This article discusses the breakdown voltage characteristics of gas‐filled transformers and reactors. To determine their insulation characteristics, this study has used models simulating their insulating elements: interturn and section‐to‐section insulation. The study measured 50% breakdown voltages in the section‐to‐section and interturn models under standard lightning impulse voltages that were 1.13 and 1.06 times higher than those under the standard switching impulse voltages, respectively. The mechanism of the above‐mentioned facts is discussed, in terms of the formative time lag of the discharge. © 2020 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.
Ultra high voltage (UHV) systems are increasingly being planned and constructed, hence studies are promoted on the standard for high-voltage test techniques for UHV-class equipment. For the lightning impulse voltage test, a study is being conducted on the application of a method of evaluating the test waveform through conversion using the test voltage function (k-factor function) that was adopted in IEC 60060-1. The existing k-factor function was established based on the experimental results for more compact models, as compared with the insulating structure of UHV-class equipment, mainly with a breakdown voltage of about 100 kV. To determine whether this k-factor function can also be used for the test of UHV-class equipment, the experimental results for large-sized models were needed. In the present paper, to address this issue, the authors initially obtained k-factor values experimentally using the largest possible model (UHV model) assuming UHV-class equipment. Substantially, a study was conducted on a new k-factor function based on these experimental results. First, in the study, several ideas for the k-factor function were shown and applied to various waveforms to clarify their advantages and disadvantages. Next, in addition to these results, a study was conducted on a k-factor function suitable for UHV-class equipment with considering the actual UHV facilities. Consequently, it was concluded that the form of the function should be the same as that of the existing one but that it would be reasonable to adopt a relatively lower k-factor function for UHV-class equipment by revising the constant. Further, this new function could replace the existing one in 60060-1 for all voltage classes to consider the breakdown voltage ranges as a basis and LIWV (Lightning Impulse Withstand Voltage) values.
Thanks to its excellent dielectric strength and arc interruption performance, SF6 gas has been widely used for power equipment since the 1960s. However, since SF6 has a high global warming potential (GWP), it was designated at COP3 in 1997 as a greenhouse gas to be reduced; hence the search for an alternative insulating gas to SF6 to be widely used for power equipment. Aiming to reduce the global warming potential (GWP) of existing power equipment using SF6, the present study investigated breakdown characteristics when various alternative gases to SF6 were mixed. Alternative gases to SF6 handled in this report were based on four kinds of gases, namely c-C4F8, 1-C3F6, C3F8 and C2F6, from among per-fluorocarbon gases with relatively high dielectric strength. However, since these base gases have relatively high liquefaction temperature, despite their advantage in terms of GWP and dielectric strength, a gas with a low liquefaction temperature was mixed as an additive. Three kinds of gases were used as additive gases, including CF4, a fluoride gas with a relatively low liquefaction temperature, in addition to N2 and CO2. Accordingly, in this report, various combinations of the four kinds of base gases and three kinds of additive gases were used to obtain the respective insulation characteristics while varying the mixture ratio of the respective gases. Consequently, it emerged that relatively high and positive synergism was likely to be obtainable when the mixture ratio of the base gas was 20%. Furthermore, a comparison of dielectric strength with respect to the GWP revealed that the GWP was likely to be reduced to 10% or less while maintaining about 80% of dielectric strength, under conditions that the upper limit of the liquefaction temperature assuming indoor equipment was 0°C.
Thanks to its excellent insulation and cutoff performances, SF 6 gas has been applied to power equipment since the 1960s and is now widely used for several kV to 1,000 kV-class GIS, GCB, and GIL. However, since 1997, when SF 6 was designated at COP 3 as a greenhouse gas to be reduced, there has been a wish to use an alternative insulating gas to pure SF 6 gas. In the present study, alternative gases were selected from among mixtures excluding SF 6 with the need to reduce GWP (global warming potential) in mind. Gas mixtures containing such substances and with a boiling point of -20°C or less, chemically stable, non-toxic, and not ozone-depleting were prioritized. Furthermore, the availability and environmental performance were taken into consideration when deciding on component gases. Consequently, to launch this series of studies, four types of gas mixtures were used combining a gas in Group A (C 2 F 6 , C 3 F 8 ) - electronegative gases with relatively high dielectric strength - and a gas in Group B (N 2 , CO 2 ) - gases existing in the natural world. The GWP of SF 6 is 22,800 whereas that of C 2 F 6 is 12,200 and that of C 3 F 8 is 8,830, or several times smaller than that of SF 6 . In the present paper, insulation characteristics were experimentally obtained while varying the mixture ratio under a quasi-uniform electric field assuming GIS. Consequently, compared to the GWP of pure SF 6 , the GWP was about 12% to 38% for gas mixtures with C 3 F 8 /N 2 or C 3 F 8 /CO 2 and 18% to 70% for gas mixtures with C 2 F 6 /N 2 or C 2 F 6 /CO 2 . Consequently, it emerged that, while assuming breakdown voltage proportional to gas pressure, the GWP was likely to be reduced by 30% to 90% while maintaining dielectric strength. In addition, a study was conducted on the synergism of a gas mixture through analysis using the Boltzmann equation. Consequently, the synergism was confirmed while its degree varied depending on the type of each gas mixture, and the mechanism thereof was clarified.
Now that gas insulated switchgear (GIS) for ac systems is becoming increasingly compact as specifications are rationalized, more consideration of their insulation characteristics for residual dc voltage is required. Furthermore, with dc power transmission technology drawing more and more global attention, clarifying the insulation characteristics of GIS for dc voltage is increasingly important. The insulating portions for which the influence of dc voltage must be taken into consideration are solid insulators, such as insulating spacers. Under dc voltage, since the electric field distribution in an insulator differs from that under ac or impulse voltage and is governed by the resistance characteristics, clarifying its characteristics is crucial to study the GIS dc insulation design. As a solid insulator, focusing on fiber-reinforced plastics (FRP) used for GIS, for example, insulating rods, as well as partially treated epoxy resin; this paper experimentally investigated the bulk and surface electric conductivity under dc voltage, using the electric field, temperature, and other factors as parameters. As a result, the bulk electric conductivity of FRP in an edgewise direction exceeded that in the penetrating direction by one digit. It emerged that the electric conductivity of an insulating material with orientation like FRP varied depending on its direction. It was also found that, despite the fact the bulk and surface conductivity depended on the electric field for both FRP and epoxy resin, the variation width was relatively narrow within the range of the actual GIS operating electric field. The bulk and surface electric conductivity were also temperature-dependent, which meant the variation width was relatively wide. Furthermore, the surface electric conductivity was measured in SF 6 gas and in the air to investigate the influence of the ambient atmosphere, whereupon it emerged that the electric conductivity was higher in air due to the adherence of moisture. As mentioned above, the electric conductivity of an insulator varies due to various factors, such as the influence of the material orientation, electric field, temperature, and moisture. Consequently, the electric field distribution inside the insulator also changes, meaning these electric conductivity characteristics must be taken into consideration to study the GIS dc insulation characteristics.
To operate power transformers long-term, as well as ensuring their insulating reliability, it is important to study age-related decline in various insulating oil characteristics and the method used to evaluate the same adequately. Previous studies showed that age-related decline in insulating oil characteristics was caused by trace components in oil produced during oxidation degradation. To maintain and manage aged insulating oil rationally, a specific diagnostic method must be studied, based on measurement of the trace components in oil that cause degradation in these insulating oil characteristics. The present study evaluated the sensitivity to detect various components based on the amount of trace components produced in oil during oxidation degradation for field-aged insulating oil for 34 transformers in various degradation conditions. A study was also conducted to evaluate the degradation condition based on the ratio of trace components produced in oil. Consequently, the detection sensitivity levels were in the order of carbonyl value, saponification value, peroxide value, and total acid value. As the ratio of the saponification value - the so-called final product increased, the degradation of insulating oil developed further. With the above study results, methods of evaluating the dissociation property in the preceding study and the breakdown voltage with the degree of water saturation taken into consideration were combined to establish a comprehensive method of evaluating aged oil. This was then applied to field-aged insulating oil as an example, whereupon the need to replace insulating oil could be evaluated.
The paper investigated the behavior and partial discharge (PD) characteristics of a metallic particle under residual dc voltage, using the setting conditions and length of the particle, the applied dc voltage, and other factors as parameters and a gas insulated switchgear (GIS) bus bar model equivalent to that of an actual 300 kV GIS. A metallic particle repeated reciprocating movement at a relatively high frequency of about 5 times per second between the high-voltage conductor and the sheath when the electric field at the bottom surface of the tank exceeded its lift-off electric field. It emerged that, in the process of this reciprocating movement, PD occurred at the moment when the particle collided with the high-voltage conductor. In addition, when a particle collided with the electrode, the electrical charges moved and the residual dc voltage was damped. Conversely, where the sheath side was insulation-coated, the particle did not move at the normal operating voltage level, and even though it moved if vibration was applied, its movement stopped relatively soon. If a metallic particle exists in GIS and moves, there is concern that the insulating performance may decline significantly. When a metallic particle moves under dc voltage, a relatively stable PD is generated and consequently, for example, the PD measurement is considered an effective way to detect a particle. It is also considered effective to take physical measures, such as using an insulating sheath at the bottom surface of the tank to restrain the behavior of particles.
The IEC-60060-1 "High-voltage test technique" was revised in 2010 and an evaluation method using the k-factor function (test voltage function) was introduced for the lightning impulse voltage test, which involved the overshoot waveform being converted into the test voltage waveform. Test standards for UHV-class equipment are also presently being studied. To date, authors have measured k-factor values with the largest possible model, assuming actual UHV-class power equipment and proposed a new k-factor function for large-scale equipment. The present paper reports on the evaluation results of expanded uncertainty when measuring a k-factor value on the basis of the ISO/IEC Guide 98-3. As a result, it was clarified that significant uncertainty in measuring a k-factor value is brought about by the definition of the k-factor itself because small differences in the 50% breakdown voltage between the smooth lightning impulse and overshoot waveforms with respective uncertainties at a certain level were calculated. The expanded uncertainty when measuring the k-factor value in the present study for an overshoot rate of 10%, which is particularly important for practical use, was a large value of +/-0.23 to +/-0.36, or 0.46 to 0.72 in top-to-bottom width compared to k-factor values themselves ranging from 0.2 to 1.0. However, this is actually considered a minimum level of uncertainty, which is inevitable in principle, and the best achievable measurement result rather than something caused by any problem in measurement itself.
To study the dc insulation design of gas insulated switchgear (GIS), the insulation characteristics under lightning impulse (LI) voltage with a superimposed dc voltage (superimposed voltage) must be clarified. The paper experimentally examined the GIS breakdown characteristics under this superimposed voltage. The test models simulating an insulator creepage surface were used for which consideration of the influence of dc voltage among various other GIS insulating elements is particularly important. To be specific, a cylindrical model made of epoxy resin or fiber-reinforced plastic (FRP) as the material and a conical epoxy spacer model were tested. For the cylindrical model, a cap-shaped electrode was placed on the insulator and a small gap was established between the end of the electrode and the insulator. When the dc breakdown voltages were measured using these samples, they were higher for the applied voltage of positive polarity than that of negative polarity for all samples. The post-test observation of the electrification condition revealed greater electrification on the insulator surface for the applied voltage of positive polarity. The electrification charges are considered to have relaxed the electric field and increased the breakdown voltage. Subsequently, the breakdown test was conducted using a superimposed voltage, whereby a foregoing dc voltage was applied to samples for a certain period, whereupon a LI voltage was applied with the dc voltage continually applied. The breakdown voltage when the LI voltage and dc voltage had equivalent polarity was approximately same to the LI alone breakdown voltage. Conversely, when they were opposite in polarity, the breakdown voltage under the superimposed voltage obviously tended to decrease from the LI alone breakdown voltage. It is considered attributable to the fact that the insulator surface was electrified by the foregoing dc voltage and applying LI voltage opposite in polarity to this electrification intensified the electric field where the breakdown started to occur. Accordingly, it emerged that the GIS breakdown characteristics changed significantly depending on the polarity combinations of LI and dc voltages. The influence of these polarities must be taken into consideration when studying the GIS dc insulation design.
Now that gas-insulated switchgear (GIS) for ac systems are becoming increasingly compact as specifications are rationalized, more consideration of their insulation characteristics for residual dc voltage is required. Furthermore, with dc power transmission technology drawing more and more global attention, clarifying the insulation characteristics of GIS for dc voltage is increasingly important. For GIS insulating epoxy spacer, the present paper experimentally and analytically studied the influence of spacer surface condition on dc withstand voltage performance. In specific terms, initially, an insulating spacer model with an area of non-uniform resistivity present in the surface layer was created and the dc voltage breakdown characteristics were obtained. As a result, it was clarified that the time delay to breakdown increased with lowering the applied voltage, or in other words, even if a low voltage was applied, breakdown may occur after an extended period. Subsequently, using the same non-uniform resistivity model, the change in the electric field distribution over time under dc voltage was investigated through transient electric field analysis. Consequently, it was found that the electric field distribution varied from a capacitive to a resistive field and the maximum electric field was generated in the boundary between the high- and low-resistivity areas. It was further clarified, based on these breakdown characteristics and electric field analytical results, that the presence of an area of nonuniform resistivity on the insulating spacer creepage surface caused electric field concentration over time and breakdown occurred if the electric field in the creepage surface direction exceeded 40 kV/mm or so. Moreover, the breakdown characteristics were investigated where a lightning impulse voltage was superimposed over a dc voltage. Consequently, the creepage breakdown electric field where a dc voltage was applied alone was almost identical to that where the superimposed voltage was applied. It means that the application of the superimposed voltage may potentially allow the insulating spacer withstand voltage performance to be evaluated by a shorter time test for a dc voltage alone or the dc voltage with a superimposed VFTO.
Gas insulated switchgears (GIS) have been increasingly introduced as main substation equipment since the late 1960s thanks to their high reliability, safety, and compactness. The UHF method to diagnose partial discharge (PD) is broadly employed in the field as an advanced insulation diagnostic technology. There are three primary factors related to the PD signal detection characteristics using this UHF method; namely the frequency spectrum of the PD signal itself inside GIS, the propagation characteristics of the PD signal inside the GIS tank, and the detection characteristics of the sensor measuring the PD signal. Accordingly, to further improvement of the diagnostic technique based on the UHF method, the PD phenomenon itself and its propagation characteristics should be clarified and the measurement system refined. From these perspectives, the present study reviewed the recent and latest findings concerning the PD phenomenon and its propagation characteristics inside GIS as well as the PD measurement system verification methods and diagnostic technologies, and summarized typical examples. Firstly, with regard to the PD phenomenon, measurement up to the high frequency band exceeding 10 GHz (even 30 GHz) clarified that the rise time of the PD current waveform is several tens of picoseconds, shorter than previously known. In the case of PD in micro-defects inside the epoxy insulator, however, the rise time is relatively long, in the order of nanoseconds, and a crack is the most critical defect. Subsequently, with regard to UHF PD diagnostics/monitoring technology, a novel technique using PD current waveform characteristics has been developed, and a more advanced PD diagnostic algorithm has been established by clarifying the influence of the shape and GIS internal structures on the electromagnetic wave propagation characteristics. Less uncertain and simplified calibration and verification technologies are also proposed both for a single UHF sensor and the entire UHF measurement system. These new technologies and further advanced studies in future are expected to make the UHF method more convenient and sophisticated.
Now that gas insulated switchgear (GIS) for ac systems are becoming increasingly compact as specifications are rationalized, more consideration of their insulation characteristics for residual dc voltage is required. Furthermore, with dc power transmission technology drawing more and more global attention, clarifying the insulation characteristics of GIS for dc voltage is increasingly important. In this paper, to understand the insulation characteristics of epoxy resin, which is widely used for GIS insulating spacers, factors determining the resistivity of the epoxy insulator surface layer under dc voltage were initially investigated on an experimental basis. Consequently, it emerged that the bulk resistance was more dominant than the surface resistance for the dc resistance of epoxy resin due to the dependency of the test sample resistance value on their radius. Since the electric field might be concentrated if some part of this insulator surface layer showed non-uniform resistivity, the influence of the curing agent, one of the potential causes of this non-uniformity, was subsequently investigated with its content as a parameter. As a result, the volume resistivity in the long-term region was likely to decline or vary for epoxy resin containing less curing agent due to the presence of numerous polarized components unreacted with curing agent. In addition, the presence of micro protrusions or similar, if any, on the insulator surface or electrode is considered to cause electrification due to the concentration of electric field on the surface layer. Accordingly, the relationship between their surface roughness and electrification level was investigated using gaps between insulators or an electrode and an insulator facing each other, respectively. Consequently, where the surface roughness of the insulator or electrode was high, a current component with a large damping time constant, considered attributable to electrical charges moving across the gap, appeared after the charging current components and an electrification condition was observed.
Generally, insulating spacers of gas insulated switchgear (GIS) have outstanding durability and are not considered prone to insulation failure within a design life of about 30 years, provided the products conform and have passed the partial discharge (PD) test. However, to assume operation over 30 years, degradation characteristics are important in cases where an extremely microscopic defect below the detection level in the PD test is present or produced inside the insulator. Accordingly, to date, the authors have obtained the breakdown voltagetime (V-t) characteristics using epoxy insulators provided with three types of micro-defect (crack, void, and delamination); the apparent PD of which is equivalent to 1 pC for an actual 550 kV-GIS spacer. It emerged based on these V-t characteristics that the breakdown risk peaked for a crack defect among those three types of shape and that it was essential to obtain data on long-term breakdown characteristics at the actual operating electric field level to rigorously evaluate the breakdown risk under long-term operation. Therefore, this paper includes a study on the equivalence of accelerated degradation by frequency acceleration as a means of determining the long-term breakdown characteristics. In the study, the dependency of various PD characteristics on frequency was initially examined by changing the frequency of the applied voltage within the range 60 to 3000 Hz. Consequently, the increase in the number of PDs per unit time proportional to frequency could be confirmed without any abnormal phenomena, such as suspended discharge, observed within the frequency range this time. In other words, it was clarified that the degradation due to PD could be accelerated by increasing the frequency. Subsequently, breakdown tests were conducted with frequency as a parameter to evaluate the frequency up to which the equivalence of accelerated degradation could be maintained based on breakdown times. As a result, it was found that the equivalence of acceleration degradation by frequency acceleration was valid in acceleration at up to 1500 Hz because the breakdown time was shortened in reverse proportion to the frequency.
Since SF 6 gas, an insulation medium used for gas insulated switchgear (GIS), has a high global warming potential, an effective alternative is sought from an environmental perspective. The authors are focusing on CO 2 gas as a potential alternative, given its relatively good insulation characteristics among gases with a low environmental impact (natural gases). The present study obtained and evaluated the insulation characteristics for complex waveforms, such as double-frequency oscillation waveforms, generated in an actual substation under a non-uniform electric field typically represented by metallic particle. Consequently, it was emerged that the breakdown voltage for a positive polarity waveform was 1.1 to 1.6 times higher than that for the standard lightning impulse waveform; even if parameters such as the frequency and damping rate were changed within the range assumed in the field. Furthermore, a study was conducted on a means of evaluating the insulation characteristics of the CO 2 gas gap for non-standard lightning impulse waveforms via an approach involving comprehensive handling of various waveforms, as was applied to the experimental results accumulated to date. Consequently, using a parameter of so-called duration, the breakdown characteristics for various waveforms, including complex oscillation waveforms generated in actual systems, could be expressed using a single characteristic line.
The lightning impulse voltage test waveform of power equipment is prescribed in the IEC 60060-1 "High-voltage test techniques" revised in 2010. This revision introduced an evaluation method by k-factor function (test voltage function), which converts an overshoot waveform to a test voltage waveform. At present, the test standard for UHV-class equipment is being studied and the authors have experimentally acquired k-factor values for gas-insulated switchgears, oil-immersed transformers and air insulation using the largest possible model assuming an actual UHV-class equipment to date. This paper reports the experimental result on the breakdown characteristics of lightning impulse voltage test waveforms for the SF6 gas insulation model under a nonhomogeneous electric field. The k-factor value was calculated by measuring the breakdown voltage and time using the superimposed oscillation frequency and the overshoot rate as parameters. The change with respect to the superimposed frequency of this experiment was similar to that of the existing k-factor function, and the effectiveness of evaluation by the k-factor function was confirmed.
To operate power transformers long-term, as well as ensuring their insulating reliability, it is also important to study the age-related decline in various insulating oil characteristics and the method used to evaluate the same in an adequate manner. In the present paper, focusing on the oxidation degradation process of insulating oil, an experimental study was conducted on the influence of the hydrophilic and dissociative properties of oxidation degradation products on the insulating oil characteristics. Consequently, it emerged that, in the presence of highly hydrophilic degradation products, the influence of the water content on the breakdown voltage was likely to be less significant because the saturated water content increased. In the presence of oxidation degradation products with dissociation properties, the characteristic dissipation factor and volume resistivity values were likely to decline due to the influence of water content and heating. Based on the above results, the data in preceding studies were organized. As a result, there was a correlation between the breakdown voltage and the water saturation rate (actual water content in oil/saturated water content). In addition, it emerged that the dissociation property of insulating oil could be potentially determined based on correlation between the 80 °C values of the dissipation factor and the zero-minute values of volume resistivity at room temperature. These results pertain to the evaluation method used to diagnose the degradation condition of actual field-aged oil.
The insulation performance of an oil-immersed power transformer against lightning surges is verified by applying a lightning impulse (LI) voltage alone. However, the surge voltage generated at the transformer terminal in an actual system is superimposed over the ac operating voltage. Particularly for UHV- and 500 kV-class, where the ratio of ac operating voltage to the LI withstand voltage is high, the insulation characteristics for this LI and ac superimposed voltage are crucial factors to study the insulation design. Accordingly, this paper reviewed the research results of insulation characteristics of transformers for the superimposed voltage. Subsequently, these results were evaluated from perspectives of an insulation coordination and transformer insulation test. The insulation elements of transformers covered were the turn-to-turn and section-to-section insulation, for which the LI test is key for insulation design. As a result, for the section-to-section insulation for a shell-type transformer and the turn-to-turn insulation, the breakdown (BD) voltage for the superimposed voltage remained unchanged from the result when the LI was applied alone. Conversely, for the section-to-section insulation of a continuous disk winding for a coretype transformer, the BD voltage attributable to the superimposed voltage declined by about 10% to 20% from that when the LI was applied alone if the oil-impregnated paper component ratio compared to the oil-gap length was low. This is because the oil gap was primarily responsible for the BD and flaws were generated there when the ac voltage was applied. On the other hand, when this ratio was increased, the oil-impregnated paper part was primarily responsible for BD and the decline in BD voltage could be suppressed. For the section-to-section insulation for the interleaved disk winding, the BD voltage attributable to the superimposed voltage remained constant from that when the LI was applied alone. This is because the BD was determined by the LI partial discharge inception at the turn-to-turn part of the section-to-section insulation, where the superimposed voltage made little impact on BD. Based on the above, the LI test is considered adequate to verify the insulation performance of the section-to-section insulation for a shell-type transformer and for an interleaved disk winding and the turn-to-turn insulation against surge overvoltage. Conversely, for the section-to-section structure of a continuous disk winding, the influence of the ac voltage must be taken into account. These results are key findings for rationalizing transformers while maintaining their insulation reliability.
The degradation characteristic of solid insulators is considered to be one of the key factors to perform a risk assessment of highly aged gas insulated switchgear (GIS). The present study experimentally obtained the insulation characteristics of epoxy insulator mainly with an internal microscopic delamination-shaped defect. In the experiment, a short-time breakdown electric field (EF) was initially obtained and comparison was made with the breakdown EF for the void and crack defects to evaluate the breakdown risk with respect to the defect shape. These defect models (delamination, void, and crack) were designed and produced so that an apparent partial (PD) might be 1 pC in an actual 550 kV-GIS spacer. Consequently, it emerged that the area where the EF was concentrated influenced the breakdown. Of the three types of defect, the degree of EF concentration was the lowest for delamination defects, meaning the lowest breakdown risk for the same. Subsequently, the V-t characteristics were obtained through an EF acceleration test. The V-t characteristics obtained were extrapolated to estimate the breakdown risk in 30 to 50 years operation, which is considered the design life of GIS. As a result, it was determined that the potential for eventual breakdown was extremely low under the normal operating EF, even if a delamination defect equivalent to an apparent PD of 1 pC was present in an actual 550 kV-GIS spacer. Hence, controlling a defect using 1 pC for actual GIS is considered to ensure reliability under long-term operation.