The Current Drive and Heating Systems for the ARIES-RS Tokamak Power Plant

T.K. Mau, and the ARIES Team

Fusion Energy Research Program, UC-San Diego, 9500 Gilman Drive., La Jolla, CA 92093-0417

Abstract -- The ARIES-RS reversed-shear tokamak power plant was designed with b=5% and a high bootstrap current fraction of 88%. Three RF current drive systems are required to maintain the plasma equilibrium. They are ICRF fast waves, high-frequency fast waves and lower hybrid waves for seed current drive in various regions of the plasma. Based on a set of top-level requirements, the three RF systems have been envisaged and their launchers fitted into a special blanket sector. Structural materials chosen for the RF in-vessel components do not appear to pose a waste disposal issue for the blanket lifetime. The RF system for mid-plasma current drive is a topic for future research.


I. INTRODUCTION

The ARIES-RS tokamak power plant [1] is based on the reversed shear concept that has moderately high beta (5%) and a high bootstrap current fraction of 88%. The machine parameters are: R=5.5 m, A=4.0, k=1.89, d=0.77, Ip=11.3 MA, Bo=8.0 T, and an average neutron wall loading of 5.7 MW/m2. At the design point, a seed current of 1.3 MA needs to be driven by noinductive means. Detailed alignment between the driven and target equilibrium current profiles requires that three RF current-drive (CD) systems, each operating at a distinct frequrency range, be deployed. The power from these systems can be used to heat the plasma from startup to its final operating conditions, perhaps by launching different wave spectra.
This paper describes the required systems for heating and CD on ARIES-RS. Section II describes the reference current drive scenario and the system requirements, followed by a summary of the design approach. In Section III, parameters for the three systems and their designs including material selection for the components are presented. The ICRF system will be described in detail. In Sec. IV, results of the design study are discussed, and areas where future work will be required are identified.


II. SYSTEM REQUIREMENTS AND DESIGN

For a typical ARIES-RS equilibrium and plasma parameters for optimum plasma performance, the self-driven bootstrap current is not well aligned with the equilibrium profile, as shown in Fig. 1, with 12% of the current to be driven externally. To match the equlibrium profile, three RF systems are required to drive the seed curent throughout the plasma. [2] Near the magnetic axis (y<0.2), the seed curent is to be driven by ICRF fast waves. In the mid-plasma region (0.2<y<0.7), high-frequency fast waves (HHFW) will be used, while the current in the outer plasma region
(0.7<y<1.0) will be driven by lower hybrid (LH) waves. Here y is the normalized poloidal flux, ranging from zero at the magnetic axis to unity at the separatrix. In the reference ARIES-RS design, the operating frequencies are 98 MHz, 1.0 GHz, and 3.5-4.6 GHz for the ICRF, HFFW, and LH systems, respectively.

Fig.1. Matching of the driven (T=RF+BS) current density profile to the target (E) profile, using ICRF, HFFW and LH wave power, for a typical ARIES-RS equilibrium.

Several criteria were considered in the conceptual design of the RF system in-vessel components close to the first wall. In a typical RF system, this component is the wave launching structure with its transmission line feed. The first design criterion requires that the first wall area (and the volume behind) occupied by the launchers be small in order to maximize the tritium breeding ratio. This implies that the power density limit, measured in MW/m2, for the launchers should be maximized. The second guideline calls for the choice of structural materials to not violate the Class-C waste disposal rating for the plant. The third criterion requires all RF launchers to be located in the least number of special blanket sectors for ease of maintenance and replacement.

Because of the limited resources in the ARIES project, only selected RF systems have been designed in detail. A feasible current ITER-EDA design [3] will be used as a starting point and extrapolated to ARIES-RS operating requirements. When there is no ITER basis, an indpendent design or vision for the target RF system will be given, drawing on innovations available in the literature. Experimental data base will be invoked in support of the design, and issues are identified.


III. RF SYSTEM DESIGN

Based on the above criteria, the design parameters for the RF systems are summarized in Table I. The total power delivered to the plasma is 102 MW, of which 81 MW is used in driving useful currents, the difference being due to the less-than-unity antenna directivity. The overall system efficiencies from wall plug to first wall are projected to be 75%, 61%, and 46% for ICRF, HFFW, and LH waves, respectively. The wave launchers and transmission systems are selected and configured so as to minimize intrusions into the blanket and shield structures. Special blanket sectors are designated for locating all the RF launchers that can be fitted into movable 2mx2m plug-in units, as shown in Fig. 2. The transmission lines and waveguides are routed behind the shielding structures to minimize neutron streaming and irradiation of out-vessel components. With aggressive assumptions for maximum power handling, the total first-wall penetration due to the RF launchers is 2.53 m2, which is 0.58% of the first-wall area, requiring only one special blanket sector.

Table I
Heating and Current Drive System Design Parameters

System ICRF HFFW LHW
Frequency 98 MHz 1.0 GHz 3.5, 4.6 GHz
Efficiency 75% 61% 46%
Wall-Plug Power 23.2 MW 61.6 MW 102.4 MW
Delivered Power 17.4 MW 37.6 MW 47.1 MW
First-wall Penetration 0.78 m2 0.80 m2 0.95 m2
Launcher Unit Folded Waveguide Combline Waveguide
Launcher Location
Relative to Midplane
Above Below Above & Below
Transmission Line Coax Waveguide Waveguide
Source Tetrode Klystron Klystron

A. ICRF Fast Fave System

Fast wave power in the ICRF is used to drive seed currrents on axis, for which 15.7 MW is required. Assuming a 90% launcher directivity, the power delivered at the first wall is 17.4 MW. The operating frequency is set at 98 MHz, to place the 2fcT resonance outside the outboard midplane.
The entire ICRF launcher module consists of a toroidal array of six quarter-wavelength folded waveguides [4], having a width of 1.53 m and a height of 0.51 m, as shown in Fig. 3. The module is located above the outboard midplane with its aperture flush with the first wall. Each of the six waveguide cavities is fed by a coax with a 60o phase shift between adjacent elements, resulting in a launched spectrum centered at N||=2.0, where N|| is the wave refractive index parallel to the magnetic field. The corresponding wavelength is 1.53 m.

Fig. 2. Blanket sector containing the RF launchers as viewed from inside the plasma.

Fig. 3. ICRF folded waveguide launcher module.

Each folded waveguide is designed to have ten folds (9 vanes), with a 4 cm gap between vanes, and a toroidal width of 26.3 cm. The wall and vanes are 1.0 cm thick. As displayed in Fig. 4, each cavity is composed of two sections. The front part is the l/4 resonant structure that contains the vanes and is 56 cm thick. The rear part is an extension of the cavity (without vanes) that serves as the coupling region from the coax to the launcher. The coax center conductor is extended inside the rear section and forms a return path with the side wall. The radial thickness of the coupling section is roughly twice the decay length of the evanescent field of the region, and is equal to 32 cm.

Fig. 4. Isometric view of single waveguide cavity and coax.

To provide space at the back of the cavity for neutron shielding, a capacitive diaphragm is inserted at the high field point of the resonant structure, that acts like an impedance transformer and reduces the effective waveguide wavelength. The diaphragms can also serve as structural supports for the thin vanes, coolant routing and additional shielding. For a diaphragm height of 2.4 cm and thickness of 16.0 cm, the effective l/4 is reduced to 56 cm, resulting in a cavity radial thickness of 88 cm and sufficient room at the back for the shielding material.

The nominal power flux through the cavity is 22.3 MW/m2, which is much lower than the projected power density limit (40 MW/m2) for folded waveguides. This safety margin is necessary to account for the enhanced field intensity in the diaphragm region.
A vanadium alloy (V-4Cr-4Ti) is selected as the cavity structural material, which is also the material for the first wall, blanket amd shield. [1] To minimize RF wall losses, the structure is coated with a 60 µm layer of GlidCop AL-25 [5], an oxide dispersion strengthened copper alloy, at an operating temperature of 500oC. These alloys have shown good swelling resistance and retention of both electrical conductivity and yield strength under irradiation.

The effect of neutron irradiation on the coating resistivity, r, and the RF wall dissipation, PW, has been estimated. Using a conservative peak wall loading of 6.95 MW/m2 near the outboard midplane, and a damage rate of 10 dpa/MW-y/m2, the increase in r per FPY is estimated to be 2.20x10-8 W-m. In the absence of detailed field calculations in the complex cavity geometry, an order-of-magnitude estimate of the RF dissipation has been made. It is assumed that Eo=44 kV/cm, the highest E-field measured to date in a folded waveguide, corresponds to a maximum power density of 40 MW/m2. For a reference peak field of 33 kV/cm in ARIES-RS, and using the normal rectangular cavity as a model, it is found that PW=Co(fr)1/2Eo2, where Co is a geometric factor. The results of this investigation are listed in Table II, where the average wall dissipation, WW, and launcher coupling efficiecny, h, are estimated for up to 10 FPYs. Under the present assumptions, the wall dissipation can double in ten years.

Table II
Waveguide Performance as a Function of Full-Power Year

Year 0 1 2 3 5 10
r (10-8 W-m) 5.3 7.5 9.7 11.9 16.3 27.3
PW (MW) 0.91 1.08 1.23 1.36 1.60 2.07
WW (MW/m2) 0.056 0.066 0.076 0.084 0.098 0.127
[eta] 0.950 0.942 0.934 0.928 0.916 0.890

Located near the frist wall, the waveguide cavity is heated by neutrons, Bremsstrahlung radiation, and RF dissipation, so active cooling of the structure is required. The coolant is liquid lithium, for which a simple assessment for its flow velocity and volume fraction of the structure has been made. The launcher is modeled as a flat plate with the same volume and poloidal height, with poloidal coolant flow, and a temperature drop of 36oC. Using a structure and coolant volumetric neutron heating of 35 and 25 MW/m3, and surface heat loads of 0.42 and 0.1 MW/m2 from radiation and RF waves, respectively, reasonable coolant flow requirements are obtained. For example, with a volume fraction of 0.4, the required coolant flow speed is 2.1 m/s, which is manageable.

B. Lower Hyrid Wave System

The lower hybrid launcher system consists of five modules, each radiating a different spectrum for current profile control in the outer plasma region. A total delivered power of 47.1 MW is required, with two modules at 4.6 GHz located below the midplane and radiating the N||=1.8, 2.0 spectra, as shown in Fig. 5. The other three modules operate at 3.5 GHz, radiate the N||=2.3, 2.5, and 3.0 components, and are located above the midplane. Each module couples about 10 MW of power to the plasma.

The base unit of the LH system is the passive active multijunction [6], modelled after the current ITER-EDA design. For example, a typical unit will consist of a toroidal array of eight passive/active TE60 rectangular waveguide pairs (mouth piece) that couple to the plasma. There is a phase difference of 240o between adjacent active waveguides that corresponds to two thirds of the launched parallel wavelength. Each base unit is fed by a 3x2 array of TE30 waveguides (emitter), with a 180o phase shift between rows, via the overmoded TE60 hyperguide unit. Each LH module is then composed of a toroidal array of four or five of these base units, with a toroidal width of 68-79 cm, and a height of 24-27 cm. From the literature, the directivity of these structures is projected to not exceed 0.7, at an edge density greater than 4x1011 cm-3. An aggressive power density capability is assumed, using the ITER-CDA guideline of Pmax~20f2/3 (GHz) in MW/m2, which translates to 46 (55) MW/m2 for f=3.5 (4.6) GHz. As a result, the total first wall penetration for the LH system is a modest 0.95 m2.

The merit for this LH launcher design is the structures behind the passive waveguides that can be used for the purpose of active cooling and also for neutron shielding. As in the ICRF case, the strucutral material is V4Cr4Ti alloy coated with 9 µm thick GlidCop AL-25, and the coolant is liquid lithium. For the passive waveguides at the front to withstand the neutron wall load and the radiative heat flux, it is suggested that they should be made of copper which is bonded to the vanadium structure behind. Cooling will be by conduction; however, the effect of thermal stresses on the bonding of the copper waveguides to the rest of the launcher remains to be assessed.


Fig. 5. Schematic of two adjacent base units in a typical LH launcher module.

C. High-Frequency Fast Wave System

About 38 MW of HFFW power at 1.0 GHz is required for driving currents inside the shear reversal location. The launcher will be located at the outboard midplane. However, there is much uncertainty in the launcher configuration, and relevant experimental data base in this frequency range is scant. It is envisaged that this might take the form of a matrix of traveling-wave combline structures. [7] For this concept, each launcher unit would consist of a toroidal array of coupling loops, each of which terminates in a capacitor forming a resonant circuit. Power is fed from the end loop and propagates by mutual coupling to the adjacent loops until all the power is coupled to the plasma. For this reason, the coupler always sees a matched load, thus eliminating the need for matching circuits and resulting in lower line voltages. Using an aggressive power density limit of ~50 MW/m2, the projected first-wall penetration is 0.80 m2. Experimental data base already exists for this launcher at 120 MHz and 250 kW of power. The key here is to develop this concept further up the frequency range.

D. Nuclear Considerations

In the present conceptual design, the three RF launcher systems have been fitted into a 2mx2m plug-in unit. The combined transmission lines and waveguides emerging from the back of the RF modules are surrounded by a 1.0 m thick shield. On the other hand, a rough estimate indicates that the amount of copper (coatings and passive waveguides) is about 0.2% of the entire vanadium launcher structure volume. It is therefore concluded that this amount of copper should not negatively impact the Class C waste disposal rating for the power plant, as long as the RF launchers are replaced together with the blankets.


IV. DISCUSSIONS

Three RF current drive systems are required for adequate profile control and seed current generation in ARIES-RS. The engineering impacts on the fusion power core due to the RF launchers appear to be modest, while effects on shielding and waste disposal are manageable. It is possible to fit all the launchers in a special blanket/shield sector, which occupies only 0.58% of the first-wall area, when aggressive power density limits are used for the launchers. These limits will need to be demonstrated in future experiments.

Power handling and core integration concerns necessitate the use of RF launchers that have modest data base in a tokamak environment, and extrapolation to power plants is therfore difficult. The largest uncertainty here is the launcher for the HFFW system, for which the combline is a plausible candidate. Alternative schemes for off-axis current drive such as mode-conversion CD and electron cyclotron waves should also be studied. Detailed wave field analysis inside the folded waveguide array and proper assessment of power handling need to be carried out. Cooling and thermal stress analyses for the RF launchers should also be carried out to substantiate the results here.


V. ACKNOWLEDGMENTS

The author acknowledges useful inputs from members of the ARIES Team, especially L. El Guebaly, X. Wang and C.P.C. Wong. This work is supported by USDOE grant DE-AC03-95ER-54299.


REFERENCES

[1] F. Najmabadi, and the ARIES Team, "Overview of the ARIES-RS Reversed-Shear Tokamak Power Plant Study," to be published in Fusion Eng. and Design, 1997.

[2] S.C. Jardin, C.E. Kessel, C.G. Bathke, et al., "Physics Basis for a Reversed Shear Tokamak Power Plant," ibid.

[3] G. Bosia, F. Elio, M. Makowski, and G. Tonon, "Ion Cyclotron, Electron Cyclotron and Lower Hybrid Heating and Current Drive in ITER," Proc. 16th IAEA Fusion Energy Conf., Montreal, 1996, Paper IAEA-CN-64/FP-17.

[4] T.S. Bigelow, M.D. Carter, C.H. Fogelman, et al., "A Folded Waveguide ICRF Antenna for the PBX-M and TFTR," AIP Conf. Proc. 355, 1995, pp. 389-392.

[5]"Glidcop-Copper Dispersion Strengthened with Aluminum Oxide," SCM Metal Products, Inc., 1994.

[6] P. Bibet, X. Litaudon, and D. Moreau, "Conceptual Study of a Reflector Waveguide Array for Launching Lower Hybrid Waves in Reactor Grade Plasmas," Nucl. Fusion 35, 1995, pp1213-1223.

[7] C. Moeller, R.W. Gould, D.A. Phelps, and R.I. Pinsker, "Combline Antennas for Launching Traveling Fast Waves," AIP Conf. Proc. 289, 1993, pp. 323-326.