1 FIP/P8-24
The dynomak: An advanced fusion reactor concept with imposed-dynamo current drive and next-generation nuclear power technologies
D.A. Sutherland, T.R. Jarboe, K.D. Morgan, G. Marklin and B.A. Nelson
University of Washington, Seattle, WA, USA.
Corresponding Author: [email protected]
Abstract: A high- spheromak reactor concept called the dynomak has been designed with an overnight capital cost that is competitive with conventional power sources. This reactor concept uti- lizes recently discovered imposed-dynamo current drive (IDCD) and a molten salt (FLiBe) blanket system for first wall cooling, neutron moderation and tritium breeding. Currently available materials and ITER developed cryogenic pumping systems were implemented in this design from the basis of technological feasibility. A tritium breeding ratio (TBR) of greater than 1.1 has been calculated using a Monte Carlo N-Particle (MCNP5) neutron transport simulation. High temperature superconducting tapes (YBCO) were used for the equilibrium coil set, substantially reducing the recirculating power fraction when compared to previous spheromak reactor studies. Using zirconium hydride for neutron shielding, a limiting equilibrium coil lifetime of at least thirty full-power years has been achieved. The primary FLiBe loop was coupled to a supercritical carbon dioxide Brayton cycle due to attractive economics and high thermal e ciencies. With these advancements, an electrical output of 1000 MW from a thermal output of 2486 MW was achieved, yielding an overall plant e ciency of approximately 40%.
1 Introduction
An advanced spheromak reactor concept called the dynomak was formulated around the recently discovered imposed-dynamo current drive (IDCD) mechanism on the HIT-SI ex- periment at the University of Washington [1]. Previous sustained spheromak experiments largely utilized axisymmetric, coaxial helicity injection for current drive. This method of sustainment led to poor energy confinement in spheromaks during sustainment due to the excitation of plasma instabilities that led to the destruction of closed flux surfaces [2]. Instead, HIT-SI uses non-axisymmetric, steady inductive helicity injection to sustain a spheromak configuration [1]. Recent results on HIT-SI suggest the driven spheromak con- figuration is stable to ideal kink modes with evidence of pressure confinement, and thus is being sustained without the requirement for non-axisymmetric plasma instabilities to FIP/P8-24 2 drive current via dynamo action [3]. These results indicate a significant improvement in the viability of a spheromak configuration for controlled magnetic fusion energy. These promising results on the HIT-SI experiment motivate a formulation of a com- mercial reactor concept called the dynomak that uses IDCD as the method of sustaining athermonuclearspheromakplasma.Goodconfinementqualityisassumedsuchthatthe plasma is able to Ohmically heat to the Mercier beta limit. A guiding philosophy of this study was to minimize costs through engineering simplicity, enabled by the use of a spheromak plasma that does not require externally linking superconducting coil sets for steady-state operation; both toroidal and poloidal plasma currents provide the stabi- lizing and confining magnetic fields, with only one superconducting coil set required for steady-state equilibrium. Also, conventional materials were implemented in this medium- temperature (peak material temperature < 700 oC) reactor unit in an e↵ort to reduce materials development time and costs. 2 Previous spheromak reactor concepts exploited high- ( (2µop)/B )) plasmas in ⌘ 2 very compact configurations with neutron wall loadings of upwards of 20 MWm ,[4] which are unreasonably aggressive values when using more recent fusion reactor studies as comparison that benefit from a more substantive understanding of degradation of materials in a deuterium-tritium (DT) fusion environments [5, 6]. The dynomak reactor 2 concept is designed for 4.2 MWm neutron wall loading, which is near the optimal 2 economic value for a 1 GWe power plant of 4 MWm as was found in the ARIES-AT study [5]. The dynomak is a high- spheromak reactor concept that uses six inductive helicity injectors to sustain a spheromak equilibrium, depicted in Fig. 1. An operating point with key reactor parameters is listed in Table I. A highly shaped flux conserver enables high beta operation, with a wall averaged beta defined by Eq. 1 of 16.6%. Eq. 1 is supported by the definition of wall provided in Eq. 2. A molten salt eutectic of LiF and BeF2 commonly referred to as FLiBe is used as the first wall coolant, neutron moderator, and tritium breeding medium. Due to the assumption that the plasma heat load is uniformly distributed on the first wall due to the lack of a diverted magnetic topology while using inductive helicity injection, a simple, single working fluid blanket design was able to be implemented in the dynomak concept. The primary FLiBe loop, with a peak o coolant outlet temperature of 580 CiscoupledtoasupercriticalCO2 Brayton cycle for electricity generation. The choice of secondary cycle was made due to the high e ciency (> 45%) in the desired range of blanket operating temperatures (480 580oC) [7, 8]. Additionally, operating at these somewhat conservative coolant temperatures should put less demanding requirements on heat exchanger designs when compared to designs for very high temperature reactors (i.e. > 900 oC) [9].
dV < >= wall (1) wall 2⇡2R a2 r o
2µop wall = 2 (2) Bwall 3 FIP/P8-24
FIG. 1: A sliced rendering of the dynomak reactor concept, excluding the secondary power conversion cycle [10].
2 Core design
2.1 Equilibrium
µoj In accordance with IDCD requirements [1], a stepped ( B ) profile is assumed in the dynomak concept, with one value of being that of the injectors⌘ and the other the value required for current amplification within the separatrix. The required toroidal current gain for the dynomak reactor concept is 1280. The IDCD profile is robust because it is maintained simply by keeping the magnetic fluctuation amplitude created by the helicity injectors above a threshold value for IDCD. An enhanced Grad-Shafranov equilibrium code was used to determine the required currents in a prescribed coil set, depicted in blue in Fig. 1. This calculation imposed marginal Mercier stability on each flux surface with a = 2.4 and an aspect ratio of 1.5. Additionally, this code calculated the required currents in two copper coils within the blanket on the outboard mid-plane to exclude magnetic flux from the helicity injector region; this ensures satisfactory operation of the six, inductive helicity injectors located on the outboard mid-plane of the dynomak concept. The dynomak equilibrium and corresponding coil set are depicted in Fig. 2.
2.2 Current drive One of the key economic advantages of a spheromak configuration is eliminating the neces- sity for a toroidal field coil set by relying solely on toroidal and poloidal plasma currents for stability and confinement. As a result, a spheromak fusion plasma will have a substan- tial higher plasma current than a similarly scaled tokamak; the toroidal plasma current in the dynomak is 41.7 MA, with approximately the same magnitude of poloidal plasma cur- rent as well. Use of conventional current drive methods, such as radio-frequency current drive (RF) or neutral beam injection (NBI), is not an option in a spheromak configuration due to the very low e ciencies of these methods. Additionally, due to postulated smaller bootstrap current in a spheromak configuration due to less pronounced neoclassical ef- FIP/P8-24 4
Parameter Symbol Value Major radius [m] Ro 3.75 Minor radius [m] a 2.5 Toroidal plasma current [MA] Ip 41.7 20 3 Number Density ( 10 m ) n 1.52 ⇥ e Wall-averaged (%) < wall > 16.6 Peak temperature (keV) Te 20 2 Neutron Wall loading (MW m ) Pn 4.2 2 First wall heat flux (MW m ) q00 1.05 Helicity Injector Power (MW) PCD 58.5 o FLiBe Inlet Temperature ( C) Tin 480 o FLiBe Inlet Temperature ( C) Tout 580 3 1 Global blanket flow rate (m s ) U˙ 5.17 Thermal Power (MW) Pth 2486 Fusion power (MW) Pfus 1953 Electrical power (MW) Pe 1000 Plasma gain Qp 33 Engineering gain Qe 9.5 Thermal e ciency (%) ⌘ 45 th Global e ciency (%) ⌘ 40
TABLE I: Key parameters of the dynomak reactor operating point [10]. [%]
Z [m]
wall wall β
Major R ad ius [m]
FIG. 2: The dynomak equilibrium and corresponding superconducting coil set [10]. 5 FIP/P8-24 fects, a very e cient current drive method such as IDCD is almost certainly required for a reasonable recirculating power fraction. The current drive power requirements for the dynomak concept will be strongly dependent on the resistivity profile, especially in the cooler resistive edge that contains a large plasma volume. Using the pressure profile derived from the Grad-Shafranov equilibrium depicted in Fig. 2, and assuming constant density, a temperature profile can be obtained provided with an assumed separatrix tem- perature. A separatrix temperature of greater than 100 eV will most certainly be required for a reasonable current drive power requirement. A separatrix temperature of 200 eV is taken as a likely value to expect from high performing tokamak discharges [11]. Assum- ing a 200 eV separatrix, along with a 41% power coupling e ciency for IDCD observed on HIT-SI, the current drive power requirement is 58.5 MW. Also, using the observed 80% wall plug e ciency on HIT-SI, the total electrical current drive power requirement is estimated to be 73 MW.
2.3 Feedback Afeedbacksystemwillbeimplementedinthedynomakreactorconceptinane↵ortto maintain desired flux surface locations and ensure operation at desired plasma param- eters. As is seen in the Grad-Shafranov equilibrium in Fig. 2, it is desired to have a limited plasma with a nearly circular poloidal cross section; this configuration will require approximately 10-20 mm accuracy in flux surface position. The YBCO equilibrium coils will be feedback controlled to keep the plasma at the desire distance away from the wall at all times. This method of feedback was successfully implemented on the HIT-II machine at the University of Washington [12].
3 Blanket design
3.1 Blanket material To achieve a closed DT fuel cycle, the blanket system implemented in the dynomak re- actor concept must be capable of achieving a tritium breeding ratio (TBR) of greater than 1. More practically, to account for imperfect extraction of generated tritium, and tritium losses in the processing system, a TBR of greater than 1.1 is desired. A Monte Carlo N-particle (MNCP5) neutron transport simulation found that the TBR was highly dependent on first wall thickness, and thus it is advantageous to minimize the thickness of high-Z, non-tritium-breeding material between the fusion neutron source and the tritium breeding medium. FLiBe contains the necessary lithium to breed tritium, and also con- tains beryllium that serves as a neutron multiplier. Additionally, it is possible to enrich FLiBe with excess Li-6 (above the natural occurrence) in an e↵ort to increase the TBR.
3.2 First wall design and thermal hydraulics The primary vacuum vessel contains the DT plasma, as shown in Fig. 1. Moving minor radially outward, a thin alumina insulation layer is plasma-sprayed on a 1 cm thick copper FIP/P8-24 6
flux conserver. The copper flux conserver serves the primary purpose of providing a stabilizing conducting shell that allows for the induction of stabilizing eddy currents should the plasma be displaced from its desired placement. Also, the use of copper as a flux conserver provides good thermal contact to the first wall cooling system and serves as a additional neutron multiplier. A series of 316 stainless steel (SS) cooling tubes are bonded to the cooper first wall to ensure good thermal contact. These first wall cooling tubes are linked to a dual-chambered, pressurized FLiBe blanket system. These pipes deliver relatively cool FLiBe (480 oC) from the thin (25 cm thick) blanket to the first wall at 1 8ms . The FLiBe flows toroidally 4.3 m before exhausting to inner, hot blanket (50 cm thick) where it is neutronically heated to the reactor outlet temperature of 580 oC. The FLiBe cooling tubes that are integrated as part of the primary vacuum vessel are welded to a 1 cm thick, 316 SS surface for structural support, which interfaces with the 50 cm thick hot blanket. A 2 cm thick, 316 SS structural shell separates the hot from the cold blanket. Lastly, a 7.5 cm thick 316 SS structural shell connects to the blanket separator via a truss system (not depicted in Fig. 1), and serves as the final pressure vessel boundary before biological containment.
3.3 Tritium breeding and shielding Using natural lithium content in FLiBe, a TBR of 1.125 0.022 was calculated using MCNP5. This TBR is above the practical requirement of 1.1,± as desired. Additionally, the TBR could be increased further by enriching FLiBe with Li-6, or adding a thin beryllium layer outside the first wall apparatus to enhance neutron multiplication. Though FLiBe is an excellent moderator, with a mean free path of 7 cm for 14 MeV neutrons [13], neutron shielding is still required to protect the sensitive high temperature superconducting coil set (yttrium barium copper oxide (YBCO)) from excessive fast neutron fluxes. Zirconium hydride (ZrH2) was chosen as the neutron shielding material [14], and the use of it in the dynomak reactor concept provided for a estimated limiting superconducting coil set lifetime of 29.3 full power years.
4 Economics
4.1 Requirements Overnight capital cost is a convenient economic metric for comparing energy sources since it is the main dissuading figure for initial investment in new power plants. The U.S. Energy Information Administration reports that the cheapest energy source is natural gas, with a cost of $665 per kW for an advanced compact turbine [15]. Adding carbon capture technologies or using other methods of natural gas electricity generation provides a range of costs from $974 $2060 per kW [15]. The next cheapest conventional competitor is coal, which has a lowest overnight capital cost of $2844 per kW. Thus, it is suggested the desired capital cost of a fusion reactor unit to displace these two energy sources is $665 2844 per kW. 7 FIP/P8-24
4.2 Cost Breakdown & Conclusions Adirectandovernightcapitalcostbreakdownofthedynomakreactorconceptispresented in Tab. II. Note the use of ARIES-AT [5] inflation adjusted figures for expected similar aspects of any fusion plant in Tab. II, indicated by asterisks. With an overnight capital cost of $2713 per kW, the dynomak reactor concept is within the desired range stated previously to be competitive with natural gas and coal-fired power plants. This low cost figure, which includes conservative estimates for the pricing of bulk reactor components and superconducting coil set costs, is certainly attractive from a economic standpoint. Provided an energy e cient means of current drive such as IDCD and adequate energy confinement quality, the spheromak configuration may o↵er an economical path to fusion power. The spheromak sustained by IDCD should be developed further to confirm that high temperature sustainment with IDCD using inductive helicity injection is compatible with su cient confinement quality; this result would be a substantial advancement of the spheromak concept in overcoming confinement quality limitations of previous sustained spheromak experiments [2].
Component(s) Cost ($M) Land and land rights⇤ 17.7 Structures and site facilities⇤ 424.3 Reactor structural supports 45.0 First wall and blanket 60.0 ZrH2 neutron shielding 267.4 IDCD and feedback systems 38.0 Copper flux exclusion coils 38.5 Pumping and fueling systems 91.7 Tritium processing plant 154.0 Biological containment 50.0 Superconducting coil system 216.0 Supercritical CO2 cycle 293.0 Unit direct cost 1696 Construction services and equipment⇤ 288 Home o ce engineering and services⇤ 132 Field o ce engineering and services⇤ 132 Owner’s cost⇤ 465 Unit overnight capital cost 2713
TABLE II: Dynomak reactor concept cost breakdown. The asterisks de- note components of the cost analysis that were taken from the ARIES- AT study [5] and inflation corrected to 2013 dollars [10]. FIP/P8-24 8
References
[1] T.R. Jarboe, et al., Imposed-dynamo current drive, Nucl. Fusion 52 (2012) 8, 083017
[2] B. Hudson, et al., Energy confinement and magnetic field generation in the SSPX spheromak, Phys. Plasmas 15 (2008) 056112.
[3] B.S. Victor, et al., Sustained spheromaks with ideal n=1 kink stability and pressure confinement, Phys. Plasmas 21 (2014) 082504.
[4] R. Hagenson and R. Krakowski, Steady-state spheromak reactor studies, Fusion Technol. 8 (1985) 1606-1612.
[5] F. Najmabadi, and the ARIES Team, The ARIES-AT advanced tokamak, an ad- vanced technology fusion power plant, Fus. Eng. and Des. 80 (2006) 3-23.
[6] J. Menard, et al., Prospects for pilot plants based on the tokamak, spherical tokamak and stellarator, Nucl. Fusion 51 (2011) 10, 103014.
[7] V. Dostal, M. Driscoll, and P. Hejzlar, A supercritical carbon dioxide cycle for next generation nuclear reactors, MIT dissertation (2004).
[8] P. Hejzlar, et al., Assessment of gas cooled fast reactor with indirect supercritical CO2 cycle, Nucl. Eng. Technol. 38 (2006) 2. [9] S. Mylavarapu, et al., Investigation of high-temperature printed circuit head ex- changers for very high temperature reactors, J. Eng. Gas Turbines and Power 131 (2009).
[10] D.A. Sutherland, et al., The dynomak: An advanced spheromak reactor concept with imposed-dynamo current drive and next-generation nuclear power technologies, Fus. Eng. and Des. 89 (2014) 4, 412-425.
[11] G. Porter, et al., Analysis of separatrix plasma parameters using local and multi- machine databases, J. Nucl. Mater. (1999) 266-269, 917-921.
[12] T. Hamp, et al., Temperature and density characteristics of the Helictiy Injected Torus-II spherical tokamak indicating closed flux sustainment using coaxial helicity injection, Phys. Plasmas (2008) 082501.
[13] M. Youssef, Damage rate in V.V. as a function of convective layer thickness, In: APEX Study Memorandum, March 1998.
[14] T. Hayashi, et al., Advanced neutron shielding material using zirconium borohydride and zirconium hydride, J. Nucl. Mater. 368-388 (2009) 119-121.
[15] E.I.A., Updated capital cost estimates for electricity generation plants, U.S. DOE, http://www.eia.gov/oiaf/beck plantcosts/pdf/updatedplantcosts.