables up to 250kV. Proceedings Jicable '07. 7th International Conference on Insulated Power Cables, France, Versailles, June 2007, paper C7.2.10 [5.7] C.G.A. Koreman et al.: Development of a new 380 kV double circuit XLPE insulated cable system in The Netherlands. Cigré 2006, B1-107 [5.8] IEC 62067: 2006-03: Power cables with extruded insulation and their accessories for rated voltages above 150 kV (Um = 170 kV) up to 500 kV (Um = 550 kV) - Test methods and requirements [5.9] S. Meijer, et. al.: Advanced partial discharge measuring system for simultaneous testing of cable Accessories. CMD conference proceedings, Changwon, Korea, 2006. [5.10] S. Meijer, et. al.: Simultaneous condition assessment of accessories of power cables using a wireless VHF/UHF PD detection system. CMD conference proceedings, Changwon, Korea, 2006 [5.11] Elektrizitätswirtschaft EW JG 106, Heft 4, page 20-26, 2007 (in German) [5.12] Elektrizitätswirtschaft EW JG 106, Heft 24, page 42-53, 2007 (in German) [5.13] D. Denissov, R. Grund, T. Klein, W. Köhler, S. Tenbohlen: UHF partial discharge diagnosis of plug-in cable terminations. 7 th International Conference on Insulated Power Cables, Jicable’07, June 24-28, 2007, Paris, France, paper No. 0133 [5.14] IEEE 1434-2000: Trial-Use Guide to the Measurement of Partial Discharges in Rotating Machinery [5.15] IEC 60034-27: 2006-12: Rotating electrical machines - Part 27: Off-line partial discharge measurements on the stator winding insulation of rotating electrical machines [5.16] H.J. van Breen: Condition assessment of stator insulation using partial discharge diagnostics. ISBN 90-9020128-9, PhD Thesis TU Delft, The Netherlands, 2005 [5.17] IEC 60076-3: 2000: Power transformers - Part 3: Insulation levels, dielectric tests and external clearances in air [5.18] T. Leibfried, U. Thieß: Electrical testing and diagnostic measurements in the life cycle of power transformers. HIGHVOLT-Kolloquium ‘07, pp. 115-121, 2007 [5.19] P. Werle et al.: Repair and HV testing of power transformers on site. (in German) ETG-Fachtagung “Diagnostik Elektrischer Betriebsmittel” Kassel, 2006 [5.20] W. Hauschild, A. Thiede, T. Leibfried, F. Martin: Static frequency converters for HV testing of power transformers (in German). Stuttgarter Hochspannungssymposium, 2006 [5.21] A. Winter, P. Coors, U. Stephan: A mobile transformer test system based on static frequency converters. HIGHVOLT-Kolloquium ‘07, pp. 137-142, 2007 [5.22] T. Aschwanden, A. Schenk et al.: On-site repair and refurbishment and HV tests of large power transformers in the transmission grid. CIGRE Session Paris, paper A2-203, 2004
  • 70. High‐Voltage On‐Site Testing with Partial Discharge Measurement  - 70 - [5.23] J. C. Mendes, R. A. Marcondes, J. Westberg: On-site repair of power transformers. CIGRE Session Paris, paper 12-114, 2002 [5.24] W. Koltunowicz, R. Plath: Synchronous Multi-channel PD Measurements. IEEE Transactions on Dielectrics and Electrical Insulation (IEEE TDEI), Vol. 15, No.6, pp. 1715- 1723, 2008 [5.25] P. Werle: On-site test of power transformers. Omicron Symposium. Austria 2006 [5.26] S. Markalous, S. Tenbohlen, K. Feser: New robust non-iterative algorithms for acoustic PD- localization in oil/paper-insulated transformers. 14 th ISH Symposium Beijing, paper No. G- 039, 2005 [5.27] S. Tenbohlen, S. Hoek, D. Denissov, R. Huber, U. Riechert, S. Markalous, T. Strehl, T. Klein: Electromagnetic (UHF) PD diagnosis of GIS, cable accessories and oil-paper insulated power transformers for improved PD detection and localization. CIGRE Session Paris, paper No. D1-104, 2006