LINEAR INDUCTION ACCELERATOR WITH MAGNETIC STEERING FOR INERTIAL FUSION TARGET INJECTION Ronald Petzoldt,* Neil Alexander, Lane Carlson, Eric Cotner, Dan Goodin and Robert Kratz General Atomics, 3550 General Atomics Court, San Diego, CA 92186-5608, *petzoldt@fusion.gat.com A traveling-wave induction accelerator was designed and built to launch 1 cm diameter cylindrical aluminum tubes (surrogate IFE targets) into a vacuum chamber at speeds greater than 50 m/s. The accelerator is 0.55 m long with 300. Each coil is energized 30 degrees out of phase with the adjacent resulting in a traveling sinusoidal magnetic field that moves past the projectile with resulting accelerating force. Saddle surrounding the axial drive provide projectile spin. Four saddle were placed around the projectile s flight path at a distance of 0.4 m from the barrel. AC voltage energizes these resulting in an AC quadrupole magnetic field that provides a centering force as the projectiles pass through the. To further improve accuracy, an actively controlled, in-flight, magnetic steering system was placed after the initial passive steering. This system measured the position of the projectile at two locations, in real time and adjusted the AC current in another set of four saddle to correct the measured trajectory errors. The first set of steering improved the standard deviation by a factor of 8 and the second set by an additional factor of 3, for a total factor of 24 improvement. I. INTRODUCTION Targets must be injected several times per second into an Inertial Fusion Energy power plant. Several types of injectors have been considered for this purpose including a gas gun, rail gun, electrostatic accelerator and induction accelerator. 1-7 An induction accelerator has advantages over the other methods in that it requires no physical contact as is normally needed with the gas gun and rail gun, requires no propellant gas to be pumped out of the system as does a gas gun, requires no electrical contact as does the rail gun, and unlike the electrostatic accelerator can be used to accelerate indirect-drive targets. D. G. Elliott provides a good introduction to traveling wave induction launchers. 8 The injector consists of many closely spaced circular. The armature (target) contains a cylindrical conducting ring. The are sequentially energized to produce a magnetic field with a radial component that travels down the injector with speed Us. The traveling magnetic field passing through the armature causes a potential difference and current flow in the armature. Figure 1 shows the geometry with the. Fig. 1. A 12-phase induction accelerator. II. SINGLE BLOCK ACCELERATOR We designed and built a 12-phase induction accelerator. With 12 phases, each coil current is 30 degrees out of phase with each adjacent coil. We provided the desired polyphase voltage to the starting with six Teagam 2730A 50 MHz Function/Arbitrary Waveform Generators. These were used as input to six 5 kw power supplies (QSC RMX5050 audio amplifiers). The were wired with each 6 th coil in series and reversed to achieve 12 phases of as in Fig. 1. We performed finite element calculations 9 to evaluate the accelerating force on a cylindrical sleeve as a function of frequency and peak current. Our first accelerator were each 2.1 mm long, 1 mm thick with 15 turns per coil. Twelve glued together as shown in Fig. 2 is called a stage. Calculations indicated that with voltage amplitude of 120 V across 5 stages wired in series and frequency of 5 khz, a current 308 http://dx.doi.org/10.13182/fst14-915 FUSION SCIENCE AND TECHNOLOGY VOL. 68 SEP. 2015
amplitude of 40 A would accelerate a 0.5 g, 2 cm long aluminum tube to approximately 30 m/s. This current and voltage level can be provided by a single channel of a power supply. The power supplies and function generators provided the required 6 phases of power to the linear induction accelerator. One block of was first used to accelerate projectiles to 29 m/s as measured with high-speed video of the projectiles leaving the accelerator. We subsequently added race-track rotational-field to the single block accelerator as shown in Fig. 3. The rotational-field are installed radially outward of the linear-field for simultaneous linear and rotational acceleration. The projectiles are accelerated vertically upward. A two dimensional model of these, shown in Fig. 4, was evaluated using finite element software. 9 For 100 A/mm 2 (234 A total) peak engineering current density at 1.5 khz AC frequency, the calculated torque on a 25 mm long projectile is 130 µn*m. 1 cm Fig. 4. 2-D model of rotational field winding (rotational surround the linear acceleration ). III. MULTI-BLOCK ACCELERATOR Linear Pickup are used to sense the arrival of the projectile to various locations within the accelerator. A 20-turn coil of 0.05 mm diameter magnet wire (2 axial by 10 radial) requires just 0.1 mm space between accelerator stages. The output voltage of the circuit shown in Fig. 5 increases when the projectile is present. The output is then compared to a reference voltage set between the two levels. The comparator output goes high when the target enters the pickup coil. Fig. 2. Groups of 12 glued together form each stage of the accelerator. Linear block 1 MHz Generator Pickup Coil 3.9 nf Switching diode 22 Ω 47 kω 2 nf Vout Fig. 3. Single-block accelerator. Rotational race-track Next, the single block accelerator was installed into a vacuum chamber and the required wiring was connected through vacuum feed-throughs. The accelerator continued to operate well in a 1-10 Torr vacuum as it had in air. Fig. 5. Pickup coil projectile sensing circuit. With two sets of acceleration amplifiers, it is possible to provide power to several blocks of acceleration with a maximum of two blocks energized at any one time. Each block is switched on prior to the projectile s arrival into that block and turned off after the projectile leaves that block. Pickup are positioned within each block as shown in Fig. 6. As the target passes each pickup coil, the output from that pickup coil circuit goes high. Figure 7 shows important components of the induction accelerator control and power circuitry. Comparators for each block compare the input voltages from two pickup coil circuit outputs (one shown) and outputs high logic signals to the PLC when the pickup output exceeds an adjustable reference voltage. The PLC then sends the appropriate logic signal to the PC942 FUSION SCIENCE AND TECHNOLOGY VOL. 68 SEP. 2015 309
photocoupler. With high input, the photocoupler directs current to the six IL410 phototriac optocouplers in each block, which allow the S6020L SCRs to be turned on when forward biased. The back-to-back SCRs then allow current flow in both directions through the accelerator. 5 blocks of Fig. 8. Induction accelerator wired to control electronics. Fig. 6. Pickup coil locations with associated pickup coil circuit signals and desired block energization. The multi-block have 24 turns per coil rather than 15 turns as in the single-block configuration to be better optimized for our power supplies in the doublevoltage, single-channel power mode (rather than dualchannel mode). The stages are each shorter ~20 mm for the single-channel multi-block configuration as opposed to 25 mm for the dual-channel stages. Blocks 1 through 5 and their respective circuit boards are shown in Fig. 8. For vacuum operation, we installed the 5-block accelerator into the chamber as shown in Fig. 9. We fired projectiles at first in dual-channel mode and later in single-channel mode. The highest velocity achieved was about 57 m/s. Rotation rates up to about 80 Rev/s were common. 5 blocks of Fig. 9. Multi-block induction accelerator installed in vacuum chamber with upward acceleration. IV. PASSIVE IN-FLIGHT MAGNETIC STEERING Fig. 7. Simplified induction accelerator control and power circuit. Injection accuracy can be improved with in-flight steering of the projectiles. The method implemented utilized an AC quadrupole magnetic field midway on the path between the end of the gun barrel and a clay surface placed at the top of the vacuum chamber, 0.8 m above the barrel muzzle. Finite element software 9 was utilized to evaluate a centering force that exists as the projectile passes through the AC quadrupole. The 3-D model used is depicted in Fig. 10. Four saddle surround the projectile with 100 A/mm 2 peak engineering current density at 5 khz. The relative direction of the current flow 310 FUSION SCIENCE AND TECHNOLOGY VOL. 68 SEP. 2015
in the is shown in Fig. 11. The straight sections of the are 100 mm long and the cross-sectional area of each coil is 25 mm 2. With these parameters there is a radial restoring force on an off-center aluminum 25.3 mm long projectile of approximately 133 N/m. Fig. 10. 3-D model used to evaluate forces exerted on the projectile passing through a set of four (coil ID is 21 mm). The control and power electrical circuit for the steering is quite similar to the accelerator coil circuit in Fig. 7. The main difference is that the pickup coil sensor circuit is replaced by an LED that is installed in the slot on one side of the steering coil that directs light to an optical fiber mounted on the opposite side of the mandrel which directs the collected light to an IF-D95T Photologic Detector. The time and magnitude of the steering coil pulse can be adjusted to optimize the steering impulse. We ran successful steering tests with the pictured in Fig. 12, but about 15% of the projectiles hit this coil. So a larger coil mandrel was fabricated with 120% scale in the transverse direction and 110% scale in the axial direction. Numerous tests were run to optimize the time and magnitude of target steering to reduce the outer diameter of the clay mark pattern in the clay for a 10 shot group. We rotated the targets for 35 ms prior to beginning axial acceleration. This increased the final rotation rate to approximately 125 Hz. A factor-of-six improvement in accuracy was achieved with optimum steering strength and duration as shown in Fig. 13. Fig. 11. Relative direction of the current flow to produce the quadrupole field. We designed a steering coil mandrel at 110% scale and fabricated it with a 3D printer out of acrylonitrile butadiene styrene. We wound the on the mandrel with 45 turns per coil of AWG 22 magnet wire as shown in Fig. 12. Fig. 13. Projectile trajectory accuracy improved from a maximum deviation of ±8.2 mrad to ±1.3 mrad for a 10 shot group with passive target steering (scale is in inches). V. ACTIVE MAGNETIC STEERING Fig. 12. Quadrupole coil mandrel wound with 45 turns per coil. Coil length is 11 cm. Additional accuracy improvement can be achieved by measuring the projectiles trajectory and energizing select FUSION SCIENCE AND TECHNOLOGY VOL. 68 SEP. 2015 311
of a second steering coil set for a duration that is proportional to the desired steering correction in each of two orthogonal directions. The axial locations for these steering system components, together with a picture of the system, are shown in Fig. 14 for a system where the target trajectory is now horizontal. The position detectors field of view was increased by utilizing two photodiodes for each detector with overlap in the transverse direction as illustrated in Fig. 15. Representative voltage vs position curves for detector 1 vertical are shown in Fig. 16. Accelerator Passive steering Position measurement detectors Active Position steering measurement detectors Clay 0 m 0.5 0.6 1.4 1.7 2.35 2.5 The projectile trajectory is computed in real time from detector input and steering corrections are applied to the appropriate active steering. Detector position outputs are recorded for post-shot analysis. We achieved the steering results shown in TABLE I for 21 consecutive shots. The no steering values are extrapolated from detector 1 position measurements. The passive steering values are those that were predicted based on measurements at detectors 1 and 2. The with active steering values are position measurements for those same shots at detector station 3. The net result is a factor-of-24 improved repeatability from magnetic steering. A factor-of-8 improvement resulted from passive steering and an additional factor-of-3 improvement from active steering. A clay impression from these 21 shots is shown in Fig. 17. Note that these projectiles were fired over a distance more than 3 times farther than those in Fig. 13 (2.5 m vs 0.8 m). TABLE I. Accuracy Improvement with Magnetic Steering Fig. 14. Locations of detectors and steering for passive and active magnetic projectile steering. Quantity measured For Detector Station 3 Repeatability 1 σ (mm) Repeatability 1 σ (mrad) Horizontal -no steering 5.97 2.5 -With passive steering 0.60 0.3 -With active steering 0.24 0.1 Vertical -no steering 5.42 2.3 -With passive steering 0.68 0.3 -With active steering 0.24 0.1 Fig. 15. Position detectors utilize two photodiode sensors with overlapping field of view. Fig. 17. Clay mark photo from 21 actively steered shots at 2.5 m from the injector (scale is in inches). VI. CONCLUSIONS Fig. 16. Voltage vs projectile position for detector 1V. A multi-block linear induction accelerator was designed, built and performed as predicted. Pickup provide axial position feedback to electronic switches that control power to sequential coil blocks. Magnetic steering provided a factor-of-24 target position repeatability improvement, including a factor-of-8 improvement from passive steering and an additional factor-of-3 from active steering. 312 FUSION SCIENCE AND TECHNOLOGY VOL. 68 SEP. 2015
A linear induction accelerator has been shown to accelerate projectiles (surrogate targets) up to 57 m/s into a vacuum chamber. Magnetic steering of the projectiles in tests up to 20 m/s demonstrated high placement accuracy. Higher speeds with steering may be accomplished using more acceleration coil blocks and larger power supplies for the steering. Additional research will be required to demonstrate acceleration of real IFE targets that are made of different materials with different electrical conductivities, densities and wall thicknesses at cryogenic temperatures. ACKNOWLEDGMENTS This work was supported by General Atomics internal research and development funds. REFERENCES 1. R. W. Petzoldt, R. W. Moir, Target Injection Methods for Inertial Fusion Energy, Fusion Technology, 26, 896 (1994). 2. R. W. Petzoldt, Inertial Fusion Energy Target Injection, Tracking, and Beam Pointing, UCRL-LR- 120192, Ph.D. Thesis (1995). 3. R. W. Petzoldt, IFE Target Injection and Tracking Experiment, Fusion Technology, 34, 831-839 (1998). 4. D. T. Goodin et al., Developing the Basis for Target Injection and Tracking in Inertial Fusion Energy Power Plants, Fusion Engineering and Design, 60, 27-36 (2002); http://dx.doi.org/10.1016/s0920-3796(01)00593-2. 5. D.T. Frey et al., Rep-Rated Target Injection for Inertial Fusion Energy, Fusion Science and Technology, 47, 1143 (2005); http://dx.doi.org/10.13182/fst05-30. 6. R. W. Petzoldt et al., Target Injection with Electrostatic Acceleration, Fusion Science and Technology, 56, 1, 417-421 (2009); http://dx.doi.org/10.13182/fst09-25. 7. R. Miles et al., Challenges Surrounding the Injection and Arrival of Targets at LIFE Fusion Chamber Center, Fusion Science and Technology, 60, 61-65 (2011). http://dx.doi.org/10.13182/fst10-333 8. D. G. Elliott, Traveling-Wave Induction Launchers, IEEE Transactions on Magnetics, 25(1), 159 (1989). http://dx.doi.org/10.1109/20.22526 9. Opera Simulation Software, Cobham Inc., 1700 North Farnsworth Avenue, Aurora, Illinois, 60505-1186, http://operafea.com FUSION SCIENCE AND TECHNOLOGY VOL. 68 SEP. 2015 313