This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright
Author's personal copy Cold Regions Science and Technology 65 (2011) 23–28
Contents lists available at ScienceDirect
Cold Regions Science and Technology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c o l d r e g i o n s
Variable-resistance conductors (VRC) for power-line de-icing Victor F. Petrenko, Charles R. Sullivan ⁎, Valeri Kozlyuk Thayer School of Engineering, Dartmouth College, NH 03755, USA
a r t i c l e
i n f o
Article history: Received 6 October 2009 Accepted 17 June 2010 Keywords: Power transmission lines Thermal de-icing Anti-icing
a b s t r a c t Ice storms can result in accumulation of ice on structures, including overhead power transmission and distribution lines and associated poles and towers; this ice may reach thicknesses of many tens of millimeters. Icing can cause catastrophic damage which disrupts power transmission and is expensive to repair. Normal operation of a power transmission or distribution line entails Joule heating of the conductor as current flows through it. Lines are normally designed to have a constant, low resistance, so as to avoid excessive power losses and avoid excessive operating temperatures. Because the normal heating is low (by design), it is of limited value in preventing or recovering from an icing event. This paper describes power conductors that can switch their electrical resistance from a very low value, to transmit electric energy, to a much higher value, for de-icing. The switching in between two conductor resistances does not disturb the main conductor function, which is to provide a customer with uninterrupted electric power. A variable-resistance conductor (VRC) is built of N strands (or groups of strands) insulated from each other, where N is any odd integer greater than one. For instance, N = 3, 5 or 7, etc. In normal energy-transmission operations all the conductor strands (or strand groups) are connected in parallel, whereas in de-icing mode they all are connected in series. Switching from parallel to series connection increases the line resistance by a large factor of N2, making the resistance sufficiently high for heating the line above the ice melting point. One important advantage of the method is that it uses low-voltage and, thus, low-cost switches. The design of a VRC de-icing system is described, including considerations for switches, conductors, control and deployment strategy. We also describe safety devices to return the line to normal operation if the electronics get damaged. Laboratory and full-scale prototypes have both successfully demonstrated the capability of VRC de-icing. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Ice storms can result in an accumulation of ice on structures, including overhead power transmission and distribution lines and associated poles and towers; this ice may reach thicknesses of many tens of millimeters. Although any one line typically encounters such conditions only a few times per year, icing can cause catastrophic damage which disrupts power transmission and is expensive to repair. The usual mechanism of ice damage is through the weight of the ice imposing added mechanical stress on cables and towers: a cylinder of solid ice, 50 mm in radius and 1 km long, weighs 7.2 tonne; ice this thick can add more than 8 t of weight per km of a single cable, as well as increasing wind-induced stress on the lines. Such an ice cylinder would increase the cable tension by a number that depends on the length of a cable span and cable sag, but typically exceeds the icecylinder weight. Accumulated ice regularly causes power transmission lines and poles to break, and towers to collapse. Dramatic ⁎ Corresponding author. E-mail addresses: Victor.F.Petrenko@Dartmouth.edu (V.F. Petrenko), charles.r.sullivan@dartmouth.edu (C.R. Sullivan). 0165-232X/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.coldregions.2010.06.003
examples include the 2008 storm in the Guangdong province of China (Liu et al., 2008) and the 1998 storm in the northeast of North America (Gyakum and Roebber, 2001), but smaller ice storms cause damage every year. Ice accretion can also contribute to conductor galloping in high winds, sometimes leading to short circuits between adjacent phases and thus to outages. In addition to disrupting power transmission damage to power lines can cause serious risk to people and property on the surface. Methods of de-icing in use today have many shortcomings, as discussed in the surveys by Laforte et al. (1998), Ryerson (2008), and Farzaneh et al. (2008). The primary categories of methods in use are mechanical and electrothermal. Mechanical methods in use generally require access to the line. One exception is the use of a pulse of current from a deliberate short circuit to create electromagnetic forces that may knock ice off a line (Landry et al., 2000). Electrically, this may create significant disruptions in power quality and may even threaten grid stability (Landry et al., 2000). Other mechanical methods are reviewed by Farzaneh et al. (2008); in addition to other difficulties, most mechanical methods require adding additional stress to a system that may already be near the breaking point, and thus can precipitate the failure they are intended to prevent.
Author's personal copy 24
V.F. Petrenko et al. / Cold Regions Science and Technology 65 (2011) 23–28
Normal operation of a power transmission or distribution line entails Joule heating of the conductor as current flows through it. Lines are normally designed to have a constant, low resistance, so as to avoid excessive power losses and avoid excessive operating temperatures. As wire reaches high temperatures, due to electrical selfheating and high ambient temperatures, it tends to lengthen and weaken. This lengthening can cause the lines to sag between poles or towers, possibly causing hazard to persons or property on the surface. Because the normal heating is low (by design), it is of limited value in preventing or recovering from an icing event. Various approaches have been proposed or tested to enable increased heating for preventing ice formation or removing ice. One option is to simply change power system operations to increase the current flowing through a particular line. Although Huneault et al. (2005a,b) and Merrill and Feltes (2006) show that this can be effective in some circumstances, it requires a network with adequate configuration options, and the heating power is limited. Another option is to disconnect the line from service, short-circuit it, and apply ac or dc power to it. Dc has the advantage that the current is not limited by the reactance of the line. Applying dc requires expensive rectification equipment, although it may be possible to configure this equipment to be useful for other purposes when it is not used for de-icing (Horwill et al., 2006). The short circuit method is widely used (Farzaneh et al., 2008), even though removing a line from service is problematic, particularly during an emergency. To provide significant, controlled heating without removing a line from service, it is possible to increase the level of current in a network through the use of phase-shifting transformers and capacitors (Cloutier et al., 2007). Even with this proposed enhancement, the heating available is limited, and there may be negative effects on grid stability. It is also possible to keep a line in service, and add additional power dissipation by superimposing a higher frequency current in the form of a sinusoid McCurdy et al. (2001) or pulse (Peter et al., 2008). However, these approaches require additional expensive equipment, as well as raising issues with electromagnetic interference or dielectric breakdown. Another approach to keeping a line in service but increasing power dissipation is to make some modification to the line itself. For example, if there are parallel conductors, some of them can be switched out of the circuit, as suggested by Couture (2004) and Pierce (1954). This increases the resistance, and the heat dissipated, by a factor up to the number of parallel conductors (if each has equal resistance). In this paper, we introduce a similar method that can provide a larger increase of resistance and can simultaneously heat all parallel conductors in a bundle.
2. The variable-resistance conductor method 2.1. Introduction This paper describes power conductors that can switch their electrical resistance from a very low value, to transmit electric energy, to a much higher value, for de-icing (Petrenko and Sullivan, 2008). The switching in between two conductor resistances does not disturb the main conductor function, which is to provide a customer with uninterrupted electric power. A variable-resistance conductor (VRC) is built of N strands (or groups of strands) insulated from each other, where N is any odd integer greater than one. For instance, N = 3, 5 or 7, etc. In normal energy-transmission operations all the conductor strands (or strand groups) are connected in parallel, whereas in de-icing mode they all are connected in series. Three- and five-strand versions are shown in Figs. 1, 2 and 3. Switching from parallel to series connection increases the line resistance by a large factor of N2, making the resistance sufficiently high for heating the line above the ice melting point. One important advantage of the method is that it uses low-voltage and,
thus, low-cost switches. We also describe safety devices to return the line to normal operation if the electronics are damaged. 2.2. Power requirements The heating power required for anti-icing is primarily the power required to maintain the line at an elevated temperature (e.g., by 10 °C) by balancing convective losses. For de-icing, additional power is required to raise the temperature of the line and the ice, and to melt the ice. A small amount of additional power can then eventually, slowly melt the ice. However, as discussed in more detail by Petrenko et al. (2003), Petrenko and Sullivan (2003) and Peter et al. (2008), and in a companion article in this issue (Petrenko et al., 2010), higher power heating can reduce the energy requirement for de-icing. This is because of reduced energy loss to convection through a reduced time at an elevated temperature, and because, at sufficiently high power, the interface between the ice and the line can be melted before the whole body of the ice is brought up to 0 °C. In the worst case, the ice may encircle the line, and may thus need to be melted through to be released. If sufficient heating power is available, this can easily be accomplished when necessary. To determine the power required for anti-icing operation we estimate heat losses from a line held above 0 °C in an ambient below 0 °C. These may include wind-driven convection, radiation and natural convection. For worst-case design, we consider the case of a strong wind, in which case wind-driven convection dominates over radiation and natural convection, and we can consider only the convective loss, which may be estimated by (Admirat, 2008): 0:61
Q = 14:2ðvdÞ
ΔT
ð1Þ
where Q is heat loss (forced convection only) in watts per meter length of a cable, v is wind velocity in m/s, d is diameter of the cable in m, and ΔT is the difference between the cable and the ambient temperatures. For example, Eq. (1) predicts that holding a 10 mm diameter cable 10 °C above ambient (e.g., at 1 °C in a −9 °C ambient), would require as much as 35 W per meter of cable at high wind speeds (10 m/s), as shown in Fig. 4. Higher power is needed for de-icing, both because the additional surface area of the ice increases convective loss, and because energy is needed to melt ice. More detailed models of power requirements are discussed in Merrill and Feltes (2006) and Personne and Gayet (1988). To compare the power requirement to the power that can be produced in a typical VRC conductor, we now consider an example cable comprising 5 strands of 2.3 mm diameter aluminum with 0.7 mm insulation, such that each strand has a diameter of 3.7 mm including insulation. The overall diameter is 10 mm, and the total cross-sectional area of aluminum available for conducting current is 21 mm2. A 1.6-mm diameter steel wire could optionally be included in the center for mechanical support without affecting the overall dimensions. Although such a cable could have an ampacity over 135 A (CME Wire and Cable, Inc., 2006), we conservatively assume a nominal current of 100 A rms. In the normal conductor configuration, such a cable would have a resistance of 1.35 mΩ per meter, and would dissipate 13.5 W/m at full 100 A rms current. If we switch it into deicing mode with the 5 strands in series, the resistance and power dissipation are increased by a factor of 25, for a dissipation of 338 W/m at full current, or 84 W/m at half current. Comparing this to the anti-
Fig. 1. Variable-resistance conductor system using three parallel strands of conductor, represented by resistors. With the switches all off the resistance from the left terminal to the right terminal is 32 = 9 times the resistance when the switches are on and all the conductors are connected in parallel.
Author's personal copy V.F. Petrenko et al. / Cold Regions Science and Technology 65 (2011) 23–28
25
Fig. 2. Variable-resistance conductor system using five parallel strands of conductor, represented by resistors and a series switch arrangement. With the switches all off the resistance from the left terminal to the right terminal is 52 = 25 times the resistance when the switches are on and all the conductors are connected in parallel.
icing requirements calculated above, we see that even at half of nominal current, the power available is more than twice that required for maintaining a temperature above freezing in anti-icing operation, with the balance (50 W/m) available for the additional convection load present with an ice coating, and for melting the ice. 2.3. Switch design A power transmission line is typically operated at an ac potential hundreds of kilovolts above ground. Equipment designed to support these voltages is large and expensive. However, the voltage across a switch in a VRC de-icing system is only the extra voltage drop across the segment it switches. This voltage may be hundreds to a few thousands of volts, within the range of standard low-cost semiconductors or electromechanical contactors. The voltages with respect to ground could still pose difficulties, except that the whole switching apparatus, including its housing, can be maintained at the potential of the line, and can be mounted suspended from the towers, similarly to equipment proposed in Divan and Johal (2007) for reactive power compensation. When more than three strands are used, the switches may be configured in two different ways—series, as shown in Fig. 2, or parallel as shown in Fig. 3. The series configuration has a lower maximum voltage across any switch, whereas the parallel configuration has a lower maximum current through any switch. The total volt–ampere (VA) requirement for the switches is equal in the two configurations, as shown in Appendix A, so the choice of switch configuration depends more on the practical limitations of the particular switches used, and how those compare to the voltage and current requirements, which depend on the distance between switching units. For example, for a long distance between switching units thyristors are a good choice for the switches, and a parallel configuration (Fig. 3) is likely to be preferred, because the switch assembly will have lower total on-state voltage drop, and the higher voltage requirement is not a problem. On the other hand for a shorter distance, MOSFET switches are likely to be preferable, and a series configuration can reduce the voltage requirement for the MOSFETs, and thus allow the use of MOSFETs with lower on-state resistance. Particularly with thyristors, a hybrid switch combining semiconductors with a parallel electromechanical relay may be desirable to
Fig. 4. Calculated convective heat loss from a 10 mm diameter cable maintained at 10 °C above ambient as a function of wind speed. At high wind speeds, this loss dominates over other power losses and is approximately equal to the power required to maintain the cable at this temperature.
reduce on-state power losses and heating (Atmadji and Sloot, 1998; Polman et al., 2001). An electromechanical relay may also be used without a parallel semiconductor device. For the example parameters discussed above, the voltage drop at full current in one strand is 0.54 V rms per meter at full current. Thus, the peak voltage across a switch in the series switch configuration (Fig. 2) is 1.53 V per meter of distance between switch boxes at normal maximum current. This is compatible with MOSFETs for a single span, or with thyristors for up to several kilometers. 2.4. Cables The VRC de-icer requires an odd number of conductor strands or groups, insulated from each other. Fortunately, the voltage difference between the strands is small compared to the line voltage, and simple, thin, polymer coating as is used for low-voltage insulation or for covering on some distribution lines is adequate. The maximum voltage between strands is the product of the number of strands N and the voltage drop per strand. For the example discussed above, for 7 strands, and a 100 m distance between switch boxes, the maximum voltage is 378 V rms. The addition of the polymer may affect the ampacity of the conductor. Heat dissipated in the conductor during normal operation must be conducted through the polymer before transferring to the ambient by convection and radiation. The conduction path adds thermal resistance, whereas the slight increase in diameter increases convective area. The emissivity of the coating also affects heat loss by radiation and solar heat gain. In practice, the net result tends to be a slight increase of ampacity as reflected in manufacturers' ampacity ratings (CME Wire and Cable, Inc., 2006). It is also possible to use bare bundled conductors separated by insulated spacers. These are commonly used in transmission lines in groups of 4; for VRC an odd number (e.g., 5) can be used instead. 2.5. Deployment strategies
Fig. 3. Variable-resistance conductor system using five parallel strands of conductor, represented by resistors and a parallel switch arrangement. Compared to the arrangement in Fig. 2, this is functionally equivalent but changes the required capabilities of the switches by reducing the current through some switches, but increasing the voltage across them.
For several reasons, it is desirable to have short distances between switch boxes suspended on a line. This allows finer granularity in the choice of which sections are heated. It also allows smaller, lighter switch boxes, which are easier to suspend from a line. Finally, the disturbance to the power network when a segment is turned on or off will be very small.
Author's personal copy 26
V.F. Petrenko et al. / Cold Regions Science and Technology 65 (2011) 23–28
The disturbance can be further minimized by coordinating switching times such that when one segment is turned off, the next segment is simultaneously turned on, such that overall dissipation and voltage drop in the line are constant. When a section of the line is uniformly coated with ice, high-power heating can be applied without overheating the line, as the ice maintains the temperature near 0 °C until the ice is cleared, at which point the heating can be turned off. If the segment has sections that are covered with ice and bare sections, the bare sections will be heated above 0 °C in the time it takes to clear the iced sections. Rated operating temperatures of polyethylene insulation are 75 °C or 90 °C (CME Wire and Cable, Inc., 2006), leaving adequate headroom in ambient temperatures below 0 °C. However, it is desirable for this reason, and for anti-icing operation, to be able to adjust the heating power. Although it would be possible to use switch configurations other than all on and all off to get different heating powers, the heating would then be non-uniform, stronger in some strands than in others. Another better approach to modulating heating power is to cycle the heating on and off and vary the duty cycle to vary the power. With synchronized switching of many segments, such pulse-width modulation can be used to achieve many different average power levels while maintaining constant total power dissipation and voltage drop. 2.6. Practical implementation Although full documentation of a practical implementation is not of interest here, we briefly address some practical strategies for implementation, including powering the switch boxes, control and communications, and failure modes. A switch box equipped with line- and air-temperature sensors, ice sensors, and a microcontroller could autonomously decide when to activate de-icing. However, wireless communications systems can be easily implemented, for example using cellular networks such as GSM (Global System for Mobile Communications), and offer many advantages. Central control allows coordination of the activation of different segments, and allows activation decisions to be made with knowledge of weather forecasts. It also allows coordination with power network operation. Bidirectional communication allows data from sensors to be returned to a control center. This data is not only valuable for monitoring and addressing an ice storm, but also for monitoring line temperatures during other situations, such as a hot summer day. Switch boxes may derive power from the line current through current transformers or small photovoltaic panels with battery backup. Given the inverse correlation between icing events and solar power availability, current transformers are a preferred approach, although photovoltaics would be viable with adequate batteries. The purpose of the VRC de-icing system is to make the power network more reliable as well as to avoid expensive damage. If the VRC de-icing system were to fail and disrupt the power network, reliability would be hurt rather than helped. Thus, in addition to engineering the switch boxes for high reliability, it is important to ensure that a failure restores the power line to normal operation, rather than defaulting to the high-resistance state, which could lead to overheating and failure of the cable. Scenarios that must be considered include failure of the insulation between strands or failure of power or control components in the switch box resulting from any cause including a lightning strike. Multiple levels of fail-safe protection can be used to ensure that this is the case. The first level of protection is an electronic cable temperature sensor and logic that shuts of the system in the case of cable overheating. With an electromechanical relay as the switch, or a component of the switch, the next level of protection can be a normally-closed relay that reverts to the low-resistance state when it has no control signal. This ensures that in the case of a failure of the electronics or the power supply, the system reverts to the lowresistance state. For designs that use only semiconductor switches, a
failure of the drive electronics would, with most types of semiconductors, lead to the switch turning off and the VRC stuck in the highresistance state. Although redundant drive systems could be used to mitigate this hazard, an emerging option is the use of SiC JFET switches, which typically are on in the absence of a control signal. SiC JFETs offer very low resistance when on and very high voltage capability (Cooper and Agarwal, 2002). In typical applications, their normally-on characteristic is a problem, but in this application it is an advantage. If a third level of protection is desired, a purely mechanical backup mechanism can be triggered by a mechanical fusible link using metal or polymer that melts at a moderate temperature (e.g. 120 °C) to close a set of contacts and revert to low-resistance mode. Testing under simulated lightning strikes should be used to verify that the failure mode is normal operation even under extreme circumstances. Failure of insulation between strands, for example due to mechanical impact or lightning, will have no effect on operation in the normal mode, as the potential difference between strands is zero. However, in some scenarios, an insulation failure could lead to nonuniform heating in de-icing mode, potentially leading to damaging temperatures at some points on the line before the sensed temperature exceeds 0 °C. Although thermal damage is unlikely when the ambient temperature is low enough to lead to icing, a protective measure for such a fault would be to sense the voltage across pairs of series strands and compare this to expected values based on line current, length, and temperature, and revert to normal operation if a significant discrepancy is detected. 2.7. Applicability The VRC method is widely applicable. Because the switches see only the voltage drop along the length of the line, not the voltage from the line to ground, their rating is independent of the transmission voltage. Thus, the method is applicable to high-voltage transmission (hundreds of kilovolts) as well as medium-voltage distribution (tens of kilovolts). For higher voltage, the main additional consideration is corona, and thus possible degradation of insulation material. Where this is a problem, spacing conductor strands apart with insulated spacers may be preferable to using polymer insulation. Although the incremental cost of adding polymer coatings to conductors is low, the cost of replacing conductors in an existing installation is high. Thus, the VRC method is least expensive either in new installations or on existing lines when replacement of conductors is being done for other reasons. However, in situations where icing is a severe or recurrent problem, or in situations where extremely high reliability is desired, the cost of new conductors may be justified. 3. Test results Although the VRC de-icing system is based on well established physical principals and proven technology, laboratory and field tests have been conducted to verify its performance. 3.1. Laboratory tests A power cable with a steel reinforced core was stretched between two insulator posts set 3.5 m apart, as shown in Fig. 5. The cable comprises seven insulated aluminum strands, each 2.6 mm diameter (AWG 10), and a steel reinforcement cable in the center. The aluminum strands were connected through switch boxes at the ends of the span. The cables for this and other tests are shown in detail in Fig. 6. The switches were controlled by a radio remote control. The line was designed to carry current of 70 A rms in normal operation, with a resistance of 2 mΩ for the length of the test span. The resistance is increased to 110 mΩ in the de-icing mode. At 70 A rms, the power dissipation is 539 W, or 154 W/m of cable length, in de-icing mode, whereas it is under 5 W/m in normal operation (low-resistance) mode.
Author's personal copy V.F. Petrenko et al. / Cold Regions Science and Technology 65 (2011) 23–28
27
Fig. 5. Laboratory test setup.
The line was covered by mixture of ice and snow with a thickness of approximately 10 mm to mimic atmospheric icing conditions. Dense snow was first saturated with water and the mixture set at a temperature of 0 °C. The mixture then was placed on a plastic film and wrapped around a cold wire, forming a wet snow cylinder inside a plastic-film sheath. After the mixture was frozen the plastic film was removed. In a − 5 °C cold room, with the line and ice starting at ambient temperature, it took 3 to 5 min to clean the line completely of ice and snow in various tests. The temperature profiles from two cold-room tests, with and without 2 m/s air flow, are plotted in Fig. 7.
Fig. 7. Laboratory test results of cable temperature vs. time. The cable temperature rises quickly and then levels off when automatic temperature control engages, and finally drops off when de-icing is turned off.
lines, using cables similar to those already in use and low-cost switching systems that do not need to withstand the full line voltage. Laboratory and field tests have confirmed operation as expected.
3.2. Field tests A full-scale VRC de-icing system was tested on a 10.5 kV power distribution line in Orenburg, Russia in early 2009, as shown in Fig. 8. A sample of the cable used is shown at the left in Fig. 6, and used the same stranding as the cables in the laboratory tests. The switch boxes were powered by the line current, and controlled by a radio transmitter from the ground. Although atmospheric icing was not encountered during the test, ice was frozen onto the line, and testing was conducted with the line in active use in the power network, carrying 60 to 70 A rms. The ambient temperature was near 0 °C and wind speeds were 4 to 6 m/s. The system worked as expected, heating the line and removing the ice. Fig. 9 shows a sequence of ice beginning to melt and falling off the line. 4. Conclusion The variable-resistance conductor de-icing method offers simple, reliable de-icing for power transmission and distribution
Fig. 8. Test installation on one span of a power line. The spheres on top of the first and last towers contain the switches for the protected cable between them.
Fig. 6. Cables used in laboratory tests. Each cable has a steel core and seven 2.6 mm (AWG 10) insulated aluminum conductors, twisted around the steel core. The left cable with, with an outer jacket, was used in the field test; the center unjacketed cable was used in lab tests.
Fig. 9. Ice being removed from an active power line. The first frame shows ice beginning to melt; the second ice about to fall, the last shows ice falling.
Author's personal copy 28
V.F. Petrenko et al. / Cold Regions Science and Technology 65 (2011) 23–28
Acknowledgement The authors thank Ice Engineering, LLC for multi-year financial support of this research. Appendix A. Switch ratings Consider a section of cable made up of insulated strands, with a total load current Each strand has a resistance RS, such that section in normal operation (low-resistance mode, the current in each strand is Is = I0 = N
an odd number N of flowing through it I0. the resistance of the mode) is RS/N. In this ð2Þ
In de-icing mode (high resistance), the total resistance is N · RS. The full current I0 flows through each strand, and thus the voltage drop across a single strand is Vs = I0 ·RS
ð3Þ
For the series switch configuration (Fig. 2), the voltage across each switch when it is off (high-resistance mode) is 2 · VS. The currents when the switches are on are not all equal. At each end of the segment, there are (N − 1)/2 switches. These carry currents {2 · IS, 4 · IS, … (N − 1)IS}. We can evaluate the total switch requirement by summing the volt–ampere (VA) requirements for each switch. Because the voltage requirements are identical for each switch, we first sum the currents and then multiply by the voltage. The sum of the currents at each end is 2(N − 1)IS, for a total of 4(N − 1)IS at both ends. Multiplying by the voltage, we obtain the total switch VA requirement for the series configuration: VAseries = 8ðN–1ÞVS IS
ð4Þ
For the parallel switch configuration (Fig. 3), the current through each switch when it is on (normal operation) is 2 · IS. The voltages when the switches are off are, {2 · VS, 4 · VS, … (N − 1)VS}. To find the total VA of the requirements for each switch, we can separately sum the voltages and then multiply by current, because the current requirements are identical for each switch. The sum of the voltages at each end is 2(N − 1) VS, for a total of 4(N − 1) VS at both ends. Multiplying by the voltage, we obtain the total switch VA requirement for the parallel configuration: VAparallel = 8ðN–1ÞVS IS
ð5Þ
We see that in terms of the sum of VA ratings of the switches, the requirements are identical. References Admirat, P., 2008. Wet snow accretion on overhead lines. In: Farzaneh, M. (Ed.), Atmospheric Icing of Power Networks. Springer, Netherlands, pp. 119–169.
Atmadji, A.M.S., Sloot, J.G.J., 1998. Hybrid switching: a review of current literature. Energy Management and Power Delivery '98. . Cloutier, R., Bergeron, A., Brochu, J., 2007. On-load network de-icer specification for a large transmission network. IEEE Transactions on Powere Delivery 22 (3), 1947–1955. Couture, P., 2004. Switching modules for the extraction/injection of power (without ground or phase reference) from a bundled HV line. Institute of Electrical and Electronics Engineers Transactions on Power Delivery 19 (3), 1259–1266. CME Wire and Cable, Inc, 2006. Power and Control Catalog. Suwanee, Georgia. Cooper Jr., J.A., Agarwal, A., 2002. SiC power-switching devices—the second electronics revolution? Proceedings of the IEEE 90 (6), 956–968. Divan, D., Johal, H., 2007. Distributed FACTS—a new concept for realizing grid power flow control. IEEE Transactions on Power Electronics 22 (6), 2253–2260. Farzaneh, M., Volat, C., Leblond, A., 2008. Anti-icing and de-icing techniques for overhead lines. In: Farzaneh, M. (Ed.), Atmospheric Icing of Power Networks. Springer, Netherlands, pp. 229–268. Gyakum, J.R., Roebber, P.R., 2001. The 1998 ice storm—analysis of a planetary-scale event. Monthly Weather Review 129 (12), 2983–2997. Horwill, C., Davidson, C.C., Granger, M., Dery, A., 2006. An application of HVDC to the deicing of transmission lines. IEEE PES Transmission and Distribution Conference and Exhibition, pp. 529–534. Huneault, M., Langheit, C., Caron, J., 2005a. Combined models for glaze ice accretion and de-icing of current-carrying electrical conductors. IEEE Transactions on Power Delivery 20 (2), 1611–1616. Huneault, M., Langheit, C., St.-Arnaud, R., Benny, J., Audet, J., Richard, J.C., 2005b. A dynamic programming methodology to develop de-icing strategies during ice storms by channeling load currents in transmission networks. IEEE Transactions On Power Delivery 20 (2), 1604–1610. Laforte, J.L., Allaire, M.A., LaFlamme, J., 1998. State-of-the-art on power line de-icing. Atmospheric Research 46 (1–2), 143–158. Landry, M., Beauchemin, R., Venne, A., 2000. De-icing EHV overhead transmission lines using electromagnetic forces generated by moderate short-circuit currents. Proceedings of the IEEE 9th International Conference on Transmission and Distribution Construction, Operation and Live-Line Maintenance, pp. 94–100. Liu, P.Y., He, H.M., Pan, C.P., 2008. Investigation of 2008 frozen disaster and research on de-icing in Guangdong power grid. China International Conference on Electricity Distribution, pp. 1–5. McCurdy, J.D., Sullivan, C.R., Petrenko, V.F., 2001. Using dielectric losses to de-ice power transmission lines with 100 kHz high-voltage excitation. Conference Record of the Thirty-Sixth IEEE Industry Applications Society Annual Meeting, pp. 2515–2519. Merrill, H.M., Feltes, J.W., 2006. Transmission Icing: A Physical Risk with a Physical Hedge IEEE Power Engineering Society General Meeting. Montreal, Canada. Personne, P., Gayet, J.F., 1988. Ice accretion of wires and anti-icing induced by Joule effect. Journal of Applied Meteorology 27 (2), 101–114. Peter, Z., Volat, C., Farzaneh, M., Kiss, L.I., 2008. Numerical investigations of a new thermal de-icing method for overhead conductors based on high current impulses. IET Generation, Transmission and Distribution 2 (5), 666–675. Petrenko, V. F. and Sullivan, C. R., 2003. Pulse Electrothermal Deicer for Power Lines. US Patent Applications No. 60/545,038, 2004 and 60/497,442, 2003. Petrenko, V.F. and Sullivan, C. R., 2008. System and Method for Deicing of Power Line Cables. US Patent Application No. 12/193, 650, 2008. Petrenko, V.F., Higa, M., Starostin, M., Deresh, L., 2003. Pulse electrothermal de-icing. Proceedings of the Thirteenth International Offshore and Polar Engineering Conference, Honolulu, Hawaii, pp. 435–438. Petrenko, V.F., Sullivan, C.R., Kozlyuk, V., 2010. Pulse Electro-Thermal De-Icer (PETD). Cold Regions Science and Technology. Pierce, W.T., 1954. Deicing Apparatus for Electric Cables, U.S. Patent No. 2,797,344. Polman, H., Ferreira, J.A., Kaanders, M., Evenblij, B.H., Van Gelder, P., 2001. Design of a bi-directional 600 V/6 kA ZVS hybrid DC switch using IGBTs. Conference Record of the Thirty-Sixth IEEE Industry Applications Society Annual Meeting, pp. 1052–1059. Ryerson, C.C., 2008. Assessment of superstructure ice protection as applied to offshore oil operations safety: problems, hazards, needs, and potential transfer technologies. Cold Regions Research and Engineering Laboratory Report ERDC/CRREL TR-08-14.