- Email: [email protected]
Concepts for using trapped-flux HTS in motors and generators
Physica C 484 (2013) 104–107
Contents lists available at SciVerse ScienceDirect
Physica C journal homepage: www.elsevier.com/locate/physc
Concepts for using trapped-flux HTS in motors and generators John R. Hull ⇑, Michael Strasik Boeing Research & Technology, PO Box 3707, Seattle, WA 98124, USA
a r t i c l e
i n f o
Article history: Available online 25 February 2012 Keywords: Bulk HTS Trapped-flux Generators
a b s t r a c t We examine the expected performance of a brushless motor/generator that uses trapped-flux (TF) bulk high-temperature superconductors (HTSs) to provide magnetomotive force, where the stator windings are used to create the TF. A key feature is the use of dysprosium (Dy) for the stator and rotor cores. We also examine methods to energize TF in HTS for generators used in pulsed-power applications. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction Bulk HTSs, consisting of single-grain Y–Ba–Cu–O (YBCO) or its rare-earth analogs, exhibit high engineering critical current density Jc and have a very strong diamagnetic response. Bulk YBCO can trap large magnetic fields; maximum fields exceeding 10 T have been demonstrated at a number of laboratories [1], and 17 T at 29 K [2] is the present record. To a first approximation, these TF YBCO bulks can be considered as analogs to permanent magnets (PMs) and used in devices where PMs are normally used, e.g., to replace the PMs on a brushless motor/generator (M/G) with TF-HTSs [3]. The larger magnetic flux that these HTSs could generate would enable the generator to have higher specific power than conventional generators, even after the cryogenic requirements were considered. Magnetization of the HTSs in a practical device with trapped fields higher than the magnetization of a PM has been a daunting problem. One could imagine charging the HTSs in a separate magnet and assembling the generator cold. In addition to requiring a cryogenic environment with controlled atmosphere for assembly, the disadvantage of this method is that if the generator warms up above the operating temperature, it must be taken apart to recharge the HTSs. To use TF HTS in a practical device, it is most desirable to magnetize the HTS in situ using the field coils and regular power electronics of the device to perform the magnetization. This is the method of the hysteresis HTS motor. Another in situ method is to pulse charge the HTSs with short-duration, high current coils [4]. This method invokes a pulsed-power supply which must be counted with the cryogenic cooling in the calculation of the specific power.
2. Trapped-flux HTS motor/generator concept with dysprosium core In this section, we present a design concept for a TF HTS M/G that uses Dy for core material in both the rotor and stator components. At temperatures below 80 K, Dy is ferromagnetic and has the highest saturation magnetic field of any known material. The use of Dy in the core is a good match to the use of TF HTSs. First, at 80 K and below, the temperature at which bulk HTS is expected to operate in this application, Dy has a much higher magnetic saturation than conventional iron. This improves the performance in two ways. First, for a given magnetizing current in the stator coil, it produces more flux at the HTS, so the trapped flux in the HTS is higher. Second, once the HTSs are magnetized, the Dy concentrates more magnetic flux in the stator teeth. Dy also has some useful features in terms of safety, e.g., if the stator coils short out. At room temperature, Dy is no longer ferromagnetic but is strongly paramagnetic. Thus, by allowing the stator core (or significant portions of the stator core) to warm up, the flux passing through the coils can be greatly reduced. The design concept is illustrated by the example 4-pole configuration shown in Fig. 1. The stator core, rotor core, and rotor cap are composed of Dy, formed as thin laminations or a powdered composite to minimize eddy current loss. The stator core consists of a cylindrical element and teeth, with the teeth passing through the center of the stator coils. An air gap separates the rotor and the stator. When the HTS is magnetized, the rotor spins and changing flux in the teeth of the rotor core induces a voltage in the stator coils, thus enabling the device to function as a generator. 2.1. Charging the HTSs
⇑ Corresponding author. Tel.: +1 206 544 0368; fax: +1 206 544 0409. E-mail address: [email protected] (J.R. Hull). 0921-4534/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.physc.2012.02.020
To magnetize the HTSs, the rotor is held stationary, so that the rotor teeth line up with the stator teeth. The stator coil is energized
J.R. Hull, M. Strasik / Physica C 484 (2013) 104–107
105
Fig. 1. Schematic of basic concept with magnetization due to coils with current density of 120 A/mm2 while the 4-mm-thick HTS is warm (above critical temperature).
with direct current to produce the magnetization shown in Fig. 1, while the HTSs are still warm, i.e., their temperature is above the critical temperature. Once the HTSs cool below their critical temperature, they will magnetize in the direction of the applied flux. I.e., whatever flux is present when they obtain their operating temperature will be trapped there. If the critical current density of the HTS is high enough, the HTS will be uniformly magnetized in the center with a small perimeter region in which the magnetization falls to zero at the HTS edge faces (the four not perpendicular to the magnetization direction). Alternatively, the HTSs can first be cooled below their critical temperature. Then the stator coils can be pulsed one or more times to magnetize the HTSs. However, the field-cooling method produces the largest trapped flux. In Fig. 1, we assume that the device is cold enough that the Dy is ferromagnetic, but the HTSs are not yet superconducting. From a magnetics point of view, the HTSs are an air gap at this temperature. The coils are energized so as to provide the magnetic flux shown in Fig. 1. The HTSs are assumed to be parallelepipeds with a 40 mm by 40 mm base and a thickness that is a design parameter, in this case 4 mm. There is a Dy pole cap between the HTS element and the actual air gap. The pole cap is flat on the surface adjoining the HTS and has a curved surface at the air gap. The air gap between rotor and stator is 1 mm. We allow the coils to have an engineering current density as great as 500 A/mm2, which is now obtainable with HTS wire and can be tolerated for short pulses in non superconducting wire. In this example, the magnetic field distribution along the centreline of the HTS is 2.8 T. In all the examples studied, the field is rather uniform across the gap, except with a small fall off at the edges. For the purposes of this study, we assume that the HTS is exposed to a uniform field. Further, we assume that when the HTS cools, the HTS magnetizes at a constant value slightly less than the peak. For example, in the above example, we would take the integral of the perpendicular flux density across the surface and divide by the area to assign a constant magnetization of 2.7 T to the HTS. Assuming that the magnetization is constant simplifies the magnetic analysis and is assumed acceptable for this scoping
study. The demagnetization curve for the HTS assumes it has a relative permeability of unity. For a BH curve for dysprosium, we used the 70 K data from Behrendt et al. [5]. Next, we turn off the coil current and use the HTS to drive magnetic flux in the circuit. An example is shown in Fig. 2. Having established the general geometry, method of operation, and method of calculation, we increase the coil current density to a more state-of-the-art value and vary the thickness of the HTS. We then calculate the flux at some point in the Dy where the coil is located. The shapes of the Dy stator and rotor cores are not optimized. 2.2. Losses Dy presently has hysteresis loss that is one to two orders of magnitude greater than conventional magnetic core materials. This hysteresis arises as a result of magnetostriction, and to some degree the 4-f electron orbital geometry. The cooling burden of the hysteresis is high enough to possibly restrict its use to the rotor of the device and use room-temperature iron for the stator core. The scientific literature suggests that methods may exist to dramatically decrease the hysteretic loss by changing the magnetostrictive behavior. For example, incorporation of beta-Mg into Dy and Tb changed the crystal structure from hexagonal to facecentered cubic and had a large effect on the magnetostrictive properties without significantly affecting the magnetic saturation [6]. To be used practically, the dysprosium may need to be ‘‘tweaked’’ technologically, so that hysteresis losses are lowered. Based on Swift and Mahur [7], the loss at the fluxes shown in Fig. 2 would be almost 4 J/kg/cycle at 77 K. In the design concept studied, the stator dysprosium has a mass of 6.2 kg. It would see a loss of 48 J per revolution. At 200 Hz, this would be a loss of 10 kW, which would be too large to cool to cryogenic temperatures, even in a 1 MW generator. Fortunately, the Behrendt et al. data [5] suggest that the hysteresis loss can be made at least 20 times lower, and 500 W could easily be cooled with existing cryocoolers. Presumably, device optimization could lower this value even more.
106
J.R. Hull, M. Strasik / Physica C 484 (2013) 104–107
Fig. 2. Magnetic flux due to magnetized HTS with profile shown in Fig. 1.
2.3. Device performance We estimate the power and efficiency of the example generator of this section at a rotational frequency of 200 Hz. The work shown here is a point design and is not optimized in any sense. The estimated mass of the system is 3.56 kg, which gives a torque density of 5.4 Nm/kg. The mass could probably be reduced to 3.0 kg in an optimization study and possibly reduced to as much as 2.5 kg. Including the estimated power for cryogenic cooling of the motor, the efficiency of the motor is 99.2%. 3. Pulsed-power generation One of the largest obstacles to in situ pulse charging of HTSs in generators is that the pulse currents are usually considerably larger than the currents needed for normal operation, and the charging process requires an external energy storage device that can be larger than the generator. An application where these conditions do not hold is in pulsed-power generators. Here, starting with a
very modest magnetization of the HTSs on one generator, a pair of generators can be used to fully magnetize the HTSs by pulse charging [8], shown schematically in Fig. 3. Each pulsed-power generator includes a set of bulk HTSs mounted on its rotor and a corresponding set of stator coils. The stator coils are sized to provide a large pulsed current to a load. The HTSs of the first generator are magnetized to a small degree by either a small dc current through its stator coils or by a pulse from a small external device. The rotor of the second generator is locked in place relative to its stator coils. The first generator is brought up to speed and a power pulse from this first generator is delivered to the stator coils of the second generator. This pulse magnetizes the HTS in the second generator to a level higher than that of the HTSs in the first generator. Next the rotor of the first generator is locked and the second generator brought up to speed. A power pulse is delivered from the second generator to the first generator, and the magnetization of the HTSs in the first generator increases. This bootstrapping procedure is repeated until all HTSs in both generators are fully charged.
Fig. 3. Schematic of charging method of HTSs in a pair of pulsed-power generators.
J.R. Hull, M. Strasik / Physica C 484 (2013) 104–107
References [1] J.R. Hull, M. Murakami, Applications of bulk high-temperature superconductors, Proc. IEEE 92 (2010) 1705–1718. [2] M. Tomita, M. Murakami, High-temperature superconductor bulk magnets that can trap magnetic fields of over 17 Tesla at 29 K, Nature 421 (2003) 517–520. [3] J.R. Hull, M. Strasik, Concepts for using trapped-flux bulk high-temperature superconductor in motors and generators, Supercond. Sci. Technol. 21 (2011) 1453–1459. [4] H. Matsuzaki, Y. Kimura, I. Ohtani, M. Izumi, T. Ida, Y. Akita, H. Sugimoto, M. Miki, M. Kitano, An axial gap-type HTS bulk synchronous motor excited by
[5] [6] [7] [8]
107
pulsed-field magnetization with vortex-type armature copper windings, IEEE Trans. Appl. Supercond. 15 (2005) 2222–2225. D.R. Behrendt, S. Legvold, F.H. Spedding, Magnetic properties of single crystals, Phys. Rev. 109 (1958) 1544–1547. J.W. Herchenroeder, K.A. Gschneider, Magnetic behavior in Mg-stabilized bcc bGd and b-Dy, Phys. Rev. B 39 (1989) 11850–11861. W. Swift, M. Mahur, Cryogenic magnetic properties of secondary recrystallized thin sheet dysprosium, IEEE Trans. Magn. 10 (1974) 308–313. J.R. Hull, M. Strasik, Methods for Charging and Using Pulsed-Power Sources, US Patent 7889,035, February 15, 2011.
Contents lists available at SciVerse ScienceDirect
Physica C journal homepage: www.elsevier.com/locate/physc
Concepts for using trapped-flux HTS in motors and generators John R. Hull ⇑, Michael Strasik Boeing Research & Technology, PO Box 3707, Seattle, WA 98124, USA
a r t i c l e
i n f o
Article history: Available online 25 February 2012 Keywords: Bulk HTS Trapped-flux Generators
a b s t r a c t We examine the expected performance of a brushless motor/generator that uses trapped-flux (TF) bulk high-temperature superconductors (HTSs) to provide magnetomotive force, where the stator windings are used to create the TF. A key feature is the use of dysprosium (Dy) for the stator and rotor cores. We also examine methods to energize TF in HTS for generators used in pulsed-power applications. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction Bulk HTSs, consisting of single-grain Y–Ba–Cu–O (YBCO) or its rare-earth analogs, exhibit high engineering critical current density Jc and have a very strong diamagnetic response. Bulk YBCO can trap large magnetic fields; maximum fields exceeding 10 T have been demonstrated at a number of laboratories [1], and 17 T at 29 K [2] is the present record. To a first approximation, these TF YBCO bulks can be considered as analogs to permanent magnets (PMs) and used in devices where PMs are normally used, e.g., to replace the PMs on a brushless motor/generator (M/G) with TF-HTSs [3]. The larger magnetic flux that these HTSs could generate would enable the generator to have higher specific power than conventional generators, even after the cryogenic requirements were considered. Magnetization of the HTSs in a practical device with trapped fields higher than the magnetization of a PM has been a daunting problem. One could imagine charging the HTSs in a separate magnet and assembling the generator cold. In addition to requiring a cryogenic environment with controlled atmosphere for assembly, the disadvantage of this method is that if the generator warms up above the operating temperature, it must be taken apart to recharge the HTSs. To use TF HTS in a practical device, it is most desirable to magnetize the HTS in situ using the field coils and regular power electronics of the device to perform the magnetization. This is the method of the hysteresis HTS motor. Another in situ method is to pulse charge the HTSs with short-duration, high current coils [4]. This method invokes a pulsed-power supply which must be counted with the cryogenic cooling in the calculation of the specific power.
2. Trapped-flux HTS motor/generator concept with dysprosium core In this section, we present a design concept for a TF HTS M/G that uses Dy for core material in both the rotor and stator components. At temperatures below 80 K, Dy is ferromagnetic and has the highest saturation magnetic field of any known material. The use of Dy in the core is a good match to the use of TF HTSs. First, at 80 K and below, the temperature at which bulk HTS is expected to operate in this application, Dy has a much higher magnetic saturation than conventional iron. This improves the performance in two ways. First, for a given magnetizing current in the stator coil, it produces more flux at the HTS, so the trapped flux in the HTS is higher. Second, once the HTSs are magnetized, the Dy concentrates more magnetic flux in the stator teeth. Dy also has some useful features in terms of safety, e.g., if the stator coils short out. At room temperature, Dy is no longer ferromagnetic but is strongly paramagnetic. Thus, by allowing the stator core (or significant portions of the stator core) to warm up, the flux passing through the coils can be greatly reduced. The design concept is illustrated by the example 4-pole configuration shown in Fig. 1. The stator core, rotor core, and rotor cap are composed of Dy, formed as thin laminations or a powdered composite to minimize eddy current loss. The stator core consists of a cylindrical element and teeth, with the teeth passing through the center of the stator coils. An air gap separates the rotor and the stator. When the HTS is magnetized, the rotor spins and changing flux in the teeth of the rotor core induces a voltage in the stator coils, thus enabling the device to function as a generator. 2.1. Charging the HTSs
⇑ Corresponding author. Tel.: +1 206 544 0368; fax: +1 206 544 0409. E-mail address: [email protected] (J.R. Hull). 0921-4534/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.physc.2012.02.020
To magnetize the HTSs, the rotor is held stationary, so that the rotor teeth line up with the stator teeth. The stator coil is energized
J.R. Hull, M. Strasik / Physica C 484 (2013) 104–107
105
Fig. 1. Schematic of basic concept with magnetization due to coils with current density of 120 A/mm2 while the 4-mm-thick HTS is warm (above critical temperature).
with direct current to produce the magnetization shown in Fig. 1, while the HTSs are still warm, i.e., their temperature is above the critical temperature. Once the HTSs cool below their critical temperature, they will magnetize in the direction of the applied flux. I.e., whatever flux is present when they obtain their operating temperature will be trapped there. If the critical current density of the HTS is high enough, the HTS will be uniformly magnetized in the center with a small perimeter region in which the magnetization falls to zero at the HTS edge faces (the four not perpendicular to the magnetization direction). Alternatively, the HTSs can first be cooled below their critical temperature. Then the stator coils can be pulsed one or more times to magnetize the HTSs. However, the field-cooling method produces the largest trapped flux. In Fig. 1, we assume that the device is cold enough that the Dy is ferromagnetic, but the HTSs are not yet superconducting. From a magnetics point of view, the HTSs are an air gap at this temperature. The coils are energized so as to provide the magnetic flux shown in Fig. 1. The HTSs are assumed to be parallelepipeds with a 40 mm by 40 mm base and a thickness that is a design parameter, in this case 4 mm. There is a Dy pole cap between the HTS element and the actual air gap. The pole cap is flat on the surface adjoining the HTS and has a curved surface at the air gap. The air gap between rotor and stator is 1 mm. We allow the coils to have an engineering current density as great as 500 A/mm2, which is now obtainable with HTS wire and can be tolerated for short pulses in non superconducting wire. In this example, the magnetic field distribution along the centreline of the HTS is 2.8 T. In all the examples studied, the field is rather uniform across the gap, except with a small fall off at the edges. For the purposes of this study, we assume that the HTS is exposed to a uniform field. Further, we assume that when the HTS cools, the HTS magnetizes at a constant value slightly less than the peak. For example, in the above example, we would take the integral of the perpendicular flux density across the surface and divide by the area to assign a constant magnetization of 2.7 T to the HTS. Assuming that the magnetization is constant simplifies the magnetic analysis and is assumed acceptable for this scoping
study. The demagnetization curve for the HTS assumes it has a relative permeability of unity. For a BH curve for dysprosium, we used the 70 K data from Behrendt et al. [5]. Next, we turn off the coil current and use the HTS to drive magnetic flux in the circuit. An example is shown in Fig. 2. Having established the general geometry, method of operation, and method of calculation, we increase the coil current density to a more state-of-the-art value and vary the thickness of the HTS. We then calculate the flux at some point in the Dy where the coil is located. The shapes of the Dy stator and rotor cores are not optimized. 2.2. Losses Dy presently has hysteresis loss that is one to two orders of magnitude greater than conventional magnetic core materials. This hysteresis arises as a result of magnetostriction, and to some degree the 4-f electron orbital geometry. The cooling burden of the hysteresis is high enough to possibly restrict its use to the rotor of the device and use room-temperature iron for the stator core. The scientific literature suggests that methods may exist to dramatically decrease the hysteretic loss by changing the magnetostrictive behavior. For example, incorporation of beta-Mg into Dy and Tb changed the crystal structure from hexagonal to facecentered cubic and had a large effect on the magnetostrictive properties without significantly affecting the magnetic saturation [6]. To be used practically, the dysprosium may need to be ‘‘tweaked’’ technologically, so that hysteresis losses are lowered. Based on Swift and Mahur [7], the loss at the fluxes shown in Fig. 2 would be almost 4 J/kg/cycle at 77 K. In the design concept studied, the stator dysprosium has a mass of 6.2 kg. It would see a loss of 48 J per revolution. At 200 Hz, this would be a loss of 10 kW, which would be too large to cool to cryogenic temperatures, even in a 1 MW generator. Fortunately, the Behrendt et al. data [5] suggest that the hysteresis loss can be made at least 20 times lower, and 500 W could easily be cooled with existing cryocoolers. Presumably, device optimization could lower this value even more.
106
J.R. Hull, M. Strasik / Physica C 484 (2013) 104–107
Fig. 2. Magnetic flux due to magnetized HTS with profile shown in Fig. 1.
2.3. Device performance We estimate the power and efficiency of the example generator of this section at a rotational frequency of 200 Hz. The work shown here is a point design and is not optimized in any sense. The estimated mass of the system is 3.56 kg, which gives a torque density of 5.4 Nm/kg. The mass could probably be reduced to 3.0 kg in an optimization study and possibly reduced to as much as 2.5 kg. Including the estimated power for cryogenic cooling of the motor, the efficiency of the motor is 99.2%. 3. Pulsed-power generation One of the largest obstacles to in situ pulse charging of HTSs in generators is that the pulse currents are usually considerably larger than the currents needed for normal operation, and the charging process requires an external energy storage device that can be larger than the generator. An application where these conditions do not hold is in pulsed-power generators. Here, starting with a
very modest magnetization of the HTSs on one generator, a pair of generators can be used to fully magnetize the HTSs by pulse charging [8], shown schematically in Fig. 3. Each pulsed-power generator includes a set of bulk HTSs mounted on its rotor and a corresponding set of stator coils. The stator coils are sized to provide a large pulsed current to a load. The HTSs of the first generator are magnetized to a small degree by either a small dc current through its stator coils or by a pulse from a small external device. The rotor of the second generator is locked in place relative to its stator coils. The first generator is brought up to speed and a power pulse from this first generator is delivered to the stator coils of the second generator. This pulse magnetizes the HTS in the second generator to a level higher than that of the HTSs in the first generator. Next the rotor of the first generator is locked and the second generator brought up to speed. A power pulse is delivered from the second generator to the first generator, and the magnetization of the HTSs in the first generator increases. This bootstrapping procedure is repeated until all HTSs in both generators are fully charged.
Fig. 3. Schematic of charging method of HTSs in a pair of pulsed-power generators.
J.R. Hull, M. Strasik / Physica C 484 (2013) 104–107
References [1] J.R. Hull, M. Murakami, Applications of bulk high-temperature superconductors, Proc. IEEE 92 (2010) 1705–1718. [2] M. Tomita, M. Murakami, High-temperature superconductor bulk magnets that can trap magnetic fields of over 17 Tesla at 29 K, Nature 421 (2003) 517–520. [3] J.R. Hull, M. Strasik, Concepts for using trapped-flux bulk high-temperature superconductor in motors and generators, Supercond. Sci. Technol. 21 (2011) 1453–1459. [4] H. Matsuzaki, Y. Kimura, I. Ohtani, M. Izumi, T. Ida, Y. Akita, H. Sugimoto, M. Miki, M. Kitano, An axial gap-type HTS bulk synchronous motor excited by
[5] [6] [7] [8]
107
pulsed-field magnetization with vortex-type armature copper windings, IEEE Trans. Appl. Supercond. 15 (2005) 2222–2225. D.R. Behrendt, S. Legvold, F.H. Spedding, Magnetic properties of single crystals, Phys. Rev. 109 (1958) 1544–1547. J.W. Herchenroeder, K.A. Gschneider, Magnetic behavior in Mg-stabilized bcc bGd and b-Dy, Phys. Rev. B 39 (1989) 11850–11861. W. Swift, M. Mahur, Cryogenic magnetic properties of secondary recrystallized thin sheet dysprosium, IEEE Trans. Magn. 10 (1974) 308–313. J.R. Hull, M. Strasik, Methods for Charging and Using Pulsed-Power Sources, US Patent 7889,035, February 15, 2011.