Shunt Capacitance Influences on Single-Phase Transformer FRA Spectrum

Transcription

1 213 Electrical Insulation Conference, #25 Ottawa, Ontario, Canada, 2 to 5 June 213 Shunt Capacitance Influences on Single-Phase Transformer FRA Spectrum Mehdi Bagheri *, B.T. Phung *, Trevor Blackburn *,and Ali Naderian ** * University of New South Wales, Sydney, Australia ** Kinectrics, Transmission and Distribution Technologies, Toronto, ON, Canada Abstract Frequency response analysis (FRA) is considered as an accurate, fast, effective and sensitive method for detection of mechanical defects within power transformers, providing worthy information on active part structure. FRA measurement has been industrialized and employed since the last decade while FRA spectrum interpretation is still under development. The interaction of distributed resistances, series and shunt capacitances as well as self- and mutual-inductances in transformer winding leads to FRA spectrum oscillations. In fact, each and every parameter has its own impact on this spectrum. This study has concentrated on shunt capacitance influences on frequency response spectrum oscillations in Bode diagram. Shunt capacitance impacts on low-, mid- and highfrequency bands of FRA spectrum are discussed. A practical study on a single phase air-core glassy model transformer is carried out to explore the influence of shunt capacitances of HV and LV windings with respect to the transformer tank on FRA spectrum. In addition, a single phase 11/.25 kv, 25 kva core type distribution transformer is also tested to evaluate the shunt capacitance impacts on FRA fluctuations. Keywords- Frequency response analysis, Shunt capacitance, Transformer diagnosis, Transformer winding deformation. I. INTRODUCTION Power transformers are supposed to be and remain in service in various environmental circumstances under different electrical and mechanical stresses [1], [2]. Experience has shown that transformer failure especially in high voltage level like 23 kv or 4 kv can cause irrecoverable harms to pertaining power system. Based on historical data of failure in power transformers, one of the major problems in transformers is mechanical defect [3]. A number of monitoring and diagnostic methods have been introduced to recognize transformer active part displacement and winding deformation [4], [5]. Frequency Response Analysis (FRA) and Short Circuit Impedance (SCI) measurement have been employed as two common diagnosis methods in large power transformer winding deformation recognition [6]. Since SCI measurements are made only at the fundamental frequency, it is expected that they provide less detailed information on the state of the windings than the more comprehensive FRA measurements. Hence, FRA is supposed to be superior to SCI in the case of transformer mechanical defect recognition. Although FRA measurement has been utilized for some years now, FRA spectrum interpretation is still under development [7]. To reach to an acceptable level of FRA result interpretation, the influence of each distributed parameter should be investigated and discussed precisely. To date, a number of studies have been conducted and provided worthy information [8]- [14] on distributed parameters influences. Mathematical and practical studies reveal that more investigation is essential to reach appropriate level of FRA interpretation. Transformer winding resistance, self- and mutual-inductance as well as series and shunt capacitances have their own impacts on the spectrum response [15]- [16]. This study concentrates on shunt capacitance influences on frequency response spectrum oscillations in Bode diagram. Previous mathematical approach on the FRA case [5] is presented to explore shunt capacitance contribution in FRA spectrum. The contribution of this study is more focused on the transformer tank influence on frequency response fluctuations. For precise investigation, a single phase air-core glassy model transformer involving concentric HV and LV windings has been taken as the test object and practical study has been performed on it through different setup configurations. The results are then discussed from a physical perspective. Practical study is also continued through study on the shunt capacitance of a single phase 11/.25 kv, 25 kva core-type distribution transformer and the result is interpreted. II. MATHEMATICAL BACKGROUND Technical specifications and physical characteristics of the transformer under test are major factors in determining the limits of the low-, mid- and high-frequency bands of FRA spectrum. The most appropriate limits for each still not widely agreed. In [17], it has been stated that low-frequency band of FRA spectrum is influenced by the transformer core, mid-frequency band controlled by winding structure and high-frequency band dominated by connection leads. From a physical point of view, winding inductances as well as capacitances play a major role in formation of the frequency response trace. At low frequencies, transformer winding inductances are more dominant as inductive reactance of the transformer winding is considerably greater than capacitive reactance. Hence, frequency response magnitude reported in Bode diagram follows a falling trend as frequency increases. Indeed, self-inductance of transformer winding is greater than mutual inductances. Therefore, low frequency behavior is affected mainly by transformer winding self-inductance (L m ). The winding self-inductance is a function of transformer magnetic core reluctance as well as number of winding turns. Magnetic reluctance of the transformer could be changed due to the core deformation. Any changes in the magnetic reluctance or number of winding turns will lead to self-inductance is changed and followed it the inductive reactance is changed, accordingly. As frequency increases, inductive reactance will take values high enough to be neglected. Synchronously, capacitive reactance /13/$ IEEE 225

2 becomes gradually a major player in the formation of frequency response trace. On the other hand, mutual inductances are considerable in mid frequencies as their inductive reactance becomes significant. Hence, the first anti-resonance (minima) in FRA trace is given as (1) [12], f r( low frequency ) 1 (1) 2 LC m where f r(low-frequency) is the resonance frequency of low-frequency band of FRA spectrum, L m is the winding self-inductance and C denotes total series and shunt capacitances (equivalent capacitance). Series capacitances consist of turn-to-turn and discto-disc capacitances whereas shunt capacitance is the capacitance of winding with respect to the core and to the electrostatic screens/tank. Mid frequency fluctuations of the frequency response are controlled by winding structure including series and shunt capacitances as well as inductances. The resonance frequencies of mid-frequency band within FRA spectrum is given as (2) [5], f k 1 2 ( c k ( ) c ) l r( mid frequency ) g s where f r(mid-frequency) is the resonance frequency of mid-frequency band of FRA spectrum, c g represents shunt capacitance per unit conductor length, c s is series capacitance per unit conductor length, l denotes inductance per unit conductor length, λ is the winding length and k is variable coefficient (k=1, 2, 3 ). According to (2), significant value for c g will increase resonance frequencies, while considerable amount of c s will lead to fewer resonance frequencies within the mid-frequency band of FRA trace [5], [12], [18], [19]. (2) #25 Figure 1. Glassy model transformer For the first experiment, the frequency response trace of the test object has been recorded for HV winding when the terminals of LV winding were left open circuit (end to end measurement). A frequency response analyzer with 2 V pp swept sinusoidal signals as the voltage source was used for entire measurements. Input and output impedances of the measurement cables were 5 ohm. Input signal was injected at the line terminal of the HV winding and the response was recorded at the neutral end. For this measurement, the test object did not have the aluminum wrapping (without tank). Similar measurement was carried out for LV side when the HV winding was left open circuit. Afterwards, the aluminum foil was wrapped over the glass casing to serve as a metal tank. The tank was isolated from the ground and the frequency response traces for HV and LV windings were then recorded. Frequency response traces under such circumstances are shown in Fig. 2 and Fig. 3. A. Case Study 1 III. PRACTICAL STUDY In order to study the impacts of shunt capacitance on FRA trace, a model transformer with air-core concentric continuous disc type HV and LV windings is used as the test object. The HV winding consists of 8 discs including 7 conductor turns per disc. The LV winding has 1 discs with 6 conductor turns per disc. The tank was manufactured by plexiglass. Line and neutral leads of the windings were brought out from the tank through appropriate HV and LV bushings. An aluminum foil has been wrapped over the glass casing to simulate the metal transformer tank. The test object is shown in Fig.1. Four different experiments have been performed on HV and LV windings of the test object. Significant value of the air-core magnetic reluctance of the test object results in small self-inductances for the windings. Hence, inductive reactance due to the self-inductance experiences low value and the resonance frequencies in FRA traces for HV and LV windings would be considerably shifted to higher frequencies. Therefore, the upper band limits for FRA measurements were extended from 2 MHz to 2 MHz to observe entire oscillations. -7 HV-Without Tank HV-With Tank 1.58 MHz, db Figure 2. FRA traces of HV winding in the presence and absence of aluminum tank LV-Without Tank LV-With Tank MHz, db 1.63 MHz, db 1.74 MHz, 1.93 MHz, db Figure 3. FRA traces of LV winding in the presence and absence of aluminum tank 226

3 According to Fig. 2, moving from left hand side to right hand side on the FRA traces, frequency response trace for HV winding with metal tank has shifted to higher frequency at compared to HV without tank. Moving to higher frequencies, oscillation trends in both graphs remain essentially the same. Similar pattern occurred for LV side as seen in Fig.3. The first anti-resonance (minima) and resonance (maxima) have shifted to higher frequencies. Since self-inductance and series capacitance of the HV and LV winding have not changed, the deviations of FRA traces in Fig.2 and Fig.3 come through shunt capacitance changes of the HV and LV winding with respect to the tank. Deviated frequencies are provided in Table I and Table II for HV and LV windings respectively, and shunt capacitance variations are calculated by considering (1). C 1 and C 2 denote the shunt capacitances for f 2 and f 1 respectively. In the case of frequency response trace of HV winding, shunt capacitance variation was estimated 52%, while LV winding reports 59 % approximately. TABLE I. CAPACITANCE RATIO AND ANTI-RESONANCE FREQUENCIES FOR THE SPECTRA OF FIG.2. Minima MHz 1.58 MHz TABLE II. CAPACITANCE RATIO, ANTI-RESONANCE AND RESONANCE FREQUENCIES FOR THE SPECTRA OF FIG.3. Minima 1.74 MHz MHz 1.59 Maxima 1.93 MHz 1.63 MHz For the second measurement, the aluminum tank of the test object was grounded through laboratory earth and FRA trace for HV winding was measured when LV terminals were left open circuit. Similar trace was measured for LV side. HV and LV FRA traces for isolated and grounded tank are shown in Fig.4 and Fig. 5 respectively. According to Fig.4 and Fig.5, numbers of resonance frequencies for grounded tank have been increased as compared to the un-grounded case, in particular the mid-frequency band. It can be interpreted through shunt capacitance increment as it was stated earlier for mid-frequency band. In addition, the magnitude of the first anti-resonance in HV trace and second one in LV trace have changed considerably due to the changes of conductance between HV and LV windings with respect to the tank. This can be interpreted by winding loss factor (conductivity). In fact, any changes in winding loss factors can result in FRA magnitude of resonance/anti-resonance points are changed accordingly. For the third measurement, the aluminum tank has been deformed deliberately as shown in Fig.6. FRA trace for HV winding was recorded when LV winding was left open circuit. Similar measurement was performed for the LV side. HV and LV FRA traces for grounded normal and deformed tanks are illustrated in Fig.7 and Fig. 8. According to Fig.7 and Fig.8, moving from low frequencies towards mid frequencies the first anti-resonances of HV and LV windings have slightly moved to lower frequencies. The same has happened for the first resonance point in LV winding. Based on (1), shifting from higher to lower frequency may be due to the shunt capacitance increment. # HV-Grounded Tank HV-Isolated Tank MHz, db Figure 4. FRA traces of HV winding for isolated and grounded tank. LV-Grounded Tank LV-Isolated Tank 1.74 MHz, MHz, dB MHz, 79 db MHz, db Figure 5. FRA traces of LV winding for isolated and grounded tank. Figure 6. Schematic of the deformed aluminum tank of the test object (top view) HV-Normal Grounded Tank HV-Deformed Grounded Tank MHz, db MHz, db Figure 7. FRA traces of HV winding for normal and deformed tank. 227

4 # MHz, db MHz, db MHz, db MHz, db MHz,.79 db 4.59 MHz, db -45 LV-Normal Grounded Tank MHz, MHz, LV-Deformed Grounded Tank db db Figure 8. FRA traces of LV winding for grounded normal and deformed tank. TABLE III. CAPACITANCE RATIO AND ANTI-RESONANCE FREQUENCIES FOR THE SPECTRA OF FIG.7. Minima MHz MHz 1.21 TABLE IV. CAPACITANCE RATIO, ANTI-RESONANCE AND RESONANCE FREQUENCIES FOR THE SPECTRA OF FIG.8 Minima MHz MHz Maxima MHz MHz Measured frequencies for the first HV and LV anti-resonances are provided in Table III and Table IV. Table IV also shows the frequency of the first resonance point of LV trace. Based on the calculated capacitance ratio in Table III and Table IV, shunt capacitance deviation for HV was 2.1%, while the values for LV side were estimated at 16.8%. As the last study on glassy test object, the deformed aluminum tank was isolated from the ground and HV FRA trace was measured when LV was left open circuit. Similar measurement was performed for LV winding. HV and LV FRA traces for isolated normal and deformed tank are shown and compared in Fig.9 and Fig. 1. Table V and Table VI provide frequency variations as well as capacitance ratio for the FRA traces of HV and LV windings. According to Fig.9 and Fig.1, the first anti-resonances of the HV and LV windings have slightly shifted to lower frequencies in low-frequency band, while mid-frequency oscillations have perfectly remained the same. Capacitance ratios in Table V and Table VI show around 14% changes for anti-resonances and 11.6% for the first resonance point in LV winding. It may be concluded that the shunt capacitance for HV and LV windings with respect to tank have increased 14% approximately. LV-Normal Isolated Tank LV-Deformed Isolated Tank MHz,.69 db 1.74 MHz, Figure 1. FRA traces of LV winding for isolated normal and deformed tank. TABLE V. CAPACITANCE RATIO AND ANTI-RESONANCE FREQUENCIES FOR THE SPECTRA OF FIG.9. Minima MHz MHz TABLE VI. CAPACITANCE RATIO, ANTI-RESONANCE AND RESONANCE FREQUENCIES FOR THE SPECTRA OF FIG.1 Minima 1.74 MHz MHz Maxima MHz MHz B. Case Study 2 In order to study shunt capacitance impact of HV winding respect to the LV winding, a single phase 11/.25 kv, 25 kva core-type transformer was taken as another test object and FRA traces for its HV winding were recorded for different configurations. At first, FRA trace was recorded for HV side when the line and neutral terminals of LV winding were shortcircuited and isolated from the ground. Afterwards, the LV winding was grounded and the FRA trace for HV side was remeasured. Test setups were configured as shown in Fig. 11(a) and 11(b). In fact, the configurations in Fig. 11 can identify turn-toturn or disc-to-disc short circuits which are isolated or directly connected to transformer tank or core. (a) (b) Fig. 11. FRA test setup (a) LV is short-circuited (b) LV is short-circuited and grounded khz, db HV-Normal Isolated Tank HV-Deformed Isolated Tank MHz, Figure 9. FRA traces of HV winding for isolated normal and deformed tank. -7 HV-LV is shortened HV-LV is shortened and grounded 9.3 khz, db Fig. 12. FRA of HV winding for different configurations of LV. 228

5 #25 In Fig. 12, frequency response traces of HV winding for grounded and isolated internal short circuit are illustrated and compared to explore why FRA trace has deviated. According to Fig. 12, frequency response trace for grounded internal short circuit is not deviated from isolated one in very low frequencies. In fact, self-inductance of the HV winding does not change for grounded and isolated internal short circuit. Also, moving from very low frequencies to low frequencies shows that minimal peak has slightly moved to right. In addition, in mid and high frequencies grounded internal short circuit trace displays greater absolute magnitude compared to isolated one, while oscillation trends in both graphs remain approximately the same. Since HV winding structure has not changed and short circuit has apparently happened in LV winding, series capacitance of HV winding has not varied, accordingly. Based on this fact, minimal peak movement in Fig. 12 can be interpreted due to the winding shunt capacitance variations. Indeed, resonance point movement in the grounded internal short circuit trace to higher frequency indicates a slight reduction in shunt capacitance of the winding. On the other hand, capacitive reactance increases when capacitance of the winding decreases. This in turn results in greater absolute frequency response magnitude in mid and high frequencies. Therefore, self-inductance does not change during isolated and grounded internal short circuit while total capacitance has been decreased. IV. CONCLUSION Mechanical defects and transformer frequency response measurement were discussed. It was elaborated that transformer winding inductance, series and shunt capacitances have their own influence on FRA trace. Practical study indicated that the winding shunt capacitance can specifically influence low- and mid- and high-frequency bands of FRA spectrum. The shunt capacitance of HV and LV windings with respect to the tank had more impact on the first anti-resonance in FRA trace and generally impacted on low-frequency band. The shunt capacitance of the HV winding with respect to the LV winding influenced more the mid- and high frequency band. Grounding of the test object tank led to total capacitance variation and also increasing numbers of resonance frequencies within the mid-frequency band. In addition, it caused considerable damping of the resonant peaks of mid-frequency band of FRA spectrum. It could be interpreted through conductivity variation of the HV winding respect to the tank. Since this study concentrated on single-phase transformers, further investigations will be conducted on three-phase transformers. REFERENCES [1] M. Bagheri, Mohammad S. Naderi, T. Blackburn, B.T. Phung, Dean-Stark vs FDS and KFT methods in moisture content recognition of transformers, IEEE Int I Conf. Power Energy (PECON 12), Kota Kinabalu, Malaysia, 212, pp [2] M. Bagheri, Mohammad S. Naderi, T. Blackburn, B.T. Phung, Transformer efficiency and de-rating evaluation with non-sinusoidal loads, IEEE Int l. Conf. Power Sys. Tech. (POWERCON 12), Auckland, New Zealand, 212, pp [3] M. Bagheri, Mohammad S. Naderi, T. Blackburn, B.T.Phung, Bushing characteristic impacts on on-line frequency response analysis of transformer winding, IEEE Int I Conf. Power Energy (PECON 12), Kota Kinabalu, Malaysia, 212, pp [4] M. Florkowski, J. Furgal,P. Pajak, Analysis of fast transient voltage distributions in transformer windings under different insulation conditions, IEEE Trans. Dielectr. Electr. Insul., vol. 19, 212, pp [5] M. Bagheri, Mohammad S. Naderi, T. Blackburn, D. Zhang, Transformer frequency response analysis: A mathematical approach to interpret midfrequency oscillations, IEEE Int I Conf. Power Energy (PECON 12), Kota Kinabalu, Malaysia, 212, pp [6] M. Bagheri, Mohammad S. Naderi, T. Blackburn, B.T. Phung, Frequency response analysis and short circuit impedance measurement measurement in detection of winding deformation within power transformers, IEEE Electr. Insul. Mag.,vol. 29, issue 3, 213, pp. 33. [7] M. Bagheri, Mohammad S. Naderi, T. Blackburn, B.T. Phung, Z.Liu Frequency response analysis to recognize inductance variation in transformer due to internal short circuit, IEEE Int l Power Energy Conf. (IPEC 12), Ho Chi Minh City, Vietnam, 212, pp [8] M. Bagheri, Mohammad S. Naderi, T. Blackburn, Advanced transformer winding deformation diagnosis: moving from off-line to on-line, IEEE Trans. Dielectr. Electr. Insul., 212, pp [9] E. P. Dick and C. C. Erven, Transformer diagnostic testing by frequency response analysis, IEEE Trans. Power App. Syst., vol. PAS-97, no. 6, 1978, pp [1] M. Bagheri, Mohammad S. Naderi, T. Blackburn, T. Phung, FRA vs. short circuit impedance measurement in detection of mechanical defects within large power transformer, IEEE Int l. Sympos. Electr. Insul. (ISEI 12), Puerto Rico, USA, 212, pp [11] K. G. N. B. Abeywickrama, Y. V. Serdyuk, S. M. Gubanski, Exploring possibilities for characterization of power transformer insulation by frequency response analysis (FRA), IEEE Trans. Power Del., vol. 21, 26, pp [12] Z. Wang, J. Li, D.M. Sofian, Interpretation of Transformer FRA responses Part I: Influence of Winding Structure, IEEE Trans. Power Del., vol 24, issue 2, pp.73-71, 29. [13] M. Bagheri, Mohammad S. Naderi, T. Blackburn, B.T. Phung, Frequency Response Analysis vs. Flux Division Measurement in detection of transformer winding internal short circuit, IEEE Int l. Conf. Power Sys. Tech. (POWERCON 12), Auckland, New Zealand, 212, pp [14] M. Bagheri, Mohammad S. Naderi, T.R. Blackburn and B.T. Phung, Practical challenges in online transformer winding deformation diagnostics, Int l. Conf. Electric Power and Energy Conversion Syst. (EPECS 11), Sharjah, UAE, 211, pp [15] M. Bagheri, Mohammad S. Naderi, T. Blackburn, T. Phung and H. Haeri, On-line transformer winding deformation diagnosis: A profound insight to methods, 26 th Int l Power Sys. Conf. (PSC 11), Tehran, Iran, 211, pp [16] M. Bagheri, Mohammad S. Naderi, T. Blackburn, T. Phung, Case study on FRA capability in detection of mechanical defects within a 4MVA transformer, CIGRE, Paris, France, 212, pp [17] IEC Standard ; Measurement of frequency response, Edition 1., [18] M. Bagheri, A. Hekmati, R. Heidarzadeh, Mohammad S. Naderi, Impulse voltage distribution in intershield disk winding vs interleaved and continuous disk winding in power transformer, IEEE Power and Energy Conf. (PECON 8), pp [19] M. Bagheri, M.Vakilian, A.Hekmati, R.Heidarzadeh, Influence of electrostatic shielding of disc winding on increasing the series capacitance in transformer, IEEE PowerTech 7, Lausanne, Switzerland, 27, pp