EVALUATION OF LIGHTNING-INDUCED VOLTAGES ON LOW-VOLTAGE DISTRIBUTION NETWORKS

Transcription

1 IX International Symposium on Lightning Protection 6 th - th November 7 Foz do Iguaçu, Brazil EVALUATION OF LIGHTNING-INDUCED VOLTAGES ON LOW-VOLTAGE DISTRIBUTION NETWORKS Fernando H. Silveira Silvério Visacro LRC Lightning Research Center UFMG Federal University of Minas Gerais silveira@cpdee.ufmg.br visacro@cpdee.ufmg.br Av. Antônio Carlos 667 Pampulha 7-9 Belo Horizonte MG Brazil phone/fax: LRC@cpdee.ufmg.br Abstract - This paper presents a set of evaluations concerning lightning-induced voltages on low-voltage networks. Such evaluations were performed by means of the Hybrid Electromagnetic Model (HEM), a frequency-domain code based on electromagnetic field equations. Sensitivity analyses were developed in order to investigate the role played by the most relevant line parameters on the voltages induced along the electrical system and at the consumer service entrance. The performance of two different lowvoltage line configurations (conventional x multiplexed) usually adopted in Brazil was also investigated. The obtained results constitute a contribution to this theme and also provide elements to the development of protection practices to the electrical system and to the consumer against such phenomenon. INTRODUCTION Lightning-Induced Voltages are known as one of the most important phenomena capable to cause severe damages to electrical systems and telecommunication networks []. Despite the lower severity of their effects in relation to those associated with direct strikes, the probability of occurrence of lightning-induced voltages is much more pronounced. As a result, this phenomenon is the main responsible by faults and outages on low-voltage distribution networks. Furthermore, its effects can be transferred to consumer service entrances that are connected to the system, causing damages on electronic equipments with low isolation level. The knowledge of the characteristics of the voltages induced on low-voltage networks allows the specification of requirements to protect the electrical system and the consumer, as well. This scenario justifies the development of this paper. The amplitude and waveform of the voltages induced along the electrical system is closely related to the adopted network configuration. Features like line geometry, grounding resistance value, distance between adjacent grounding points and consumer load type affect the resultant induced voltage. Investigations concerning this subject have been presented in the literature along the years. Several approaches are normally adopted, including measurements using reduced scale-models to represent the involved system [], as well as numerical calculations with powerful computational models (in most cases, adopting a time domain-approach) []. The objective of this paper is to investigate the characteristics of the voltages induced on low-voltage networks due to nearby lightning, evaluating the role played by the most relevant line parameters. The authors of this paper aim to present a contribution to the theme by performing such evaluations using the Hybrid Electromagnetic Model HEM [4], a frequency-domain code based on electromagnetic field equations. The obtained results can provide elements that allow the development of protection practices to the electrical system and to the consumer against lightning-induced voltages. METHODOLOGY The evaluations presented in this work are based on systematic simulations using the Hybrid Electromagnetic Model HEM [4]. This model was developed on LRC Lightning Research Center, Brazil, for application in lightning-related problems [5-7]. The application of HEM for lightning-induced voltages evaluations incorporates simultaneously the current distribution along lightning channel and the electromagnetic coupling between line conductors and channel. Both stages are automatically computed without the requirement of theoretical assumptions concerning either the behaviour of the lightning current distribution or the behaviour of the electromagnetic interaction between the elements of the analysed system. The HEM has physical appeal and reproduces, even in a simplified way, physical features and associated process of the lightning channel. It can also model conductors in any

2 arbitrary angles, allowing evaluations related with inclined lightning strokes, tortuosities along the channel and complex line configurations. The HEM is also capable to compute the effect of a lossy ground on lightning-induced voltages. The details of the assumed approach are presented in [8] and are based on the work proposed by Norton [9]. Such approach was adopted in order to modify the electromagnetic coupling between lightning channel and line conductors. The lightning-induced voltages results generated by the HEM present an outstanding agreement with results associated with other computational models usually adopted in literature and some experimental results, as can be seen in [8] and []. DEVELOPMENTS A typical configuration of a Brazilian low-voltage network was adopted in the simulations. This condition is illustrated in Fig.. The neutral conductor is located 7. m above level and the phase conductors are installed below it. In order to avoid undesirable reflection effects, the conductors were matched in their ends. neutral phase phase phase 7. m. m. m. m Phase conductors 5 m Fig. Line configuration. Neutral Conductor The evaluated system considered the discharge occurring 5 m away from line center. It was simulated the injection of a /5 µs triangular waveform with ka peak value at the base of the lightning channel. Such condition is similar to the occurrence of a subsequent stroke. The current propagates along the lightning channel with light velocity (c = x 8 m/s). The resultant induced voltage is expressed in kv/ka. In this work, sensitivity analyses were performed relating the effect of several line parameters with the resultant induced voltage along the electrical system and at the consumer service entrance. Thus, for the sake of simplicity and to make clear the presented analyses, the simulations did not take into account the effect of resistivity. The was represented as a perfect conductor. The performed investigations were organized in two parts: first, evaluations related to the performance of the electrical system are presented remarking the effect of several parameters like the neutral conductor positioned above the phase conductors, number of grounding points, value of grounding resistance on the voltages induced at the low-voltage network. The second part is dedicated to evaluate the induced voltage levels at the consumer service entrance. The role played by the presence of the load as well as the adopted line configuration (conventional x multiplexed) on the resultant induced voltage was investigated. 4 RESULTS AND ANALYSES 4. Effect of Line Parameters on Lightning-Induced Voltages at Low-voltage Networks 4.. Influence of the Neutral Conductor The effect caused by the presence of neutral conductor above the phase conductors was investigated adopting the line configuration depicted in Fig.. Such configuration assumes two vertical down-conductors connecting the neutral conductor to the. The grounding was represented as 8 Ω resistances. The voltages induced at the center of the phase conductors are illustrated in Fig.. 4 ph ph ph ph ph ph Line without neutral conductor Line with neutral conductor Fig. Lightning-induced voltages at phase center - Influence of the neutral conductor. Obs: line configuration with neutral conductor has two grounding points R g = 8 Ω (ph = phase). As shown in Fig., the neutral conductor above the phase conductors promotes the reduction of the induced voltage peak values. In phase (positioned 7 m above level), it is observed a decrease around 8%, from 4.4 kv/ka to.5 kv/ka. Such reduction is due to the negative voltage reflection at the grounding points of the neutral conductor. This effect acts in order to diminish the induced voltage amplitude at phase conductor. Consequently, the neutral conductor works as a shielding conductor. The presence of the neutral conductor also imposes an oscillatory behaviour on the induced voltage waveform at phase conductors, as remarked in Fig.. Such behaviour is associated with the voltage induced at the neutral conductor that is also oscillatory. Due to the

3 electromagnetic coupling between neutral and phase conductors, such oscillatory behaviour is repeated on the induced voltage waveform at phase conductors. Fig. also presents similar induced voltages waveforms for the three phases. Therefore, for the sake of simplicity, the following evaluations of this paper will be related with the highest phase conductor (phase ) represented in Fig Number of Grounding Points Down-conductors were added to the line configuration illustrated in Fig. in order to investigate the influence of the number of grounding points on the voltages induced at phase conductor. Fig. remarks the line configuration with 4 and 6 grounding points. For this last case, the span length equals m. The grounding resistance (R g ) was fixed in 8 Ω. The voltages induced at the center of phase conductor are presented in Fig. 4.. m 7 m. m 7 m m 5 m 9 m (a) 4 grounding points m m m m 5 m m m Neutral Phase Neutral Phase (b) 6 grounding points Fig. Low-voltage network configurations with 4 and 6 grounding points grounding points 4 grounding points 6 grounding points Fig. 4 Lightning-induced voltages at the center of phase conductor (phase ) as a function of the number of grounding points - R g = 8 Ω. The obtained results denote a decrease of the induced voltages amplitudes with the increase of the number of grounding points. Such behaviour is explained by the reduction of the distance between adjacent grounding points (responsible by negative voltage reflections). Consequently, the subtractive effect, associated with the first negative reflection, happens in early times of the induced wave. The induced voltage peak value at phase center of the line configuration with 6 grounding points presents a decrease around 7% in comparison with the one induced at the phase center of the line configuration with grounding points (.5 to. kv/ka). 4.. Grounding Resistance Value Systematic simulations adopting five grounding resistance values (R g =, 5, 8, and 4 Ω) were performed in order to evaluate the relevance of this parameter on the induced voltages at low-voltage networks. Two overhead line configurations were considered with and 6 grounding points as previously illustrated in Figs. and (b). Such cases were named as "5 m span length" and " m span length" cases, respectively. The " m span length" case represents a typical condition presented in Brazilian urban areas, where the distances between adjacent grounding points are generally very short. Fig. 5 presents the resultant phase-to-ground induced voltages at the center of the phase conductor for the assumed line configurations. Table summarizes the induced voltage peak values Ω 4 Ω Ω 8 Ω 5 Ω 4 5 (a) grounding points Ω 4 Ω Ω 8 Ω 5 Ω 4 5 (b) 6 grounding points Fig. 5 Lightning-induced voltages at the center of phase conductor (phase ) for several values of grounding resistance and assuming different distances between grounding points: (a) 5m, (b) m. Table : Phase-to-ground induced voltage peak values at the center of phase conductor (phase ) considering the variation of grounding resistance value (R g ). grounding points (5 m span length) 6 grounding points ( m span length) R g (Ω) V p (kv/ka) Reduction V p (kv/ka) Reduction % % %..97% %. 4.77%.5 8.5%.85.94% As remarked by the results of Fig. 5 and Table, the effect of the grounding resistance value is closely related with the distance between adjacent groundings. The "5 m span length" case presented a reduction around 8% on

4 the induced voltage maximum amplitude at the phase center when the grounding resistance value was reduced from 4 Ω to Ω. The reduction observed for this same set of grounding values, but assuming the " m span length" case, is more relevant, around %. Fig. 6 and Table illustrate the phase-to-neutral induced voltages at the center of the low-voltage network assuming the above conditions. Such overvoltage has an opposite behaviour in comparison with the one presented by the phase-toground induced voltages remarked in Fig. 5. The obtained results denote the increase of the phase-to-neutral induced voltage amplitudes as the grounding resistance value is reduced. Such behaviour is explained below. Since the neutral conductor has a direct connection to the ground, the reduction of the induced voltage amplitude on this conductor due to the decrease of the grounding resistance value is more relevant. Moreover, the voltages induced at the phase conductor experience the effect of the grounding indirectly. Hence, the decrease of the grounding resistance values results in a trend of rise of the induced voltages amplitudes between phase and neutral conductors. For instance, evaluating the " m span length" case, whereas the phase-to-ground induced voltage presents a decrease around % when the grounding resistance is reduced from 4 to Ω, the phase-to-neutral induced voltage has an increase around 5% for the same grounding resistance variation. Similar results and conclusions are found in [] Ω 5 Ω 8 Ω Ω 4 Ω 4 5 (a) grounding points.5.5 Ω 5 Ω 8 Ω Ω 4 Ω 4 5 (b) 6 grounding points Fig. 6 Phase-to-Neutral induced voltage at line center for several values of grounding resistance and assuming different distances between grounding points: (a) 5 m, (b) m. Table : Phase-to-neutral induced voltage peak values at the center of phase conductor considering the variation of grounding resistance value (R g ). grounding points (5 m span length) 6 grounding points ( m span length) R g (Ω) V p (kv/ka) Increase V p (kv/ka) Increase %.65 % % % %.77 8%. 9.8%.88 5.% 4. Evaluation of Induced Voltage Levels at the Consumer Service Entrance Evaluations concerning the phase-to-ground and phaseto-neutral induced voltage levels at the service entrance of a low-voltage consumer are the scope of this section. Fig. 7 depicts the adopted line configuration that aims to represent a low-voltage network extending from the transformer secondary terminals to the consumer service entrance. The performed investigations considered a 5 m span length between the grounding points of the neutral conductor. 4.. Presence of the Load The modeling of the consumer load under transient conditions was not the purpose of this study, so the loads illustrated in Fig. 7 were modeled as simple resistors. The obtained conclusions are related with such assumption. In point A, a Ω load value was assumed in order to simulate a stage after the operation of a lowvoltage surge arrester installed at the secondary terminals of the transformer. In point B, two conditions were simulated. Initially, a Ω load connecting neutral and phase conductors was considered [-]. Following, such load was removed (so, the termination of the phase conductor was considered open). This last assumption aims to represent the consumer load disconnected from the electrical system, a condition that can occur if all the equipments of the consumer were assumed turned off. In all these cases, the assumed grounding resistance (R g ) were 8 Ω.. m 7 m Point A Point B Neutral Phase 5 m Fig. 7 Line configuration simulated in order to evaluate the induced voltages levels at the consumer service entrance. Fig. 8 shows the voltage induced on both neutral and phase conductors for the two specified cases. Fig. 9 illustrates the induced phase-to-neutral voltage for the above conditions. According to the obtained results, the induced voltage value at the phase conductor is always high (.8 kv/ka and. kv/ka for Figs. 8(a) and (b), respectively). Assuming a discharge with 6 ka current peak [4], that corresponds to the median value of the maximum current amplitude of subsequent strokes measured at Morro do Cachimbo Station, Brazil, the phase-to-ground

5 overvoltage can exceed kv and 7 kv, respectively. Such values are extremely severe, constituting a serious risk for the consumer s safety and protection, as well. The presence of the load connecting neutral and phase conductors acts in order to approximate the values and waveforms of the induced voltages on both conductors. For such condition, the stress level of the load (phase-toneutral voltage) is very small, as shown by the phase-toneutral induced voltage waveform presented in Fig. 9. Otherwise, when the consumer load is considered disconnected from the electrical system, the phase-toneutral induced voltage value is extremely intense (.5 kv/ka). Assuming the occurrence of a discharge with 6 ka current peak, such value becomes near 4 kv, certainly causing serious damages to the consumer load Phase-to-Ground Neutral-to-Ground Phase-to-Ground Neutral-to-Ground (a) with Ω load between (b) without load between phase phase and neutral and neutral Fig. 8 Lightning-induced voltages at the consumer service entrance (Point B) assuming R g = 8 Ω. - - Point B: no load - Point B: Ω load Fig. 9 Phase-to-neutral induced voltages at the consumer service entrance (Point B) assuming R g = 8 Ω. 4.. Performance of Line Configurations: Conventional x Multiplexed In Brazil, the conventional networks are constituted by a 7. m height neutral conductor positioned above three phase conductors vertically and equally spaced. The multiplexed configuration presents the three phase conductors (isolated cables) curled around the neutral conductor. To investigate the performance of this two line configurations, the same conditions previously simulated in section 4.. for the conventional configuration were here applied for the multiplexed configuration. Fig. remarks the phase-to-neutral induced voltages at the consumer service entrance (Point B) for both network configurations, considering the Ω load and the absence of such load in Point B, as well. The obtained results show that the adoption of the multiplexed configuration is responsible to cause a relevant reduction on the induced voltage levels between neutral and phase conductors. It is observed a decrease around 7% (considering absolute values) on the phaseto-neutral voltage peak value when the presence of load connecting neutral and phase conductors in Point B is considered (,64 kv/ka for the conventional configuration, and,47 kv/ka for the multiplexed configuration). Such decrease is higher for disconnected loads (.5 kv/ka for the conventional configuration and.9 kv/ka for the multiplexed configuration). The distance between neutral and phase conductor justifies such behaviour. For the multiplexed configuration, this distance is extremely small, acting in order to maintain both the phase-to-ground and neutral-to-ground induced voltage waveforms very similar. Consequently, it promotes the reduction of the resultant phase-to-neutral induced voltage values, contributing to diminish the stress level of the load at the consumer service entrance Multiplexed Configuration Conventional Configuration Multiplexed Configuration Conventional Configuration (a) with Ω load between (b) without load between phase phase and neutral and neutral Fig. Phase-to-neutral induced voltages at the service entrance (Point B) assuming R g = 8 Ω. 5 CONCLUSIONS This work presented evaluations concerning lightninginduced voltages on low-voltage distribution networks using the Hybrid Electromagnetic Model (HEM). The authors of this work intend to give a contribution to this subject by performing such evaluations. The obtained results can provide elements that allow the development of protection practices to the electrical system and to the consumer against lightning-induced voltages. The investigation comprised sensitivity analyses relating line parameters with the resultant induced voltage along the electrical system and at the consumer service entrance. The most relevant conclusions are summarized below.

6 The increase of the number of grounding points caused a decrease of the induced voltage levels at phase conductor. Such behaviour is explained by the reduction of the distance to grounding points of the neutral conductor, responsible by negative voltage reflections. As a result, the subtractive effect associated with the electromagnetic coupling between neutral and phase conductors occurs in early times of the induced voltage wave. The reduction of the grounding resistance value was able to cause a decrease of the phase-to-ground and an increase of the phase-to-neutral induced voltages at the low-voltage network. Both effects became more intense with the diminution of the distance between adjacent grounding points. The results related with the induced voltage levels at the consumer service entrance remarked small phase-toneutral induced voltage values when it was considered the presence of resistive loads connecting neutral and phase conductors. On the other hand, the phase-to-ground induced voltage was always intense, constituting a serious risk to the consumer s safety and also requiring special attention to the protection of consumer s installations. The multiplexed configuration had a better performance regarding the stress level of the load (phase-to-neutral induced voltage). Such overvoltage was always lower than the one presented at the conventional configuration, independently if the load were connected or not to the electrical system. 6 REFERENCES [] C.A. Nucci, F. Rachidi, M. Ianoz, and C. Mazzetti, "Lightning-induced Voltages on Overhead Lines", IEEE Trans. on Electromagnetic Compatibility, vol.5, no., pp.75-86, Feb. 99. [] A. Piantini, J.M. Janiszewski, A. Borghetti, C.A. Nucci and M. Paolone, "A Scale Model for the Study of the LEMP Response of Complex Power Distribution Networks", IEEE Trans. On Power Delivery, Vol., No., Jan. 7 [] A. Borghetti, J. A. Gutierrez, C.A. Nucci, M. Paolone, E. Petrache and F. Rachidi, "Lightning-induced Voltages on Complex Distribution Systems: Models, Advanced Software Tools and Experimental Validation", Journal of Electrostatics, Vol. 6/-4, pp. 6-74, Elsevier, 4. [4] S. Visacro and A. Soares J., "HEM: A Model for Simulation of Lightning Related Engineering Problems", IEEE Trans. on Power Delivery, Vol., No., Apr. 5. [5] A. Soares J., M. A. O. Schroeder and S. Visacro. "Transient Voltages in Transmission Lines Caused by Direct Lightning Strikes", IEEE Trans. Power Delivery, Vol., No., pp , Apr. 5. [6] S. Visacro and F.H. Silveira, "Lightning Current Waves Measured at Short Instrumented Towers: The Influence of Sensor Position", Geophysical Research Letters, vol., L884, doi:.9/5gl55, 5. [7] S. Visacro and F.H. Silveira., "Evaluation of Current Distribution along the Lightning Discharge Channel by a Hybrid Electromagnetic Model", Journal of Electrostatics, Vol. 6/ 4, pp.-, 4. [8] F.H. Silveira, "Modelagem para Cálculo de Tensões Induzidas por Descargas Atmosféricas", Doctoral Thesis, Supervisor: S. Visacro, LRC/PPGEE Universidade Federal de Minas Gerais, Belo Horizonte, Brazil, Dec. 6. [9] K. Norton, "The Propagation of Radio Waves Over the Surface of the Earth and in the Upper Atmosphere, Part II The Propagation from Vertical, Horizontal, and Loop Antennas over a Plane Earth of Finite Conductivity", Proceedings of the Institute of Radio Engineers IRE, vol. 5, no. 9, Sep. 97. [] F.H. Silveira, J. Herrera, S. Visacro and H. Torres, "Comparison between NEC and HEM for the Calculation of Lightning-Induced Voltages", Proceedings of GROUND 6 and nd LPE - International Conference on Grounding and Earthing & nd International Conference on Lightning Physics and Effects, pp. 5 8, Maceió, Brazil, Nov. 6. [] A. Piantini and J.M. Janiszewski, "Lightning Induced Overvoltages on Low-Voltage Lines", Proceedings of V International Symposium on Lightning Protection (SIPDA), pp. 4-9, São Paulo, Brazil, May 999. [] A. De Conti, "Modelos para Definição de Ondas de Corrente e Tensão Representativas das Solicitações de Sistemas de Distribuição por Descargas Atmosféricas", Doctoral Thesis, Supervisor: S. Visacro, LRC/PPGEE Universidade Federal de Minas Gerais, Belo Horizonte, Brazil, Aug. 6. [] A. De Conti and S. Visacro, "Evaluation of Lightning Surges Transferred from Medium Voltage to Low Voltage Networks", IEE Proceedings on Generation, Transmission and Distribution, vol. 5, no., doi:.49/ipgtd:46, p. 5-56, May 5. [4] S. Visacro, M.A.O Schroeder, A. Soares J., L.C.L. Cherchiglia and V.J. Sousa, "Statistical Analysis of Lightning Current Parameters: Measurements at Morro do Cachimbo Station", J. Geophys. Res., vol. 9, No. D5, -, 4.