This action might not be possible to undo. Are you sure you want to continue?
1, January 1996
MODELING GUIDELINES FOR FAST FRONT TRANSIENTS
Report Prepared by the Fast Front Transients Task Force of the IEEE Modeling and Analysis of System Transients Working Group Contributing Authors: Ali F. Imece, Daniel W. Durbak, Hamid Elahi, Sharma Kolluri, Andre Lux, Doug Mader, Thomas E. McDemott, Atef Morched, Abdul M. Mousa, Ramasamy Natarajan, Luis Rugeles, and Eva Tarasiewicz
Abstrucz This paper summarizes the works of an IEEE T a s k Force focused on development of modeling guidelines for fast front transients (i.e., frequency range from 10 kHz up to 1 M a ) . These guidelines are intended t o be used by engineers involved in digital simulations of electro-magnetic transients, with particular emphasis on lightning surge analysis of overhead lines and substations. Modeling philosophies, simplified mathematical relationships, typical data, and examples are given for various power system components. To illustrate the overall modeling procedure, a case study has been presented. Keywords - Lightning, EMTP, Insulation Coordination, Overvoltages
The fast front transients in power systems cover a frequency range from 10 kHz up to 1 MHz. One of the primary causes of such transients is the lighming strokes to the transmission lines and associated backflashovers. The objective of this paper is to develop modeling guidelines for digital simulations involving fast front waveforms. Background
lightning surge protection. Determine minimum phase-to-ground and phase-tophase clearances. - Calculate Mean Time Between Failure (MTBF) for the substation. - Determine optimum location of capacitances to reduce steepness of surges. The lightning overvoltages are caused by either shielding failures or backflashovers of the tower insulation on the transmission lines. Direct strokes to the phase conductors may also become a concem if the transmission line andor substation is not shielded due to relatively low lightning activity in the area. Direct strokes to the substations are generally ignored, since it is commonly assumed that the substation is perfectly shielded, via lightning masts or other means. is one The Electro-Magnetic Transients Program (EMTP) of the most widely used digital transient programs for lightning simulations. Although the information presented in this paper generally refers to EMTP, the modeling requirements are general enough to be equally applicable to other digital transient programs.
Power system studies involving lightning surge 2. MODELINGGUIDELINES phenomena are performed to design transmission lines and The main emphasis of this section is to identify the substations, and for the protection of power system equipment [l]. Substation voltage uprating and models of power system components to be used in digital transmission line compacting are some of the specialized lighming studies. For each component, the important studies that require lightning surge analysis . The model parameters will be described, and typical values will objectives of these studies are to characterize the be provided. The general trends and rules of thumb that magnitude of the lightning overvoltages for insulation should be followed in the model development will also be requirements, andor to find the critical lightning stroke addressed. Finally, the sensitivity of the study results with respect to model parameters will be discussed. current that causes insulation flashovers. Specific study objectives for transmission lines may be Overhead Transmission Lines to: - Determine transmission or distribution Line The overhead lines are represented by multi-phase Flashover Rates (LFOR). models considering the distributed nature of the line - Establish line arrester application guidelines. parameters due to the range of frequencies involved. Phase conductors and shield wires are explicitly modeled between For substations the objectives may be to: towers and only a few spans are normally considered. - Establishlverify surge arrester ratings. models include the effects of tower geometry and - Find optimum location for surge arresters for Tower tower grounding resistance with special emphasis on its lightning current magnitude dependent characteristics due to soil ionization. Insulators are also modeled with their 95 WM 278-2 P W R D A paper recommended and approved flashover characteristics. AI1 incoming lines to the by t h e IEEE Transmission and D i s t r i b u t i o n Committee substation should be represented in substation design of the IEEE Power Engineering Society f o r presentation a t t h e 1995 IEEE/PES Winter Meeting, January 29, t o February 2 , 1995, New York, NY. Manuscript submitted August 22, 1994; made a v a i l a b l e f o r p r i n t i n g November 23, 1994.
studies, although the worst case tends to be with only one
line in-service due to voltage doubling effect from open circuit breakers. Depending upon the relative position of open circuit breakers inside the substation, the overvoltages may be higher for specific line outages, thus requiring the evaluation of different line contingencies. Fig. 1 shows the
0885-8977/96/$05.00Q 1995 IEEE
the stray capacitance of the insulator to wood and to ground may be neglected. 1.100 to 300 ohms. The matrix can be determined by a line constants Towers and Footing Impedance el Towers .In transmission line ~nfluenced by reflections from the adjacent towers. y while evaluating the insulator flashovers with propagation method. Some typical values of the capacitance and o t e that resistance of wood are given in Reference . This can be achieved by selecting f line spans modeled such that the travel time tween the struck tower and the farthest tower is more f of the lightning surge front time. ~ u f ~ c ~number e n t of adjacent towers at both sides of the struck tower should be modeled to determine the accurately. The number s modeled must be increased when the effects tail of the lightning surge is considered. = Tower footing resistance at low current and low frequency. Typical values range from . and the velocity of propagation can be assumed to be equal id the speed of light. Tower. = Soil ionization gradient (about 300 kV/m) [3. a similar approach can be provided that all the line spans and towers the struck tower and the substation are modeled i n transmission line and substation design studies. The flashover characteristics of wood can be modeled by a voltage-controlled switch.The steel towers are usually represented where: p = Soil resistivity. This results in reflections from the matched termination. Since the leakage resistance across the insulators is normally much greater than that of the wooden arms or poles. ohm Ig = the limiting current to initiate sufficient soil ionization. A I = the lightning current through the footing impedance. N the values of the capacitance and resistance are very sensitive to moisture contents.W d poles and crossarms can be modeled by a parallel combination of a resistor and a capacitor. The surge impedance for the tower depends on the structure details. The limiting current is a function of soil ionization and is given by: (2) and Termination .494 Phase Conductors ancl ShieM Wires I lnsulators I I I I by a resistance representing the tower footing impedance. The typical tower grounding resistance is between 10 and 100 ohms.9] The counterpoise can be represented as a nonlinear resistance with values calculated by Equation (1). and is determined according to the procedure outlined in Ref. 5 ] . The influence of the apparent tower footing impedance on the tower top voltage is determined by its response time and current dependence The response time is usually only important in cases where counterpoise is installed covering distances greater than 30m from the tower base. A. W o o d Poles and Cross-arms . ohm-m E . Within 35m of the tower base. Overhead Transmission Line. and Insulator ~e~esen~t~on. Ref. Similarly. model of transmission line and tower used for lightning where: R. ohm RT = Tower footing resistance. In substation design studies. the suucture is also modeled in a similar fashion to the cross- as a single conductor distributed parameter line terminated a r m .  Shows the characteristic per-unit volt-time curve with its upper and . it may be neglected. (1) Fig. TRe peak overvoltage occurring on the tower is mainly determined by the apparent footing impedance since reflections from the tower base will arrive much sooner at the tower top than reflections from adjacent towers. the time response can generally be neglected and the tower footing impedance is determined by using the current dependence of the tower footing resistance as follows E 4 . e extended beyond the last tower can be represented with a matrix of self and mutual resistances equal to the ng line surge impedances. In the case of wood pole. [l].
For a simplified analysis. The breakdown occurs when the leader length “L” is equal to gap length “g” and the corresponding time specifies the t t in Equation (4). The leader development method is recommended as described below. t. assumed zero ceramic insulators.e. Typical capacitance values for From Ref. The method is partially based on experimental results which. Then this equation is combined with the flashover switches in parallel with capacitors connected circuit equation determined by the experimental condition between the respective phases and the tower. Accurate representation of the short and long air gap (insulator strings and spark gaps) breakdown is necessary to overcome the shortcomings of volt-time characteristics. and a gap length up to 7m. It can be applied to a very wide range of experimental conditions such as gap configuration (rodplane and rod-rod). This method can yield the breakdown volt-time characteristics of a long gap with any impulse wave in extremely short time (e. Insulators leader length.3 I:[ EO 600 670 520 600 - Using the differential equation of (6). impulse voltage polarity (positive and negative). = Streamer propagation time The backflashover or flashover mechanism of the insulators can be represented by volt-time curves. Ten pin t..o 12 1.e.8 1. the tower structure. in tum.V(t) . K 1 = 4OO*L K2 = 710*L Vv-t = Flashover voltage. zero time) for the volt-time characteristics must be synchronized to the instant that where: lightning stroke hits the shielding wire or the tower top. = ti + ts + tt (4) insulators in a string has an equivalent capacitance of 100/10 = 10 pF/suing. The experimental volt-time characteristic is only adequate for relating the peak of the standard impulse voltage to tbe time of flashover. gap geometry and electric field intensity with two contributing factors.. gap length.g. psec the insulator voltage exceeds this voltage. (around 2011s) and is determined by the E50 = Average gradient at CFO voltage. a detailed arcing model for flashover is not necessary. In the numerical algorithm. It bas been reported [11. the leader length can be solved as a function of time. and compared to the actual insulator voltage. i. leader core and The insulators are represented by voltage-dependent corona cloud. first the leader velocity is empirically formulated as a function of the applied voltage.Eo] dl g-L (3) where: lower limits for a wood cross-arm subjected to a 12/50 ps voltage waveshape. kV = Leader length. The and solved numerically in a time step loop until the capacitors simulate the coupling effects of conductors to breakdown as described below. By solving a set of differential where: V(t) = Voltage across gap. The impulse flashover voltage of wood as function of its length for different moisture contents is also provided in the same paper. The insulator flashover For ts: voltage can be calculated using the Quation (3) during the . Capacitance values for non-ceramic where: insulators are an order of magnitude less for comparable 4 = Corona inception time. .0. any impulse voltage waveform. kV/m physics of air gap breakdown.12] that calculations of . h e insulator. it is inaccurate to assume that flashover will occur when a voltage wave just exceeds the volt-time curve at any time. ps Note that knowledge of the performance of insulation under the stress of the standard impulse is not sufficient to predict the performance of insulation exposed to any nonstandard lightning impulse.495 equations leader development is calculated. while for pin expressed as: insulators. kV L = Insulator length. If ts = Streamer propagation time. and a For t c: short-circuit (ideal switch) representation will be adequate. The voltt c = Leader propagation time time characteristics of insulators can be represented as a function of insulator length [l]. Furthermore.. lead to some analytical formulations. the capacitance is around 100pFhnit. The Leader Development Method The leader development method gives special consideration to breakdown parameters and to physical aspects associated with the discharge mechanism. less than 1 psec impulse rise time).95 simulation. m t = Elapsed time after lightning stroke. The start-up time (i. the time to breakdown (tc) can be suspension insulators are 80 pF/unit. and then volttime breakdown characteristic is obtained in a numerical process which can be related to physical phenomena. The front time for the arcing occurs across t kV/m can be quite steep. m L g = Gap length. a flashover E = Maximum gradient in gap before breakdown. m and Configuration ~ ia p r s . Post I Polarity L I I Positive Negative Positive Negative Insulators Cap and Pin Insulators kV sec J 0. dL = K V(t)[.
due to multiple reflections). For decreasing voltages. A single-phase corona model has been implemented in EMTP [lo]. The views from all three ordinal directions are very beneficial to locate the exact position of substation equipment. Below corona inception. caution must be exercised 1. Typical surge impedance valves are between 3060 ohms with velocity of propagation equal to 1/3 to 1/2 of speed of light. 2 summarizes circuit breaker The corona constant K varies with conductor diameter circuit diagrams and minimum capacitance values used in and the number of subconductors as well as the polarity of lightning studies for different types of substation the applied surge voltage. However. Corona In this equation. These the constant varies from 4.8 x 10-3 pF/kVm for a single data are based on supplier information. typically about 60Q traveling wave reflection at the voltage is increasing. For disconnector switches or breakers have more than one negative polarity surges. and Buswork. such as closer than 3m to 5m.The bu conductors between the discontinuity points the substation. The corona-inception voltage (Vi) for a single conductor above earth is given by the streamer approach as follows the geometric coupling factor. and K = Coronaconstant instrument transformers.e. Furthermore. the q/V curve is a straight when performing lightning overvoltage calculations for Gasline determined by the geometrical capacitance. These modeled with distributed parameters to increase the q N curves can be divided into three parts: simulation time step. busbar sections longer than 15m are curves of charge (9) versus impulse voltage (V). are represented by their stray Vi = Corona inception voltage. This variation with voltage can be expressed as a voltagedependent capacitance (Ck) which is added to the geometrical capacitance of the transmission line. kV also be represented. model. substation transformers.20%. appropriate capacitances should be added to the of these values. otherwise. The time delay introduced by corona takes effect above the coronainception voltage (Vi) and varies with surge magnitude. as well as. Conductors and Cables . Although corona effects may reduce the peak of lightning related overvoltages by 5 .0 pH/m is used. if they are longer than 3m. to are close to each other. r = Conductor radius. critical not to assume a lumped model for the GIB. Kc is the coupling factor in corona. The capacitance of CVTs should V = Instantaneous impulse voltage. The excess capacitance (Ck) to be added during the regardless of the length. the q/V curve again is overvoltages at the open disconnect position within the GIB practically determined by the geometrical (i.496 leader development and breakdown volt-time characteristics were carried out and results were compared with experimental results obtained by authors. For a 3 1 " conductor diameter. previously published by other researchers. if some of the substation equipment The coupling factor between phases increases due. equipment. Fig. cm Note that the minimum conductor length with distributed h = Conductor height. when the actual data is not available. constant such as circuit breakers. The complexity of corona modeling and associated burden on the computer CPU time are the other reasons for such an approach.. their corona and is given by: capacitances can be grouped together for simplification. in some studies corona is neglected in order to be on the pessimistic side. a lumped parameter inductance of 1. Kg is The essence of the corona effect on the impulse is the introduction of a time delay to the front of the impulse corresponding to the loss of energy necessary to form the corona space charge around the conductor. and Cg is the geometric capacitance [31. Ci = Initial capacitance jump. Above corona-inception Vi. and only the lowest conductor and positive polarity surges to 2. kV capacitances to ground. the qN curve shows an the gas-insulated buswork (GIB) is considerably smaller initial capacitance jump (Ci) plus an increase in than the surge impedance of the air-insulated line or capacitance which is voltage dependent as long as buswork. In such scenarios it is capacitance. the GIB entrance could rapidly result in sizable 3. The modeling details of corona can be expressed by In some studies.4 x 10-3 pF/kVm value are reported as a pessimistic assumption. the constant is approximately half support. Further investigation of associated physical phenomena is expected in the future. Insulated Substations (GIs). and connections between the substation Ell: equipments are explicitly represented by line sections. These line sections are modeled by untransposed Vi = 23 (1 + (7) distributed parameter sections with modal surge (r)ln(r g) impedances. second stage of the q N curve is given by : Cables are represented by distributed parameter line sections. cm parameter representation dictates the simulation time step. where: Substation Equipment A The substation equipment. Very good agreement was obtained for various gap configurations with various voltage waveforms. The line where: parameters can be calculated using a line constants routine. Substation The overall substation models are derived from the substation layout drawings. for 8 subconductors and positive polarity surges. Since the surge impedance of 2. .
The arresters can be modeled as nonlinear resistors with 8 x 20 ps maximum voltage-current characteristics. Capacitances can be determined from the geometry of the coil and core structure. and the difficulty in choosing the parameters of the model. the capacitance to ground of all insulators should be represented. Insulators and Bus Support Structures . I 2700pF I 5000pF I Capacitance also depends on MVA. . the most conservative and therefore pessimistic approach would be to represent the transformer as an open circuit. To increase the level of accuracy when setting up a terminal model of a transformer. supplied by IEEE Working Group on Transient Recovery Voltages. and with velocity of propagation equal to the speed of light. Fig. and can be represented by ideal switches. especially in simulations that involve high frequencies or backflashovers. Surge Arresters The voltage-current characteristics of metal-oxide surge arresters are a function of incoming surge steepness. The drawbacks of this model are the increased computational time due to several nonlinear elements. Circuit Breaker Representations and Minimum Capacitance-to-Ground Values for Various Substation Equipment. Note that the nonlinear arrester characteristics need to be modeled up to at least 20 to 40kA. This model duplicates the transformer behavior. windings to core and windings to ground as well as bushing capacitances. can accurately determine how a voltage applied to one set of terminals is transferred to another set of terminals. In lightning studies.The bus support structures are represented by a distributed parameter model with surge impedances calculated from the structure geometry. described in [161. This is due to the fact that the solution algorithm allows for an incorrect operating point for one time-step while switching from one segment to another on the piecewise linear representation of the current versus voltage characteristics. The arrester lead lengths at the top and at the bottom must be considered to account for the effects of additional voltage rise across the lead inductance. The model presented in  allows the simulation of any type of multi-phase. The model is used to calculate the surge transfer from winding-to-winding as in the case of generator (or motor) protection studies. The winding and core capacitances constitute together a crude potential divider in which the voltage divides inversely to the capacitance.1 to 1 ohms. The pseudo non-linear resistor should be avoided if possible. A resistance equal to the surge impedance of the winding can be placed across the capacitance. 2. conducted to design transformer protection at the terminals against high voltage.497 Circuit breakers. However. high frequency disturbances.O pH/m will be sufficient. live tank 1 DisconnectorSwitch 5PF I CircuitBreaker(DeadTank) Bus Support Insulator Capacitive Potential Transformer I lOOpF 8OpF 8000pF I 150pF 120pF 5000pF I 600pF 150pF 4000pF I I I Magnetic Potential Transformer Current Transformer Autotransformer" I I 500pF 250pF 3500pF 1 550pF 680pF I 6OOpF 8OOpF I Recently developed transformer model. together with capacitance values for outdoor bushings are presented in [lS]. Protective characteristics for surge arresters showing crest discharge voltage versus time-to-crest of discharge voltages are available from manufacturers [71. An iterative non-linear resistor is preferable. similar to the transmission line bus tower models [l]. Values of the winding capacitance. There have been few attempts to produce frequency dependent surge arrester models such as the model suggested by IEEE Surge Arrester Committee which is fairly accurate for a wide range of surge wavefronts . over a wide range of frequencies and acts like a filter suppressing some frequencies and passing others. dead tank 5OPF / [ ' Circuit breakers. The opedclose status of circuit breakers/switches should be considered. The representative grounding resistance inside the substations is between 0. since high current surges initiated by close backflashovers can result in arrester discharge currents above 10 kA. Comparative simulations showed that the support structures do not have much impact on the simulation results. or from manufacturers and tables. The frequency dependent characteristics of arresters may be of significant importance when excited by fast transient surges. since the substation capacitance is one of the critical parameters that modify lightning surge waveshapes. The model can be enhanced by adding the inductive transformer model and relevant capacitances between windings. it is recommended to account for the capacitance of the winding and the capacitance of the bushing to which the winding is connected. A lumped element representation with an inductance of 1. and can be neglected. multi-winding transformer as long as its frequency characteristics are known either from measurements or from calculations based on the physical layout of the transformer. The use of frequency dependent transformer model is important when determining the surge that appears on a generator bus when a steep-fronted surge impinges on the high voltage terminals of the generator step-up transformer.
It is well recognized that the flashover voltage is also dependent upon the steepness of the voltage. The lightning stroke is captured by the shield wire at the point of connection to the tower (i. In this case. 3. This assumption was justified since the lightning is assumed to have stricken the first tower outside the substation for the base case. and duration. the lightning is assumed to strike the fmt tower outside the substation on 230 kV line #2 at tower T 2 . and to establish the extent of the overvoltage control required to achieve the desired lightning performance. A rigorous approach would require the front of the lightning current source to be upwardly concave. if its voltage reaches the critical flashover voltage (CFO). however.90%)range from about 3 p at 20 kA to about 8 ps at 200 kA. The detailed calculation procedure for these parameters is shown in the CIGRE Guide. When the insulator flashover occurs. maximum current steepness. a negative triangular wave shape for the lightning current source can be selected. Overvoltage control consists of selecting optimum rating and location for the surge atresters inside the substation. median front times (30% . The lightning stroke parameters used in this study are given below: Peak Current = . its highest point) and is represented as a current source with a negative triangular waveshape. The severity of the overvoltages at the substation depends on how far the struck tower is away from the substation. corona effects are neglected. The parameters of a lightning.111. The time step selection depends upon the steepness of the surge. Maximum current steepness ranges t 200 kA from about 20 kA/p at 20 kA to about 48 k A / p a [3. Note that the double exponential impulse model given in EMTP should be used only with caution since this model does not accurately reflect the concave wave shape of the wave front.e. After the lightning stroke hits the shield wire. The substation is uprated from 115 kV to 230 kV maintaining 550 kV BIL insulation. The parameters of the lightning stroke. are determined considering the line dimensions and lightning incidence data for the site of the substation . Further distances result in less severe overvoltages due to stronger distortion and attenuation effects due to line parameters and also corona. the Time Step and Sitnulation Time The accuracy of the digital simulation can be affected by the use of a time step that is too large or too small. Backflashover .Lightning strokes of amplitude below the critical shielding current. if a lightning more than 110 kA would strike the first tower of either of the two incoming lines.typical 40 ps . Ref. In the computer simulations. The total simulation time was chosen to be sufficiently large to observe the decaying of overvoltages and reflections from the incoming lines. In this study. Note that the peak current is statistically related to the steepness or time to crest of the current wave form. The maximum lightning current that can strike the phase conductors of any given overhead transmission line can be predicted by using the methods recommended by the IEEE Working Group 0 x 1 the Estimation of Lightning Performance of Overhead Lines . . protective levels for line entrance arresters are below conductor corona starting voltage.Lightning strokes of high magnitude. resulting in voltage rise across the insulator strings supporting the phase conductors. rise time. can bypass the overhead shield wires and strike directly on the phase conductors.. They are generally characterized by log-normal distributions. The instant of lightning stroke with respect to the instantaneous steady-state ac voltage must be coordinated to maximize its impact for worst case conditions.110 kA Time to Crest = 3 ps HalfTime = 75p These parameters are based on the MTBF rate of 200 years for the substation. Computer Model The lightning stroke is usually represented by a c m n t source of negative polarity. in the range of 20 kA up to values rarely exceeding 200kA. such as crest. The steepness increases as the peak current increases. In this current range. As a general guide: range of time step = 1 to 20 ns. Direct Lightning Stroke . a surge with a very steep front propagates through the multi-conductor trammission line toward the substation. 550kV BIL substation ana its incoming transmission lines were developed as shown in Fig. the minimum length of traveling wave models. tower top voltage increases. that is. front time. however. Based on the guidelines described in this paper and using the substation layout drawings. cause the backflashovers. There is a 50% probability of insulator flashover. are determined by a statistical approach considering the ground flash density at the location. plus the use of flashover gaps and surge arresters with significant lead lengths. The strikes further away from the substation are less severe.498 Lightning Stroke 3. and duration. and are also used in calculation of LFOR and MTBF. These parameters are all statistical in nature. insulator flashover voltage has volt-time dependency. for practical purposes. CASESTUDY The modeling guidelines described in this paper are applied in a system study where a lightning surge analysis of an uprated substation is performed. the front time increases as well with peak current. generally less than 20 kA. a linearly rising front at the selected maximum current steepness is sufficient. an EMTP model for the 230kV. Also. This can be achieved by properly selecting the magnitude and phasing of the three-phase sinusoidal voltage sources at the terminating point of the transmission lines. such as crest. The specific objectives of the study are to evaluate the severity of lightning overvoltages stressing the substation. In this study. typical = 5 ns range of simulation length = 20 to 50 p. This assumes that substation would fail. Typical lightning current values for backflashovers and direct strokes are discussed below.
The effects of the lightning overvoltages are performed with a 5 ns time-step. 4 and 5. The simulations . Additional simulation Simulation Results results showed that by placing line entrance arresters overvoltage control inside the substation can be The simulation results are given in terms of plots of accomplished. and surge arrester energy duties that the selected time step is sufficiently small for the are evaluated for several contingencies. Substation EMTP Model Details. Modeling philosophy. A comprehensive set of lightning overvoltage studies are Although emphasis has been on lightning surge analysis of performed to establish the minimum insulation transmission l i e s and substations. simplified mathematical current and energy duties. and tower footing BIL substation from 115 kV to 230 kV. In some of the transients of interest. In overall conclusion. in addition to arrester voltage. cases simulated. phase-@ground and phase-to-phase voltages at dl equipment locations. Decreasing the time on the equipment BIL. application of 180kV waveforms for the important system variables.12' #I b + i 38 ^ . such as l i e rated metal oxide arresters at both line enrrance and all insulator voltages.499 -$ i 41 #3 I #2 A. modeling guidelines are maximum simulation times i s set to 20 ps. it has been observed that the BIL rating o f the substation could be exceeded. Inside the substation. at 4. tower top voltages. 3. CONCLUSIONS minimum clearance points. power system components in studies of fast front transients. Sample waveforms for the base relationships and typicaI data are presented for various case are provided in Figs. phase-to-ground and phase-to-phase transformer locations makes it feasible to uprate the 550 kV voltages at towers. -23OkVllne#l Fig. indicating phase-to-phase clearances. and at midway between arresters are provided.l Y I A. minimum phase-to-ground and step down to 111s' gives almost identical results. resistance voltages. requirements.
. .S Time i n Microseconds' Fig. * .0 . . . TOWER 41-1 INSIILATOR IkV] Fig. . S ad. d.5 5. Tower # I 2 1ooo7 - TOWER 11-2 PHASE VOLTAGE [kV] I Lightning Stroke lwo.. . . .. .0 14. . Struck Tower Voltage Wavefonns. ~ . 0... 5.500 \ . 4. ____-__________ 5oo ..0 . TOWER I1-2 INSULATOR IkVI . . d. . . * . . . . . o o o ~ . . 7!5 2.0 1 j . 230 kV Line Voltage Wavefonns. * . . .
121. NO. Altogether. this paper provides a working reference document for engineers involved in electromagnetic transients simulation of transmission and distribution systems. Volumes I and 11. Suliciu. as well as air-insulated and gas-insulated substation. and I Suliciu. A. Report Prepared for Bonneville Power Administration. 1988.. 5. Note that field test validation of the mbdeling guidelines is reported in Ref. 1982. M. Electromagnetic Transients Program Reference Manual (EMTP Theory Book). L. 1993. August 1981. Vol. Weck. pp 1909-1915. “Modeling of Metal Oxide Surge Arrester. pp. Rioual. Eriksson. “A New Calculation Method of Breakdown Voltage Time Characteristics of Long Air Gaps. EPRI EL-6474. M.” IEEE Trans. Application of Surge Protection Devices Subcommittee. 1556-1563. 1989. California. June 1985.” CIGRE 33-16. A. c) calculation of flashover rates or MTBF..501 applicable for similar phenomena having the same frequency bandwidth (i. REFERENCES Transmission Line Reference Book.W. Electric Power Research Institute. 1991. “Guide to Proceiures for Estimating the Lightning Performance of Transmission Lines.4. 1991. 9. on P WRD.M.” IEEE Trans. PAS-98. Paris. on PWRD. PAS-100. et. al. pp. A. August 1986. Greenwood. 1992. “A High Frequency Transformer Model for the EMTP. Darvenniza and R.M. Vol. California. “A Simplified Method for Estimating Lightning Performance of Transmission Lines. K. No. 1. “Simplified Procedures for Determining Representative Substation Impinging Lightning Overvoltages. No. Marcel Dekker Inc. 1994.H. PAS 104. 3. “Electrial Transients in Power Systems. T. 1985.. Portland. “The Soil Ionization Gradient Associated With Discharge of High Currents Into Concentrated Electrodes. “An Improved Method for Calculating the Impulse Strength of Wood Porcelain Insulation. H. Suzuki. ADrill992. Product and Application Guide. Examples where such modeling guidelines are applicable include: a) selection of surge arrester location and rating for protection of transmission lines. “Short and Long Air Gaps (Insulator Strings and Spark Gaps) Modeling for Lightning Studies with the EMTP Program. Further illustration of the modeling guidelines has also been demonstrated through a “case study”. 3. Vol.J. NO. pp.H. IEEE Working Group on the Estimation of Lightning Performance of Overhead Lines.” Second Edition. Surge Protection Devices Committee. PD-8.” CIGRE Working Group 33-01 (Lightning) of Study Committee 33 (Overvoltages and Insulation Coordination). PAS-104. b) selection/verification of clearances within substations and transmission towers (i. NO. Dommel. Still”. Mousa.e.” IEEE Trans. 7. Shindo and T.4. pp 1669-1677. Palo Alto. 10 kHz up to 1MHz). phase-to-ground and phase-to-phase). on PWRD.” IEEE Trans. New York. 8. . 1986. 345 kV and Above. “A Rate TyDe Constitutive Equation for the Description of -&e Corona Effect. John Wiley and Sons. M.11. Ottevangers.. Morched. Vol. High Voltage Engineering. Vol. Vol. EPRI Substation Voltage Uprating. 6. No. 1979. Marti and J. 919-932. etc.” IEEE Transactions on PAS. M.e.” IEEE Transactions on Power Apparatus and Systems. p]   [io] [ill [I21 U31 141 151 [I61 GE Tranquello XE Station Surge Arresters. Oregon.” EPRI. A. Palo Alto. Second Edition. IEEE Working Group 3. Khalifa.. on PAS. 1990. 3681-3685.” IEEE Transactions on PAS.
and in specific situations where adjacent span lengths to stations or stricken Breakdown Time ( p s ) Figure AI: Calculated Volt-Time Curve.A. to describe Case 2. positive polarity The CIGRE estimate with the recommended factor of K=1.6 and AI Gap Length = 2 m. This seems almost adequate. This circuit generates a lightning impulse with (1. un-ionized footing impedances. CHISHOLM (Ontario Hydro Technologies) and YAMA (Central Research Institute of Electric Power With advances in availability and ease-of-use. rod-plane gap. Maximum stress on insulation occurs just before return of reflections from nearest adjacent towers. nonstandard (partially-chopped) lightning impulse wave insulation strength. The use of an equivalent front time describes the co-incidence of the maximum steepness with the peak current. Case 2: 1 . it is more important to have a good nonstandard-wave flashover model. such as the description given by the leader progression (LP) model. Equation (3). but the 2-ps time is generally before the time of maximum stress. The LFOR problem can be broken into at several time periods.8 m*/(kV-s) matches the IEEE volt-time curve much better. The balance gives adequate LFOR estimates. or instead should the value of K be increased to non-physical values? What range of adjustments would the authors consider reasonable to achieve this match? . This model uses a correct application of the standard lightning impulse volt-time curve. This class is also important at time of any insulator flashover. since the chop time is steeper than 300 ns and the voltage collapse will regenerate similar transient effects on other stressed insulation. Should the CIGRE LP model be recommended in spite of the fact that it overstates the cap-and-pin insulator volt-time curve? Is the additional model of leader current important. The Fast-Front Transients Task Force is to be commended for developing suitable recommendations for EMTP models of several important components in the Line Flashover Rate (LFOR) problem. Line parameter. Results are sensitive to estimates of stroke current rate-of-rise and current. A calculation was carried out to compare the CIGRE LP model with the reference volt-time curve of Equation 3. the CIGRE LP model shown is not as good for describing standard-wave flashover strengths on porcelain capand-pin insulators. corona deformation models important. 300pH and 400Q in series. However.2 m*/(kV-s) and gradient of Eo=520 kV/m is still too high by about 13%. adopted by the IEEE Working Group [Sa. LFOR problem associated only with very-large current events. However.502 Discussion W. This improvement tends to bring the CIGRE predictions about 8% closer to the IEEE VT curve than if the current is not modelled (using only Equations 5 and 6). the ss of programs are increasingly suitable for analysis of lightning problems on power systems. which has a high withstand strength at short times. at an equivalent front time of 2ps. Results in this case are sensitive to probability of high stroke current magnitude. 3500 h h 3000 v +-' 0 g2500 2000 K=l . The CEGRE method was expressed by the authors' Equations (5) and (6). An over-estimate of insulation strength by 13% would lower the LFOR estimate by more than 45%. adjustments to the LP model should also match observed characteristics of leader velocity from experiments. We would like the authors' view of this approach. Equations 5. In these cases. The fixed rise-time tends to overstate the steepness of large strokes. Possible worst case for energy absorption. ionized footing impedance. This 13% uncertainty also reduces our confidence in the LP description for nonstandard waves on porcelain insulator strings.revised] and updated recently in Reference [A]. when reflections from adjacent towers return. standard lightning impulse wave insulation strength. - additional term models the drain of current from the impulse generator into the gap. One rough division could be: towers are documented. EMTP programs should be especially useful in study of longerduration flashover problems in Cases 3 and 4.2E-6 C IGRE K=l . transmission line coupling parameters. > -o Y L I5Oo 1000 m 0 1 2 3 4 5 6 1 Case 4: 10-1OOp Stroke current flows into seven or more tower footings. The authors have adopted the CIGRE LP model . front-of-wave flashover strength model. The calculation circuit has a source impedance of 26nF. especially considering the improvement in line description needed to provide parameters for better estimates. We explored other values of K . One reasonable model for the LFOR is the simplified method developed by Anderson [ 11. electromagnetic coupling from stricken OHGW to unstricken phases. A 1-3m gap was modelled in series with a 400R front resistor and shunted by 7713Q in parallel with 1 nF. Figure A1 shows the remaining error. which is optimized for air-gap strengths. but LFOR estimates are very sensitive to the insulator strength model. for the case where the stress is a sum of both resistive (RE) and inductive (dUdt) effects.Y45)ps rise and fall times. five or possibly seven tower footings through series impedance of OHGW system. and found for example that a value of K=1.along with a third expression: Case 1: 300 ns lp Fast-rising components in the stroke current lead to high transient stresses on line insulation.8E-6 C IGRE Case 3: 3-lops Stroke current flows into three.3 p Stroke current flows either into one tower footing or into surge impedance of overhead groundwires. tower surge impedance. Sensitive to stroke current magnitude and rate of rise.
pp 102-115. Manuscrip. Therefore. and corresponds to the corona-inception field strength given by the streamer approach considered in equation (7). Correia de Barros. H. current magnitudes and possible stress on insulators. Menemenlis. However. and improving success. Manuscript received February 23. which plays an important role for fast front surges and can be neglected for slow front transients [ 2 ] . received March 8. As indicated already in the paper. However. Inc. Pittsburgh. NY). Eva Tarasiewicz (Ontario Hydro. Ontario. Schenectady.3. Sharma Kolluri (Entergy Services. we . Atef Morched (Ontario Hydro. the equations giving the two different-meaning excess capacitances can be found.S. which is shortly presented in this paper. ground conductors and tower for different time periods. Malewski. it is most suitable to describe corona by a charge-voltage curve. Abdul M. dependent on the voltage rate of rise.%dies< IEEE Paper 94 WM 043-0-PWRD. Jan. in order to avoid misusing of the information given in equations (7) and (E). PAS-96. de Jesus. 1995. NC). M. pp 919-932. [ l ] P.Correia de Barros. "Modelling of corona Dynamics for Surge Propagation. the following concepts should be clarified: Excess CaDacitancQ. Raleigh. Experimental charge-voltage curves (11 show that for slow front transiects. Members of the IEEE Task Force on Fast Front Transients wish to thank the discussers for their interest and valuable comments on this paper. Andre Lux (ABB Power T&D. Ramasamy Natarajan (EBASCO. especially where there is two-layer soil. We would like the authors to comment on the high values of I# associated with typical fourfooting tower electrodes. The authors'  and Reference [A] provide a healthy respect for some of the uncertainties. "Estimating Lightning Performance of Transmission Lines 11: Updates to Analytical Models". The authors recommend a simple model for tower impedance.. NJ). an excess capacitance appears. 1995. PWRD-8. Motoyama on line flashover rate gives a detailed accounting of the rate of rise of surge current. Metairie. We would finally ask the authors to clarify two minor points.4. or the derivative of the charge with respect to the voltage. which is more representative of bellshaped insulators. "Corona Characteristics of conductor bundles under impulse voltages: IEEE Trans.. MA). the latest is of practical interest. we agree that the authors have assembled a good working reference for EMTP simulation of transmission systems. which increases above corona threshold. No. Vancouver. the discussions on the IEEE Working Group method versus the CIGRE leader propagation model highlights the need for an improved model for flashover mechanism. 1977. Imece (Morrison Knudsen Corp.. also two different excess capacitances can be defined. NY). can be defined in the presence of corona: the ratio between the charge and the voltage. and this is also sensible for Cases 3 and 4. IEEE Trans. This can be described by a voltage-dependent capacitance. typical capacitance values for 25-kip standard cap-and-pin porcelain insulators are usually measured to be 30-35 pF range. + K (V-Vi) ~ Cdyn =dQ IdV = Cg + Ci K 4 + 2K V from where. it should be noticed that two different quantities. PA). Hamid Elahi (General Electric Company. My comments concem the paragraph on corona. Canada). Nova Scotia. Doug Mader (Nova Scotia Power Corp.V. Vol. Equating peak currents and rates of rise for a sine pulse and the standard lightning impulse suggests a representative frequency of 80-120 kHz for these effects. Dedham. IST I Technical University o f Lisbon. Durbak (Power Technologies. Maruvada. Daniel W. "A Simplified Method for Estimating Lightning Performance of Transmission Lines". the model may not be as important in Cases 3 and 4. PAS-104. the authors recommend 500 kHz for models of skin effect in the line constants section. The first is the one shown in equation (8). Chisholm and Dr. and can be identified by the sudden change of the charge-voltage curve slope. First. Schenectady. McDermott (AnsoR Corp. Second. W. For computation of surge attenuation and dlstortlon by corona. It is usually referred as the threshold voltage. No. In closing. the experimental curves show that the threshold voltage ~ has an higher value. a good LFOR estimate will also require a suitable stroke incidence model. Halifax. Venezuela). Caracas. According to the authors' reference . and Luis Rugeles (C. page 35: C = Q l V = Cg + Ci Ali F. Also. I understand the difficulty on accommodating on a IEEE paper the extensive work performed by the Task Force. Mousa (British Colombia Hydro. Standing upon the two different capacitance concepts. Toronto. Further.503 The footing impedance ionization model selected in Equations (1) and (2) is appropriate for Case 2. for fast front transients. in describing these factors in the LFOR problem. and look forward to seeing wider use of this work. A. [AIIEEE Working Group on Estimating the Lightning Performance of Overhead Transmission Lines. We agree with the discussers that adjustments to the leader propagation model should match first and foremost observed characteristics of the leader velocity from experiments. Canada). The latest is the one to be used for surge propagation studies. Toronto. Edelca. Ontario.T.. Lyndhurst. April 1985./Fev. where some guidance to the readers may be missing. a shielding failure model and an estimate of local lightning ground flash density. M. IEEE Trans.  C. R.For the computation of surge propagation on transmission lines. July 1993. Corona-inceotion voltage A distinction should be made between what can be said the microscopic and macroscopic inception of corona. Canada). pp 1254-1267. We will address the issues raised in the discussions by order of appearance. [8a]IEEE Working Group on Estimating the Lightning Performance of Overhead Transmission Lines. BC. subtracting the geometric capacitance . Portugal : I wish to congratulate the Fast Front Transients Task Force members for the excellent work that has been done. The discussions by Dr.T. H. the threshold voltage is independent from the front slope. Canada). not the recommended 80pF. with the physical dimension of a capacitance. On the contrary.G. LA). In all cases. Thomas E. This may be explained by the dynamics of the space charge generation.
However.700 540 5000 ' * Rock footing are usually not used where pois low.000 5. 181. and are of the order of 'bome 10 pF" [17.504 100 500 1000 I -* -* 755 151 27.400 2. The stroke currents of interest in backflash The discussers' recommendations concerning typical capacitance values for 25 kip standard cap-and-pin porcelain insulators generally agree with the values found in the literature. the following insulator length versus capacitance curve for Lapp station post insulators indicates slightly higher values if extrapolated considering the dimensions for typical .
. In this recommended considering the frequency of surge connection.505 suspension insulators . We wish also to report some minor revisions to the text which arose from discussions within the Task Force following publication of the preprint of this paper: T O J 0 10 20 30 40 50 60 Capacitance [pF] The section entitled ‘Lightning Stroke” discusses the typical current amplitudes involved in For calculation of line constants. good agreement was obtained between the test and computer simulation results . The selection critical flashover voltage of the insulation. and determine the line parameters based be taken equal to 1. resulting in time delays observed as increases in the value of threshold voltage. which can be modeled as an RL component in series with the excess capacitance. some proposals  explain this apparent increase physically in terms of the mechanical inertia of the particles and mobility losses. the following should be noted: propagation after backflashovers and reflections from adjacent towers. 500 kHz is backflashovers and in direct strikes.+ c. She is correct steeper fronts and may also be higher for some nonconcerning her comments on the concept of dynamic standard waveforms. Cdyn= c Professor Correia de Barros is also correct in pointing out the usual distinction between the corona-inception voltage.2 x CFO. More rigorous approach a) Regardless of mechanism by which a lightning voltage surge is generated (backflash or direct would be to perform an FFT on the model step strike). The value of 80 pF per unit was used in a study conducted by several of the authors in which transient study calculations were compared with high voltage impulse tests. impulse waveform. For users who need hence the insulation can carry higher voltages accurate line models for surge propagation studies. The discussion by Professor M.T. The increase in threshold voltage with increasing dV/dt has been accommodated by selecting higher values for Vi in Equation (8) than would be predicted by Equation ( 7 ) . and.2 x 50 ps k€h in Reference .KVi + 2KV (8) and the sum of the excess capacitance and geometric capacitance is the dynamic capacitance: 1 . obtained from the streamer approach as in Equation (7) and the threshold voltage corresponding to the sudden change in QNcurve slope. disconnect switches and suspension and post insulators [2. The tests were performed on a full-scale substation mock up consisting of strain and rigid bus work. refer to the exact line parameters presented at 400 ii) the CFO is based on the standard 1. values are typically chosen based on organizations’ experiences or rule-of-thunib approximations. The of 500 kHz is also supported by the error analysis on 1. Station Post Insulator Capacitance loo capacitance for use in surge propagation studies. Figure 1. where CFO is the on the observed resonant frequencies. 50 % of the time. the maximum amplitude of the surge may response. Equation (8) should read: C k = Ci .2 multiplier accounts for two effects: the line constants at various frequencies shown in i) the fact that the CFO is a median value and Reference [lo] of the paper. 201. As in the case of modeling studies. However. Correia de Barros on corona model is very helpful to clarifjl the definition of excess capacitance for steady-state The withstand voltage will hence be higher for versus transient conditions. With the selected capacitance values.
Lapp Insulator Company. H. 1994.Int. Torres. Marti. Paper 33-207. on Lightning and  J. varying p with voltage produces results which are inconsistent with the physics of the problem. Oklahoma. A.506 b) In determining the maximum stroke current which can cause shielding failures. where S is the striking distance to the wires.Private Communications. EPRI EL4651. Castellanos & J..S.’: EPRI EMTP Workbook 11.. J. .” . Oklahoma City. 1988. Insulators for High Voltages.0 is a factor which is a hnction of the voltage. and P I 1.  F. REFERENCES  A. April. As noted in the discussion of  by A. J. Peter Peregrinus Ltd.W. ‘A Revised Electrogeometric Model for the Termination of  E* Kuffel’ w*s*Zaeng17 High Lightning Strokes on Ground Objects. Looms.  Panek. 1995. 1988.. Static Electricity... D. on Large High Voltage Electric Systems (CIGRE).T. Reference  presents a revised electrogeometric model which gives predictions which match field observations. Engineerinn Fundamentals. Srivastava. Pergamon Press. Aerospace and Ground C o d . “Dynamic Corona Modeling in the EMTP with Simple Constant Parameter Circuit Components.It Canadian Electrical Association Meeting. Nilsson. 5. Maroh 1993. [ 191 Zoltan Szilagyi. 1984. 2. 1992. of Int.. Elahi. Mousa. Dommel & C. ‘Simple Overhead Line Models for Lightning Surge Studies. Reference  uses a model in which the striking distance to ground is taken equal to PS.R. 342-352.  H. R.. Cod. Manuscript received April 17. M. LaPanse. . Lux.. R. Stewart.”Proc. 30 . J. Aug. 1989. “Substation Voltage Uprating. S. Mousa and K.Sep. pp. Vol. Montreal. Porter.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.