SAFETY ASPECTS AND NOVEL TECHNICAL SOLUTIONS FOR EARTH FAULT MANAGEMENT IN MV ELECTRICITY DISTRIBUTION NETWORKS

Transcription

1 SAFETY ASPECTS AND NOVEL TECHNICAL SOLUTIONS FOR EARTH FAULT MANAGEMENT IN MV ELECTRICITY DISTRIBUTION NETWORKS A. Nikander*, P. Järventausta* *Tampere University of Technology, Finland, Fax Keywords: High-impedance earth fault, fault indication, distribution network, safety regulations, hazard voltage. Abstract The paper considers the safety aspects of earth fault management and needs for indication and location of highresistance earth faults in medium voltage (MV) electricity distribution networks. Applying the methods available the safety of electricity distribution can be improved and the number and duration of outages to customers can be reduced. Methods for very high-resistance earth fault identification and location in neutral isolated or compensated distribution systems have been developed in Tampere University of Technology. These indication methods are able to detect and locate faults up to some hundreds of kilo-ohms. The indication methods have proven to be very appropriate for the implementation and the results from the field installation and field experiments are promising. Aspects for achieving a compromise between the safety and reliability of the electricity distribution are also highlighted. 1 Introduction The overall target of earth fault management is to indicate, locate and isolate a fault and to restore the healthy part of the network as quickly as possible in order to reduce the outage costs and to maintain the quality of supply and safety requirements. WHY? Safety aspects Protection of network Minimising of the costs due to interruptions HD 637 S1 Fundamental frequency overvoltages - risk of secondary faults Costs for electricity companies and customers Figure 1: Why earth fault management is needed in neutral isolated or compensated system. Unlike short-circuits, earth faults include the ground connection. This matter creates the unbalance in phase-toearth voltages and phase currents during a fault and causes the zero sequence components of the phase-to-earth voltages and phase currents. From the safety point of view the main quantity is the fault current flowing from phase to earth. The flowing path normally includes the fault resistance and earthing resistance. In neutral isolated system the fault circuit is closed via phase-to-earth capacitances. In high reactance earthed system a parallel resonant circuit is composed between earth capacitances and the neutral-to-earth resistor. The fault current produces the earthing voltage (U E ) when flowing through the earthing impedance (Z E ). Part of the earthing voltage appears as a touch or step voltage. These hazard voltages are regulated by pan-european hazard voltage regulations [1]. The touch or step voltage causes the current passing through the human body which is the cause of danger. Medium voltage networks in the Nordic Countries have mainly isolated or compensated neutrals, which leads to very low fault current with high-resistance earth fault. Thus the detection of such faults is a challenging task. In this study a high-resistance earth fault is defined as a fault with fault resistance from 10 kω up to several hundreds of kilo-ohms. Conventional feeder protection cannot even detect an earth fault with very high fault resistance and thus it cannot trip a feeder. Because the phase-to-earth fault current is very low the touch voltage regulations does not presuppose the tripping of the faulty feeder [1]. The major safety risk appears in the situation where the conductor or other live part of the system has an earth contact e.g. directly or via a tree and the fault current is not high enough to lead to tripping. These high-resistance faults tend to evolve gradually into a full-scale earth fault. Thus, early identification and location of such faults is of increasing importance in improving the safety and reliability of electricity distribution. Continuous and accurate monitoring of each separate MV feeder in order to evaluate the electric insulation state of the feeder and enable the early detection of the earth faults becomes more and more important. In the past few years Tampere University of Technology has being studying methods for identifying and locating earth faults [3]. The methods of the paper utilize the same measuring information of modern numerical feeder terminals which is used for normal protection function. All methods

2 operate independently as relay functions. Thus the electric state of all the feeders can be monitored separately. The identification and location of a high-resistance earth fault also include the information on the faulty feeder and faulty phase. This, in addition to better sensitivity, is one major improvement on conventional methods. With permanent low resistance earth faults the overall aim is to indicate the presence of an earth fault and the faulty feeder, to determine the distance to a fault, to isolate and remove a fault in order to reduce outage costs and to ensure as high safety level as possible. The sound part of the feeder is restored as quickly as possible. 2 Safety aspects of earth fault management Standard IECTR gives guidance on the effects of current flowing through the human body due to its magnitude and duration. In practice it is more convenient to refer to touch voltages. Touch voltage limits due to earth faults are given in Figure 2 [1]. The curve represents the value of voltage (U TP ) that can appear across the human body, bare hands to bare feet. No additional resistances have been considered in the calculations. current (R F = 0 Ω) and the worst earthing (highest earthing impedance Z E ). U E IEZE 2 UTP = (1) 3 Safety aspects of earth fault protection and location versus quality of supply Earth faults or secondary faults developing from earth faults form the most common fault group in the MV networks of the Nordic Countries. In overhead line networks automatic reclosing usually clears over 90 % of the faults and below 10 % of the faults remain permanent. Permissible touch voltages defined in [1] depend on the duration of fault. The typical tripping delay used in neutral isolated or compensated system is s. From the safety point of view the short tripping delay is advisable because of the short current-flow duration. Due to the short tripping delay savings in earthing costs can also be achieved. In neutral isolated and especially in neutral compensated system the self-extinction of the earth fault arc is much more likely if the tripping delay and thus the current flow duration is long enough. The lower steepness of the recovery voltage is the main reason why the possibility of arc extinction is much higher in a neutral compensated than in an isolated system [4]. Thus the safety and quality of supply are partly contradictory objectives. In Central Europe it is a long tradition to use neutral compensated MV systems during an earth fault. Operating the MV system during phase-to-earth fault includes a considerably higher risk in Finland than in Central Europe. This is a consequence of considerably higher soil resistivity and extensive usage of overhead line networks. In Central Europe MV systems are maily of underground cable construction. Figure 2: Permissible touch voltages U TP for limited currentflow duration [1]. Every earth fault will be disconnected automatically or manually. Thus touch voltages of very long or indefinite duration do not appear as a consequence of earth faults. The novel residual current compensator systems makes it possible to carry on the electricity distribution during an earth fault because the residual fault current is completely compensated [5]. With these applications the insulation state of the system must be assured with an eye to the risk of a cross-country fault. If the current flow duration is much longer than shown in Figure 2, a value of 75 V can be used for U TP. The highest permissible earthing voltage (U E ) is double compared to voltage U TP (Eq. 1) or quadruple if the touch voltage U TP is proved by measurements to stay below the values of the curve (Fig. 1). Then the longest tripping delay of the feeder relay is determined according to maximum fault Residual current compensation (RCC) devices makes it possible to compensate completely the fundamental frequency fault current [5]. This makes it possible to use a network during an earth fault because no touch or hazard voltage appears at the fault location. In neutral isolated or compensated system phase-to-phase voltages do not change as a consequence of an earth fault. The vector group of 20/0.4 kv transformers is Dyn11. Thus undisturbed power distribution to customers can continue during a fault. The sustained overvoltages due to the rising potential of the healthy phases increase the risk of secondary failure (crosscountry fault or short-circuit), however. With cross-country fault the fault current flowing via earth between fault locations can be of the order of magnitude of a phase-to-phase short-circuit current. This type of fault must be tripped out very quickly because of touch voltage consideration [1] and danger of damage to communication cables and telephone systems which can cause very high costs. By improving the safety and quality of the supply by compensating the residual earth fault current the risk of secondary failure exists,

3 however. The longer the duration of an earth fault is, the more likely is the occurrence of a secondary failure, especially a cross-country fault. Nowadays the fault distance of short-circuit faults can be calculated very accurately using the measured fault current. Permanent phase-to-earth faults are still problematic in neutral isolated and compensated networks. The longitudinal impedance of the line between the fault location and the substation has practically no influence on the magnitude of the fundamental frequency fault current. One possibility to arrange the earth fault distance calculation is to increase the fault current by connecting a low impedance to the neutral point of the system temporarily during a fault. When the safety aspect is considered the following question arises. Which is more important from the safety point of view: Low duration of high fault current and quick calculatory earth fault distance estimation, or Low fault current and several sequential current-flow periods (tripping delays and trial switchings used for locating the earth fault). Using trial switchings the fault location is determined by dividing the faulty feeder into sections by a disconnector and closing the substation circuit breaker against the fault until the faulty line section is found. This is laborious and time consuming. The other disadvantage of trial switchings is that the voltage is connected against the fault. This causes switching overvoltages and temporary overvoltages between healthy phases and the earth and may lead to a secondary failure e.g. to a cross-country fault. Every opening of the circuit breaker with the trial switchings causes an interruption to customers supplied by the feeder concerned. Trial switchings may also cause unsafe situations when the fault location is situated in an urban environment. In addition to fundamental frequency current the fault current normally includes harmonic components. In neutral compensated systems harmonics are generated by saturated components of the fault circuit such as neutral-to-earth coil and main transformer. In the case of temporary faults the earth fault arc generates harmonics to the fault current. Figure 3 presents the fault current measured with a real fault case in neutral isolated system. The fault is of the arc type. The fault current interrupts and restrikes causing high harmonic content of a fault current. Figure 3 presents wave shape of the fault current. The horizontal uniform lines represent the fundamental frequency component, 3 th, 5 th, 7 th and 9 th harmonics from top to bottom. The harmonic content of the fault current may be significant compared to the fundamental frequency component with earth fault arc. This fact should be noted when touch voltages are considered. In unearthed and neutral compensated systems the possible overvoltages due to successive current interruptions and resrikings should also be considered. These may happen in the case of low fault currents, typically below 10 A. These overvoltages may cause a secondary failure. Figure 3: Earth fault current and harmonic content with restriking fault. 4 Identification and location of very highresistance earth faults In overhead line networks high-resistance earth faults may occur due to trees leaning against a conductor or when the conductor falls to the ground with very high resistivity. Such faults also develop due to conductor break when the load side end of the conductor has an earth contact. Faults with covered conductors or other network component faults (e.g. metal oxide surge arresters, overhead line pin insulator) often have very high impedance. These faults tend to evolve gradually into a full scale earth fault. Thus, early identification and location of such faults is of increasing importance in improving the reliability of electricity distribution. Although the touch voltage regulations do not require the tripping of the faults with very high fault resistance such faults may cause unsafe situations. Falls and breaks in the conductor when the load side end of the conductor has an earth contact are especially dangerous if the live parts can be touched by the public. Faults with covered conductors may be latent for a long time before a tripping of the feeder protection. The typical resistance of an unseasoned tree is in the range kω [2]. These resistance values are valid in those seasons when the earth is not frozen. In wintertime much higher resistances, ranging up to several hundreds of kiloohms or some megohms can be found. Most faults of this type are out of neutral voltage overvoltage relays or monitoring of the neutral voltage. 4.1 A method for monitoring the shunt resistance of a feeder The method detects the magnitude of the fault resistance of the faulty phase of the faulty feeder. In the case of a healthy feeder the magnitude of the fault resistance is theoretically infinite.

4 The measuring accuracy does not allow the accurate determination of the phase-to-earth conductance of the feeder. Instead, according to experiences from field tests, the calculated phase-to-earth susceptance has a reliable value representing the electric length of the feeder. Thus more reliable values for leakage resistances and thereby more sensitive fault indication compared to earlier known methods can be achieved if the phase-to-earth admittances are assumed to be pure capacitances. The following presents a method for the monitoring of the shunt resistance of the MV feeder. Every feeder is monitored separately. Only neutral voltage and sum currents due to an earth fault are considered. The method can be applied with isolated and compensated systems. 3 U 3 ' U 2 ' U 2F U 3F U 3 U 0M U 2 U 0 U 0 ' U 1 ' U 1 U 1F 1 The most sensitive fault indicator is the change of the neutral voltage of the network. The indication process starts when the absolute value of the change of the neutral voltage U 0 exceeds the specified threshold value. If the change of the neutral voltage is due to an earth fault, U 0 refers to the neutral-to-earth voltage caused by an earth fault (Fig. 4). When the specified threshold value U th is exceeded, the determination process of the faulty feeder and calculation of the fault resistance start in every feeder. The phasor U 0 is used as a reference. I 0i represents the sum current of the feeder due to an earth fault. The influence of the capacitive unbalance on the sum current is eliminated. Phase-to-earth voltages U ν are also referenced to U 0. The phase angle of U ν is equal to the phase angle of the earth fault current if the fault impedance is supposed to be purely resistive. ' U = U 0M U 0 > U th 0 (2) where U 0 ' is measured neutral voltage before a change U 0M is measured neutral voltage after a change U th is specified threshold value of the neutral voltage. When Formula 2 has been fulfilled, the faulty phase is determined. The inference of the faulty phase occurs in a slightly different way in isolated and compensated systems. 2 Figure 4: Voltage phasors of an isolated system when Phase 1 is faulty. Calculation of fault resistance. When information on the faulty phase is available, the fault resistance of Feeder i can be calculated using Formula 3. Neutral voltage, phase-toearth voltages and sum currents of feeders are known as measured values at all 110/20 kv substations. UνF sinϕu F R Fi = ν I 0i sin (3) ϕ I + B i U 0i 0 0 where U 0 is change of neutral voltage U νf is phase-to-earth voltage of a faulty phase ϕu νf is phase angle of phasor U νf referenced to U 0 I 0i is change of the sum current of Feeder i ϕ I 0i is phase angle of phasor I 0i referenced to U 0 B 0i is phase-to-earth susceptance of Feeder i. The fault resistance can also be calculated also during an earth fault which is presented in the following Figure 5. Indication of a faulty phase in isolated system. Figure 4 illustrates balanced phase-to-earth voltages (U 1, U 2, U 3 ), unbalanced phase-to-earth voltages of healthy state (U 1 ', U 2 ', U 3 '), unbalanced phase-to-earth voltages during an earth fault (U 1F, U 2F, U 3F ), neutral voltage due to a healthy state (U 0 '), neutral voltage caused by an earth fault ( U 0 ) and neutral voltage measured during an earth fault (U 0M ). In isolated system neutral voltage ( U 0 ) and the phase-to-earth voltage of the faulty phase (U 1F ) draw a semicircle as a function of the fault resistance. This means that with an isolated system phasor U νf leads 90 phasor U 0 during an earth fault irrespective of the fault resistance. Phase-to-earth voltages of healthy state (U 1 ', U 2 ', U 3 ') are unbalanced and result from capacitive unbalance of the system. Figure 5: Sum current and fault resistance of faulty feeder.

5 5 Conclusions Methods for very high-resistance earth fault identification and location in neutral isolated or compensated distribution systems have been developed in Tampere University of Technology. Applying these patented methods the safety of electricity distribution can be improved and the number and duration of outages to customers can be reduced. By applying earth fault current compensation and residual current compensation the touch voltages can be considerably reduced. These applications do not eliminate the risk of secondary failure (cross-country faults) as a consequence of sustained overvoltages. In neutral compensated system the risk may be even higher compared to neutral isolated system because the fault must not be tripped as quickly as with higher fault current. In many cases the safety and quality of supply are partly contradictory objectives with earth faults. The harmonic content of the fault current may be significant compared to the fundamental frequency component with earth fault arc. This fact should be noted when touch voltages are considered. References [1] Anon. Harmonization document HD 637 S1: Power installations exceeding 1 kv a.c. European Committee for Electrotechnical Standardization (CENELEC). 107 p. (1999) [2] S. Hänninen, M. Lehtonen, Method for detection and location of very high resistive earth faults. European Transactions on Electrical Power Engineering (ETEP), Vol. 9, No. 5, pp , (1999) [3] A. Nikander, P. Järventausta, J. Myllymäki. Novel algorithms for earth fault indication based on monitoring of shunt resistance of MV feeder as a part of relay protection. Proceedings of the Seventh International Conference on Developments in Power System Protection, IEE Conference Publication, No. 479, Amsterdam, The Netherlands, pp , (2001). [4] A. Nikander, E. Lakervi. A philosophy and algorithms for extinguishing earth fault arcs in suppressed medium voltage networks. Proceedings of the 14th International Conference on Electricity Distribution (CIRED), Birmingham, UK, June 1997, pp , (1997) [5] K. Winter. Swedish distribution networks A new method for earth fault protection in cable and overhead systems, Proceedings of the Fifth International Conference on Developments In Power System Protection, IEE Conference Publication, No.368, UK, 3 p., (1993).