Overview of Non-Solenoidal Startup Studies in the Pegasus ST

5 downloads 45442 Views 5MB Size Report
Nov 2, 2016 - Injection Campaign. Highlights ... Initial HFS injector campaign in progress. – Development to .... System automation. – Intra-shot beam ...
Overview of Non-Solenoidal Startup Studies in the Pegasus ST M.W. Bongard, J.L. Barr, G.M. Bodner, M.G. Burke, R.J. Fonck, J.L. Pachicano, J.M. Perry, J.A. Reusch, N.J. Richner, C. Rodriguez Sanchez, D.J. Schlossberg

58th Annual Meeting of the APS Division of Plasma Physics

University of Wisconsin-Madison

2 November 2016 San Jose, CA

PEGASUS Toroidal Experiment

Layout 8.5” x 11”

8’ W x 4’ H

Title Banner Technology and Diagnostic Development

Local Helicity Injection (LHI) Provides Robust Non-Solenoidal Startup on the PEGASUS ST

Three HFS Injection Systems Implemented and Tested Since April 2016

Power Balance Model Provides Predictive Tool for Ip(t)

NIMROD Describes Current Helical Current Stream Reconnection as Drive Mechanism

LFS Local Helicity Injection Produces Core Te > 100 eV

Ip > 0.15 MA Achieved Via HFS Injection To Date

LHI Startup Scenarios Grow From Helical Current Streams to Quality, High Ip Plasmas

Multi-Year Technology Development has Produced Robust, High Performance Current Injectors

Analytic Formulation of Power Balance Model Elements Allow Partitioning of Energy Flow

Current Stream Interaction Manifests as Edge-Localized MHD Burst

Te (R, t) Remains Peaked for LFS Injection Geometry and Minimal VIND

HFS Helicity Injection Provides Non-Solenoidal Sustainment at High IN

High-Field-Side (Divertor) Injection Experiments Provide Confinement Tests & Higher Ip

Large-Ainj Injector Design Provides Enhanced Performance, Simplified Geometry

EquilibriumCalibrated Inductance Model Improves Estimates of NonSolenoidal VIND

ReconnectionDriven Ion Heating Gives Ti > Te During LHI

Technical Challenges Arise for LHI Startup With HFS Injection Geometry

LHI Provides Access to High-T at A ~ 1 with Non-Solenoidal Sustainment and Anomalous Ion Heating

Hierarchy of Physics Models Provide a Predictive Understanding for LHI Startup

Thomson Scattering Enhancements to Measure Te and ne Profiles During LHI

0-D Power Balance Model Used to Explore Projections for NSTX-U Startup

Different MHD Activity Observed Between LFS and HFS Injection Geometry

Poloidal Field Shaping Facilitates Relaxation at Full Toroidal Field (Binj = 0.23 T)

Progress in NonSolenoidal Startup on Pegasus

M.W. Bongard, APS-DPP 2016

Progress Toward Predictive Models of LHI

2016 Helicity Injection Campaign Highlights

Non-Solenoidal Startup via Local Helicity Injection

Reprints

Non-Solenoidal Startup via Local Helicity Injection

Technology and Diagnostic Development

Progress Toward Predictive Models of LHI

2016 Helicity Injection Campaign Highlights

Local Helicity Injection (LHI) Provides Robust Non-Solenoidal Startup on the PEGASUS ST Ip ≤ 0.18 MA via LHI (Iinj = 5 kA)

Helicity Injectors

LFS System HFS System

Anode

Shaped W Cathode/Anode

Mo PMI Suppression Guard Rings

VINJ

IINJ

+

Mo Cathode/Anode Shield

IArc

Mo Arc Cathode

Mo / BN Washer Stack

Varc

D2 gas feed

M.W. Bongard, APS-DPP 2016

+

Plasma Parameters Ip ≤ 0.23 MA shot ≤ 0.025 s BT 0.15 T A 1.15–1.3 R 0.2–0.45 m a ≤ 0.4 m κ 1.4–3.7

Injector Parameters  Iinj ≤ 14 kA Iinj ≤ 4 kA Vinj ≤ 2.5 kV Ninj ≤ 4 Ainj = 2-4 cm2 Iarc ≤ 4 kA Varc ≤ 0.5 kV

LHI Startup Scenarios Grow From Helical Current Streams to Quality, High Ip Plasmas Three-Injector Array

LHI

OH (H-mode)

Injector Shutoff

Null Formation

Relaxation

Unstable injected current streams M.W. Bongard, APS-DPP 2016

Reconnect, relax to Tokamak-like state

Subsequent OH-Driven Tokamak

High-Field-Side (Divertor) Injection Experiments Provide Confinement Tests & Higher Ip •

Initial HFS injector campaign in progress – Development to minimize PMI as BTF increases



Configuration minimizes VIND



3-4x increase in HI drive: Veff ~ AinjVinj/Rinj



Test reconnection mechanisms at higher Ip, BTF



Injectors at longer pulse, high-BTF

NSTX-U Projected Performance Ohmic L-mode

R-R Stochastic

Fixed Te=150 eV

Fixed Te = 75 eV

: AinjVinj/Rinj [V-m]

Centerstack Lower DIV Strike Plate

DIV injectors M.W. Bongard, APS-DPP 2016

Hierarchy of Physics Models Provide a Predictive Understanding for LHI Startup Taylor Relaxation

1. Taylor relaxation, helicity conservation

I p  ITL ~

– Steady-state maximum Ip limits

ITF I inj w

Helicity Conservation

VLHI

Ainj B ,inj  Vinj 

I p VLHI VIR VIND   0 ; I p  ITL

2. 0-D power-balance Ip(t) – VLHI for effective LHI current drive

Reconnecting LHI Current Stream

3. 3D Resistive MHD (NIMROD) – Physics of LHI current drive mechanism

M.W. Bongard, APS-DPP 2016

D.J. Battaglia, et al. Nucl. Fusion 51 (2011) 073029. N.W. Eidietis, Ph.D. Thesis, UW-Madison, 2007. J. O’Bryan, Ph.D. Thesis, UW-Madison, 2014. J. O’Bryan, C.R. Sovinec, Plasma Phys. Control. Fusion 56 064005 (2014)

Three HFS Injection Systems Implemented and Tested Since April 2016 • Two injectors at toroidally opposite positions in lower divertor region • Design point leverages high Ainj – 3–4 increase in VLHI over prior systems

Configuration 1

Configuration 2

– Ainj = 8 cm2 total – Vinj ≤ 1.2 kV – Iinj ≥ 8 kA total Configuration 3

• Systems vary Rinj, Zinj, local limiter geometry – Latest design incorporates floating, electropolished divertor shield plates M.W. Bongard, APS-DPP 2016

Multi-Year Technology Development has Produced Robust, High Performance Current Injectors Three-Injector Array

• Washer-stack arc source: – Jinj ~ 1kA/cm2

• High-voltage in SOL: Vinj > 1kV – Frustum cathode – Floating cathode shield

1.6 cm

Cathode Shield Injection Cathode Arc Source

• PMI control: 1-2 cm from LCFS

Shield Rings

– Cascaded shield rings – Local limiter – Mo, W PFCs

M.W. Bongard, APS-DPP 2016

Gas Feed

Clean, High-Vinj Operation

Large-Ainj Injector Design Provides Enhanced Performance, Simplified Geometry •

New injectors designed for HFS system – Doubled Ainj (2 cm2  4 cm2) – Compact design for lower divertor region

• Modular internal assembly – Permits in-vessel maintenance/repositioning – Exterior PFC components rapidly adjusted about common arc chamber / fueling system • Changes to Ainj, shield structures

– Integrated hypodermic gas feed alleviates field sensitivity from previous

• Refractory materials for resilience to harsh environment – W for high-Vinj cathode/anode – Mo for external shield assemblies M.W. Bongard, APS-DPP 2016

New: 4 cm2 Old: 2 cm2

Thomson Scattering Enhancements to Measure Te and ne Profiles During LHI • Improved timing / synchronization

Thomson Viewing Locations and A ~ 1 Plasma

– Higher realized laser power – Lowered beam scrape-off losses

• System automation – Intra-shot beam alignment – Data acquisition

• Stray light mitigation – Baffling, electronic gating

• Background signal reduction – Wire grid polarizers – High speed shutters M.W. Bongard, APS-DPP 2016

80 cm viewing region

Power Balance Model Provides Predictive Tool for Ip(t) • Model reasonably recreates Ip(t)

I p VLHI VIR VIND   0 – VLHI: effective drive – VIR: resistive dissipation – VIND: analytic, from shape(t) – Taylor relaxation limit: Ip ≤ ITL

• VIND dominates current drive with LFS mid-plane injection M.W. Bongard, APS-DPP 2016

System Voltages [V]

Shape Evolution

= 60eV

Eidietis et al., J. Fusion Energ. 26, 43 (2007) S.P. Hirshman and G.H. Nielson 1986 Phys. Fluids 29 790 O. Mitarai and Y. Takase 2003 Fusion Sci. Technol. Battaglia et al., Nucl. Fusion 51, 073029 (2011)

Analytic Formulation of Power Balance Model Elements Allow Partitioning of Energy Flow I p VPF Vgeo VWm VIR VLHI   0



Recent Improvements

VIND Inductive Drive from Poloidal Fields

d  VPF     PF     M V  R20 BV t coils dt

 R0 

0 I p  1 Le ℓ i 1 BV     p    4 R0   0 R 2 2 2 c( ) 1 0.98 2  0.49 4 1.47 6 1    M V (,  )  2 1   c( )  d( )  d( )  0.25 1 0.84 1.44 2 

Revised Lp, BZ models*



Moving plasma boundary



Neoclassical resistivity

LHI Drive

Ainj B ,inj VLHI Vinj eff  

Resistive Dissipation

  2 R0  VIR  I p Rp  I p    Ap  

Inductive Drive from Shape(t)

Plasma Magnetic Energy Change

dI p d dL  Vgeo    Le I p   Le  Ip e dt dt a( )  11.81   2.05  ln  8  dt   a( )(1  )   2.0  9.25  1.21  Le  0 R0 b( )  0.73  1 2 4  6 5  3.7 6  1    b( ) M.W. Bongard, APS-DPP 2016



VWm  

1 d 1 2  Li I p   I p dt  2

* S.P. Hirshman and G.H. Nielson 1986 Phys. Fluids 29 790 O. Mitarai and Y. Takase 2003 Fusion Sci. Technol. S. Ejima et al 1982 Nucl. Fusion 22 1313 J.A. Romero and JET-EFDA Contributors 2010 Nucl. Fusion 50 115002

C p2 ℓi  Li  0Vp

Equilibrium-Calibrated Inductance Model Improves Estimates of Non-Solenoidal VIND •

Maintaining radial force balance provides VIND –

Originally calculated via H-N formulae



Important to quantify contributions from shape, PF drive in LHI system design



Model equilibium database generated to test analytic formulae in realistic magnetic geometries – –



3

Poor partitioning of VIND between shape, VPF components found –



N = 331; 1.15 < A < 8; 1 0 1; 0.2 ℓ 0.75

However, total flux estimates in better agreement

Revised VIND model developed – –

Derived new coefficients in H-N formalism via fit to equilibrium database Weak dependence on , ℓ introduced

M.W. Bongard, APS-DPP 2016

S.P. Hirshman and G.H. Nielson 1986 Phys. Fluids 29 790

0-D Power Balance Model Used to Explore Projections for NSTX-U Startup •



Helicity dissipation (VIR) dependent on Te, realized electron confinement

Shape evolution for LFS LHI on NSTX-U

NSTX-U: Projected = 150 eV

Importance of VLHI, VIND depends on injector geometry, plasma growth scenario – Final plasma depends strongly on full time evolution



Injector geometry emphasizes different drive terms – LFS injection: VLHI early, VIND late – HFS: injection mainly VLHI



Need to explore plasma evolution with different dominant drive terms – Informs predictive model – Future: High Ip tests in both geometries

M.W. Bongard, APS-DPP 2016

LFS PEGASUS

HFS PEGASUS

NIMROD Describes Helical Current Stream Reconnection as Drive Mechanism



Divertor injection → minimal inductive drive

Divertor LHI Startup Shows suggestive commonality between experiment and NIMROD modeling

Current Ring

t = 2.93 ms

t = 2.91 ms

1. Streams follow field lines

2. Adjacent passes attract

M.W. Bongard, APS-DPP 2016

t = 2.95 ms

3. Reconnection pinches off current rings

NIMROD Simulation [O’Bryan PhD 2014]

J. O’Bryan, et al., Physics of Plasmas, 19, 080701 (2012) J. O’Bryan, C.R. Sovinec, Plasma Phys. Control. Fusion 56 064005 (2014)

May 2016 PEGASUS High-speed Imaging

– Infers NIMROD streams in edge

Ip [kA]

Reconnection-drive edge ion heating Internal Bz Measurements



NIMROD

Any stochastic reconnection region may be localized to edge

External Bz Phase Correlation 0.4

Inferred Stream Location Z [m]



PEGASUS

~b[mT]

Magnetics localize coherent streams in edge

~b [T/s]



Ip [kA]

Current Stream Interaction Manifests as Edge-Localized MHD Burst

-0.4 0.0

M.W. Bongard, APS-DPP 2016

R [m]

0.85

Reconnection-Driven Ion Heating Gives Ti > Te During LHI

Continuous ion heating from reconnection between collinear current streams – –

No effect on current drive efficiency

Ip [kA]



Impurity Ti(0) ~ 100 – 500 eV > Te routinely observed during LHI

(a)

100 0

dBz/dt [T/s]



Ion heating correlated with high frequency MHD fluctuations, not with discrete reconnection between helical streams

400

(b)

0 -400

Significant ion heating (~ few 0.1 MW)

16

18

20

22

24

26

Ion heating consistent with 2-fluid reconnection theory

HeII Ti [eV]

200 100

(c)

0 3%

(d)

Ti,⊥

Ti,||

5 - 70 kHz

2 1 0

(e)

0.02 %

0.2-0.4 MHz

0.01 0

25

26 Time [ms]

M.W. Bongard, APS-DPP 2016

27

28

Different MHD Activity Observed Between LFS and HFS Injection Geometry •

LFS (outboard) injection: – MHD initially continuous, large amplitude, n = 1 – Transitions to intermittent bursts later in the discharge – Burst spacing increase with Ip – Similar to NIMROD simulation



HFS (inboard) injection: – Continuous, large-amplitude n = 1 activity early on – Abrupt cut-off in large amplitude activity – Reduced n = 1 magnitude for remainder of discharge

• Differences suggest multiple current drive mechanisms present M.W. Bongard, APS-DPP 2016

LFS LHI

HFS LHI

LFS Local Helicity Injection Produces Core Te > 100 eV • Plasma shape grows inward from LFS injectors

Peaked Te(R) while Connected to Injectors 23.0 ms

24.2 ms

– Shape evolution generates VIND – VIND > VLHI during high-Ip phase

• Peaked Te(R) during drive phase (connected) – Not strongly stochastic – After disconnect radial compression drives skin current

• Core ne > 1019 m-3, Te ≥ 100 eV provides target for subsequent CD M.W. Bongard, APS-DPP 2016

Ip=0.08 MA

Connected

Ip=0.10 MA

Disconnected

Te (R, t) Remains Peaked for LFS Injection Geometry and Minimal VIND •

Plasmas with same LFS LHI system and static geometry evolution – Lower performance due to shape constraint •

Te(R) > 85 eV with majority LFS LHI-drive 23.03 ms

28.03 ms

Ip=46.7 kA

Ip=48 kA

High R0, reduced Aplasma

– VIND ~ 0 < VLHI; Te(0) ~ 80 eV



Te(R) peaked while driven by outboard LHI Contrast-enhanced high-speed image and fast boundary reconstructions

M.W. Bongard, APS-DPP 2016

Technical Challenges Arise for LHI Startup With HFS Injection Geometry •

Initial relaxation to tokamak state – More difficult for low Rinj, high Binj – Magnetic geometry constrained by injector clearance requirements





Current source behavior at increased Binj Plasma-material interactions – PMI on injector surfaces • inhibits Vinj • can damage injectors

– PMI on machine surfaces • Impedes reproducibility • More severe for HFS injection M.W. Bongard, APS-DPP 2016

Above: LHI plasma before and after relaxation Below: example of PMI on injector (left), eventually leading to insulator failure (right)

Poloidal Field Shaping Facilitates Relaxation at Full Toroidal Field (Binj = 0.23 T) •

Milestone for HFS LHI system achieved



Technical challenge with HFS injectors: –

Lower Rinj → higher BTF with respect to LFS system •



BTF increased ~ 10× over previous experiments •



→ more Bz for injector clearance (~ Bz/BTF) → Relaxation at constant Iinj more difficult

Poloidal field shaping key to full-field relaxation –

Reduces midplane |B| and maintains injector clearance



Limited by Iinj-deformed streams contacting vessel

M.W. Bongard, APS-DPP 2016

Vacuum Field

Deformed Field

Ip > 0.15 MA Achieved Via HFS Injection To Date •

VLHI ~ 1 kV increased 2 over previous HFS LHI experiments



Most operations at low field: – Binj = 0.046-0.092 T •

(20-40% of Pegasus maximum)

– Reduced PMI, easier relaxation



Full BTF scenarios developed – Binj = 0.23 T, ITF = 0.288 MA – Ip ≈ 0.1 MA – PMI more prevalent at high BTF



Injector geometry variants addressing observed PMI – Improvements found in each iteration

M.W. Bongard, APS-DPP 2016

HFS Helicity Injection Provides Non-Solenoidal Sustainment at High IN Constant geometry: minimal VIND



Low ITF 0.6 Ip



IN > 10 accessible

Access to IN > 14, ne ~ 1x19 m-3 with HFS Injection, BTF Rampdown

Ip (kA)



– Constant or ramped-down BTF

Potential for high T

IN



Ne

M.W. Bongard, APS-DPP 2016

Te

Ti_OV

ne (1019 m-3)

– Aided by anomalous ion heating

LHI Provides Access to High-T at A ~ 1 with Non-Solenoidal Sustainment and Anomalous Ion Heating •



Equilibrium reconstructions with kinetic constraints used to determine ≡2 / –

Matches external magnetics, ptot(0), and edge in Te(R)



Includes anomalous Ti(0)



Some caveats for these initial results Assumes closed flux surfaces inboard of injectors



Role of SOL edge current



Magnetics-only reconstructions scaled via comparison to those with kinetic constraints



Need full kinetic profiles in future

High T plasmas often terminated by disruption –





n = 1, low-m precursors

Expands accessible high IN, T space for tokamak stability studies at extreme toroidicity –

Campaign underway to document, extend to higher Ip



Improving LHI injector hardware to increase Ip, BTF access

M.W. Bongard, APS-DPP 2016

Initial Exploration of High-T Space

Progress in Non-Solenoidal Startup on Pegasus • LHI provides high Ip, non-solenoidal tokamak startup

– Flexible injection geometry balances VLHI and VIND drive, engineering constraints

– Improved power balance model suggests technique is scalable to larger devices – Questions remain on confinement and reconnection dynamics • Thomson scattering: Peaked Te, ne suggest favorable realized confinement

• New high-field-side injector systems exploring strong VLHI limit – Injector operation and relaxation to tokamak demonstrated at full TF (Binj ~ 0.25 T) – Completely VLHI driven startup and sustainment realized – Non-solenoidal Ip(t) via LHI enables access to stability tests at extreme toroidicity • Sustained operation at high IN, high T

• Present campaign: – Optimize HFS injector implementation to mitigate PMI at high BTF – Develop high Ip scenarios to test scalings in LFS, HFS geometries – Design CHI system for comparison studies (with PPPL, U. Wash) M.W. Bongard, APS-DPP 2016

Reprints Reprints of this and other PEGASUS presentations are available online at http://pegasus.ep.wisc.edu/Technical_Reports