You are on page 1of 9

IEEE Transactions on Dielectrics and Electrical Insulation

Vol. 19, No. 3; June 2012

803

Study on Lightning Impulse Test Waveform for UHV-class Electric Power Equipment
Shigemitsu Okabe, Toshihiro Tsuboi, and Genyo Ueta R&D Center, High Voltage & Insulation Group, Tokyo Electric Power Company 4-1, Egasaki-cho, Tsurumi-ku, Yokohama, 230-8510, Japan ABSTRACT
At present, a study is being conducted on the lightning impulse (LI) voltage test standard for UHV-class electric power equipment at WG 19 Adaptation of TC 42 standards to UHV test requirements in IEC TC42 High-voltage and high-current test techniques. In the present paper, the past LI voltage test results for UHV-class equipment were initially investigated to determine test waveforms that could be generated. As a result, for a UHV-class transformer and gas insulated switchgear assembly, the front time T1 exceeded 2.2 s or more, or even considerably 3.0 s in some cases in order to generate a waveform with an overshoot rate of 10% or less using existing test facilities. Subsequently, experiments were conducted with regard to the extension of the T1, which is a prospective way to revise the LI voltage test waveform. Breakdown characteristics with respect to the T1 were measured using the largest possible models assuming UHV-class equipment for three types of insulating media, namely an SF6 gas gap, an oil gap, and a long air gap. Consequently, it was confirmed that extending the T1 up to about 3.6 s had a minor influence on the breakdown characteristics. In addition, a study was also conducted on the influence of the deregulation of the standard value for the T1 from a viewpoint of insulation properties of equipment for the wave front steepness. As a result, it was concluded that the equipment insulation performance could be appropriately verified even if waveforms with an extended T1 were used. Based on these data, the idea of the standard for lightning impulse test waveforms for UHV-class equipment was proposed as follows: The standard voltage waveform shall be the same as the conventional one for less-than-800 kV equipment with a T1 of 1.2 s and a time to half value of 50 s but the upper allowable range of the T1 may be up to 3.6 s; provided, however, that a note shall be added besides the exceptional provision that, whenever possible, the test should be conducted using a waveform of which the T1 is adjusted to 1.56 s or less, the same as that for less-than-800 kV equipment. Index Terms Lightning impulse voltage test, front time T1, gas insulated switchgear, oil-immersed transformer, bushing, IEC 60060-1.

1 INTRODUCTION
NOW that plans and the construction of ultra-highvoltage (UHV)-class power transmission systems are being promoted, studies are also underway on the establishment of the standard at the relevant IEC. IEC 60060-1 High-voltage test techniques [1] is regulated by IEC TC42 High-voltage and high-current test techniques. In this regulation, the test standard is specified to determine the dielectric strength of electric power equipment, including the lightning impulse (LI) voltage test. In 2010, IEC 60060-1 was revised mainly for 800 kV-or-less equipment, whereby a waveform conversion method using the K-factor function (test voltage function) was introduced and standard values such as the
Manuscript received on 12 January 2012, in final form 28 February 2012.

overshoot rate (hereinafter ) were clarified [2-5]. At present, WG19 Adaptation of TC 42 standards to UHV test requirements is promoting a study on the establishment of the standard for UHV-class electric power equipment [6, 7]. Included in the challenge is the difficulty in generating waveforms that meet the standard due to, for example, the increase in series inductance based on the larger test circuit as well as the increased capacitance of the equipment to be tested. Consequently, in the study on the standard for LI voltage test waveforms, it is necessary to consider the possibility of waveform generation as well as the test capability in cases where waveforms deviate from the present standard waveform [1]. The present paper initially reports the results of an investigation into the LI voltage test results in Japan and China as well as simulation results with regard to waveforms

1070-9878/12/$25.00 2012 IEEE

804

S. Okabe et al.: Study on Lightning Impulse Test Waveform for UHV-class Electric Power Equipment

that can be generated. Subsequently, for the extension of the front time T1, which is a prospective way to revise the LI voltage test waveform, the experimental results of the breakdown characteristics regarding the T1 for the largest possible models assuming UHV-class equipment (three types, namely an SF6 gas gap, an oil gap, and a long air gap) are indicated. However, there is concern that the test waveform with an extended T1 may not appropriately verify the insulation performance with respect to wave front steepness. In this respect, the influence of the deregulation of the standard value for the T1 is evaluated from a viewpoint of equipment insulation properties. Based on these data, the prerequisites for the study conditions and standardization in accordance with the market needs are discussed to finally establish the idea of the revision of lightning impulse test waveforms.

2 LIGHTNING IMPULSE WAVEFORMS THAT CAN BE GENERATED FOR UHVCLASS EQUIPMENT


2.1 WAVEFORMS IN LIGHTNING IMPULSE VOLTAGE TEST RESULTS FOR UHV-CLASS ELECTRIC POWER EQUIPMENT Figure 1 summarizes the values of the front time T1 and the overshoot rate presently available for LI voltage test waveforms for actual UHV-class equipment [8]. In addition to the actual test results in Japan and China, this investigation also includes calculation results according to the circuit simulation used for the waveform adjustment for LI voltage test facilities based on the latest design data of three UHVclass equipment manufacturers in Japan. According to the test results for a UHV-class transformer, if of the test waveform is adjusted to 10% or less, the T1 far exceeds its upper allowable range of 1.56 s in the present standard. As for gas insulated equipment, a test that conforms to the present standard can be conducted for single gas circuit breaker (GCB); however, it is difficult to conduct such a test for gas insulated switchgear (GIS) assembly. Consequently, as things stand, the LI voltage test for UHV-class equipment often cannot be conducted under the test conditions specified in the present standard, and the test is conducted in accordance with the standard for individual equipment or the explanatory note (exceptional provision) attached assuming test results that exceed the standard value.
30

2.2 TRIAL CALCULATION OF THE CAPACITANCE OF THE TESTABLE LOAD A trial calculation is performed based on the circuit simulation to obtain the capacitance of the equipment to be tested that can generate the waveform conforming to the present LI voltage test standard. Figure 2 illustrates a simplified equivalent circuit for the LI voltage test facility [9]. In the test for actual large-sized equipment, the capacitance Ct of the test equipment is considerable, and thus likely to cause a long T1. Consequently, in order to shorten T1 to within the standard value, the series resistance Rs, for example, must be small. This eventually causes a relatively large residual inductance Ls of wiring or such compared with Rs, and a waveform with overshoot is generated due to the influence of this inductance. Accordingly, T1 and are in a mutually contradictory relationship, making it important to understand their inter-relationship to determine whether or not a waveform that conforms to the standard can be generated. In this regard, the relationship between T1 and was investigated using an equivalent circuit in Figure 2.
Impulse generator Rs Ls

Wiring Cs Rp Test equipment Rs: Series resistor Rp: Parallel resistor Ls: Residual inductance Cs: Power source Ct: Test equipment Ct

Figure 2. Image of the equivalent circuit of LI voltage test [9].

Overshoot rate, (%)

25 20 15 10 5 0 0.5
0.84 1.2 1.56

GIS GCB

Transformer of JPN Transformer of CHN GCB of JPN GCB of CHN GIB of JPN Assembly GIS of JPN

Transformer GIS
5 s, <10%

<some %
8 s, <som e %

2.4

3.0

3.6

Upper allowabletolerance range of of the front timeT T ( s) Upper-limit Front time, s) 11(

1.5

2.5

3.5

Figure 1. T1 and in the actual LI voltage tests for UHV-class electric power equipment.

Assuming the LI voltage test for UHV-class equipment, the value of each element was determined as follows: The impulse generator capacitance Cs was fixed at 25 nF assuming an actual UHV-class lightning impulse voltage generator (rated voltage: 6 MV, rated capacity: 450 kJ). The residual inductance Ls was set to 150 H, a relatively large value among the actual test results. For the test equipment capacitance Ct, the parameters were varied widely, assuming various types of equipment, while the parallel resistance Rp was adjusted so that the time to half value T2 may be about 50 s. Under such conditions, the series resistance Rs was varied to obtain the relationship between T1 and . The values T1 and T2 are those before the K-factor process. The use of a compensation circuit to improve the overshoot waveform was not considered. As examples, in the case of Ls = 150 H and Ct = 10,000 pF, assuming a UHV-class oil-immersed transformer, Figure 3a displays the respective waveforms generated corresponding to four levels of Rs and Figure 3b exhibits the change in T1 and with respect to Rs. If Rs is increased, the oscillation superimposed over the wave front diminishes, resulting in smaller but longer T1. In the case of Figure 3b, if Rs is increased to 220 or more, the present standard of

IEEE Transactions on Dielectrics and Electrical Insulation


1.6 Test Voltage (standadized) 1.4 1.2 1 0.8 0.6 0.4 0.2 0 0 5 10 Time (s) (a) Generated waveform 15 20
Rs = 200 Rs = 50 Rs = 100 Rs = 150

Vol. 19, No. 3; June 2012

805

Rs = 200 Rs = 150 Rs = 100 Rs = 50

Natural frequency (50ohm): 151.5kHz

3.0 2.5 Front time T1 ( s) 2.0 1.5 1.0 0.5 0.0 0 50 100 150 200 Series resistance Rs () 250 300 0.84s 1.56 s

30% 25% 20% Overshoot rate

T1
No allowed test condition 10%

15% 10% 5% 0%

satisfies the conditions of expands. If Ct and Ls are further decreased, Rs may satisfy the conditions of both T1 and , making it possible to generate a waveform that conforms to the standard. Figure 4 summarizes the calculation results of the relationship between T1 and with Ct as a parameter. The values of the actual results shown in Figure 1 are also plotted in Figure 4. The T1 - plots for these actual values are considered to move alongside the curve obtained by the calculation if the waveform is adjusted by Rs. As seen in the value of Ct, the smaller its value, the smaller those of T1 and . According to a comparison with the criteria in the present standard, it is found difficult to generate waveforms that conform to the standard if Ct exceeds about 2,300 pF when Ls is 150 H. Consequently, for UHV-class equipment, particularly the full GIS assembly and high capacitance transformers, it is actually difficult to generate waveforms that conform to the standard. If the residual inductance Ls is decreased by elaborately changing the test circuit configuration, Ct that can generate the waveform conforming to the standard is increased. However, such a practice has limitations, as indicated by the plots for the actual test results. 2.3 DIRECTION TO DEREGULATE THE STANDARD VALUE Under such circumstances as described in Section 2.2, it will be an idea to deregulate the standard values for T1 and in the test standard for UHV-class equipment. In this section, the values of Ct that can be tested with the standard waveform are compared using Figure 4 on condition that the standard for either T1 or is deregulated. However, this study is only from a viewpoint of circuit constants apart from, for example, the equivalence of the breakdown mechanism. (In Sections 3.2 and 4.3, a study is conducted from a viewpoint of the equipment insulation properties.) First, where is determined to be 20% or less, Ct that conforms to the standard can be up to about 3,000 pF (the Ct of the curve that traverses the point of T1 = 1.56 s and = 20% in Figure 4), which differs only slightly from the original value of 2,300 pF for = 10%. Conversely, where
GIS 8000 pF 10000 pF 12000 pF Transformer

(b) Changes of the front time T1 and the overshoot rate in relation to the series resistance Rs Figure 3. Calculation results for the LI voltage test waveform (Ls = 150 H, Ct = 10,000 pF).

10% [1] can be satisfied; however, even if Rs is minimized, T1 cannot be shorter than 1.78 s, which cannot satisfy the standard of 0.84 s T1 1.56 s. Consequently, in the case of Ls = 150 H and Ct = 10,000 pF, even if the series resistance Rs of the test facility is adjusted, waveforms that conform to the standard cannot be generated. If Ct and Ls are decreased, the curves of T1 and in Figure 3b shift downwards and leftwards, respectively, and the area of Rs that satisfies the conditions of T1 emerges and the area of Rs that
Other small apparatus GCB 3000 pF 5000 pF Ct=2000 pF 4000 pF

6000 pF

30
Overshoot rate, (%)

Ls=150 H
Transformer of JPN Transformer of CHN GCB of JPN GCB of CHN GIB of JPN Assembly GIS of JPN 5s, <10%

25 20 15 10 5 0 0.5
0.84 (-30%) 1.2 1.56 s (+30% 2.4 s (+100%) 3.0 s (+150%)

8s, <some %

3.6 s (+200%)

1.5

2.5

3.5

Time to peak, T1 (s) Figure 4. Relation between the front time T1 and the overshoot rate .

806

S. Okabe et al.: Study on Lightning Impulse Test Waveform for UHV-class Electric Power Equipment

T1 is determined to be 2.4 s (1.2s + 100%) or less, Ct that conforms to the standard can be significantly increased to up to 6,000 pF (the value of Ct of the curve that traverses the point of T1 = 2.4 s and = 10% in Figure 4). If T1 is extended to 3.0 s (+150%) or 3.6 s (+200%), Ct that conforms to the standard can be further increased to 10,000 pF or more, which will cover almost all equipment. Consequently, it is considered effective to extend T1 to increase the number of equipment that conforms to the standard for UHV-class equipment.

3 BREAKDOWN CHARACTERISTICS FOR THE LIGHTNING IMPULSE WITH THE EXTENDED FRONT TIME
3.1 EXPERIMENTAL CONDITIONS In order to study the possibility of extending the front time T1, a breakdown test was conducted for LI waveforms with a long T1 using the largest possible models assuming UHVclass equipment [10-12]. While there are some previous experiments where the V-t characteristics were obtained for the LI standard waveform, this was the first experiment, as far as the authors know, where an experiment was directly conducted to obtain the breakdown characteristics with the extended T1 by varying T1 of LI waveforms using a largesized model. The experimental conditions are summarized in Table 1. In this experiment, tests were conducted on three types of typical insulating elements, namely an SF6 gas gap assuming GIS, an oil gap assuming an oil-immersed transformer, and an air gap assuming an air insulation part such as a bushing. The test bus bar used for an SF6 gas gap had a coaxial cylindrical geometry assuming an actual GIS. Four electrodes 100 mm long and with a diameter of 240 mm simulating the insulating spacer shield part where the electric field was
Test object Test parameters: Front time T1 Test voltage Voltage level SF6 gas gap (Gas Insulated Switchgear) 1.2, 2.4, 4.8 s Negative LI voltage 1000 - 1200 kV
Gas pressure: 0.50 MPa-abs Electrode material: Stainless steel (Surface roughness: 25 m) Electric field utilization factor: 0.59 Electrode effective area: 4.5 104 mm2
GIS tank Electrod

maximized were placed inside the tank with a diameter of 340 mm. The electrode effective area (where the electric field strength is 90% or more of the maximum strength) was 4.5 104 mm2 and the field utilization factor was about 0.6, which was within the typical range of the field utilization factor for GIS. The electrodes were made of stainless steel and processed to a surface roughness of 25 m or less, the same as that for actual GIS. The SF6 gas pressure during the test was set to an absolute pressure of 0.50 MPa, the same as that when the equipment was in operation. The basic insulation configuration for a transformer is a composite insulation using insulating paper and insulating oil. However, since the composite insulation model is not self- restoring, it is difficult to obtain sufficient breakdown data. Meanwhile, the experimental results for turn-to-turn insulation have been already reported with the voltage distribution in transformer windings taken into consideration [13]. The breakdown characteristics of an oil gap mainly dominate in the partial discharge inception characteristics of oil-pressboard insulation. Experiments performed in the European Project [14, 15], which is a basis for the existing Kfactor function, were also conducted for an oil gap. From these viewpoints, the present experiment was conducted for an oil gap. For test electrodes constituting an oil gap, aluminum disk electrodes 1,000 mm in diameter and processed to a surface roughness of 50 m or less were used. They formed parallel planes featuring a gap length of 40 mm. The volume of space in these test electrodes, where the electric field intensity was 90% or more of the maximum, was 3.1 104 cm3, which represents about one third of the volume of the main insulation part between the windings of an actual 500 kV transformer. For the test model for an air gap, a ring shield and an earth plane constituted the gap simulating the corona shield of an
Air gap (Air insulation such as bushing) 1.2, 2.4, 3.6, 4.8 s Positive LI voltage 1500 - 1700 kV
Electrode material: Stainless steel (Surface roughness: 50 m) Electrode diameter: 560 mm Gap length: 3000 mm Electric field utilization factor: 0.04
Ring shield 400mm 60mm

Table 1. Experimental conditions for insulation characteristics under extended T1 waveform with large-sized model assuming UHV-class equipment.

Oil gap (Oil-immersed transformer) 1.2, 2.4, 3.6, 4.8 s Negative LI voltage 800 - 1000 kV
Tank volume: 8000 L Electrode material: Aluminum Electrode diameter: 1000 mm Gap length: 40 mm Electrode effective volume: 3.1 104 cm3
Bushing Acrylics window 40 mm gap 1000 Disk electrode 1m diameter

Test equipment

GIS tank: 340mm Electrode: 240mm 100mm width 4

560mm

3000mm

Test electrode Circulating pump Micro filter

Earth plate

IEEE Transactions on Dielectrics and Electrical Insulation

Vol. 19, No. 3; June 2012


1800 50% breakdown voltage (kV) . Air gap (positive) 1500 1200 900 600 300 0 0.0 1.0 2.0 3.0 4.0 Front time (s) 5.0 6.0 SF6 gas gap Oil gap

807

air bushing in GIS facility. The gap length of the test model was set to 3,000 mm. The ring shield shape was designed using electric field analysis so that the field utilization factor may be equivalent to those of actual equipment with both the central conductor and porcelain bushing taken into consideration. The field utilization factor was about 0.04. For the ring shield, an aluminum electrode processed to a surface roughness of 50 m or less was used. In all cases, an up-and-down method was used for the test, and the breakdown voltages and times were recorded. The test was conducted so that breakdown might occur 20 times or more. The use of this small step voltage V of 1-2% was intended to evaluate the influence of the overshoot part in detail in the whole waveform. A video camera was also used to check and ensure that the breakdown paths were not concentrated in any particular locations. 3.2 BREAKDOWN VOLTAGE CHARACTERISTICS WITH RESPECT TO THE FRONT TIME Figure 5 summarizes the change in the 50% breakdown voltage (hereinafter the 50% BDV), where the T1 varied from about 1.2 s to 4.8 s ( 0%). The error bars in Figure 5 represent the respective standard deviation values. Figure 6 displays examples of applied voltage waveforms and the plots of the breakdown time with the applied voltage. These plots are indicated using different plot shapes for the respective front times. The results for the SF6 gas gap are indicated using open plots to avoid likely confusion between the test results for an SF6 gas gap and those for an oil gap. For all the SF6 gas, oil, and air gaps tested, the longer the T1, the lower the 50% BDV tended to be. Accordingly, the insulation performance is considered to be more severely verified if the T1 is extended for the LI voltage test. The change in the 50% BDV with a T1 up to 3.6 s was roughly within 6%, or the width of the tolerance 3% of the test voltage. According to this result and the study results in Section 2.3, when verifying insulation performance, it is considered more suitable to allow an extension of the T1 rather than to increase to conduct the test. However, for oil-immersed transformers, the 50% BDV declined to 94.1% when the T1 was 4.6 s, showing an aspect that differed from that under conditions of 3.6 s or less, due to an increase in the breakdown ratio before the peak of the waveform as indicated in Figure 6. Also for the air insulation, the 50% BDV declined to 94.2% when the T1 was 5.0 s. Therefore, it is considered appropriate to set the T1 to about 3.6 s or less, otherwise the extension of the same may lead to excessively severe test conditions.

Figure 5. Changes in the breakdown voltage with respect to the front time.
1.2s 1.2s 2.4s 2.4s 3.6s 4.8s 4.8s 3 Test voltage waveform. (Standardized).

1800 1600 Breakdown voltage (kV) .. 1400 1200 1000 800 600 400 200 0 0

Air gap, T1 = 1.3, 2.6, 3.8, 5.0 s SF6 gas gap, T1 = 1.3, 2.4, 4.8 s

2.5

1.5

Oil gap, T1 = 1.2, 2.4, 3.5, 4.6 s

Standardized test voltage waveform: T1 = 1.3, 2.6, 3.8, 5.0 s

0.5

0 10 20 Time (s) 30 40

Figure 6. Characteristics of the breakdown voltage and time with respect to the front time T1.

of LI test waveforms leads to relatively severe test conditions in the verification of the insulation performance. Conversely, there is concern that the insulation performance for steep waveforms may not be appropriately verified if the T1 is extended. A discussion is made here on these points. For reference, the image of the relation between the elements of LI waveform and the insulating elements of a transformer verified by LI voltage test was indicated in the appendix. 4.1 GAS INSULATED SWITCHGEAR The insulation characteristics of GIS can be broadly classified into those when GIS is clean and those when a metallic particles exist. Firstly, Figure 7 displays the insulation characteristics for the standard LI waveform and waveforms with an extended T1 in the clean case [16]. According to Figure 7, the less steep the rising part of the waveform, the lower the breakdown voltage, and the breakdown voltage declines less at around 1.2 s to 3.0 s, meaning that the characteristics follow general V-t characteristics from 1.2 s to 3.0 s with no peculiar

4 INFLUENCE OF STEEPNESS AT FRONT PART ON EQUIPMENT INSULATION


This section discusses the extension of T1, which was indicated as a prospective measure to deregulate the test standard, from a viewpoint of equipment insulation characteristics. According to the experimental results in Section 3.2, the longer the T1, the lower the breakdown voltage tended to be. In other words, the extension of the T1

808

S. Okabe et al.: Study on Lightning Impulse Test Waveform for UHV-class Electric Power Equipment
Section No. #1 (1) Breakdown voltage (kV) . Electrode: Coaxial cylinder (50 / 130 mm) SF6 Gas pressure: 0.15 MPa-abs #2 (5) #3 (4) #4 (3) (2) U

(a) Measurement points in the insulation windings


1.2 1.0 0.8

0.3/50s 1.2/50s 3.0/50s


(5)

Potential (pu)

Time to breakdown (s) Figure 7. V-t characteristic of a coaxial cylinder arrangement for standardized impulse voltage (1.2 s / 50 s, negative polarity) [16].

(3)
0.6

phenomena observed. Subsequently, the characteristics under the existence of a metallic particle in GIS are represented in Figure 8 [17]. It can be confirmed that the breakdown voltage varies little within the range 1.0 s to 10 s and remains relatively low. Consequently, extending the T1 up to about 3.6 s is not considered to influence the verification of GIS insulation performance. 4.2 OIL-IMMERSED TRANSFORMER Figure 9 exhibits the actual measurement results of the potential distribution within the winding with the T1 varied assuming a UHV core-type transformer. According to this experimental result, the potential distribution inside the winding of an oil-immersed transformer was almost the same for waveforms with the T1 of 1.2 s and 3.0 s [18]. Furthermore, since the discharge formative time lag is longer than T1 for the oil-impregnated paper insulation structure of a UHV-class transformer, the influence of the steepness of the rising part is minor. Consequently, extending the T1 up to about 3.6 s is not considered to influence the verification of the insulation performance of an oil-immersed transformer.
Electrode: Coaxial cylinder (42 / 150 mm) Metallic particle: 10 mm needle electrode SF6 Gas pressure: 0.1, 0.2, 0.45 MPa-abs Oscillation impulse voltage, T1: 0.1 to 1000 s

(2), (4)
0.4 0.2 0.0 0.0 0.2 0.4 0.6 0.8 Number of turns (pu) 1.0 [3 sections]

(1)

(b) Potential distribution between sections Figure 9. Potential distribution between sections measured using a real-scale winding model [18].

1.0 s 10 s Time to breakdown (s)

Figure 8. V-t characteristic of a coaxial cylinder arrangement with metallic particles (10 mm needle electrode, positive polarity) [17].

4.3 DISCUSSION For a long air gap, considering the equipment insulation structure, the influence of T1 extension in a time domain of several s is considered almost null. Consequently, the insulation performance is considered to be appropriately verified, even if the T1 is extended to about several s. However, an excessive extension of the T1 leads to a severer insulation test, meaning equipment is subject to unnecessary stress. The test should be basically conducted with a T1 close to that of the standard LI waveform, and the insulation performance is considered verified under test conditions with the T1 extended, which are severer than the standard conditions, only in unavoidable cases due to the realities of the test facilities. As mentioned above, even if T1 is extended, there is little concern from a viewpoint of verifying the insulation reliability of equipment. Consequently, as schemes to deregulate the test standard, the extension of T1 is also superior from a viewpoint of the equipment insulation characteristics.

Breakdown voltage (kV) .

IEEE Transactions on Dielectrics and Electrical Insulation

Vol. 19, No. 3; June 2012

809

5 STUDY ON LIGHTNING IMPULSE TEST WAVEFORMS


5.1 CONDITIONS OF STUDY Based on the study in the previous Sections, the basic principles for the study on lightning impulse test waveforms will be as follows: (1) More focus may be placed on the possibility of waveform generation than the system aspect or significance, in the case of the lightning impulse test waveform. (2) This standard should treat, as a horizontal one, a common divisor among various apparatuses. Consideration of phenomena specific to the equipment, such as the inner electric potential oscillation of a transformer, should be entrusted to each apparatus standard. (3) Maintaining the maximum overshoot value to be 10%, consideration will be given to the possibility of increasing T1. The above items (1) and (2) are considered understandable according to the realities of the LI voltage test of UHV-class equipment and its status as an international standard (Sections 1 and 2). Item (3) is the result of the study of the deregulation, the standard value of which, or T1, is superior to others from a viewpoint of the realities of test facility (Section 2.3) and the equipment insulation (Sections 3.2 and 4.3), on the condition that the standard value must be deregulated as pointed out in (1) and (2). According to the above basic principles, the range of equipment conforming to the test standard when T1 is extended is summarized in Table 2. Table 2 is a result of evaluation through the circuit simulation and the actual test results summarized in Figure 4. In addition to the present T1 of 1.56 s (1.2 s + 30%), as an upper allowable range of the T1 extension, 2.4 s (+100%), 3.0 s (+150%), and 3.6 s (+200%) were considered. Table 2 also indicates the capacitances of the equipment to be tested that can generate waveforms conforming to the standard when the T1 is extended. Based on this, for GCB, GIS, a transformer, and

other apparatuses as specific equipment to be tested, a study was conducted on whether the equipment could be covered by the range whereby it could be tested within the standard value. Based on Table 2, the circumstances regarding the extension of the T1 are summarized as follows: I. A transformer and GIS assembly have the largest electrostatic capacitance and are followed by a single GCB in capacitance values. II. Concerning a transformer and GIS assembly, test waveforms are supposed to reach the T1 of 3.0 s, even exceed 3.6 s in some cases. With regard to other single pieces of equipment, they are expected to fall within the specified range (T1 1.56 s, and 10%) in the present standard 60060-1-2010.

III.

5.2 PREREQUISITES FOR STANDARDIZATION Followings summarize the prerequisites for standardization to narrow the conditions of study described in Section 5.1 down to the idea of the standard to be proposed: (a) Since the standard discussed specializes in UHV-class equipment, values in conformity with realities (market relevance) are desirable. A front time T1 of 3.0 s may not be potentially attained in cases of major equipment, such as transformers and GIS, and the T1 may even exceed 3.6 s in some cases. Since utilities and manufacturers of UHV-class equipment are limited, it is desirable that all apparatuses be covered, except for exceptional cases.

(b)

(c)

Item (a) is considered to be a prerequisite desired for the international standard that is directly connected to the application in the field. Item (b) points out the realities where a relatively significant extension of T1 is required for major equipment to be tested. Item (c) determines that, since the manufacturers and utilities of UHV-class equipment are limited to only certain companies having such advanced technology, it is desirable that the equipment of such limited companies be covered as a minimum.

Table 2. Relation between the front time and the range of equipment to be covered.

Upper allowable range of the front time T1 Capacitance range covered under the conditions of 10%, [Ls = 150 s] GCB Equipment GIS Transformer Other apparatus

1.56 s (+30%)

2.4 s (+100%)

3.0 s (+150%)

3.6 s (+200%) Much larger than 12,000 pF OK OK OK OK

2,300 pF Almost OK No No OK

6,000 pF OK Partially OK Partially OK OK

11,000 pF OK Almost OK Almost OK OK

810

S. Okabe et al.: Study on Lightning Impulse Test Waveform for UHV-class Electric Power Equipment

5.3 PROPOSAL OF THE LIGHTNING IMPULSE TEST WAVEFORM FOR UHV-CLASS EQUIPMENT Based on the study in Sections 5.1 and 5.2, the LI voltage test waveform for UHV-class equipment is proposed as follows; provided, however, that the standard for the waveform below shall be applied to the test voltage curve converted by the K-factor function. As a prerequisite, of the recorded curve must be 10% or less. (A) The standard voltage waveform shall remain unchanged, having the front time T1 and the time to half-value T2 of 1.2 s / 50 s. The upper tolerance of T1 shall be set to + 200% (3.6 s). (B) Provided, however, that a note sentence shall be added. Note: * There may be cases in which a transformer or GIS, etc. with a higher earth electrostatic capacitance cannot comply with the standard. In that case, each piece of apparatus shall be subject to the individual apparatus standards. * Whenever possible, the upper tolerance should remain confined to +30% (1.56 s), and even if this tolerance limit is exceeded, it is desirable that the new limit be close to it. The first point of (B) Note explains as a supplement that high capacitance equipment, such as GIS full assembly, should follow the standard for individual equipment. This type of the exceptional provision is considered necessary irrespective of the possibility of extending the T1. The second point of (B) Note is a supplement so that the test may not be conducted without adjusting the waveform to the standard (T1 = 1.56 s), despite though appropriate standard waveform (T1 = 1.56 s) being applicable. There is concern that extending the T1 in the present study may result in an excessive deregulation of the standard value except for GIS and a transformer. Extending the T1 also applies to the conditions where the equipment insulation performance is verified relatively severely and there is concern that unnecessary stress may be applied to the equipment. Consequently, the second point of the Note states as a supplement that, whenever possible, the test shall be conducted with a waveform adjusted to the standard one.

The actual LI voltage test results were initially investigated to determine test waveforms that could be generated. As a result, according to the test results and trial calculation for a UHV-class transformer and GIS assembly, the front time T1 was 2.2 s or more or even considerably exceeded 3.0 s in some cases in order to obtain a waveform with an of 10% or less with existing test facilities. Subsequently, with regard to the extension of the T1, which is a prospective way to revise the LI voltage test waveform, experiments were conducted for breakdown characteristics with respect to the T1 using large-sized models assuming UHV-class equipment. Consequently, it was confirmed that extending the T1 up to about 3.6 s had a minor influence on the breakdown characteristics. In addition, and T1 were compared from the aspect of insulation properties, and it was concluded that the extension of T1, which could verify the insulation performance appropriately, was superior as a test condition. Based on the study results from the above two perspectives, a draft standard for LI test waveforms for UHV-class equipment was proposed as follows: The standard voltage waveform shall be the same as the conventional one for less-than-800 kV equipment with a T1 of 1.2 s and a time to half value of 50 s but the upper allowable range of the T1 shall be up to 3.6 s; provided, however, that a note shall be added that the test should be conducted using a waveform of which the T1 is adjusted to 1.56 s or less, whenever possible, the same as that for less-than-800 kV equipment, together with the exceptional provision.

APPENDIX
To verify insulation performance against a lightning surge, a LI voltage test waveform is to be discussed mainly in terms of (1) crest value, (2) wave-front steepness, and (3) wave-tail. By conducting a LI voltage test, the following transformer insulation characteristics are verified: (a) insulation between turns in the winding (mainly near the line terminal), (b) insulation between sections in the winding, and (c) insulation between the windings or windings to ground. Regarding (a) and (b), (1) crest value and (2) wave-front steepness are dominant factors for determining the generated voltage. For (c), (1) crest value and (3) wave-tail are very influential due to the differential voltage between the windings or windings to ground. [18]

6 CONCLUSION
The present paper reported on the LI voltage test waveforms for UHV-class electric power equipment from two points. One is the possibility of generating test waveforms based on investigation of the actual test condition. The other is the influence of the deregulation of the test standard considered from the aspect of the insulation properties of equipment to be tested. Based on those investigations, the idea of LI voltage test waveforms was studied.

ACKNOWLEDGMENT
The authors express their sincere gratitude to the persons concerned of Tokyo electric power company and the three Japanese UHV manufacturers, Mitsubishi Electric Corporation, Toshiba Electric Corporation and Japan AE Power Systems Corporation, for their cooperation in the experiments, calculations and discussions.

IEEE Transactions on Dielectrics and Electrical Insulation

Vol. 19, No. 3; June 2012

811

U Ue (2)

(1) crest value (3) wave-tail wave-front steepness

(a) Important factors in lightning impulse waveform (a) Insulation between turns in the winding (b) Insulation between sections in the winding (c) Insulation between the windings or windings to ground

(a) (c)
Core

(b)

Tertiary Secondary winding winding (147kV) (500kV)

Primary winding (1000kV)

[10] T. Tsuboi, G. Ueta, S. Okabe, M. Miyashita, and K. Inami, Insulation Breakdown Characteristics of UHV-class Gas Insulated Switchgear for Lightning Impulse Withstand Voltage Test Waveform - K-factor Value and Front Time Related Characteristics, IEEE, Trans. Dielectr. Electr. Insul. , Vol. 18, pp. 1734-1742, 2011. [11] T. Tsuboi, G. Ueta, S. Okabe, Y, Shimizu, and E. Hino, Insulation Breakdown Characteristics of UHV-class Oil-immersed Transformer for Lightning Impulse Withstand Voltage Test Waveform - K-factor Value and Front Time Related Characteristics, IEEE, Trans. Dielectr. Electr. Insul. , Vol. 18, pp. 2049-2058, 2011. [12] G. Ueta, T. Tsuboi, S. Okabe, Y, Shimizu, and E. Hino, K-factor Value and Front Time Related Characteristics of UHV-class Air Insulation for Positive Polarity Lightning Impulse Test, IEEE Trans. Dielectr. Electr. Insul., Vol. 19, No. 3, pp. 877-885, 2012. [13] S. Okabe, T. Tsuboi and J. Takami, Evaluation of K-factor based on insulation characteristics under non-standard lightning impulse waveforms, IEEE Trans. Dielectr. Electr. Insul., Vol. 16, pp. 1124-1126, 2009. [14] S. Berlijn, F. Garnacho, P. Simn, E. Gockenbach, P. Werle, K. Hackemack and M. Watts, Digital Measurement of Parameters Using for Lightning Impulse Tests for High Voltage Equipment, Final Report, Contract PL951210-SMT4-CT96-2132, 1999. [15] P. Simn, F. Garnacho, S. M. Berlijn and E. Gockenbach, Determining the Test Voltage Factor Function for the Evaluation of Lightning Impulse with Oscillation and/or an Overshoot, IEEE Trans. Power Delivery, Vol. 21, pp.560-566, 2006. [16] W. Knorr, K. Moller, and K. J. Diederich, Voltage-time characteristics of slightly non-uniform arrangements in SF6 using linearly rising and oscillating lightning impulse voltages, CIGRE paper, 15-05, 1980. [17] CIGRE WG.C4.302, Insulation Coordination related to Internal Insulation of Gas insulated Systems with SF6 and N2/SF6 Gas Mixtures under AC Condition, CIGRE Technical Brochure No.360, 2008. [18] Y. Yamagata, and S. Okabe, Utility's experience on design and testing for UHV equipment in Japan, Intl. Sympos. Standards for Ultra High Voltage Transmission, Session 3, No. 3.6, 2009. Shigemitsu Okabe (M98) received the B.Eng., M.Eng. and Dr. degrees in electrical engineering from the University of Tokyo in 1981, 1983 and 1986, respectively. He has been with Tokyo Electric Power Company since 1986, and presently is a group manager of the High Voltage & Insulation Group at the R & D center. He was a visiting scientist at the Technical University of Munich in 1992. He has been a guest professor at the Doshisha University since 2005, at the Nagoya University since 2006, and a visiting lecturer at the Tokyo University. He works as a secretary/member at several WG/MT in CIGRE and IEC. He is an Associate Editor of the IEEE Dielectrics and Electrical Insulation. Toshihiro Tsuboi received B.Eng. and M.Eng. degrees in electrical engineering from the Tokyo Institute of Technology in 1997 and 1999, respectively. He joined the Tokyo Electric Power Company in 1999 and is presently a member of the High Voltage & Insulation Group at the R & D center. His main research interest is the insulation design of power systems.

(b) Main insulating elements to be verified Figure 10. The image of the relation between the insulating elements of a transformer and the factors in LI voltage test waveform.

REFERENCES
[1] [2] IEC 60060-1, High-voltage test techniques Part 1: General definitions and test requirements, 2010. F. Garnacho, E. Gockenbach, S. Benda, P. Simn, K. Hackemack, and P. Werle, New proposal for evaluation of lightning impulses based on experimental research. Revision of the IEC60060-1 and IEC 61083-2, Electra Paper, No. 204, pp. 30-39, 2002. J. Hllstrm, S. Berlijn, M. Gamlin, F. Garnacho, E. Gockenbach, T. Kato, Y. Li, and J. Rungis, Applicability of different implementations of K-factor filtering schemes for the revision of IEC60060-1 and -2, 14th Intern. Sympos. High Voltage (ISH), Paper No.B-32, 2005. Y. Li and J. Rungis, Analysis of Lightning Impulse Voltage with Overshoot, 14th Intern. Sympos. High Voltage (ISH), Paper No.B-08, 2005. J. Hllstrm, A. Bergman, Ding, F. Garncho, R. Gobbo, T. Kato, Y. Li, A. Nilsson, G. Pesavento, S. Sato, and A. X. Yu, International comparison of software for calculation of lightning impulse parameters based on a new processing algorithm, 15th Intern. Sympos. High Voltage (ISH), Paper T10539, 2007. M. Hinow, W. Hauschild, E. Gockenbach, Lightning impulse voltage and overshoot evaluation proposed in drafts of IEC 60060-1 and future UHV testing, IEEE Trans. Dielectr. Electr. Insul., Vol. 17, pp. 1628-1634, 2010. U. Riechert, R. Pietsch, S. Okabe , F. Garnacho, M. Hanai, M. Hinow, W. Hauschild, A. Pigini, J. Rickmann, and X. Jiang, Experiences Concerning Dielectric Testing of Ultra High Voltage (UHV) Equipment, CIGRE SC A2 & D1 Joint Colloquium 2011, Kyoto Japan, PS3-O-4, 2011. S. Okabe, T. Tsuboi, G. Ueta, Y. Yamagata, and E. Zaima, Study of Lightning Impulse Test Waveforms for UHV Equipment, CIGRE Paris session 2012, D1-204, 2012. S. Okabe, T. Tsuboi, and J. Takami, Basic study of possible waveforms generated in lightning impulse withstand voltage test on UHV equipment, IEEE, Trans. Dielectr. Electr. Insul., Vol. 16, pp. 1127-1133, 2009.

[3]

[4] [5]

[6] [7]

[8] [9]

Genyo Ueta received the B.S. and M.S. degrees from Doshisha University in 2000 and 2002, respectively. He joined Tokyo Electric Power Company in 2002. Currently, He is a researcher at the High Voltage & Insulation Group of R & D Center and mainly engaged in research on insulation characteristics of electric power equipment.

You might also like