JEEE Transactions on Dielectrics and Electrical Insulation Vol. 9, No. 4; August 2002 569 Experience With Development and Evaluation of Corona-Suppression Systems for HV Rotating Machines Hassan El-Kishky Department of Electrical Engineering The University of Texas at Tyler 3900 University Blvd. Tyler, TX 75799, USA Beant S. Nindra National Electric Coil 800 King Avenue (43212) Columbus, OH 43216, USA Mazen Abdel-Salam Department of Electrical Engineering Assiut University Assiut, 71516, Egypt and Eugene Williams California Department of Water Resources Divisron of Engineering 1416 9th St Sacramento, CA 95814, USA ABSTRACT This paper presents the results of a project aimed a t the development of reliable corona-suppression systems for high voltage rotating machines. These systems are based on both conducting and semi-conducting dry and B-stage tapes a s well as paints of different resistivity. Three groups of anti-corona systems were developed; namely, paint (on the slot portion) paint (on the slot exit) system, paint (on the slot portion) tape (on the slot exit) system, and tape (on the slot portion) tape (on the slot exit) system. Sample results of standard evaluation and acceptance tests are presented. The accelerated life-aging test was run on test coil samples at the National Electric Coil HV testing facility. A numerical 2-dimensional finite difference model for the electric potential and field analysis along the end-turn zone was developed and the model results are used a s guidelines through the design process of the grading system. Laboratory test results on randomly selected sample coils were confirmed by an independent testing facility. The proposed corona-suppression systems were applied on thousands of production coils. - - 1 INTRODUCTION T H E failure of a generator or a large motor can cause considerable financial losses to the utilities and industries due to non-planned outage costs in addition to the cost of repairing and/or rewinding the machine. The reliability of HV machines depends mainly on the soundness and integrity of the insulation system [1-141. Exposed to complex electrical, mechanical and environmental stresses, the stator winding’s main insulation is the most vulnerable to deterioration and failure. Long exposure to internal and surface discharges due to high localized concentration of electric stress can directly accelerates the deterioration - and ultimately, complete failure of the ground-wall insulation. In particular, intensified electric stress in the overhang portion of the high voltage stator winding of hydroelectric machines is of special concern. Breakdown of the gas adjacent to the insulation in the immediate vicinity of the slot exit leads to the development and propagation of discharges over the end-turns [1,5-7,13-201 which in tum results in gradual degradation of the ground-wall insulation. Therefore, the application of stress-grading systems along the end-turn is considered essential for high voltage machines. It is a common practice that corona-suppression systems are designed such that a high voltage winding passes a set of standard acceptance tests [1,13,14,21-241 as well 1070-9878/1/$17.00 0 2002 IEEE El-Kishky et al.: Experience With Development and Evaluation of Corona-Suppression Systems 570 as meets customer’s special requirements. It is normally required that a HV winding passes the high potential blackout test at a level ranging from 100 to 150 percent of the nominal line-to-ground voltage with no visual corona discharges [1,14,22-241. Yet, it is not uncommon that special requirements are set up, such as passing a voltage level that exceeds 200% of the nominal line-to-ground in the high voltage black-out test without detecting any visual corona discharge on the stress-grading system. Moreover, a randomly selected, sample of coils has to pass the standard voltage-endurance test with no repairs on the corona-suppression system. Proper design, analysis, and development of corona suppression systems can significantly reduce the possibility of surface discharge inception and hence, extend the life of the H V stator winding which in turn, significantly enhances the reliability of H V machines. This paper presents the results of a research project carried out at National Electric Coil and University of Texas at Tyler towards the development of reliable corona-suppression systems for HV machines. The main drive behind the project was the development of coronasuppression systems, which can withstand elevated thermal and electrical stresses. Moreover, the trend of thinning ground-wall insulation for the purpose of machines up-rating and subsequent stress intensification, i.e. use of higher design electrical stress has become a major challenge to the research and development of insulation and corona suppression systems in HV rotating machines. New insulation and anti-corona systems that can withstand higher electrical and thermal stresses have to be developed. Moreover, new test requirements and standards have to be generated for the development, evaluation, and acceptance of these systems. Within this work, several anti-corona systems were developed, tested and applied to thousands of production coils. pared into flat rectangular samples, which are usually left to dry at the ambient temperature in case of paint or oven-cured in case of tapes. After taking multiple resistivity measurements, a mean value is usually checked against the required design range. The process may be repeated with adjustments if deemed necessaly. It is quite normal that measuring on flat samples may yield different results from that along the corner of a finished coil. Perhaps this may be attributed in part to the non-uniform application of the material in addition to the difficulty of mounting the measuring probe over the coil surface. Therefore, it is important to take measurements on finished coils and establish a range of surface resistivity for the corona-suppression system. Nevertheless, visual inspection of the corona-suppression systems on the slot portion and the end-grading zone is of considerable importance during the development stage and before undergoing electrical testing. Surface roughness, foreign material, and local damage of coatings and tapes can lead to electric field intensification and SUIface discharge. Based on the extent of the system irregularities, a simple touch up repair or reapplication of the system maybe performed before proceeding into further testing. Although, it is evident that the immediate part of the stress-grading zone next to the overlap area plays the most decisive part in the grading process, it is insightful to assess the potential distribution along the whole grading area. Figure 1 shows a schematic of the end-grading zone of a HV coil. The potential distribution along a paint system is presented in Figure 2. The potential distribution along the stress-grading system was measured using a small sphere contact probe tied to a standard 100 kV electrostatic voltmeter [l].The effect of stray capacitance was significant on the measured potential distribution along the stress-grading system. .~ 2 2.2 EVALUATION AND ACCEPTANCE EXPERIMENTAL SET-UP AND TESTING Three groups of corona-suppression systems were developed and investigated; (a) cell (that portion of the system within the slot) conducting paint and stress-grading paint (paint-paint), (h) cell conducting paint and stressgrading tape (paint-tape) and (c) cell conducting tape and stress-grading tape (tape-tape). It is worth mentioning that all tapes and paints used in this project are either silicon carbide, graphite, or silicon carbide and graphite based. In the mixing stages, necessaly amounts of graphite are added to control the paint resistivity. 2.1 TESTING A set of sample coils of each system are prepared for both evaluation and acceptance testing phase. Moreover, each coil with stress-grading system applied on all corners V f Ground +tress-grading s y s t e v DEVELOPMENT TESTING Mixing the ingredients to the required viscosity and low voltage surface resistivity range is the primer of developing corona-suppression systems. Normally, those are pre- Copper conductor (HV electrode) Figure 1. Schematic of the end-turn region of a high voltage ma- chine. IEEE Transactions on Dielectrics and Electrical Insulation Vol. 9, No. 4; August 2002 571 sional problem. (2) The stress-grading system is approximated by a linear resistive layer of finite thickness. ( 3 ) No free charge exists at the interface between the gradient system and the adjacent media. The electrostatic field distribution is obtained by solving the boundary-value problem formulated in the region under consideration V.( a*) =0 .I= iE I 0 2 4 Msvurlng pdrd fmm sk4 (1) (2) where, 6 is the electric potential and 2 is the complex permitivity of the medium, which is given by 6 an I Figure 2. Measured potential distribution along the stresr-grading of a 13.8 kV coil at room temperature Ill. I can he treated as a sample of four test specimens provided that consistency was maintained through the application process and a minimum leads’ clearance does exist. This phase of testing consists mainly of the high voltage blackout test and the accelerated aging voltage endurance test [13,21-301. All tests are performed at room temperature. Systems with reproducible results of the high voltage blackout test on all samples under similar conditions are subjected to the accelerated aging voltage endurance test. The integrity of the anti-corona systems under elevated thermal and electric stresses can he assessed through the study of several parameters and characteristics. This includes in particular, the recording of amount of discoloration and powder generation as well as any surface discharge or arcing spotted on the corona suppression system with reference to location and elapsed time under test. It is hoped that a following paper that will shed some light on the interdependence of thermal and electric stresses and the dynamic nature of degradation process in the end-turn zone of HV machines will he available to the reader towards the conclusion of this project. 3 MODELING AND ANALYSIS The model is based on numerical solution of the hounda l y value problem governing the electric.field distribution along the end-turn region using the finite difference method (FDM) [31-331. It is worth mentioning that although the model is not meant to provide extensive analysis of the corona-suppression system, it provides some necessary guidelines throughout the design process. Referring to Figure 1 , x-, y-. and z-coordinates extend along the axial coil length outside the slot, the coil height, and the coil width, respectively. To simplify the modeling process, the following assumptions are made: (1) Electric field is not z-dependent which reduces the model to a 2-dimen- where p is the volume resistivity of the medium. The potential solution is unique when Jr satisfies a set of specified boundary conditions. Referring to Figure 1, Dirichlet boundary condition must be imposed on the copper conductor where the voltage is given by the nominal line to ground value, V *=V (4) Similarly, on the grounded screen, Figure 1, the potential boundary condition is given by *=0 (5) In the absence of a perfect conductor, no surface current exists at the insulation-stress grading, insulation-air, or the stress grading-air interface. Using the complex perimitivity notation given by equation (3), the Neumann boundary condition applies at all interfacial boundaries. At the interface between the ith and kthmedium Extending the modeled region between ( x = 0) which is located at nearly 4“ deep inside the slot, Figure 1, to the end of the corona-suppression system ( x = L ) , the horizontal component of the electric field across both ends ( x = 0, x = L ) can be ignored and the boundary conditions on both sides of the insulation may be approximated by -= dX 0 (7) Considering the geometrical complexity of the domain and the type of boundary Conditions, the solution of this boundary-value problem can only he obtained numerically. In this study FDM with non-uniform grid is used to calculate the electric potential and field distribution along the modeled end-turn portion of a high voltage coil, Fig- El-Kishky et al.: Experience With Development and Evaluation of Corona-Suppression Systems 572 ure 1 , A versatile FORTRAN.90 program was built for analysis of the problem under consideration. The program is capable of simulating both linear and non-linear resistive stress-grading systems, however, only the results of a simple resistive grading system are presented here. In a nonlinear resistive grading system, the resistivity can he expressed empirically as a function of both the electric field and temperature, p = f ( E , T ) . The resistivity is updated each iteration using the most recent value of the electric field and temperature. If $PI, E f " , p p l , TP' and * .: .+I, El""', p. ..l " + ' ) , Tpt" are the corresponding.values of the electric potential, field, and temperature at node k after iterations n and n + 1, respectively, this iterative process can he given by Table 1. Low voltage resistivity measurement of corona-suppression coatings at room temperature. .. s,,$~~s~~n system Resistancc of cell conducting coating Paint-paint 1-5.0 Paint-tape kW2/sq 1-5.0 k Qsq Taoe-taoe L 1 1-40 k n/sq Surface resistance of end-grading system 0.5kV 1.0kV 15.0 to 250Gn 20.0 to 300Gn. 20 to 300Gn 2.0 t" SOGR 3.5 to 80GR 3.5 to 80GR Table 2. Experimental test resuits of corona-suppression systems on 13.8 kV Epoxy coils. Voltaee C0rO"a. endurance test at suppression ac high voltage 30kY and system black out test 100°C I Both electric field and thermal analysis of nonlinear resistive grading systems are yet to he presented in a manuscript under preparation. 4 RESULTS AND DISCUSSION Paint-oaint systcm-11 3 four-coils samples no visual corona spotted at 16kV Paint-paint system-l 3 four-coils samples no visual corona spotted at 16kV spotted after 400Hrs I Paint-tane r~ 6 four-coils samples no visual corona spotted at 22kV no deterioration spatted after 470 Hrs Tape-tape 2 four-coils samplcs (12kV and I6 kV level) no visual corona spotted at 12kV and 16kV level not applied ~ The project was focused on developing corona suppression systems for both Epoxy and Polyester vacuum pressure impregnated (VPI) coils taking into consideration the variation in the design and manufacturing process of both systems such as the average electric stress in the insulation or volts-per-mil (VPM) and the curing cycle requirements. Figure 2 shows the measured potential distribution along the stress grading system of a 13.8 kV, Epoxy Coil at room temperature. The salient non-hearity may be attributed to the electric field intensification in the region adjacent to the slot exit. Both conducting and stress grading paints and tapes are classified according to the supplier and the initial resistivity level. Several samples of each group were manufactured and the low voltage surface resistivity at 500 V and 1000 V (0.5 and 1.0 kV/cm) levels was measured at different locations along the stress-grading zone as well as on the cell portion. The average low and average high values are used to determine the range of variation of surface resistivity. Samples with anti-corona systems' surface resistivity values drastically different from the established range are corrected or reworked. All samples are subjected to ac high voltage blackout test, then only passing systems in which no sign of visual corona during black-out of repetitive samples are subjected to accelerated aging through voltage endurance test. Table 1 shows the average measured surface resistance along the cell portion and along the stress-grading zone ~~~~~~ ~~~ dctcrioratian spotted in a 2" wide zone next to the overlap area after 130 Hrs no deterioration for the accepted anti-corona systems according to the criterion mentioned above. The results of the ac HV blackout and voltage endurance testing on Epoxy coils with different corona suppression systems are displayed in Table 2. Although some of the systems showed better performance than others, generally, all the anti-corona systems were able to meet the basic requirements of application on high voltage rotating machines rated 6.6 kV and above. Despite the fact that it is economic and easier to apply and repair, the paint-paint system suffers from application inconsistency which may he attributed to low-skilled workmanship in addition to its vulnerability to different levels of visual damage and its volatility during installation. On the other hand, anti-corona systems based on B-stage tapes showed better consistency, durability and discharge resistance. Yet, the lack of bonding of anti-corona tapes to the ground-wall insulation could he a major problem, which may he attributed to the lack of resin contents in the Bstage tape in addition to other factors such as pressure and other curing cycle parameters. Repair of such systems IEEE Transactions on Dielectrics and Electrical lnsulotion Vol. 9, No. 4; August 2002 573 Figure 3. Stress-grading paint deterioration under elevated electrical and thermal strcsses. is difficult and may become even impossible without damaging the upper most layers of the ground insulation. Figure 3 shows one of the coil samples that passed a 16 kV ac HV blackout test undergoing the voltage endurance test . There is an obvious discoloration across a 5 cm wide area of the stress-grading system past the overlap zone. The discoloration may be attributed to relatively high discharge activity developed in the region after about 130 h in voltage endurance, which demanded intenention to repair the paint. Systems based on cell conducting paint along with Bstage stress-grading tapes have shown the best performance through the ac HV blackout test on Epoxy coils, Table 2. Test voltages exceeding 22 kV line to ground were applied with no trace of visual corona along the anticorona system. Figures 4a and 4b show the B-stage stress-grading system applied on a 13.8 kV Epoxy coil before and after the voltage endurance test. No visual corona or damage of the stress-grading system was detected within or after the voltage endurance test, Figures 4a and 4b. Moreover, the surface resistivity of the stress grading structure remained within the established design range, Table 1, after exposure to elevated thermal and electrical stresses for more than 400 hours. Nonetheless, no deterioration was noticed on the end-grading portion of the anti-corona system after taking the coils out of the voltage endurance test, Figures 4a and 4b. This may be attributed in part to the inherent capability of the resin in the B-stage tape to withstand elevated thermal stresses. Therefore, we recommend the paint-tape system for application on machines with up rated windings or running under nonfavorable environmental conditions. According to a contract between National Electric Coil and one of its customers, the paint-tape system applied on the customer's 13.8 kV coils has to be tested by an independent laboratoty selected by the customer in addition to all contract testing performed at National Electric Coil's (b) Figure 4. (a) Paint-tape corona suppression system before voltage endurance testing; (b) Paint-tape corona suppression system after the voltage endurance testing. HV testing facility. The system has to pass a voltage level of 16 kV in the high potential blackout test with no visual corona and also pass the voltage endurance test with no repairs on the system. Two 10-coils samples of the painttape anti-corona system were sent to a Canadian independent testing laboratoty for evaluation of the new system along with the assessment of the integrity of the groundwall insulation. Sensitive night vision cameras were used throughout the high potential blackout test where no visual corona was spotted at a level of 16 kV. The laboratory reported no repairs done on the system through the voltage endurance test. The tape-tape corona suppression system applied on Epoxy coils showed good performance under ac HV blackout test where two stress-grading system designs were made to meet a minimum of blackout test requirements of 12 kV and 16 kV, respectively. The main experience was the difficulty of controlling the surface resistance of the cell tape as well as adherence of the tape to the groundwall insulation. The latter may depend to some extent on the porosity of the tape as well as on the VF'I cycle parameters. Table 3 shows the results of testing Polyester coils with tape-tape corona suppression systems. Cell conducting tapes as well as stress grading tapes from different suppli- El-Kishky et al.: Experience With Development and Eualuation of Corona-Suppression Systems 574 Table 3. Experimental tcst results of corona-suppression systems O n 13.8 kV Polyester coils. ac high voltage ac high voltage Tapc-tape black out test black out test corona(B- stage stress (dry stress-grading suppression tape) system grading tape) Supplier A 2 four-coils samples Supplier B 2 four-coils samples faint glow spotted at lOkV not applied no visual corona spotted at 16kV not applied Supplier C 2 four-coils sam des faint daw spotted at lOkV no visual corona spotted at 12kV 02 01 ers were experimented including dry VPI and resin-rich B-stage tapes. Some contribution is claimed in successfully applying resin-rich B-stage stress-grading tapes to low cure-temperature Polyester coils. That system is based on the slow curing of the stress-grading tape at low temperature during the post-bake cycle of the Polyester. The system based on tapes from supplier B gave the best results that were easily reproducible on different samples of Polyester coils, Table 3. Sample of the FDM model results is presented in Figures 5 to 8. The effect of the gradient system resistivity on the total electric field distribution along the end-turn of a modeled 13.8 kV coil is shown in Figures 5 and 6. Considerable local stress intensification is noticed at the slot exit for a high resistivity stress-grading system. Lowering the resistivity of the gradient system significantly lowered the local field intensification, Figure 6. Figures 7 and 8 show the total electric field distribution along the modeled end-turn region of a 26 kV class high voltage coil with different applied voltage stress values, E,. The intensification is relatively less than that on a 13.8 0.3 0.4 0.5 Distance from slot exit, m Figure 6. Total electric field distribution along the end-turn of a S/m and E. = 2.36 kV/mm. 13.8 kV coil with u = 4x 0 0 01 02 03 04 05 Distance from slot exit Figure 7. Total electric field distribution along the end-turn of a 26 kV coil with u = 4 x lo-'' S/m and E. = 2.36 kV/mm. "1 I 0 1 I 0 0 0.1 0.2 0.3 Distance from slot exit, m Figure 5. Total electric field distribution along the end-turn of a 13.8 kV coil with u = 4x lo-'" S/m and Em= 2.36 kV/mm. 574 0.1 0.2 0.3 0.4 0.5 0.4 Distance from slot exit, m Figure 8. Total electric field distribulion along the end-turn of a 26 kV coil with u = 4x lo-'' S/m and E, = 3.15 kV/mm ZEEE Transactions on Dielectrics and Electrical Insulation kV class coil which may be attributed in part to a higher thickness of the main insulation in the 26 kV coil which enhance the non-linearity of the y-component of the main insulation electric field. In other words, it means less electric field at the interface between the main insulation and the end-grading system. Thinner insulation may lead to a higher local field intensification and subsequent discharge along the end-tum zone and within the insulation and hence, overall shorter insulation life. This may he of special interest to the utilities and coil manufacturers in regards to the issue of up rating electrical machines output power. Although thinner insulation simply means more space for copper and hence higher output and may improve heat transfer on the one band, it does lead on the other hand to higher electrical stresses in the main insulation and along the gradient system, which may lead to surface discharge and faster degradation of the insulation system. The model can be used to predict the optimal up rating level of a HV winding without overstressing the main insulation. 5 CONCLUSIONS 1. Several corona suppression systems based on conducting and stress-grading paints and tapes were developed. The new systems withstand voltage levels ranging from 110% to well above 275% of the nominal line to ground voltage on the HV blackout test. 2. The paint-tape anti-corona system with superior performance on both the ac HV blackout test and the voltage endurance test is recommended for application on high voltage machines with up rated windings. 3. Stress-grading system based on resin-rich B-stage tapes was successfully applied on low curing-temperature Polyester high voltage coils. j 4. A 2-dimensional FDM model is introduced to predict the electric stress distribution along the end-turn region of high voltage machines. 5. Local intensification of the electric field is significant1y:affected by the resistivity of the gradient system. 6. The model predicts the resistivity range of the gradient system in order to keep the maximum surface stress below a threshold value. 7. i h e model can he used to predict the optimal up rating level of a HV winding w‘ithout overstressing the main insulation. ACKNOWLEDGMENT The authors would like to acknowledge the support from National Electric Coil, Inc. and University of Texas at Tyler throughout this work. Special thanks go to Dan Bucklew and Fred Dawson of National Electric Coil, Inc. for their great support of the project. Vol. 9, No. 4; August 2002 575 REFERENCES I l l B. S. Nindra, V. Kogan, and F. Dawson, “Surface Corona Suppression in High Voltage Stator Winding End-turns,” Proc. of EEIC7&EM Conf., pp. 411-415, 1995. I21 A. Campbell, “Thc Refurbishment of a 60 MVAR Hydrogen Cooled Synchronous Compensator,”IEE Colloquium on Refurbishment of Machincs, Digest No. 410, PP. 3/1-3/7, 1998. I31 R. C. Arbor and B. Milano, “Diagnosing High-Potential Test Failures in Large Water-Cooled Hydrogeneratars,” Proc. of EEICflCWA, pp. 228-235, 1989. I41 G. K. Ridley, “Refurbishment and Uprating of Hydrogenerators,” Proc. intern. Conf. on Refurbishment of Power Station Electrical Plant, pp. 165-169. 1988. IS1 J. Rivenc, S. Dinculescu, and T. Lebey, “Suitable Properties of Stress Grading Materials,” Proc. IEE HV Engineering Symposium, pp. 22-25, 1999. [61 S. M. Cargill, D. G. Edwards, “Corona Screcn Effectiveness in Large Rotating Machines Under High Voltage, High Frequency Transient Conditions,” IEE Proc. Electric Power Applications, Vol. 145, pp. 469-474, 1998. I71 1. Rivenc, P. Bidman, and T. Lebey, “Stress Grading Matcrials: A Discussion on Lumped Elements Circuits Validity,” Proc. IEEE Intern. Conf. on Conduction and Breakdown in Solid Dielectrics, pp. 524-527, 1998. 181 Z. Yeo, F. Buret, and L. Krahenbuhl, “Electric Field Computation in Insulation Structures with Non-linear Conducting Laycr,” Proc. of the 10Ih Intern. Symposium on HV Engineering, pp. 17-20, 1997. 191 8. Hemalatha and M. C. Ratra, “Field Equalization at Coil Ends in High Voltage Rotating Machines - Numerical Approach,” Proc. of the 6Ih Intern. Symposium on HV Engineering, pp. 1-4, 1989. [lo] K. Kimura and S. Hirabayashi, ”Improved Potential Grading Methods with Silicon-Carbide Paints for High Valtage Coils,” IEEE Trans. El, Vol. 20, pp. 511-517, 1985. [lll C. Pinto, “A Generalized Approach for the Study of the Nonlinear Behavior of Stator Winding Insulation,” Proc. IEEE Intern. Conf. on Conduction and Breakdown in Solid Dielectrics, pp. 528-832, 1998.  K. Kimura, M. Tsukiji, T. Tani, and S . Hirabayashi, “Suppression of Local Heating on Silicon-Carbide Lzyer by Means of Divided Potentials,”lEEE Trans. El, VoI. 19, pp. 294-302, 1984. 1131 A. I. Harris? “Identification of Discharges in Electrical Machines-A Manufacturing Perspective,” IEE Colloquium a n Discharges in Large Machines, Digcst No. 264, pp. l/l-1/6, 1998. [I41 H. El-Kishky, W. Hoover and B. S . Nindra, “Electrostatic Field and Potential Distribution along Gradient Systems for High Voltage Machines.” Proc. IEEE Conf. on Electrical Insulation and Dielectric Phenomena, CEIDP, Victoria, BC, Canada, pp. ~02-~ns,2000. [I51 M. Belec, C. Hudon, D. Jean. C. Guddemi, and S. Lamothe, “Relative Risks of End Arm Discharges on Stator Bars,” Proc. IEE HV Symposium, pp. 107-111, 1999. I161 F. T. Emery and D. C. Johnson, “Voltage Grading Model for High Voltage Electric Generator Stator Coil End Turn Region,” Conf. on Electrical Insulation and Dielectric Phenomena, CEIDP, pp. 589-592, 1998. L171 E. Kuffel and W. S. Zaengl, High Volrage Engineering Fundamenlab, Book, Pergamon Press, 1988. (181 R. T. Waters, “Breakdown in Nan-uniform Fields,” IEE Proc., Vol. 128, part A, pp. 319-325, 1981. 1191 A. I. Davis, ”Discharge Simulation,” IEE Proc., Vol. 133, part A, pp. 217-240, 1986. 1201 J. M. Meek and J. D. Craggs, Editors, Elecrricol Breokdown of Cases, Book, John Wiley & Sons, 1978. 1211 IEEE Std 1043-1996, Recommcnded Practice for Voltage Endurance Testing of Form Wound Ban and Coils. -~ ~ ~~ ~ 5 76 El-Kishg et al.: Experience With Deuelopment and Eualuation of Corona-Suppression Systems I221 IEEE Standard 275-1981, Proposed Test Procedure far Evaluation of Systems of Insulating Materials for AC Electric Machinery Employing Form-Wound Pre-insulated Stator Coils. [231 IEEE Standard 434-1973, Guide for Functional Evaluation of Insulation Systems for Large High Voltage Machines.  IEEE Standard 56-1977, Guide for Insulation Maintenance of Large Alternating Current Machines. [251 ti. L. Moses, ElectnCal hulalion - Its Appliculion lo Shipbourd Elecm'cal Equipment, Book, McGraw-Hill, 1951.  IEEE Std 1147-1991, Guide for the Rehabilitation of Hydroclcctric Power Plants.  IEEE Std 56-1991, Guide for Insulation Maintenance of Large AC Rotating Machinery (10000 kVA and Larger.) [ZSl IEEE Std 95-1991, Recommended Practice for Insulation Testing of Large AC Rotating Machinely with High Direct Voltage. [291 IEEE Std 433-1991, Recommended Practice for Insulation Testing of Large AC Rotating Machinely with High Voltage at Very Law Frequency. [301 IEEE Std 434-1991, Guide for Functional Evaluation of Insulation Systems for Large High Voltage Machines. [311 R. H. Good, Jr. and T.J. Nclson, Classical Themy of Elecmic ond Mognelic Fields,Book, Academic Press, 1971. 1321 C. T. A. Johnk, Engineering Electromagnetic Fields and Wnws, Book, John Wilcy & Sons, 1975. [331 M. N. Sadiku, Numecca1 Techniques in Elecrromagnetics, Book, CRC press, 2000.