Strike point control dynamics
Download
1 / 35

Strike Point Control Dynamics - PowerPoint PPT Presentation


  • 62 Views
  • Uploaded on

Office of Science. Supported by. Strike Point Control Dynamics. College W&M Colorado Sch Mines Columbia U Comp-X General Atomics INEL Johns Hopkins U LANL LLNL Lodestar MIT Nova Photonics New York U Old Dominion U ORNL PPPL PSI Princeton U SNL Think Tank, Inc. UC Davis

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about 'Strike Point Control Dynamics' - jariah


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
Strike point control dynamics

Office of

Science

Supported by

Strike Point Control Dynamics

College W&M

Colorado Sch Mines

Columbia U

Comp-X

General Atomics

INEL

Johns Hopkins U

LANL

LLNL

Lodestar

MIT

Nova Photonics

New York U

Old Dominion U

ORNL

PPPL

PSI

Princeton U

SNL

Think Tank, Inc.

UC Davis

UC Irvine

UCLA

UCSD

U Colorado

U Maryland

U Rochester

U Washington

U Wisconsin

Culham Sci Ctr

U St. Andrews

York U

Chubu U

Fukui U

Hiroshima U

Hyogo U

Kyoto U

Kyushu U

Kyushu Tokai U

NIFS

Niigata U

U Tokyo

JAEA

Hebrew U

Ioffe Inst

RRC Kurchatov Inst

TRINITI

KBSI

KAIST

ENEA, Frascati

CEA, Cadarache

IPP, Jülich

IPP, Garching

ASCR, Czech Rep

Egemen Kolemen,Princeton University

in colloboration with

D. Gates, C. Rowley, J. Kasdin, K. Taira

NSTX 2009 Advanced Scenarios and Control


Strike point control dynamics

Motivation: Density Reduction via Strike Point Control

• Density reduction depends on proximity of outer strike point to LLD-1

High d : ne reduced by 25%

Low d : ne reduced by 50%

LLD

LLD

Density

Reduction

R=0.65

R=0.84

R=0.84

R=0.65

Density

Reduction

10cm

15cm

20cm

10cm

15cm

20cm

Henry Kugel,

R. Maingi, ORNL

V. Soukhanovskii, LLNL

Shown for different LLD-1 widths

Shown for different LLD-1 widths


Strike point control dynamics

Preliminary Study: SISO System No Dynamics System

PF2L=2 kAmp/MAmp

PF2L=3 kAmp/MAmp

PF2L=4 kAmp/MAmp

  • The change in the strike point with different PF2L current (isolver)


Strike point control dynamics

PF2L Scan: Shot 115495

  • The change in the strike point for 115495.00501 with different PF2L current (isolver)

  • Change strike point location by changing PF2L while keeping the shape and PF1B, PF1AL, PF1AU, PF2U constant (let PF3L,PF5,PF3U vary to compensate).

  • While achieving (in order of importance):

    • Low Inner Gap, High Kappa

  • Scan by changing PF2L only:


Strike point control dynamics

PF2L Scan: Shot 120001

  • The change in the strike point for 120001.00650 with different PF2L current (isolver)

  • Change strike point location by changing PF2L while keeping the shape and PF1B, PF1AL, PF1AU, PF2U constant (let PF3L,PF5,PF3U vary to compensate).

  • While achieving (in order of importance):

    • Low Inner Gap, High Kappa

  • Scan by changing PF2L only:


Strike point control dynamics

PF1B Scan: Shot 120001

  • The change in the strike point for 120001.00650 with different PF1B current (isolver)

  • Change strike point location by changing PF2L while keeping the shape and PF1B, PF1AL, PF1AU, PF2U constant (let PF3L,PF5,PF3U vary to compensate).

  • While achieving (in order of importance):

    • Low Inner Gap, High Kappa

  • Scan by changing PF1B only:





Strike point control dynamics

PF1 versus PF2

For PF1B = 0.003

For PF2L = 0.0057

PF2L is 3-4 times more effective then PF1B in controlling the strike point


Strike point control dynamics

Aim: Design a Real Time Controller for the Strike Point Motion

  • Use the insight from the non-dynamic model to design a PID controller to keep the strike point at the center of LLD, with ~1 cm variation from the reference value.

  • Experiment:

    • Put perturbations in the PF1/PF2 requests & measure the strike point response.

    • Test and tune the strike point controller.

  • Study the compromise with respect to the loss in control for shape control and other control aims.

  • In this case, s=position and r=reference position of the strike point.


Strike point control dynamics

Experiment Procedure: Step Response and PID controller Motion

P

PF1/2

  • L = lag in time response

  • ΔCp (%) = the percentage change in output signal in response to the initial step disturbance

  • T = the time taken for this change to occur

  • N =         ; where N is the reaction rate

  • R =

  • Then, for a given perturbation (P)

    Ziegler–Nichols

    PID controller              


Strike point control dynamics

Experiment: Starting Condition Motion

Shot 115495.00501 is given as the requested profile.

This is an old shot.

Instead choose from shot 120001 with similar profile and same strike point.


Strike point control dynamics

Experiment Current Request: li Dependence Motion

  • Depending on li and strike point distance choose the control input.

    • Experiment runs ~ 8-10 shots for PF2L and~ 8-10 shots for PF1B

  • Use shot120001 with strike point between 0 and 25 cm for PF2L and 0 and ~15 cm for PF1B.

  • Strike Point distance start with 0 then 25 cm (15 cm for PF1B) then interpolate as much points as possible in the experiment day. 0, 25, 12, 6, 18...(0 15 7 3 12 … for PF1B)


Strike point control dynamics

Extra Motion


Strike point control dynamics

Extra Motion

dsf


Strike point control dynamics

Further Study: Multiple Control Input Variation Motion

Example x-point scan: Move it horizontally (Stefan Gerhardt)

The following scan varies multiple variables at the same time.

Try to keep the basic shape constant except for lower outer squareness.


Strike point control dynamics

Experiment Procedure: Step Response and Time Constant, Motion.

PF2L

PF2L

Choose two plasma shapes (and strike points) from the scan.

Stabilize the plasma for the 1st shape.

We know the coil currents need to take us to the 2nd equilibrium.

Put a step input for the coil currents that gives the 2nd equilibrium.

Measure the time constant of the strike point motion to the change in current.

Repeat for many other equilibrium.


Strike point control dynamics

Aim: Design a Real Time Controller for the Strike Point Motion

PID

controller

plant

sensor

  • Use the insight from the non-dynamic model to design a PID controller to keep the strike point at the center of LLD, with ~1 cm variation from the reference value.

  • Experiment:

    • Put perturbations in the PF2 request & measure the response of the strike point.

    • Test and tune the strike point controller.

  • Study the compromise with respect to the loss in control for shape control and other control aims.

  • In this case, s=position and r=reference position of the strike point.




Strike point control dynamics

Preliminary Study: No-Dynamics Model

PF3L=1 kAmp

 Example Shot and the change of the strike point with different PF3L current.

PF3L=2 kAmp

PF3L=4 kAmp



Strike point control dynamics

Experiment Procedure Motion

  • Choose two plasma shapes (and strike points) from the scan.

  • Stabilize the plasma for the 1st shape.

  • We know the coil currents need to take us to the 2nd equilibrium.

  • Put a step input for the coil currents that gives the 2nd equilibrium.

  • Measure the time constant of the strike point motion to the change in current.

  • Repeat for many other equilibrium.


Strike point control dynamics

Aim: Design a Real Time Controller for the Strike Point Motion

PID

controller

plant

sensor

  • Use the insight from the non-dynamic model to design a PID controller to keep the strike point at the center of LLD, with ~1 cm variation from the reference value.

  • Experiment (0.5 Day):

    • Put perturbations in the PF2 request & measure the response of the strike point.

    • Test and tune the strike point controller.

  • Study the compromise with respect to the loss in control for shape control and other control aims.

  • In this case, s=position and r=reference position of the strike point.





Strike point control dynamics


Strike point control dynamics

  • Liquid lithium divertor (LLD) on NSTX, enables experiments with the first complete liquid metal divertor target in a high-power device in 2009.

  • The location in the vacuum vessel is shown schematically in figure 1.

  • Reduced recycling with LLD.

  • The most important parameter that defines the density reduction is strike point position.

Background: Liquid Lithium Divertor & Edge Density

1. Schematic of NSTX showing location of Liquid

Lithium Divertor inside vacuum vessel

2. Edge density profiles calculating with UEDGE for NSTX Liquid Lithium Divertor assuming different recycling coefficients


Strike point control dynamics

  • A liquid lithium divertor (LLD) is being installed on NSTX, to enable experiments with the first complete liquid metal divertor target in a high-power device in 2009. The location in the vacuum vessel is shown schematically in Fig. 9. The LLD is a conic section with four 90-degree segments, each consisting of a 1.9 cm-thick copper plate with a 0.02 cm-thick stainless steel liner that is isolated toroidally with carbon tiles. Molybdenum will be plasma sprayed onto the liner in vacuum, to form a 0.01 cm-thick layer with 50% porosity. This will become the plasma-facing surface when filled with lithium, which will be kept liquid by resistive heaters in the plates. The present outer divertor (Fig. 1) consists of concentric rows of ATJ graphite tiles on copper baseplates. Lithium evaporated onto the tiles prior to a shot would solidify, and pump only while the hydrogenic atoms can react with the surface layer of the lithium coating. The LLD will replace part of these tiles with lithium that will be kept molten. Because the lithium will continue reacting with hydrogen or deuterium until it is volumetrically converted to hydrides,[12] the LLD is expected to provide better pumping than lithium coatings on carbon PFC’s. FIG. 9. Schematic of NSTX showing location of Liquid Lithium Divertor inside vacuum vessel Detailed edge plasma modeling has begun with the UEDGE transport code to simulate the effects of reduced recycling expected from the LLD.[13] The simulations start with adjusting the transport coefficients until the edge temperatures and densities match the data from the multipoint Thomson scattering diagnostic for existing NSTX plasmas. New profiles are then generated for a variety of recycling coefficients (Fig. 10). The results of the simulations have the same nonlinear radial dependence as the SOL measurements during high lithium evaporation. The simulations, however, do not show the linear density rise observed at the low lithium evaporation rate during NSTX experiments. This suggests that more work needs to be done on the transport modeling before further conclusions can be drawn from the UEDGE calculations.

Background: Liquid Lithium Divertor & Edge Density

Schematic of NSTX showing location of Liquid

Lithium Divertor inside vacuum vessel

FIG. 1. (left) Elevation of NSTX showing position of LITERs. (right) NSTX plan viewing indicating toroidal

location of LITERs and coating regions blocked by center stack.

FIG. 10. Edge density profiles calculating with UEDGE for NSTX Liquid Lithium Divertor assuming

different recycling coefficients


Lld 1 overview
LLD-1 Overview to enable experiments with the first complete liquid metal divertor target in a high-power device in 2009. The location in the vacuum vessel is shown schematically in Fig. 9. The LLD is a conic section with four 90-degree segments, each consisting of a 1.9 cm-thick copper plate with a 0.02 cm-thick stainless steel liner that is isolated toroidally with carbon tiles. Molybdenum will be plasma sprayed onto the liner in vacuum, to form a 0.01 cm-thick layer with 50% porosity. This will become the plasma-facing surface when filled with lithium, which will be kept liquid by resistive heaters in the plates. The present outer divertor (Fig. 1) consists of concentric rows of ATJ graphite tiles on copper baseplates. Lithium evaporated onto the tiles prior to a shot would solidify, and pump only while the hydrogenic atoms can react with the surface layer of the lithium coating. The LLD will replace part of these tiles with lithium that will be kept molten. Because the lithium will continue reacting with hydrogen or deuterium until it is volumetrically converted to hydrides,[12] the LLD is expected to provide better pumping than lithium coatings on carbon PFC’s.

• Geometry:, 0.01cm thick, 50% porous Mo flame-sprayed on 0.02 cm SS brazed to 1.9 cm Cu. 20 cm wide, Ri = 0.65 m, Ro = 0.85 m. Li loading via LITER.

• Operating temperature: typ. =205°C

• Power Handling: SNL thermal analysis for the cases for the strike point on the

LLD with peak Li temperature set at 400 °C,

- can sustain a peak of ~2MW/m2 for 10s and 4 MW/m2 for ~3s.

- Less Li, higher heat transfer.

• Interlocks: similar to LITER (Heater Power off during fields, vacuum event, if temperature >set point,…..)

- presently no automatic LLD-1 thermal interlock to interrupt plasma position or NBI

- Fast and slow IR cameras will monitor LLD-1 front-face temperature.

- Visible cameras will monitor divertor region.

Li thermal conductivity is low. (~ W/m-°K 400 Cu, 150 Mo, 45 Li, 15 SS)


Strike point control dynamics

Day-1: Slowly increase Li and Test Pumping by LLD-1 on Outer Divertor to Provide Density Control for Inner Divertor Broad SOL Dα Profile, High δ Plasmas

• Density reduction will depend on proximity of outer strike point to LLD-1

High d : reduce ne by 25%

Soukhanovskii LLNL

Density

Reduction

LLD

10cm

15cm

20cm

R. Maingi, ORNL

Shown for different LLD-1 widths


Strike point control dynamics

LLNL Divertor to Provide Density Control for Inner Divertor Broad SOL D

Day-2: Test Pumping by LLD-1 with Strike Point on Outer Divertor to Provide Density Control for Low δ Plasmas

Low d : reduce ne by 50%

Density

Reduction

LLD

10cm

15cm

20cm

R=0.65

R=0.85

R. Maingi, ORNL

Shown for different LLD-1 widths


Strike point control dynamics
Due to High Flux Expansion, Pumping by 20 cm Wide LLD-1 on Outer Divertor Will Provide Density Control for Both High and Low δ Plasmas

• Density reduction will depend on proximity of outer strike point to LLD-1

High d : ne reduced by 25%

Low d : ne reduced by 50%

LLD

LLD

Density

Reduction

R=0.65

R=0.84

R=0.84

R=0.65

Density

Reduction

10cm

15cm

20cm

10cm

15cm

20cm

R. Maingi, ORNL V. Soukhanovskii, LLNL

Shown for different LLD-1 widths

Shown for different LLD-1 widths