COMPARATIVE ASSESSMENT OF ABNORMAL VOLTAGE STRESSES IN AN ISOLATED MUTUALLY COUPLED TRANSFORMER MODEL WINDING WITH AND WITHOUT VARISTOR

Transcription

1 International Journal of Electrical Engineering & Technology (IJEET) Volume 7, Issue 6, Nov Dec, 2016, pp.25 35, Article ID: IJEET_07_06_003 Available online at ISSN Print: and ISSN Online: Journal Impact Factor (2016): (Calculated by GISI) IAEME Publication COMPARATIVE ASSESSMENT OF ABNORMAL VOLTAGE STRESSES IN AN ISOLATED MUTUALLY COUPLED TRANSFORMER MODEL WINDING WITH AND WITHOUT VARISTOR Javid Akhtar Associate Professor, Department of Electrical and Electronics Engineering, Ghousia College of Engineering, Karnataka, India Dr. Mohd. Zahed Ahmed Ansari Member IEEE, Professor and Head, Department of Electrical and Electronics Engineering, Ghousia College of Engineering, Karnataka, India ABSTRACT The dynamic performance of high voltage power transformer (HVPT) windings owing to the impact of high voltage surges has evoked the attention of insulation engineers for almost a century. As these very fast rising transient over voltages generate huge peak voltage stresses along the windings and may perhaps threaten its insulation. Accordingly a profound understanding of this is crucial for insulation engineers to design the winding insulation optimally and thinking of novel protective devices to be used to ensure that insulation failure does not take place when such high voltage surges strike the terminals of HVPT s. In this paper, a HVPT winding is being represented by an isolated single phase 12 coil section and 18 coil section mutually coupled model winding and shunted with metal oxide varistors to suppress the huge peak voltage stresses and their complex oscillations occurring due to these extremely fast rising lightning surges. A simple electrical model of a Zno varistor is presented and its excellent properties are being utilized for the transformer winding insulation protection against such fast rising lightning surges. Simulation studies have been conducted for comparative assessment of abnormal voltage stresses in an isolated mutually coupled transformer model winding energized with full a periodical and chopped a periodical lightning surge voltages without and with varistor for a distribution winding constant α=10. Key words: Transformer model winding, HVPT, dynamic performance, impulse voltage, peak voltage stresses, complex oscillations, winding insulation, varistors. Cite this Article: Javid Akhtar and Dr. Mohd. Zahed Ahmed Ansari. Comparative Assessment of Abnormal Voltage Stresses in an Isolated Mutually Coupled Transformer Model Winding with and Without Varistor. International Journal of Electrical Engineering & Technology, 7(6), 2016, pp editor@iaeme.com

2 Javid Akhtar and Dr. Mohd. Zahed Ahmed Ansari 1. INTRODUCTION Power transformers are one of the most essential equipment s used in ac electric power transmission systems. During service HVPT windings are exposed to mainly two kinds of transient over voltages, lightning surges and switching surges. Lightning surges being the most hazardous abnormal voltages and when strike the transformer windings it may breakdown the insulation thereby causing heavy damage to the windings and primarily responsible for outages of power transformers and hence loss of revenue to the electric utilities and also power interruption to the valuable customers [1-3]. The very objective of this paper is to bring down effectively the huge peak voltage stresses and suppress the complex oscillations due to lightning voltage surges across the transformer winding sections (chiefly focusing on the initial part of the model winding which is most responsive to such fast rising surge voltages) by means of suitably modeled varistors shunted to the coil sections to keep them within the satisfactory limits and to evade the collapse of winding insulation AND 18 COIL SECTION MUTUALLY COUPLED TRANSFORMER MODEL WINDING WITH VARISTOR Figure 1 Single-phase circuit model of a 12/18 coil section mutually coupled transformer model winding for the determination of peak voltage stresses Figure 1.shows a 12 and 18 coil section mutually coupled transformer model winding shunted with varistor to assess its dynamic performance. Here the distribution winding constant α is defined as α= Cg/Cs, where Cg is the ground capacitance and Cs is the series capacitance of the winding. 3. DESIGN OF 12 AND 18 COIL SECTION MUTUALLY COUPLED TRANSFORMER MODEL WINDING TO CALCULATE THE WINDING PARAMETERS The transformer model winding is designed for 12-coil section and 18 coil sections with 60 turns per section and also the mutual inductance between the different coil sections is considered. A. Calculation of self-inductance of the coil (Ls) Self-inductance L s of each coil section is determined using the Nagaoka s inductance formula for circular cross section [6] as given in (1) editor@iaeme.com

3 Comparative Assessment of Abnormal Voltage Stresses in an Isolated Mutually Coupled Transformer Model Winding with and Without Varistor L s = 0.002π²a (2a/b) N²k µh (1) Where a = mean radius of the coil b = length of the winding k = factor that takes account of the effect of the ends and is a function of the shape ratio (2a/b) N = number of turns B. Calculation of mutual inductances between the different coil sections (M jk ) Mutual inductances between different coil sections of the model winding, was found using (2) through a MATLAB source code program M = 0.002π²a²n 1 n 2 (r 1 B 1- r 2 B 2- -r 3 B 3+ r 4 B 4) µh(2) Where n 1 and n 2 are winding densities and the functions B n depend upon the parameters ρ² n = A²/r² n and are obtained by interpolation [6]. 4. MODELING OF A VARISTOR Metal oxide varistors have steep nonlinear voltage current characteristics that make them most suitable for protection of HVPT windings against fast rising transient over voltages. Varistors are nonlinear bipolar resistors which offer a very highh resistance to power frequency sine wave voltages and offer very low resistance for fast rising over voltages thereby protecting the equipment that it shunts and prevent costly system damages. Varistors can pass wide varying currents in the order of µa to tens of ka over a narrow band of voltage. Since the inception of metal oxide varistor several models have been proposed by different researchers including varistor models recommended by IEEE Working Group , for lightning studies. In all these models the hard point is the identification of the model parameters, have more than one non-linear block and requires iterative procedures, interpolation formula (polynomial equations with several coefficients to be preset) to determine acceptable parameter values [7]. In this paper a simplified model of varistor is been incorporated due to the following facts compared to the other models: It comprises of only one non-linear block, eliminates problems of numerical convergence and reduces simulation time by about 30%. The number of parameters to be preset to create a new model is reduced. The accuracy is good, as it allows obtaining a good fitting to the measured characteristics [8]. Good model with precise approximation of device characteristics is indispensable to quality of simulation results. In the model considered here, R and L represent the characteristics of the leads of the component and C represents the properties of the material and packaging process. Figure 2 Simplified electrical model of a metal oxide varistor 27 editor@iaeme.com

4 Javid Akhtar and Dr. Mohd. Zahed Ahmed Ansari 4.1. Investigating the Dynamic Performance of the 12-Coil Section Mutually Coupled Isolated Transformer Model Winding A. Sine voltage response with varistor shunted to the first coil section in the vicinity of the hv line terminal for α =10 1 X-Axis: Time (ms) Y-Axis: Voltage and Current (pu) Figure 3 Sine voltage response of a 12-coil section model winding for α = 10 Figure 3 shows the sine voltage response for power frequency sine wave voltage with peak value of 1pu applied to the single phase transformer model winding with varistor shunted to the first coil section about 8.2% portion of the model winding in the vicinity of the line terminal. It can be seen from the waveforms that the coil node voltages with respect to ground are less than the peak value of the input voltage and the current through the varistor is almost zero. This implies the varistor does not conduct for normal power frequency sine voltages and the transformer windings behave in normal steady state condition. B. Impulse voltage response with no varistor for α =10 Figure 4 Impulse voltage response of a 12-coil section model winding due to a 5pu peak full lightning impulse voltage forα = 10 and with no varistor 28 editor@iaeme.com

5 Comparative Assessment of Abnormal Voltage Stresses in an Isolated Mutually Coupled Transformer Model Winding with and Without Varistor Figure 5 Impulse voltage response of a 12-coil section model winding due to a 5pu peak full impulse lightning voltage forα = 10 and with no varistor Figure 4 and Figure 5 illustrate impulse response with no varistor. The highest peak voltage with respect to ground at first coil is pu rising very fast at 15.11µs containing only positive decaying oscillations. But the voltage across the first coil section (8.2% of the winding) is 4.295pu positive peak rising very sharp in 255ns with 2.926pu negative peak containing highly nonlinear oscillations with both positive and negative excursions. On referring Fig.5 we notice that the peak voltage across the mid-point and the last coil section of the model winding is 6.72pu and 5.41pu respectively containing both positive and negative oscillations. It is carefully noted that the negative peak at the last coil section is almost 3.312pu. C. Impulse voltage response with varistor shunted to only 8.2% portion of the winding for α =10 Figure 6 Impulse voltage response of a 12-coil section model winding due to a 5pu peak full impulse lightning voltage for α = 10 and with varistor Here the varistor is shunted to the first 60 turns (8.2%) of the model winding. The highest peak voltage with respect to ground at the first coil is smaller than the input voltage and contains no oscillations. The peak voltage across the 8.2% of the winding is reduced to a maximum steady value of 0.465pu with no oscillations (compared to a rapid rise of oscillative 4.295pu peak with no varistor). This shows the effectiveness of the varistor for suppressing huge peak voltage surges and their complex oscillations editor@iaeme.com

6 Javid Akhtar and Dr. Mohd. Zahed Ahmed Ansari D. Tail Chopped impulse voltage response for α =10 and no varistor - Figure 7 Voltage waveforms for tail chopped impulse voltage at 20µs for a 12-coil section model winding for α = 10 and no varistor Figure 7 demonstrate the impact of chopped tail impulse voltage surge on the model winding with no varistor. On examining the waveforms we observe that the first coil section takes the brunt of highest peak voltage of 7.76pu with respect to ground. The voltage across the first coil section is 4.3pu positive peak rising very sharply and the moment the tail is chopped at 20µ it takes a sharp transition from 1.56pu to 2.87pu with highly random positive and negative oscillations. We can also notice that the peak voltage across the center of the model winding is 7.26pu containing both negative and positive oscillations. E. Tail Chopped impulse voltage response for α =10 and with varistor shunted to only 8.2% portion of the winding Figure 8 Voltage waveforms for tail chopped impulse voltage at 20µs for a 12-coil section model winding for α = 10 and with varistor Here the first coil section voltage is less than the input chopped voltage with respect to ground and the voltage across the first coil section is almost reduced to a small steady value of 0.355pu (from oscillative 4.3pu peak value without varistor) and all the complex oscillations over the wide frequency range are suppressed editor@iaeme.com

7 Comparative Assessment of Abnormal Voltage Stresses in an Isolated Mutually Coupled Transformer Model Winding with and Without Varistor F. Time step voltage response for 5pu step for α = 10 and no varistor Figure 9 Time step voltage response due to a step voltage of 5pu for α = 10 and no varistor On examining the waveforms we notice that the highest voltage is at the coil-1 that reaches to 7.92pu peak value (towering 58.4%) at 15.22µs with several oscillations. The voltage across the first coil section is 4.94pu peak occurring at 39.6ns (a very fast rise) with a negative peak of -2.92pu occurring at 15us with several oscillations. G. Time Step voltage response for 5pu step and α = 10 with varistor shunted to only 8.2% portion of the winding Figure 10 Time step voltage response due to a step voltage of 5pu for α = 10 and with varistor The voltage at the first coil is 4.75 steady value less than the input step with almost all oscillations suppressed (compared to 7.92pu peak with enormous oscillations without varistor) The voltage across the first coil section is 0.6pu steady value (compared to 4.94pu positive peak and negative peak of pu without varistor). It is to be noted that the voltage across the second coil section where the varistor is not shunted is 7.4pu peak with increasing positive oscillations editor@iaeme.com

8 Javid Akhtar and Dr. Mohd. Zahed Ahmed Ansari Coil Section Mutually Coupled Isolated Transformer Model Winding H. Impulse voltage response with no varistor for α =10 Figure 11 Impulse voltage response of a 18-coil section model winding due to a 5pu peak full impulse lightning voltage for α = 10 and with no varistor The coil voltages have complex oscillations and increased in large numbers with both positive and negative excursions. The voltage at the first coil section oscillates around 5.202pu peak value. The voltage across the first coil section reaches peak value of around 4.3pu very fast at 253ns. The voltage across the last coil section is pu peak and around 4.01pu positive peak. I. Impulse voltage response with varistor shunted to only first coil section of the model winding for α =10 Figure 12 Impulse voltage response of a 18-coil section model winding due to a 5pu peak full impulse lightning voltage for α = 10 and with varistor On examining the waveforms of Figure 12we notice that the first coil section voltage is less than the input impulse voltage and the voltage across the first coil section is quickly reduced to a small steady value of 0.255pu (from 4.3pu peak value without varistor) and all the complex oscillations over the wide frequency range are suppressed editor@iaeme.com

9 Comparative Assessment of Abnormal Voltage Stresses in an Isolated Mutually Coupled Transformer Model Winding with and Without Varistor J. Tail Chopped impulse voltage response for α =10 and no varistor Figure 13 Transient voltage response for 18-coil section model winding due to an impulse lightning voltage chopped at 20µs for α = 10 The voltage at the first coil section oscillates between 6.08pu up to 20µs thereafter undergoes several oscillations. The voltage across the first coil section reaches its first peak value at 4.3 pu rise very fast at 321ns and -5.8pu negative peak with enormous positive and negative oscillations. K. Tail Chopped impulse voltage for α =10 with Varistor shunted to the first coil section Figure 14 Transient voltage response for 18-coil section model winding due to an impulse lightning voltage chopped at 20µs for α = 10 On examining the waveforms of Figure 14 we notice that the first coil section voltage is less than the input impulse voltage and the voltage across the first coil section is almost reduced to a small steady value of 0.255pu (from 4.3pu peak value without varistor) and all the complex oscillations over the wide frequency range are suppressed editor@iaeme.com

10 Javid Akhtar and Dr. Mohd. Zahed Ahmed Ansari L. Special case: Creating a fault between the first 60 turns and ground with Varistor shunted to it with α =10 Y-Axis: Voltage and Current (pu) Figure 15 Voltage and current response of a 12-coil section model winding due to 5pu peak impulse lightning voltage for α = 10 and with varistor In this case the transformer model winding is energized with a 5pu peak standard lightning impulse voltage 1.2/50 µs and a fault is created between the first coil section and ground to test the effectiveness of the varistor in discharging the lightning impulse current to ground. It is noticed that the voltage at the first coil section with respect to ground is zero and the current passing to the ground through the varistor is 1.96pu peak value and immediately falls to 0.058pu in 598ns and becomes zero in a period of 20µs. This shows the effectiveness of the varistor in very fast discharging of the lightning current to the ground and thereby safeguarding the insulation of the first coil section. 5. CONCLUSION Complex oscillations of broad frequency spectrum that present in the huge non uniform peak voltage stresses due to lightning surges may tear down the winding insulation and besides overstressing of the winding insulation may drastically reduce the transformer lifespan and often leads to internal short circuit. Simulation results demonstrate the efficacy of varistors shunted to the winding section. It can be concluded that varistors can best suppress the huge peak voltages and their complex oscillations. The zinc oxide varistor device is adding a new dimension to the technology of protecting transformer windings against the lightning and also the switching surges. This novel method is a simple and reliable technique and requires less investment, thus paving a way for its widespread use in the coming days. ACKNOWLEDGMENT The authors Mr. Javid Akhtar and Dr. Mohd. Zahed Ahmed Ansari would like to express their thanks to the authorities of Ghousia College of Engineering, Ramanagaram for all the cooperation and encouragement. REFERENCE [1] BHEL, Transformers, Tata McGraw Hill publishing company limited New Delhi, [2] Subir Ray, An introduction to High Voltage Engineering 2004,Prentice hall of India, New Delhi. [3] Mohd. Zahed Ahmed Ansari, G.R. Gurumurthy, J. Amarnath and N.K. Kishore, 2006, Surge Voltage Stresses Across Power Transformer Winding Sections Provided With Metal Oxide Surge Absorber 34 editor@iaeme.com

11 Comparative Assessment of Abnormal Voltage Stresses in an Isolated Mutually Coupled Transformer Model Winding with and Without Varistor Blocks, Proceedings of IEEE International Conference On Electrical Insulation and Dielectric Phenomena (CEIDP-2006), 16 th 19 th October, held at Kansas City, Missouri, USA, pp [4] L. F. Blume, Boyajian, G. Camith, T.C Lennox, S. Minneci, V.M Montsinger, 1, Transformer Engineering, John Wiley and Sons (Book) [5] S. B Vasutinsky, Principles, operation and design of power transformer, PSG College of Technology, Coimbatore (Book), 1962, Tamilnadu, India. [6] F.W. Grover Inductance calculation working formulas and tables Dover publications, New York. [7] Boris Zitnik, Maks Babuder et al, Numerical Modeling of Metal Oxide Varistors, Proceedings of the xith International Symposium on High Voltage Engineering, Tsinghua University,Beijing,China, August 25-29,2005, pp B49,1-6. [8] Julio Guillermo Zola, Simple Model of Metal Oxide Varistor for PSpice Simulation, IEEE Transactions on Computer Aided Design of Integrated Circuits and Systems, VOL. 23. N0. 10, October 2004, pp [9] Mr. Sumit Kumar and Prof.Dr. A.A Godbole. Performance Improvement of Synchronous Generator by Stator Winding Design. International Journal of Electrical Engineering & Technology (IJEET), 4(3), 2013, pp [10] Kirti G. More and Ramling D. Patane, Non-Isolated Soft Switching DC-DC Converter and Load at Full Range of ZVS. International Journal of Electrical Engineering & Technology (IJEET), 7(5), 2016, pp editor@iaeme.com