OVERVOLTAGE PROTECTION. Dimensioning, testing and application of metal oxide surge arresters in low-voltage power distribution systems

Transcription

1 PPLICATION GUIDELINES OVERVOLTAGE PROTECTION Dimensioning, testing and application of metal oxide surge arresters in low-voltage power distribution systems

2 Foreword Up until 1998 no international standards existed for surge arresters in lowvoltage power systems. This situation presented two difficulties: firstly it lead to specifications which were adapted from other standards, for example, IEC99-1 and IEC-99-4, which are applied for high voltage surge arresters, with and without spark gaps; secondly, declared rating, parameters and tests performed by different manufacturers were not clear, and therefore not really comparable. In the past, different committees of IEC worked (and are still working) on standards and guidelines, as in IEC SC 28A: Insulation co-ordination of low-voltage installations; SC 37A: Surge protective devices (SPDs) in low-voltage power distribution systems; TC 64: Electrical installations of buildings; SC 77B: Electromagnetic compatibility high frequency phenomena; and TC 81: Lightning protection. This did not make a clear and easy situation. Joint Working Group (JWG) 31 of TC 64 has taken the task to co-ordinate the work of the different technical committees and sub-committees under the title: Surge overvoltages and surge protection. In 1998 the standard IEC (First edition ), was released with the title: Surge protective devices connected to low-voltage power distribution systems- Part 1: Performance requirements and testing methods Mr. Bernhard Richter, Product Manager of the surge arrester division of ABB High Voltage Technologies Ltd, gladly took on the task to describe in a short and clear form the technical bases and application of surge protective devices for lowvoltage power systems, concentrating on Metal-Oxide surge arresters (MO-arresters) without gaps for outdoor and special applications. Mr. Richter is an active member in different working groups of IEC SC 37A and TC 81. His activity field includes mainly the development, testing and application of surge arresters for use in all voltage systems of power supply. We hope, that you as a reader, will find this booklet useful. We welcome amendments, suggestions and qualified hints, which may help us to cover all the demands of our customers. ABB High Voltage Technologies Ltd Wettingen, April 2001 First published: May 2001 All rights reserved. No parts of this booklet may be reproduced or translated in any manner without the express written consent of ABB High Voltage Technologies Ltd. ABB High Voltage Technologies Ltd Division Surge Arresters Wettingen / Switzerland 1

3 Contents 1 Introduction 2 Overvoltages in low-voltage supply networks 2.1 Overvoltages due to direct flashes 2.2 Induced overvoltages 2.3 Overvoltages due to coupling 2.4 Transferred overvoltages through transformers 2. Probability of overvoltages 3 Low-voltage networks 3.1 System voltages in low-voltage networks 3.2 Insulation categories 3.3 Low-voltage earthing systems 3.4 Temporary overvoltages (TOV) in low-voltage systems 4 Surge protective devices (SPDs) 4.1 Principle function of surge arresters 4.2 Definitions 4.3 Classifications 4.4 Service conditions Low-voltage MO-surge arresters from ABB.1 MO-resistors.2 MO-surge arresters.3 Technical data of the arresters 6 Tests 6.1 Type tests 6.2 Special tests 6.3 Routine tests 6.4 Acceptance tests 7 Selection of MO-surge arresters 7.1 Selection of Uc 7.2 Selection of Up 7.3 Selection of the energy capability 8 Coordination of surge arresters 9 MO-surge arresters for d. c. systems Installation of surge arresters Bibliography 2

4 1 Introduction Overvoltages in electrical supply networks result from effects of lightning strokes and switching actions, and cannot be avoided. They endanger the electrical equipment and due to economical reasons, the insulation cannot be designed for all possible cases. Therefore, a more economical and safer on-line network calls for extensive protection of the electrical equipment against unacceptable overvoltages. This applies to high voltage as well as to medium and low voltage networks. u i/2 i u i/2 Overvoltage protection can be basically achieved in two ways: Avoiding lightning overvoltages at the point of origin, for instance through shielding earth wires. Limit overvoltages near the electric equipment, for instance through surge arresters in the vicinity of the electrical equipment. overhead line i : lightning current U: generated overvoltage Z: 0 surge impedance of the line earth Z 0 In low voltage systems the earth wire protection is generally not very effective. A lightning would hit not only one wire (the earth wire), but all, including the phase wires, and induced and transferred overvoltages could not be avoided. The most effective protection against overvoltages in low voltage networks is therefore the use of surge arresters in the vicinity of the equipment. For general information, and especially with regard to medium voltage networks, we refer to our APPLICATION GUIDELINES: Dimensioning, testing and application of metal oxide surge arresters in medium voltage networks [1]. Overvoltage protection in railway facilities, a. c. and d. c., is described in: Dimensioning, testing and application of metal oxide surge arresters in railway facilities [2]. Figure 1 Lightning overvoltage caused by a direct lightning flash to an overhead line. U = Z 0 x i/2 Assuming Z 0 = 40Ω and a typicall current of i=20ka (80% probability, see Table 1), the prospective voltage will reach U = 400 kv. On low voltage lines, therefore, flashovers will occur between all the line conductors, and usually also a flashover to earth at the closest pole of the line. After flashover the effective impedance is reduced, depending on the earth resistance involved. Even with a low impedance of Ω, and the current being at ka, the voltage will still be U = 0 kv, travelling along the line. Therefore further flashovers can occur along the line. Lightning overvoltages are the greatest threat to the low voltage networks. Overvoltage protection must be arranged in such a way that the overvoltage is limited to non-damaging values. Percentage Negative downward 98% 9% 80% 0 % 20 % % 2 Overvoltages in low-voltage supply networks Current peak value > 4 ka > 6 ka > 20 ka > 34 ka > ka > 90 ka Lightning surge overvoltages in electrical systems may be classified according their origin as follows [3]: overvoltages due to direct flashes to overhead lines induced overvoltages on overhead lines due to flashes at some distance overvoltages caused by resistive, inductive and capacitive coupling from systems carrying lightning currents. In [4] is discussed in detail the case of transferred overvoltages through a distribution transformer from the medium voltage to the low voltage side. 2.1 Overvoltages due to direct flashes The overvoltage is determined by the effective impedance of the line and the lightning current. For a flash to an overhead line conductor, the impedance is in the first moments determined by the characteristic impedance (surge impedance) Z 0 of the line. The impedance Z 0 is normally in the range of 400 to 00Ω for one conductor. As shown in Figure 1 the lightning current is diverted in two, each part travelling along the line. The generated voltage is calculated Table 1 Probability of lightning peak values. 2.2 Induced overvoltages Due to the electromagnetic field changes caused by a lightning flash, overvoltages are induced in overhead lines of all kinds. As a rough approximation, the prospective overvoltage between the line conductors and earth can be estimated according to Rusck [] U max = Z 0 x I max x H / D I max is the peak value of the lightning current Z 0 is the effective impedance (assumed to be 30 Ω) H is the height of the line D is the distance of the flash location from the line Considering a height of m for low voltage overhead lines, a lightning current of 20 ka, and a distance of 0 m, the induced voltage is calculated U max = 30 kv. 3

5 With a distance of 00 m between the line and the flash location, the induced voltage has a value of U max = 3 kv. The above calculated values of induced voltages in low voltage overhead lines show that this kind of surge is of primary concern for low voltage distribution systems. Lightning induced overvoltages occur mainly between the conductors and earth. The voltage difference between the conductors is initially small, especially when twisted conductors are used. However, due to different loads on phase conductors (depending on low voltage system), interactions of surge protective devices, flashovers, etc., considerable line-to-line stresses can also occur. An example illustrating induced overvoltages line-to-line in low voltage systems is shown in Figure 2. Twisted conductors, including neutral, are assumed. The neutral is earthed on both ends of the line. The voltages at a certain point of the line show a high frequency dumped oscillation (ringing wave). The period of the oscillation corresponds to twice the travel time of a span, a span being the distance between two poles. Furthermore, it is found that the highest voltage occurs in the middle of the span. In the given example the voltage reaches up to 23 kv in the middle of the span, and up to kv directly at the pole, where consumers may be connected. 2 U(kV) U2 U1 U m 32 ka 10 m 6 7 t(µs) 8 Figure 2 Induced overvoltage line-to-line. Calculated values, assuming twisted conductors. 0. i U 0% earthing impedance Figure 3 Example of resistive coupled overvoltages in electrical systems. In the electrical installations in the building, as well as in close installations (and all conductuing parts) in the earth high overvoltages can be generated. Due to the high electromagnetic fields caused by the lightning current, inductive and capacitive coupling to electrical systems close to a lightning path can also cause overvoltages of concern, causing failures or malfunctions. 2.4 Transferred overvoltages through transformers i Overvoltages generated in the medium voltage (MV) system are transferred to the low voltage (LV) system in two ways, by capacitive and magnetic coupling through the MV / LV transformer by earth coupling (see Figure 4). The magnitude of the transferred overvoltage depends on many parameters and some important differences can exist between different countries, due to differences in the transformer design and the LV earthing systems (T T, T N, IT). Medium voltage line (MV) PE L1 L2 L3 N Lightning protection system (LPS) i 0. i U LV cable Low voltage line (LV) 2.3 Overvoltages due to coupling A2 B2 A lightning flash to earth can result in an earth potential of high value at the point of the strike, as well as in the vicinity. This phenomenon will cause overvoltages in electrical systems, using this point of earth as reference for their earthing system. Figure 3 shows the principle of this phenomenon. The potential rise of the earthing system is determined by the lightning current and the effective earthing impedance. In the first moment the earth electrode potential is determined by the local impedance, for instance Ω. This means that a high voltage is generated between the earthing system and electrical installations inside the building, or other installations close to the earthing system. With a high probability this overvoltages will cause either flashovers, insulation breakdown or operation of surge protective devices. Following such events, current impulses can flow into the various systems, mainly determined by their impedance to earth. In this way overvoltages are produced in the power supply system as well as in other services (telecommunication, data and signalling systems, etc.). Furthermore, overvoltages are transferred to other buildings, structures and installations. A1 Transformer C L1 L2 L3 A1 by direct lightning to the MV line A2 by indirect lightning to the MV line (induced voltage) B1 by direct lightning to the LV line B2 by indirect lightning in the LV line (induced voltage) C by capacitive coupling through the transformer Figure 4 Overvoltages in the Low voltage system B1 L1 L2 L3 4

6 U (kv) U (kv) b) 0 0 a) without representation of the users installation (no load assumed) b) user installations represented by lumped capacitances 1 1 a) 20 t (µs) 20 t (µs) Figure a / b Typical wave shape of overvoltage transferred to the LV line (calculated). The high frequency components of the overvoltage are transferred capacitively from the MV to the LV side of the transformer [4]. Figure a shows a typical wave shape of the overvoltage transferred to the LV line. Being the transferred overvoltage characterized by high frequency oscillations, the natural capacitance of the load can reduce very effectively the peak overvoltages, as shown in Figure b. The calculated voltages in the given example reach peak values of kv (without load, Figure a), and 3 kv (with load, Figure b). In case of direct lightning to the MV line, the surge arrester operation or an insulator flashover diverts the surge current through the earthing system, and can produce a resistive earth coupling between the MV and LV system. An overvoltage is transferred to the LV system as shown in the typical case of Figure 6a. Depending on the earthing impedance, this earth coupling overvoltage can be much higher than the capacitive coupling through the transformer. Separating the earthing electrodes, as in Figure 6b, avoids this problem. Practically it is not possible to have really separated earth systems, due to the short distance and the conductivity of the earth. 2. Probability of overvoltages The frequency of lightning flashes to an overhead line, or in the vicinity of the line, depends on the local flash density, line type (especially the height) and possible shielding effects of the surroundings [3], [4], []. For lines in an open area the number of flashes can be calculated as follows N = A x N g x -6 A = 6 x H x L MV Arrester a) MV and LV side of the transformer have same earthing point. This generates, in case of arrester operation, an overvoltage Ug on the LV system ( U g = R i + L di/dt) (no load assumed) i Arrester Earthing i (R, L) Transformer Earthing MV Arrester Soil Transformer Earthing Soil Installation Earthing Installation Earthing b) Separate earthing for MV and LV side of the transformer. U = U + R i + L di/dt 0 Equipment U = U 0 Equipment A = effective area for direct lightning to the line (in m 2 ) H = height of the line ( in m) L = length of line (in m) N g = local flash density per km 2 and year For a line of m height and assuming N g = 1, N is found to be 0,03 per km of line and year, that means three direct flashes per 0 km of line length and year. This gives a rough estimate of number of direct flashes to low voltage overhead lines. The number of induced and transferred overvoltages is certainly much higher than the overvoltages due to direct flashes in the line. Especially the local lightning density and the different possibilities of generating overvoltages, including switching, has great influence on the occuring number of dangerous overvoltages. In Figure 7 a typical low voltage system with overhead line is given. Calculated figures are presented for induced overvoltages which may be expected in this network, [3]. The ground flash density was assumed to be 2,2 flashes per km 2 per year, all loads were modelled by frequency independent resistors. Table 2 shows the calculated results. The last column (> 20 kv) shows high levels of overvoltages, but these occur only in case of direct lightning to the low voltage line. The probability of occurrence of such surges in this example is once in 22 years. But the overvoltages in the range of 1, kv to 6 kv can occur several times a year in a low voltage network, depending on the type of installation. Figure 6a / 6b Overvoltage on the low voltage side due to earth coupling.

7 20 kv line MV/LV station 230/400 V line twisted cable (3 phases + neutral) Line connection Voltage line-to-neutral derived from nominal voltages a.c or d.c. up to and including Three-phase four-wire systems with earthed neutral E Three-phase three-wire systems unearthed Single-phase two-wire systems a.c. or d.c. Single-phase three-wire systems a.c. or d.c. MV arresters 240 m 240 m 30 m 2 m Consumer s earthing V 0 V V V V 12, / Conductive parts earthing 30 ohms Neutral earthing 30 ohms Neutral earthing 30 ohms Installation earthing 0 ohms /208* 127/220 12, , Figure 7 Typical low voltage network with overhead line. Arrangement used for calculating the values in Table /380, 230/ /41, 260/ / /600, 380/ /690, 417/ / , 230, , 277, , 400, , , 77, > 1, kv > 2, kv > 4 kv > 6 kv > 20 kv , , Unloaded TT system ,04 * Practice in the United States of America and in Canada. Loaded TT system Loaded TN system ,6 1 0,3 0, 0,2 0,04 0,04 Table 3 Nominal voltages presently used world wide. Table 2 Line-to-earth prospective overvoltage levels in the LV installation, occurrences per year. Note 1: The numbers shown in the table were obtained for an overhead twisted cable distribution system. For a distribution system with overhead open conductors in air, the voltage levels can be expected to be twice as high for the same probabilities. Note 2: In this example, when performing a variation of the model to represent a TN system, it was found that the value of the earthing impedance had no significant influence because the LV neutral is directly connected to earth. 3 Low-voltage networks Around the world very different low voltage networks exist. They differ in the system voltage, the number of wires, the handling of the neutral and the protective measures. The nominal voltages of the supply systems are basically given in publication IEC ( ) and amendments: IEC standard voltages. In IEC [6] is given a good overview of the nominal voltages presently used in the world, depending on the type of network, see Table System voltages in low-voltage networks 3.2 Insulation categories The concept of overvoltage categories is used for equipment energized directly from the low voltage mains. For the different categories the insulation levels are specified. According to [6] the definitions of the categories are as follows: Equipment of overvoltage category IV is for use at the origin of the installation (e. g. overhead lines, cables, bus bars, meters, primary overcurrent protection equipment). Equipment of overvoltage category III is equipment in fixed installations and for cases where he reliability of the equipment is subject to special requirements (e. g. mainly fixed indoor installation). Equipment of overvoltage category II is energy-consuming equipment to be supplied from the fixed installation (e. g. appliances, portable tools and other household and similar loads). Equipment of overvoltage category I is equipment for connection to circuits in which measures are taken to limit transient overvoltages to an appropriate low level (e. g. protected electronic circuits). As seen in Table 3, there is world wide a variety of existing voltages. The standard voltages in Europe, for instance, are given in [7]. The system voltages, according to the harmonization document, are 230 / 400 V, where 230 V is the line to neutral voltage, and 400 V is the line to line voltage. Other existing common voltages in Europe are 240 / 41 V and 220 / 380 V. Considering an allowed tolerance of %, the highest voltages to be expected for the 400 V system are U 0max = 23 V (line to neutral voltage) and U Nmax = 440 V (line to line voltage). 6

8 Table 4 gives the four insulation categories. The rated impulse voltage gives the insulation withstand capability for the different categories, depending on the line to neutral voltage of the systems derived from the nominal voltages a. c. or d. c., based on IEC Voltage line-to-neutral derived from nominal voltages a.c or d.c. up to and including Rated impulse voltage for equipment V Insulation category LV line Electrical Power Source Combined PEN Conductor L Customer Connection Point N & E Electrical Equipment TN-C System V I II III IV Figure 8b Table 4 Insulation categories for low voltage systems. 3.3 Low voltage earthing systems TN - C-S system (Figure 8c) The supply neutral is earthed at the source and points in the network. Supply lines have a combined neutral and earth wire. Supply within the customer premises would have separate neutral and earth wire, connected only at the service position. A protective neutral bonding (PNB) arrangement may be used to provide an earth terminal connected to the supply neutral. With this arrangement, the neutral will be connected to earth at the source point only, at or near to the customers supply point. The arrangement is generally restricted to a single customer with it`s own transformer. See Figure 8d. There are a number of methods used to provide an earth connection or system. The different arrangements and standard definitions are given below. Each is defined by a coding which contains the following letters: LV line Electrical Equipment T : terre, direct connection to earth N : neutral C : combined S : separate The different principle earthing arrangements are shown in Figure 8. For simplification single line diagrams are used. Electrical Power Source Combined PEN Conductor L Customer Connection Point N E TN-C-S System TN - S system (Figure 8a) The incoming supply has a point of connection between the supply neutral and earth only at the supply transformer. The lines have separate neutral and earth protective conductors. Figure 8c LV line Electrical Power Source L Customer Connection Point N E Electrical Equipment TN-S System LV line Electrical Power Source Combined PEN Conductor L Customer Connection Point N E Electrical Equipment TN-C-S System (PNB) Figure 8a Figure 8d TN - C system (Figure 8b) The neutral and earth wire are combined within the premises, and are earthed at the supply transformer or close to it. TT system (Figure 8e) The transformer is connected directly to earth, the customers installation is earthed via a separate electrode. This will be independent of any supply point electrode. 7

9 Figure 8e LV line Electrical Power Source L Customer Connection Point N E Electrical Equipment TT System Alternative Location for Earth Terminal IT system (Figure 8f) This arrangement has no direct system connection between live parts and earth, but the exposed conductive parts of the customers installation and its equipment is earthed. SPDs connected to: Minimum UT for s: TN-systems Connected L- (PE)N or L-N Connected N-PE Connected L-L TT-systems Connected L-PE Connected L-N Connected N-PE Connected L-L IT-systems Connected L-PE Connected L-N Connected N-PE Connected L-L TN, TT and IT-systems Connected L-PE Connected L-(PE) N Connected N-PE Connected L-L Table TOV values in low voltage systems. 1,4 U0 3 U0 1,4 U0 1,4 U0 3 U0 1,4 U0 TOV values for 0,2s: 1200V + U V 1200V + U V 1200V + U V Figure 8f LV line Electrical Power Source L Customer Connection Point N E 3.4 Temporary overvoltages (TOV) in low-voltage systems In case of a failure on the medium voltage side of the MV / LV transformer, due to an internal fault of the transformer or a sparkover of a gap or insulator, a current flows through the earthing impedance of the transformer. Depending on the connection between this earth impedance and the low voltage network a temporary overvoltage with power frequency can stress the low voltage network for a given period of time, equal to the clearing time of the fault in the medium voltage network. This can be between some µs up to some hours. For a detailed discussion of temporary overvoltage conditions see IEC [8]. Depending on the earthing system of the low voltage network different TOV can occur. Table gives an overview about the considered systems and the possible TOV between the different lines. Two values are given, the minimum TOV value for sec, and the TOV values for 0,2 sec. Corresponding test procedures are described in the amendment of IEC [9]. The test procedure depends on the intended application of an SPD in a low-voltage power installation system according to the installation instructions given by the manufacturer. 4 Surge protective devices (SPDs) Electrical Equipment IT System Alternative Location for Earth Terminal SPDs are devices for surge protection against direct and indirect effects of lightning or other transient overvoltages. They contain at least one nonlinear component that is intended to limit surge voltages and divert surge currents. The discussed SPDs are typically for use in low-voltage power systems, providing protection from the low-voltage bushing of the MV / LV transformer up to the plugs in buildings. In the course of this guidelines we will talk mainly about metal oxide surge arresters (MO-arresters) without gaps for outdoor and indoor application. 4.1 Principle function of surge arresters There are two different designs for surge arresters: a voltage limiting type, and a voltage switching type. The voltage limiting type is a nonlinear resistor, generally a metal oxide resistor, without any spark gap in series. This types are sometimes called MOV, which is an abbreviation of metal oxide varistor. The voltage switching type is a spark gap, or a spark gap with a nonlinear resistor (MO or SiC) in series or parallel. Figure 9 shows the principle difference in the function of the two types. v v Gapped arrester t time scale µs/div v v t MO-arrester time scale 2 µs/div Figure 9 Difference in function of gapped arresters (left), and MO-surge arresters without gaps (right). Both types were tested with switching voltage impulses of the wave shape 20/ 200 µs. The voltage scale is the same in both cases. It is to be seen that in case of the MO-arrester the residual voltage is only half of the one given by the gapped arrester (same U c for both types of arresters). 8

10 Surge arresters which contain only spark gaps, or spark gaps with nonlinear resistors in series, have the disadvantage that the voltage collapses suddenly when the sparkover-voltage of the device is reached. This very high du/dt may cause EMC problems in data-lines which are close to the power lines, or lead to failures in inductive loads. Furthermore, the spark-over voltage depends on the steepness of the overvoltage. Because the spark gap fires only at very high voltage levels, it can happen that overvoltages bypass the surge arrester, and downstream connected instruments or installations are over-stressed. Surge arresters containing only MO-resistors have no sparkover-voltage. The turn on time is in the range of 1 ns, and the voltage is limited according to the extremely nonlinear voltage-current characteristic of the MOmaterial. A bypassing of these arresters is not possible. The advantages of MO-surge arresters are mainly the constant low protection level independent on the steepness and polarity of the incoming surge, the very good ageing behaviour, and the high energy capability. Possibilities of coordination of parallel MO-surge arresters are described in chapter Definitions In the new standard family of IEC the special requirements for surge arresters for application in low-voltage power systems are considered. In the following, the most important definitions are given with reference to [9], concentrating on MO-surge arresters without gaps. For the purpose of this guidelines some definitions with reference to [11] are added. The surge arresters addressed in this guidelines are to be connected to 0 / 60 Hz a. c. and d. c. power circuits, and equipment rated up to 00 V a. c. (rms) or 100 V d. c. Surge Protective Device (SPD) A device that is intended to limit transient overvoltages and divert surge currents. It contains at least one nonlinear component. Note: as mentioned above, in the course of this guidelines this is the same as a surge arrester, or short arrester. Nominal discharge current In The crest value of the current through the arrester having a current wave shape of 8/20µs. This is used for the classification of the arrester for classii test and also for preconditioning of the arrester for class I and II tests. Impulse current Iimp It is defined by a current peak value Ipeak and the charge Q, tested according to the test sequence of the operating duty test. This is used for the classification of the arrester for class I test. A typical waveshape that can achieve the parameters is that of a unipolar impulse current with a waveshape of / 30 µs. An other waveshape or impulse combination is acceptable, as long as they obtain the peak value Ipeak within 0µs and the charge Q within ms. Maximum discharge current Imax for class II test Crest value of a current through the arrester having a 8/20µs wave shape and magnitude according to the test sequence of the class II operating duty test. Imax is greater than In and declared by the manufacturer. It is used in the operating duty test to prove the correct function and thermal stability of the arrester. Maximum continuous operating voltage Uc The maximum a. c. (rms) or d. c. voltage which may be continuously applied to the arresters terminals. This is equal to the rated voltage. Nominal a. c. voltage of the system U0 U0 is the nominal line to neutral voltage of the a. c. system (rms value). Continuous operating current Ic The current flowing through the arrester when energized at the maximum continuous operating voltage Uc. Follow current If Current supplied by the electrical power system and flowing through the arrester after a discharge current impulse. Note: the follow current is significantly different depending on the design of the arrester. For MO-surge arresters without gaps the follow current is generally in the range of some ma in maximum. Reference current of an arrester Iref The reference current is the peak value of the resistive component of a power frequency current used to determine the reference voltage of the arrester. The reference current should be high enough to have a clear dominating resistive component, so that capacitive influences can be neglected. The reference current is specified by the manufacturer, and generally in the range of 1 ma to ma, depending on the cross section of the MO-resistor used in the arrester. Reference voltage of an arrester Uref The reference voltage of an arrester is the peak value of the power frequency voltage divided by 2 which has to be applied to the arrester to obtain the reference current Iref. The reference voltage at a given reference current is used to determine a point on the u-i-characteristic of an arrester in the low current range. Voltage protection level Up A parameter that characterizes the performance of the arrester in limiting the voltage across its terminals, which is selected from a list of preferred values. This is generally the guaranteed value given by the manufacturer. Residual voltage Ures The peak value of voltage that appears between the terminals of the arrester due to the passage of discharge current. Protection ratio Up /Uc The protection ratio gives the relation between the voltage protection level Up at In and the maximum continuous operating voltage Uc. Up is given as a peak value and Uc is given as a rms value. The lower the ratio Up /Uc, the better the protection given by the arrester. Temporary overvoltage The maximum a. c. (rms) or d. c. overvoltage that exceeds the maximum continuous operating voltage of the network for a specified time duration. Note: It has to be made a clear distinction between the temporary overvoltage UTOV occuring in the network at a given location, and the temporary overvoltage UT an arrester can withstand. The power frequency voltage versus time characteristics of an arrester (TOV-characteristic), provided on request by the manufacturer, is in low-voltage systems normally used only in case of special applications of the arrester. Combination wave The combination wave is delivered by a generator that applies a 1,2/0µs voltage impulse across an open circuit and an 8/20 µs current impulse into a short circuit. The voltage, current amplitude and waveforms that are delivered to the arrester depend on the impedance of the arrester to which the surge is applied. 9

11 Thermal runaway An operational condition when the sustained power dissipation of an arrester exceeds the thermal dissipation capability of the design, leading to an increase in the temperature of the internal elements culminating in failure. Thermal stability An arrester is thermally stable if after an energy input causing a temperature rise the temperature of the arrester decreases with time under applied continuous operating voltage. Degradation The change of original performance parameters as a result of exposure of the arrester to surges, service or unfavourable environment. Disconnector A device for disconnecting an arrester from the system in the event of arrester failure. It is to prevent a persistent fault on the system and to give visible indication of the arrester failure. Type tests Tests which are made upon the completion of the development of a new arrester design. They are used to establish representative performance and to demonstrate compliance with the relevant standard. Once made, these tests need not to be repeated unless the design is changed so as to modify its performance. In such a case, only the relevant tests need to be repeated. Routine tests Tests made on each arrester or parts of it to ensure that the product meets the design specifications. Acceptance tests Tests which are made when it has been agreed between the manufacturer and the purchaser that the arrester or representative samples of an order are to be tested. 4.3 Classification In [9] the surge protective devices (or short arresters) are classified according the number of ports (one or two) A one port device has two terminals, a two port device has four terminals. The two port device may contain internal decoupling elements. the design topology (switching type, limiting type, or combination type) the test method (class I, class II, or class III test method) the location (outdoor or indoor) the accessibility (accessible or out-of-reach) the mounting method (fixed or portable) the disconnector (with or without) the backup overcurrent protection (specified or not specified) the temperature range (normal or extended) As long as the arresters are installed at different locations in a system or installation the stresses to be expected are very different. Therefore, the arresters are classified with respect to the expected stresses, and consequently the test methods, in three classes. See Table 6. The class I test is intended to simulate partial conducted lightning current impulses. Arresters subjected to class I test methods are generally recommended for locations at points of high exposure, e.g. line entrances to buildings protected by lightning protection systems (LPS). These devices are called lightning current arresters. In addition to nominal discharge current In, information is required for the impulse current Iimp. Class I Class II Class III lightning current arresters lightning protection in connection with lightning protection structures I imp (/30 µs) 1 ka 20 ka Arresters tested according to class II test methods are generally subjected to impulses of shorter duration than class I arresters. The typical application is the overvoltage protection of low-voltage overhead lines and cables, as well as the protection of indoor installations. The expected stresses are originated by direct or indirect lightning to overhead lines or cable junctions. Required information is the nominal discharge current In and the maximum discharge current Imax. Arresters tested according, to class III test methods are subjected to impulses of lesser energy content than class I and class II arresters. They are recommended for locations with less exposure, mainly indoor. The information required is the open-circuit-voltage Uoc of the combination wave generator. 4.4 Service conditions The normal service conditions are the applied continuous voltage between the terminals of the arrester should not exceed the maximum continuous operating voltage Uc frequency between 48 Hz and 62 Hz a. c., or d. c. voltage altitude up to 2000 m operating and storing temperatures normal range: - C to + 40 C extended range: - 40 C to + 70 C relative humidity up to 90 % for indoor temperature conditions Exposure of the arrester to abnormal service conditions may require special considerations in the design or application of the arrester, and should be called to the attention of the manufacturer. Abnormal conditions may be extreme ambient temperatures (minus or plus), mechanical stresses, shock and vibration, etc. For outdoor arresters exposed to solar radiation, air pollution, bad weather conditions, additional requirements may be necessary. Low-voltage MO-surge arresters from ABB A MO-surge arrester is made of two parts: the active part, which consists of a MO-resistor, and an insulating housing including the terminals..1 MO-resistors surge arrester overvoltage protection energy supply I max(8/20 µs) > 0,0 ka 0 ka surge arrester overvoltage protection down stream U oc (2 Ω) Table 6. Classification of low-voltage surge arresters. The given values are typical ratings. The voltage-current (u-i) characteristic of a metal oxide resistor is extremely nonlinear. That is the reason why arrester designs without spark gaps are possible [1], []. Figure shows a typically u-i-characteristic of a MO-surge arrester with In = ka. The voltage is normalized to the residual voltage at In.

12 The diameter of the MO-resis-tors decides the carrying capacity of the current, the height of the voltage, and the volume of the energy capacity. Table 7 shows the main data of the MO-resistors. For low-voltage application the same high-quality MO-material is used as for distribution and high voltage application. MO-resistors are compressed and sintered in the form of round blocks of different metal oxides in powder form. The diameters of the MO-resistors from ABB for low-voltage application are between 30mm and 7mm, covering even the highest energy requirements. The height of the blocks is between 1 mm and mm, covering a voltage range from 120 V a. c. to 100 V d. c. For special applications MO-resistors with a rectangular shape can be produced. All used materials are UV-resistant and perform well under extreme weather conditions. Safety and ecological aspects are specially taken into consideration with all arresters. Figures 11 to 14 show a selection of different types of MO-surge arresters from ABB. U 4/µs [p.u.] 1/µs 8/20µs /60µs 2000µs Figure 11 MO-surge arrester type LOVOS. This type was developed for outdoor application and can be used under all weather conditions. It is available with In = ka or ka, with or without disconnector. Uc = 280 V, 440 V and 660 V I [A] Figure Normalized voltage-current-characteristic of a MO-surge arrester with In = ka. Diameter of blocks in mm Nominal current In 8 / 20 µs in ka / 20 Imax 8 / 20 µs acc. class II test in ka Ipeak ( / 3 0 µs) acc. class I test in ka 2, 4,0 4, 12,0 Energy capability in kj / k VUc Figure 12 MO-surge arrester POLIM-R. Very high energy capability. Can be used for a. c. and d. c. networks. This type is, besides other applications, used in d. c. railway networks. Uc range from 140 V d.c. to 00 V d. c., and 1 V a. c. to 780 V a. c.. Tested according test class I and test class II. Table 7 Main data of ABB MO-resistors used in ABB MO-surge arresters for low-voltage application. The values are given as tested in the operating duty test to prove the thermal stability of the respective surge arrester. Other values are possible in other arrester designs..2 MO-surge arresters As long as very different applications and ratings for low-voltage surge arresters exist, different designs are needed. ABB offers a great variety of different arrester types for all kind of applications. The main design principle is always the same: A MO-resistor, as the active part, and the terminals are moulded completely in an insulating housing. Depending on the application and rating of the arresters the physical shape and housing material may be different. The general, surge arresters for outdoor application (e. g. overhead lines, MV / LV transformers) have a housing of polyamide; arresters for outdoor and indoor applications (e. g. railway applications) have a housing of silicon, and arresters of older design have housings of PUR. All arresters are moulded to be completely sealed and waterproof. Figure 13 MO-surge arrester MVR. Used in low-voltage systems and railway equipment. For a. c. and d. c. application. Available for In = ka and ka, with Uc = 440 V, 660 V and 800 V. 11

13 The standard IEC does not mention a high current impulse with a waveshape of 4/ µs and a rectangular current with a time duration of some ms. The high current impulse 4/ µs, as known from IEC [11], was intended to represent a severe direct lightning to the line very close to the arrester location. Direct lightning, and the relevant parameters, are covered more realistically by the impulse current Iimp, which is used for testing lightning current arresters (class I test). Rectangular currents are generated by discharges of a loaded transmission line of typically some hundred km of length. Such a current waveshape, coming from a line discharge, is not relevant for low-voltage networks. 6.1 Type tests Type tests are performed after completion of the design to prove the performance and specified characteristics of the product. The type tests are described in detail in the relevant standards. In the frame of this guideline, the main electrical tests for MO-surge arresters without gaps for outdoor application are described briefly. In general each test series is performed on three new test samples. The tests are performed in free air at room temperature (20 C ± 1 C) Figure 14 MO-surge arrester MVR...ZS. For low-voltage systems. Only indoor application. Suitable for fixing on DIN racks. In = ka, Uc = 140 V, 20 V and 440 V..3 Technical data of the arresters Test procedure to measure the residual voltage with 8/20 µs current impulses The voltage-current characteristic of the MO-surge arrester is measured with 8 / 20 µs current impulses in the range 0,1 to 2 times In. The result is given in form of a table or curve to show the protection performance depending on the current magnitude. Table 8 presents main electrical data of the arresters. The ratings are given according to [9], see also the definitions in chapter 4.2. All described MO-surge arresters are of the voltage limiting type. The energy capability, as given in the table, is the value as tested in the operating duty tests to prove the thermal stability of the arrester with the maximum continuous operating voltage applied. It is not the limiting value that would destroy the arrester. Arrester Type In Test class II for a. c. systems 8/20 µs ka LOVOS - LOVOS - POLIM-R...1N POLIM-R...2N Test class II for a. c. systems and special applications MVR...- MVR...- MVR...ZS Test class I for a. c. systems and special applications POLIM-R...1N POLIM-R...2N Up /Uc I max 8/20 µs ka Operating duty test The operating duty test has two parts: the preconditioning and the evidence of the thermal stability of the MO-surge arrester. It is a test in which service conditions are simulated by the application of a stipulated number of specified impulses to the MO-surge arrester while it is energized at the maximum continuous operating voltage Uc. For the preconditioning test, 1 times In in three groups of five impulses each, are applied to the test samples which are energized at Uc. Each impulse shall be synchronized to the power frequency. Starting from 0 the synchronisation angle shall be increased in steps of 30 intervals. The interval between the impulses is 1 min; the interval between the groups is 2 to 30 min. For practical reasons it is not required that the test sample is energized between the groups. In the operating duty test itself, e.g. to prove the thermal stability, the test sample is energized at Uc, and current impulses up to Ipeak (test class I) or Imax (test class II) are superimposed. The power frequency voltage is applied for 30 min after each impulse to prove the thermal stability. The superimposed current impulses should be of positive polarity and initiated in the corresponding positive peak value of the power frequency voltage. The value of the current impulse is increased from 0,1 to 1,0 Ipeak or Imax. The intermediate values are 0,2; 0, and 0,7 Ipeak or Imax. The arresters have past the test if thermal stability was achieved and the residual voltage at In measured before and after the test sequence has not changed by more than ± %. Energy Capability kj/kvuc 20 4,1 4,1 3,1 3, , 4,0 12,0 24,0 In 8/20 µs ka Up /Uc I max 8/20 µs ka Energy capability kj/kvuc 3, 3,64 3, ,0 4, 3,0 In 8/20 µs ka Up /Uc 20 3,1 3,1 I imp (/30 µs) I peak Charge Q ka As 20 Table 8 Electrical main data of the ABB surge arresters for low-voltage systems. The arresters of type POLIM-R have been tested according both class I and class II tests. The arresters of type MVR and POLIM-R can be used in d. c. systems as well, see chapter 9. Disconnector tests Arresters with an integrated or external disconnector are tested together in the operating duty test. During the complete sequence of preconditioning procedure and operating duty test the disconnector remains nonfunctioning. 6 Tests All tests for ABB low voltage arresters follow internationally agreed upon recommendations. For low voltage arresters in power systems the international standard IEC [9] is valid. For some special cases, for instance surge arresters for railway systems with d. c. voltage, other standards are applicable [2]. Thermal stability test (of disconnector) This test shows the disconnecting characteristic and the safety performance of overstressed surge arresters with disconnectors. The arrester with the disconnector is heated electrically with constant current untill 12

14 thermal equilibrium is reached or the disconnector operates. If the disconnector functioned, there should be clear evidence of effective and permanent disconnection by the device. The surface temperature of the device during the entire test should be below 120 C, and there should be no evidence of burning or ejected parts. The pass criteria depend on the classification of the arrester, e. g. whether it is indoor, outdoor, accessable or not accessable. 6.2 Special tests Additionally to the type tests given by the applicable standard, it may be necessary to conduct tests covering special requirements, (i. e. long term behaviour of the MO-material or the behaviour of the housing material under severe weather conditions). Accelerated ageing test This test has to show that the power losses of the arrester in the network under applied continuous operating voltage does not increase with time. An increase of the power losses would lead with time to a thermal runaway, and consequently to a failure of the arrester. In the accelerated ageing test the complete arrester is to be tested under increased stress, e. g. under increased ambient temperature of + 11 C. During the whole test period of 00 h the power losses are measured. It is vital that the power losses do not increase with time, but remaining constant at the lowest reached level. Because the material around the MO-resistor may influence its long term performance, it is important that the complete surge arrester is tested and not only the MOresistor. The test has to be performed with power frequency voltage for surge arresters with a.c. systems, and with d.c. voltage for surge arresters for application in d.c. systems. Ageing tests carried out with a.c. voltage are not transferable to the application in d. c. networks. The accelerated ageing test is performed with reference to the test procedure given in [11]. All ABB MO-resistors or MO-surge arresters, which are to be installed in d. c. networks, fulfill the most strict demands towards the long-term stability under d. c. voltage stress. UV radiation test In regions with strong solar radiation it is important to determine the behaviour of polymeric materials under UV radiation stress. The energy of the radiation can crack the surface of the insulator made of a synthetic material, and as a result the insulator may erode and finally fail. ABB surge arrester housing materials (silicon, polyamide and PUR) have successfully withstood UV radiation tests with time duration of 00 h. Water immersion test This test is performed to show the tightness of design against water permeation. It is performed with reference to [12]. The test samples are kept in a vessel with deionized boiling water with 1 kg / m 3 NaCl for 42 hours. To ensure the long term stability of the MO-resistors, from each produced batch two MO-resistors are taken and tested in a time-reduced accelerated ageing test. 6.4 Acceptance tests Acceptance tests are made upon agreement between manufacturer and customer. If acceptance tests are agreed upon they are then to be performed on the nearest lower number to the cube root of the number of arresters to be supplied. If not otherwise specified, the following acceptance tests are performed: verification of identification by inspection verification of marking by inspection verification of electrical parameters, for instance repetition of routine tests. 7 Selection of MO-surge arresters For selecting a MO-surge arrester three main electrical parameters have to be evaluated: continuous operating voltage Uc voltage protection level Up energy capability Additionally we need to be informed which modes should be protected. Table 9 shows the possible modes of protection, depending on the earthing practise in the low-voltage network. Depending on the application and the environment it has to be decided whether a disconnector is needed, which mechanical requirements need to be fulfilled (vibration and shock resistant, other mechanical stresses), and which ambient conditions have to be considered (increased temperature, solar radiation, rain, saltfog, etc.). SPD connected between: TT TN-C TN-S IT Line and neutral X X X* Line and PE X X X Line and PEN X Neutral and PE X X X* Line to line X X X X * When the neutral is distributed Table 9 Possible protection modes in low-voltage systems. Power system type 6.3 Routine tests Routine tests are carried out on every arrester or parts of it (e. g. on the MO-resistors) in order to ascertain that the product meets the requirements of the design specification. The test method and the pass criteria are declared by the manufacturer. All above mentioned MO-surge arresters for low-voltage application made by ABB are tested to 0 % in the routine test. On each arrester the reference voltage Uref is measured at the declared reference current Iref. Additional the arresters are checked to be free of internal partial discharges or contact noise. 7.1 Selection of U c The maximum continuous operating voltage Uc of an arrester has to be selected with respect to the power frequency voltages which can occure in the low-voltage system. Maximum system voltage has to be considered and possible temporary overvoltages in the network. Uc shall be equal or higher than the maximum power frequency voltage Ucs occuring in the system. U c U cs 13