Eliminating SF₆ from Switchgear

Abstract

This chapter provides a brief overview of SF₆ use in medium- and high-voltage gas-insulated electrical equipment and the outsized environmental impact of SF₆ which has prompted a decades-long search for alternative gases and gas mixtures. Section 2 summarizes the key gas properties required for gas insulators, the early SF₆ gas mixture alternatives that were considered, and the recent successes in the development and testing of gas mixtures based on 3 M’s Novec™ 4710 and Novec™ 5110 gas. Section 3 briefly summarizes the life cycle challenges associated with the use of SF₆, including leak detection and end-of-life disposal. Section 4 describes the hardware development of new electrical equipment when near drop-in SF₆ replacements are not possible or desirable.

1 Introduction

Today, the electric grid in the United States is responsible for distributing over 4 trillion kWh per year of electricity from generators to consumers. It forms an integrated network that has become an indispensable asset to the nation’s economy, infrastructure, and security. The physical infrastructure of this network depends on a combination of specialized equipment including transformers, power converters, power factor correctors, and switchgear. A critical component for the safety and reliability of the electric grid is a man-made gas, sulfur hexafluoride (SF₆). In 1937, General Electric (GE) introduced SF₆ as an insulation gas to the electric industry; since then, SF₆ has become ubiquitous in medium-voltage (MV) and high-voltage (HV) equipment. Among its many key attributes are its intrinsic nontoxic, noncorrosive, and nonflammable nature, in addition to its superior stability over a wide operating window, good thermal conductivity, high dielectric strength, and excellent arc-quenching capabilities. These properties make it particularly amenable as an insulating and arc-quenching gas in electrical equipment [1]. As a result, over 90% of gas-insulated switchgear globally uses SF₆ as the insulating gas [2]. However, SF₆ emissions from the electric transmission and distribution sector pose a significant climate risk as a potent and long-lived greenhouse gas (GHG) source. One ton of SF₆ emitted to the atmosphere has an equivalent 100-year global warming potential (GWP) of 22,800–26,700 tons of carbon dioxide and has an estimated atmospheric lifetime of 3200 years [3]. As a result of its strong radiative forcing and long atmospheric lifetime, SF₆ was designated as one of the six main greenhouse gases in the 1997 Kyoto Protocol.

As countries set increasingly ambitious emission targets in accordance with the Paris Agreement, emissions of all GHGs, particularly from the electric grid, will be scrutinized. Furthermore, regulations being considered in places like California and the European Union (EU) aim to completely phase out SF₆ from electrical equipment, necessarily setting a timeline to develop alternative solutions to SF₆-insulated equipment. Alternative solutions developed today could define the market for decades to come, both in the United States and globally.

Equipment leaks are a major source of SF₆ emissions from the electrical transmission and distribution sector. This fact is particularly true for aging equipment which, due to natural deterioration, is more prone to gas leaks [4]. A study presented at the 2000 International Conference on SF₆ and the Environment suggests that 10% of circuit breakers in the United States leak; of that 10%, 15% were identified as minor leaks and 85% were identified as major leaks or leaks that required operations to schedule repairs [5, 6]. The National Electrical Manufacturers Association (NEMA) estimates leak rates of 0.1% per year, while the International Electrotechnical Commission (IEC) standard 62271-1 (2004) sets the standard for equipment leakage at 0.5% [7]. Across the entire life cycle of the equipment, however, SF₆ emissions may be as high as 15% and potentially underreported by at least a factor of two [8, 9]. In addition to the emissions associated with equipment service life, losses due to poor gas handling practices are to blame. The operation and maintenance of SF₆ gas carts are considered a major source of handling-related losses [4], and some in industry say the eventuality that all created SF₆ will ultimately end up being released into the atmosphere should be considered [10].

Today, significant effort is dedicated toward supplanting fossil fuel-derived electricity generation with wind and solar power, with the concomitant effect of the grid becoming increasingly decentralized. The electrification of transport, heating, and cooling will also require grid expansions [11]. Barring any disruptive technological advances or policy-driven trends, more gas-insulated equipment (GIE) employing SF₆ will be added to the grid, increasing the emission risks and, ironically, potential climate impacts. Between 1994 and 2022, the measured SF₆ in the atmosphere has increased 3.5-fold [12], and as more clean energy is integrated onto the grid, SF₆ emissions from the transmission and distribution (T&D) sector (0.25% of combined emissions from SF₆ and power generation, 0.07% of total GHG emissions from the United States in 2019) will likely continue to rise and constitute a larger proportion of emissions from the electric grid [13] (Fig. 18.1).

Fig. 18.1

Projected SF₆ emissions as a percentage of the emissions from power generation and T&D and in real terms, expressed as MMT CO_2e. Calculations based on emissions from power generation decreasing linearly to zero by 2035 and SF₆ nameplate capacity increasing by 4% per year (consistent with the average historical increase between 1999 and 2013) and SF₆ emission rates of 1.5% consistent with EPA reported values (recent SF₆ emission rates have plateaued since 2015) [14, 15]. Note: SF₆ emission rates may be underreported by a factor of two, as one study suggested, based on atmospheric concentrations [9]

In addition, a large portion of the US grid was built in the 1960s and 1970s, implying that the equipment currently in use is approaching or exceeding its useful life span [11]. The aging infrastructure has two important implications. First, older equipment tends to leak more SF₆ or require more volumes of SF₆ which pose a significant climate risk [4]. Second, within the next few decades, much of this equipment will be replaced and will require large investments. Precluding any market-ready alternatives, this equipment will be replaced with new equipment that still uses SF₆, potentially locking in this potent greenhouse gas in the grid for the next 20–50 years and increasing the risk of future SF₆ emissions. Because of the environmental challenges associated with using SF₆ in the electric grid, a few states, including California and Massachusetts, are updating their legislation to address SF₆ emission reporting and to set new, more stringent, emission limits. Given the age of the equipment, the investment needed to update and expand the grid, and the stricter policy measures, new technologies and/or alternative gases that minimize or eliminate SF₆ and SF₆ emissions from gas-insulated equipment (GIE) will be required.

2 The Search for SF₆ Substitutes

SF₆ is used extensively as the insulating and arc-quenching medium in MV and HV (12–720 kV) electrical power systems due to its high dielectric strength, nontoxicity, nonflammability, chemical inertness, excellent thermal interruption and heat transfer properties, high vapor pressure at low temperatures, and “self-healing” or fast recovery properties when exposed to an electrical arc. It is an excellent and reliable gas for gas-insulated equipment, were it not for its outsized impact on global warming.

The search for SF₆ alternative gases with a lower environmental impact has been an area of focus for decades [16]. The most critical property for alternative insulating gases is a high dielectric strength, implying that the gas molecules are strongly electronegative. Practically speaking, the dielectric strength of the gas is an indication of its ability to reduce the number of free electrons in an electrically stressed dielectric gas. An appropriate substitute, therefore, must be able to scavenge free electrons with a wide range of energies and over a range of temperatures, have favorable electron slowing down properties which reduce additional electron generation from electron impact ionization, and be characterized by a low ionization cross-section/high ionization onset [17].

If the only concern for an SF₆ alternative was the dielectric strength, several alternatives would have been identified decades ago. As can be seen from Table 18.1, which has been reproduced from a 1997 NIST report, there are several gasses with higher dielectric strengths than SF₆.

Table 18.1 Relative DC uniform field breakdown strengths, V_s^R, of some dielectric gases^a

However, the many gases with high dielectric strengths were ruled out because they did not meet other performance requirements [17]. In addition to dielectric strength, gas insulators must have a high vapor pressure to ensure they stay in the gas phase, even at low temperatures. They must have high thermal conductivity, high specific heat, and long-term thermal stability (i.e., no significant degradation when exposed to elevated temperatures over long periods of time) which is essential for managing the significant thermal loading during an arcing event. They must also feature high breakdown voltages under uniform and nonuniform electric fields, be robust to factors like surface roughness or moisture content, and be compatible with the materials of construction used for gas-insulated equipment. For safety reasons, gas insulators must be nonflammable and nonexplosive and must have low toxicity to minimize adverse impacts during gas handling or in the event of a gas leak. The chemical compatibility and toxicity levels of the breakdown products after a gas insulator is exposed to an arc must be considered, as these will impact performance and safety over the lifetime of the gas-insulated equipment. These stringent performance and property requirements, on top of environmental considerations like a low GWP and zero ozone depletion potential (ODP), have challenged the search for a suitable gas insulator alternative to SF₆ for decades.

Early candidates considered as possible SF₆ substitutes included carbon dioxide, nitrogen, and mixtures of carbon dioxide and/or nitrogen with SF₆ [17]. A key driver in selecting mixtures based on SF₆ was to find a “universal” drop-in replacement that could be used across all types of gas-insulated equipment with only minor equipment modifications. For some applications, like gas-insulated transmission lines, comparable performance with modest increases in gas pressures relative to pure SF₆ were observed for 50/50 and 40/60 SF₆-N₂ mixtures with the added benefit of slightly lower GWP and lower cost [18,19,20,21]. Under some conditions like nonuniform fields or in some interrupter applications, mixtures of SF₆-N₂ were even found to confer superior performance over pure SF₆ due to factors like nitrogen’s better insulating properties at high pressures or because nitrogen has complimentary thermal properties at temperatures below 3000 K relative to SF₆ which is better at temperatures higher than 3000 K [22, 23]. However, SF₆ mixtures or pure N₂/CO₂/dry air were found to not be compatible as drop-in replacements for gas-insulated substation circuit breakers and switchgear without a significant thermal derating or redesign [24]. Additional concerns were raised regarding safety issues associated with the rate of pressure rise during an internal failure arc and additional costs associated with gas mixture recycling. It is perhaps because of these and other drawbacks that rollout of these early SF₆ alternatives was limited to MV (12 kV/24 kV) in the case of N₂ and dry air and HV gas circuit breakers up to 72.5 kV for CO₂ [25].

Later research sought to identify alternatives beyond gases like nitrogen and carbon dioxide. One study by Rabie and Franck compared the dielectric strength of SF₆ to 2611 carbonyl compounds as a preselection process to identify potential SF₆ alternatives [26, 27]. They focused their search on groups of hydrofluoro-ketones and hydrofluoro-aldehydes, acyl fluorides, and perfluoro-ketones and perfluoro-aldehydes. Figure 18.2 shows the 2611 molecules grouped into three classes of C_3-, C_4-, and C_5-carbonyl compounds and their predicted dielectric strength (E_r) relative to SF₆ and T_B, the molecules’ boiling point. The most promising candidate molecules, due to their relatively high values of E_r and low values of T_B, are captured in the black box. Fluorinated compounds have been of particular interest, in terms of their favorable dielectric strengths, but most had key drawbacks like high GWPs (5000–12,000 for perfluorocarbons) and/or high toxicity (CF₃I) which excluded them from further consideration. Two of the most promising fluorinated compounds are a fluoronitrile (C₄F₇N or (CF₃)₂-CF-CN) and a fluoroketone (C₅F₁₀O or CF₃C(O)CF(CF₃)₂) which have acceptable GWPs and low toxicities [25, 28]. Both are currently marketed under 3 M’s Novec™ dielectric gases, Novec™ 4710 and Novec™ 5110, as SF₆ gas alternatives for power utility applications.

Fig. 18.2

Predicted dielectric strength E_r vs predicted boiling point T_B for 2611 carboxyl compounds. The complete list of molecules is split into carbonyl compounds containing three (triangles), four (circles), and five carbon atoms (dots) [27]

Original equipment manufacturers (OEMs) are currently testing mixtures based on the Novec™ dielectric fluids in high-voltage electrical equipment. For example, GE developed green gas for grid (g³) using CO₂ and O₂ mixed with Novec™ 4710 for gas-insulated substation (GIS) and gas-insulated line (GIL) applications [29]. Testing to assess the performance of g³ gas in 60 bays of 145 kV gas-insulated substations, over 2000 meters of 420 kV gas-insulated lines, and in 6 AIS 245 kV current transformers has produced promising results. ABB has also developed AirPlus™, which combines dry air with Novec™ 5110, for GIL and GIS applications.

Zhang et al. recently reviewed studies that have modeled and experimentally characterized the arc plasma and decomposition products of the Novec™ gas mixtures [25, 30]. The basic properties of the arc plasma produced from mixtures of C₄F₇N and C₅F₁₀O are critical for understanding arc plasma behavior and evaluating arc-quenching capability. The main takeaway from a comparison of the composition of the arc plasmas calculated assuming local thermodynamic equilibrium conditions (LTE), non-LTE but local chemical equilibrium condition (LCE), and using chemical kinetics models with no equilibrium assumptions reveals a discrepancy at low temperatures. Chemical kinetics are more critical for low-temperature conditions where the impact of energy barriers for each decomposition pathway is expected to dominate; LTE assumptions ignore the influence of energy barriers. Regarding the decomposition products, the recognition that, unlike SF₆, the decomposition by-products of the proposed gas mixture alternatives do not recombine to form the original structure has necessitated studies that characterize the decomposition products under thermal degradation, spark discharges, corona discharges, and arcing. To this end, researchers have conducted experimental studies to characterize the decomposition products directly and developed computational models to understand the chemical kinetics and decomposition mechanisms [29,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46]. Erroneous or incongruous results are still prevalent, however, due to the limitations of the chosen methods. The detection methods most often used in experimental studies that characterize the gas composition are Fourier transform infrared spectroscopy (FTIR) and gas chromatography-mass spectroscopy (GC/MS). FTIR enables in-line gas composition analysis but is limited by a relatively high detection threshold. Conversely, GC/MS is excellent for detecting trace gasses but introduces a significant time delay, so the decomposition products are detected when they are no longer under their formation conditions. In the case of computational methods, the accuracy will necessarily depend on the input assumptions which should be documented carefully and provided with adequate rationale.

A few studies from the last decade focus on the materials compatibility of C₄F₇N and C₅F₁₀O and their decomposition products with metals, metal oxides, common elastomers, and lubricating greases. Early computational results indicate, for example, that C₄F₇N may be more compatible with aluminum than copper, while experimental results from Pohlink et al. found no serious compatibility issues among the most common metals used in HV equipment over a period of several months at 120 °C [32]. The most prevalent compatibility issues for both C₄F₇N and C₅F₁₀O were associated with polymers, particularly ethylene propylene diene monomer (EPDM) which is used as a gasket/sealing material. C₄F₇N can also have an effect on the degradation of other common materials found in circuit breakers such as nitrile O-rings and synthetic zeolite molecular sieves used to remove decomposition by-products [47]. In addition to material incompatibilities, the different decomposition pathways of these gases are important in order to design molecular sieves to capture by-products and to preserve the integrity of the dielectric and cooling properties of the gas in the chamber. In summary, research regarding arc plasma behavior, decomposition products, and materials compatibility are still at the early stages for both C₄F₇N and C₅F₁₀O but are critical to understand before moving entirely away from SF₆.

Although potential gas alternatives like those previously mentioned are promising, several challenges exist in accomplishing full replacement of SF₆. Though some of the proposed alternatives can reduce the GWP by over 98% relative to SF₆, a 98% reduction still leaves many alternatives with a GWP >300 [48]. What’s more, alternative gas and gas mixtures are not necessarily drop-in solutions for SF₆ under all conditions; some gas mixtures pose a potential performance risk in colder climate zones due to their higher boiling points, while some gas mixtures require modest increases in pressure to achieve equivalent performance to SF₆ [49]. Alternative gases and gas mixtures may require new or modified equipment specially designed to use these gases, new leak detection and monitoring equipment, and new practices to address end-of-life disposal or recycling. Like the work described earlier related to arc plasma characterization and decomposition products, toxicological studies are also at the beginning stages and reflect conflicting results, though some initial results are promising [28]. Finally, there is still a degree of uncertainty around the potential of future regulations which may impact adoption. Even if these points are addressed, market adoption will require overcoming challenges associated with workforce training and spacing constraints, in addition to the need to assess the connection compatibility with existing equipment/infrastructure.

A final comment regarding vacuum dry air or vacuum solid dielectric circuit breakers as another route to SF₆-free electrical equipment will be made, which is the route that OEMs like Siemens and Mitsubishi are currently pursuing [50]. A vacuum is used as the arc-quenching medium and insulator between the contacts as part of a vacuum interrupter (VI) and either dry air or a solid dielectric as the insulator around the VI and in the bushings, eliminating the need for SF₆ or related gases entirely. Several manufacturers already have vacuum technologies for 38 and 72.5 kV, while technologies at 145 kV are expected to become available within the next year. Although these vacuum or near-vacuum technologies have a higher dielectric strength compared to SF₆ circuit breakers, OEMs consider scaling to higher voltages (i.e., 245 kV) a difficult technical challenge.

3 SF₆ Life Cycle Considerations

There are other considerations in the life cycle of the equipment beyond developing alternatives for SF₆-insulated GIE. One of the most effective strategies for mitigating SF₆ emissions in existing equipment is to detect leaks early and to fix these quickly. The International Electrotechnical Commission sets the maximum allowable leakage rate of SF₆ at 0.5% per year [51]. Most new equipment achieves leakage rates far below this limit, but as the equipment (which has a useful lifetime of several decades) ages, the leakage rates may exceed 0.5% per year. To achieve this stringent requirement, it is necessary to continuously monitor for leaks; IEEE Guide for the Selection of Monitoring for Circuit Breakers also recommends monitoring SF₆ density which can fluctuate in response to thermal fluctuations. Several technologies are currently available on the market including portable point source nondispersive infrared (NDIR) detectors, NDIR room sensors which detect ppb and ppm SF₆ concentration levels in enclosed GIE substations, and pressure gauges interfaced with alarm systems that monitor changes in pressure in GIE [52,53,54,55,56,57]. While these technologies are mature and widely available, further improvements in continuous, sensitive, early-warning detection systems are merited. Developing sensors that combine lower cost, higher sensitivity, and continuous monitoring for all equipment settings could lead to advances in early detection technologies that reduce SF₆ emissions in the short and medium term as the electric grid transitions to non-SF₆ alternatives as well as lead to more accurate accounting of emission rates. Of particular interest are cost-effective detection technologies with remote notification systems or systems sensitive enough to detect slow leaks in small-capacity gas-insulated equipment which pose unique challenges that demand low detection limits and high accuracy under a variety of environmental conditions. Today’s continuous sensors primarily rely on pressure gauges that are not sensitive enough to detect slow leaks or leaks in small-capacity equipment. As a result, even with the technology that is on the market today, several small-capacity GIE owners responded to the proposed policy changes in the California legislation by noting that achieving 1% emission rates on a consistent basis is challenging [58]. One percent emission rate still corresponds to the equivalent of at least 1 megaton of CO₂ released per year in the United States; these emissions could potentially increase fivefold by the year 2035 if SF₆ nameplate capacity continues to increase at an average of 4% per year and the emission rate stay the same.¹ NDIR-based sensors offer higher sensitivity than pressure gauges, but they currently have several drawbacks. NDIR point source sensors are more sensitive and can detect ppm or ppb SF₆ concentrations but are not continuous, are cost-prohibitive, and require a technician to manually check all equipment for leaks. NDIR room sensors can continuously monitor dilute SF₆ concentrations but are still costly and are only applicable for equipment housed in an enclosure. Density and pressure sensors carefully located at optimal locations on the equipment and potentially supplemented with CFD models may enable NDIR-level accuracy at a similar cost and robustness of pressure gauges, thereby leading to more accurate leak rate monitoring [51]. Looking longer-term, cost-effective sensors for proposed gas and gas mixture alternatives will be required because the alternative gases on the market today have a GWP potential that is lower than SF₆ but still significant when compared with other greenhouse gases [59, 60]. Of note, alternatives based on gas mixtures may demand higher accuracy or bespoke sensors to monitor the conditions of the gas-insulated equipment where risks of changes in mixture composition will have major implications on the equipment operation. The gas composition must be always known to ensure the dielectric gas is within a safe operational window.

In addition to leak detection, there are other items to consider when using SF₆ alternatives. The characteristics of SF₆ replacements dictate changes to the mechanical structure of the breaker to accommodate variations in pressure and dielectric withstand. Because of this, several studies have been published regarding the life cycle impact of the manufacture, use, and disposal of the new GIE hardware. For example, hardware changes required for Novec 4710™ mixtures for a 145-kV, 40-kA breaker were found to have minimal impact relative to the equivalent SF₆-based switchgear when accounting for the emissions and waste generation from production of material, manufacturing, distribution, construction, operation and maintenance, and end-of-life destruction [61]. In contrast, while vacuum-based breakers will not have any deleterious emissions during use, compared to a Novec 4710™ -based circuit breaker, researchers in [62] conclude that the need for larger equipment (a direct result from the vacuum’s lower dielectric strength), the GWP, among other negative factors, is higher for vacuum-based breakers at 72.5 kV.

Finally, novel end-of-life processes for SF₆ must be considered as SF₆-insulated equipment is retired and replaced with non-SF₆ alternatives. While SF₆ is available, it poses a significant environmental risk. This represents an opportunity to develop novel SF₆ destruction pathways which are less energy-intensive and do not produce toxic and/or corrosive products [63,64,65]. Alternative pathways for efficient and complete destruction of SF₆ use strong reducing agents [66,67,68], low-valent transition metal complexes [69,70,71,72,73,74,75,76], strong nucleophiles [77], monovalent aluminum reagents [78], and catalytically enhanced dielectric barrier discharge processes [79]. A more recent approach has been to use SF₆ as a safe fluorinating agent in organic synthesis, opening new valorization routes for stranded SF₆ supplies when it is no longer needed in electrical equipment [70, 80,81,82,83,84,85,86]. For alternative gas and gas mixtures, end-of-life considerations are in their nascent stage.

4 SF₆-Free Circuit Breaker Hardware

Alternative dielectric mediums in HVAC switchgear and near drop-in replacements for SF₆ require changes to the system hardware that can involve significant research, development, design, and testing. While allowing for more significant hardware redesigns offers more options for dielectric mediums, trade-offs in size and life cycle impacts must still be taken into account. Ensuring safety aspects are considered and that the dielectric withstand capability of the circuit breakers can still pass power frequency, lightning impulse, and chopped wave dielectric tests according to IEEE C37-04 or other standards is also critically important when any changes to dielectric medium or hardware are made. Dynamic tests such as those in IEEE C37-04 serve to verify fundamental performance parameters critical for any new design such as how to extinguish the arc quickly and safely and ensure a speedy and full dielectric recovery. For reference, an example of a high-voltage SF₆-based dead tank AC breaker and a corresponding bushing and tank cutaway is shown in Fig. 18.3.

Fig. 18.3

Example of three-phase HVAC dead tank breaker (left) and cutaway drawing of example chamber, bushings, and contacts for a single phase (right) [87]

One of the initially more attractive options for SF₆-free switchgear is vacuum-based circuit breakers due to their ease of maintenance and zero GWP [88]. While vacuum interrupters (VI) perform well at lower voltages, at voltages above 72.5 kV, as a result of practical design limitations, there can be diminishing returns for increased gap lengths [89, 90]. As a result, multibreak VI designs are often proposed as a solution though this is still an active area of research and considerations for voltage balancing between the sets of electrodes, such as grading capacitors, must be considered [91]. Other active areas of research for increasing a single VI’s interruption capabilities particularly beyond 72.5 kV are similar to those for other mediums such as new contact materials and geometries, improved speed and control of operating mechanisms, new solid insulator material, and improved designs for electric field control [92,93,94,95]. For dead tank breakers, an insulating medium must also be used in the tank that contains the VI as well as the bushings. While technical air is a common choice, VIs with technical air insulators could have weights and footprints 132% higher and 160% larger, respectively, than an SF₆ breaker of similar performance [62, 96]. Recent work has also looked at dielectric mediums primarily for the tank and bushings such as CF₃I-CO₂ to improve the footprint of VI-based breakers with technical air while maintaining lower GWP [97].

In [24], researchers demonstrated how non-vacuum dielectric mediums combined with circuit breaker redesigns can achieve design goals while maintaining a low GWP. Using a 72.5-kV breaker redesigned to use a 0.8-MPa CO₂, the researchers reduced the GWP of the circuit breaker significantly while still meeting the key fault and capacitive switching requirements of IEC 62271-100. The redesign included accounting for the lower heat capacity of 0.8 MPa CO₂ which causes a more rapid pressure rise and fall than SF₆ at similar starting temperatures and pressures. As a result, the CO₂ puffer pressure may not be adequately maintained for long-duration arcs, necessitating a larger puffer cylinder volume. The authors conclude that higher pressures of CO₂ as well as blends with O₂, which has better arc-quenching capability than CO₂, could possibly overcome some of these issues as well as further improve the GWP. Another option for higher-pressure CO₂, as demonstrated in [98], is to use it in its supercritical state. Supercritical carbon dioxide (scCO₂) at around 8 MPa has a dielectric strength of around three times that of SF₆ in addition to high heat transfer capability and low viscosity. scCO₂-based breakers have already been proposed for a compact, low-arcing, medium-voltage breaker [99]; however, scCO₂ breakers have their own challenges as they require the chamber, bushings, and all external seals to manage the 8-MPa pressure and a well-controlled temperature to maintain the CO₂ at supercritical levels. Additionally, during an AC fault, the great amounts of heat energy released from the arc over the course of several milliseconds or partial cycles could cause more overpressure in the switchgear enclosure and bushings, resulting in rupture. While this phenomenon has been studied at lower starting pressures for SF₆, CO₂, and N₂ [100], further analysis on the higher operating pressure effects over time and in operation during arcing faults in scCO₂ are needed.

In addition to looking to CO₂ at higher pressures, in [101], researchers demonstrated that N₂, with a combination of higher pressures and/or increased sparking distance, can achieve a dielectric insulation equivalent to SF₆ at 0.7 MPa. In [102], researchers further evaluated gaseous nitrogen at pressures up 2.6 MPa and with various mixtures of SF₆ and showed that, for example, with a working pressure of 1 MPa and only 5% SF₆ and 95% N₂, a similar insulation performance could be achieved to SF₆ at its typical working pressure of 0.7 MPa. As has been discussed, for application to circuit breakers, a suitable dielectric medium is dependent on many other variables besides dielectric strength, such as gas flow characteristics and various thermal properties, that should be considered to definitively evaluate the equivalency of the insulating medium as a replacement for SF₆ in switchgear.

Other work has also looked at nongaseous forms of nitrogen. If future substations have available liquid nitrogen (LN₂), then concepts proposed in [103] such as a 126-kV LN₂-cooled superconducting fault current limiter (SCFCL), which reduces the fault current energy, could enable a smaller, cryogenically cooled vacuum interrupter in series which would have the added benefit of an “on” resistance of less than a third of a traditional AC breaker. Alternatively, since LN₂ can have a higher breakdown voltage than SF₆ at its standard operating pressure, LN₂ has been proposed as the dielectric medium in the interrupter itself where it can both serve to greatly reduce steady-state losses due to reduced resistivity of the cryo-cooled contacts and extinguish the arc. In [104], the arc-extinguishing capability of LN₂ was tested up to nearly 3-kA peak with 50-mm-diameter contacts, opening at 1 m/s to reach a 25-mm gap. The LN₂ was shown to help successfully extinguish the arc at the zero crossing without reignition for root mean square currents of less than 2085 A. The two concepts of SCFCL and dielectric medium are combined in [105] where researchers have also proposed and tested at lower voltages with LN₂ as the dielectric medium in the interrupting mechanism combined with a SCFCL. The results indicate the system is fast-acting and results in lower current and therefore lower-energy faults, though many other items, including the hardware redesign, the logistics of cooling the LN₂, and how to deal with any GN₂ that is generated, must be considered. For those that envision liquid nitrogen and cryo-cooled elements as part of the solution for the future grid, demonstrating the value of fault current limiters and liquid nitrogen in switchgear could play a part but may require rethinking of traditional fault coordination systems.

Another area of research as it relates to SF₆ alternatives is in the materials, speed, and movement of the mechanisms that separate the contacts, determine the movement of the arc, and control gas flow. In [106], researchers demonstrate in a low-voltage circuit breaker the use of permanent magnets to help guide the arc toward a splitter stack that elongates and cools it. However, physical prototypes to prove the efficacy of these approaches on high-voltage equipment are costly, and theoretically evaluating the range of possible solutions also proves challenging. The simulation of such systems are complex due the multi-physics phenomena related to flow, pressure, contaminants, electrical operating points, mechanical structures, the shape and movement of contacts, and dielectric medium management mechanisms. Much research has been done to make the simulation problem more accurate and manageable such as defining the ignition of discharges and their lifetime transitions from corona to streamer or to Townsend avalanches by evaluating the discharge’s stability mathematically as an eigenvalue problem [107] or defining an iterative calculation of streamer propagation with a dependence on pressure into an algorithm that helps simplify finite element simulation. A similar effort to simplify simulations in [108] uses computational fluid dynamics and detailed enthalpy flow simulations to determine the discharge coefficients of various geometries of gas valves during the circuit breaker operation that can then be plugged into enthalpy flow models of the full circuit breaker system. The need for high accuracy simulations and the multidisciplinary nature of the problem is apparent even in research related to existing high-voltage switchgear. For example, some 550-kV HVAC SF₆-based breakers must now manage higher short-circuit ratings than ever, such as 80 kA versus a previous maximum of 63 kA, as a result of new grid conditions. This increase in peak fault current is compelling researchers to explore electromagnetic- and mechanical-based solutions that can reduce contact travel time and guide the arc so that it extinguishes more effectively so as to not cause significant changes to equipment size and operation as a result of the increased fault current rating [109, 110].

5 Concluding Remarks

The risks of using SF₆ as the insulating medium on the grid due to its outsized GWP have long been recognized by researchers and regulators. The search for suitable alternatives to SF₆-insulated GIE has spanned decades and only recently has it achieved gains in the form of unique fluorinated gas mixtures, vacuum dielectrics pushing higher voltages, and cryo-insulators. Though early results are promising, there remain many open questions related to the life cycle impacts of new GIE or vacuum-based equipment, the performance of the dielectric medium when exposed to an arc, the safety and monitoring aspects that will be required, and the end-of-life handling practices. With much of the current grid infrastructure approaching end of useful life, the introduction of new renewable energy generation to the grid, and upcoming policy mandates, the need for alternatives to SF₆-insulated GIE has never been greater. Addressing these questions, analyzing the findings from ongoing pilot studies, and anticipating future needs for SF₆ alternative technologies are sure to be a top priority for stakeholders in this space and will be critical for achieving a zero-emission grid.