Charles R. Poole1, Tanvir Baig1, Robert J. Deissler1, David Doll2, Michael Tomsic2, and Michael A. Martens1
1Department of Physics, Case Western Reserve University, Cleveland, OH, United States, 2Hyper Tech Research Inc., Columbus, OH, United States
Synopsis
To reduce the usage of liquid helium, conduction cooled MRI
magnets using high temperature magnesium diboride (MgB2) have been
considered. Because the thermal normal zone propagation velocity (NZPV) of MgB2
is much slower compared to NbTi wire, active quench protection is needed. The
temperature rise in the magnet design was modeled using Douglas-Gunn method to
solve the governing heat equation for operating temperatures ranging from 10 to
18 K. It was shown that the temperature rise is slower for higher operating temperatures,
and thus better for quench protection.Introduction
Conventional MRI magnets, consisting of low
temperature superconducting (LTS) niobium titanium (NbTi) wire, require around 2000
liters of liquid helium (LHe) to operate. With the rising demand for LHe [1],
alternative magnets designs requiring less LHe are worth considering. One
option is a conduction cooled magnet using magnesium diboride (MgB
2)
wire and operating at temperatures above 4.2 K [2]. The feasibility of such
magnets have been studied [3], and designs exist for whole body magnets with
fields greater than 1.5 T using second generation MgB
2 wire [2]. An
essential part of any superconducting magnet design is the protection of the
magnet during a quench where local hot-spots and large temperature rises can
destroy the magnet [4]. Compared to NbTi magnets, the normal zone propagation
velocity (NZPV) in MgB
2 is much slower, which presents challenges in
the design of a quench protection system [4,5]. Hence, high temperature
superconducting (HTS) magnets must rely on active rather than passive quench
systems [4]. The effectiveness of the protection depends, in part, on how
quickly the quench can be detected and the temperature at the local hot-spot at
the time of detection [4]. In this study, the initial propagation of a quench
due to a hot-spot for a 1.5 T MgB
2 magnet design is studied with
operating temperatures in the range 10 to 18 K.
Methods
A quench is simulated for a 1.5 T MRI magnet
design (Figure 1) with first generation MgB
2 wire. The governing
heat equation is solved in MATLAB (MathWorks) using the implicit Douglas-Gunn
method [6,7]. The wire, based on one from Hyper Tech Research Inc., consists of
MgB
2 filaments surrounded by a matrix of niobium and copper and a
sheath of Monel [8]. The quench is also simulated with Glidcop Al-60 as the
sheath material, which increases the thermal conductivity and decreases the
electrical resistivity of the wire [9]. The wire has a cross sectional area of 1.47
mm
2 and an operating current of 256 A. In this model, the quench is
initiated by a small disturbance heater (2 cm x 0.18 cm), which deposits energy
for 0.5 seconds. The quench is initiated in one of two locations: 1) the center
of the outer surface of Coil 1 with a transition temperature of 27.2 K, and 2)
in Coil 4 at the location of the maximum magnetic field (2.84 T) with a transition
temperature of 20.3 K. In the first set of simulations, the wire sheath is Monel,
and in the second set, Glidcop. For all these simulations, the maximum
temperature, resistive voltage, and NZPV were recorded.
Results
Figure 2 shows that the azimuthal NZPV increases
as the operating temperature increases for both types of wire and both
disturbance heater locations. The maximum temperature versus resistive voltage
for Coil 1 using the Monel sheathed wire is depicted in Figure 3 for various
operating temperatures. Finally, the maximum temperature inside the coil at the
time the resistive voltage reaches a 200 mV threshold is shown in Figure 4 for a
range of operating temperatures for both types of wire. The use of Glidcop has
the largest impact when the quench starts at a location with both a lower
magnetic field and operating temperature.
Conclusion
The rate of temperature rise and resistive
voltage increase is a consideration in the design of a quench protection
system. The maximum temperature should be as small as possible when the
resistive voltage reaches a detectable threshold [10]. The maximum temperature
at a 200 mV threshold decreases when both the operating temperature increases
and when the quench disturbance occurs at a location with a higher magnetic
field. Thus, a quench can be detected at a lower hot-spot temperature when the
magnet operates at a higher temperature. However, as mentioned by Ye, a
trade-off exists between the stability of the magnet and the protection of the
magnet [11]. For instance, when operating at 18 K, the minimum quench energy is
reduced to 150 mJ, (but is still an order of magnitude larger than the 10 mJ
needed to quench a NbTi magnet [2]). Another reason to operate at a lower
temperature may be to keep the entire magnet superconducting if a non- uniform
temperature profile related to the conduction cooling is present. Nonetheless,
if the temperature profile is uniform, an increased operating temperature has
the advantage in making the magnet easier to protect during a quench.
Acknowledgements
The authors would like to acknowledge the
support of the Ohio Third Frontier and an NSF grant PFI:BIC 1318206 for this
research. This work made use of the
High Performance Computing Resource in the Core Facility for Advanced Research
Computing at Case Western Reserve University.References
[1] http://minerals.usgs.gov/minerals/pubs/commodity/helium.
[2]
Lvovsky, Y. et al. Supercon. Sci.
Tech. 26, 093001 (2013).
[3]
Baig, T. et al. Supercon. Sci. Tech.27
125012 (2014).
[4] Iwasa,
Y. Case Studies in Superconducting Magnets (Springer, New York, 2009).
[5] Ristic, M. et al., IEEE Trans. Appl. Supercond.21(5), 3501, (2011).
[6]. Poole, C. et
al., Proc. Intl. Soc. Magn Reson. Med.23, 1845 (2015).
[7] Douglas,
J., Numerische Mathematik 4, 41 (1962).
[8]http://www.hypertechresearch.com.
[9]
Malagoli, A., et al. IEEE Trans. Appl.
Supercond. 19, 3670, (2009).
[10]
Huang, X. et al., IEEE Trans. Appl.
Supercond. 20(3), 2091, (2010).
[11] Ye L. et al. Supercon. Sci. Tech.
28, 035015 (2015).