Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out

RF Optimization of the ITER ICRF Antenna Plug

VIEWS: 14 PAGES: 44

									Validation of the Electrical Properties of the
ITER ICRF Antenna using Reduced-Scale
Mock-Ups


              P. Dumortier, F. Durodié, D. Grine,
              V. Kyrytsya, F. Louche, A. Messiaen,
              M. Vervier, M. Vrancken

              LPP-ERM/KMS, Brussels, Belgium, CYCLE



Work supported by F4E-2009-GRT-026 grant
P. Dumortier et al.       Slide 1          IC IDRM2: 18 May 2011
                                     Outline
   ITER ICRH antenna - RF requirements
   Actual reference antenna
       Design choices & features
       Optimization of the antenna
       Frequency response
   Reduced-scale mock-ups
       Rationale for the use of reduced-scale mock-ups
       Phase 1: Optimization of one triplet
            Validation of antenna box optimization
       Phase 2: Validation of optimized model
            Validation of optimized front face and 4-port junction
            Broadbanding by service stub
       Phase 3: Mock-up of full antenna
            Performance evaluation
            Grounding
      Matching and Decoupling system test
   Conclusions


        P Dumortier et al       Slide 2   19th RF Top Conf: 3 June 2011
                        What does IO request ?
   ITER ICRH antenna - RF requirements
   Actual reference antenna
       Design choices & features
       Optimization of the antenna
       Frequency response
   Reduced-scale mock-ups
       Rationale for the use of reduced-scale mock-ups
       Phase 1: Optimization of one triplet
            Validation of antenna box optimization
       Phase 2: Validation of optimized model
            Validation of optimized front face and 4-port junction
            Broadbanding by service stub
       Phase 3: Mock-up of full antenna
            Performance evaluation
            Grounding
      Matching and Decoupling system test
   Conclusions


        P Dumortier et al       Slide 3   19th RF Top Conf: 3 June 2011
      ITER ICRH antenna – key RF requirements
   Nominal power: 20 MW per antenna (2 antennas)
   Frequency range: 40 – 55 MHz
   Phased antenna array (6 poloidal x 4 toroidal array) for radiated
    power spectrum control:
      Control of toroidal phase differences

      Control of current ratio between columns of straps

   Maximum allowed voltage: Vmax=45kV
   Maximum allowed electric field:
      Torus vacuum: Emax=2kV/mm perpendicular to Btor ;
                          Emax=3kV/mm parallel to Btor
      Private vacuum: Emax=3kV/mm

   Quasi CW operation
   Location: equatorial port plug




        P Dumortier et al   Slide 4   19th RF Top Conf: 3 June 2011
    What is the ITER antenna looking like ? Why ?
   ITER ICRH antenna - RF requirements
   Actual reference antenna
       Design choices & features
       Optimization of the antenna
       Frequency response
   Reduced-scale mock-ups
       Rationale for the use of reduced-scale mock-ups
       Phase 1: Optimization of one triplet
            Validation of antenna box optimization
       Phase 2: Validation of optimized model
            Validation of optimized front face and 4-port junction
            Broadbanding by service stub
       Phase 3: Mock-up of full antenna
            Performance evaluation
            Grounding
      Matching and Decoupling system test
   Conclusions


        P Dumortier et al       Slide 5   19th RF Top Conf: 3 June 2011
          Actual Reference Design – ICRH Antenna
      24 straps grouped in triplets → 6x4 array                                 B17 M. Nightingale
                                                                Port plug wall
                                                                                   Port Plug Flange
                        4-Port Junction             RF grounding
                       (arms: Z01=15Ω)                                                 Neutron shield
Antenna box



                                                                                          Feeding line
  Faraday                                                                                  (Z02=20Ω)
   screen
                                                                                          Service stub
                2160




                                                                                          (Z0SSt=15Ω)


Short current
   straps

Short circuit

                                                                   RF vacuum windows

            P Dumortier et al             Slide 6      19th RF Top Conf: 3 June 2011
                   Design choices and features
   Short low-inductance straps
       Lower voltage on straps, better radiation efficiency
        → high power density
   Z0F=15Ω
       Trade-off between maximizing coupling, minimizing VSWR and
        minimizing Emax
   Segmentation (3 straps)
       Minimizing Emax and Vmax
   Passive 4-port junction (4PJ)
       Connects 3 straps in parallel to one feeding line
       Reduction of the number of feeding lines
       No active/moving component in the antenna
       Currents are in phase
        → triplet of straps is seen as a long strap with uniform current by plasma
   Service stub
       Broad-banding of the RF response curve
   Outside antenna:
       20Ω-50Ω transition at Vmax to reduce VSWR
       Decoupling and matching network (Double Stub Tuner)
            Reduction of mutual coupling effects
            Control current distribution of array to impose required current spectrum
        P Dumortier et al         Slide 7    19th RF Top Conf: 3 June 2011
                        Antenna triplet RF model




                                      2
                               Vmax               1
   In all regions: P  Gmin               and Gmin 
                              2                 S Z0
   Assumption for optimization: critical parameter is Vmax
     Improve achievable Vmax by design (rounding edges,…)
     For given Vmax and Imax,lines → Maximize Gmin
      to maximize P and minimize S (SWR) for given Z0
   Example: if Gmin ↑ by 20% → S  by 20%
    For given |Vmax|:                                   For given P:
    → P ↑ by 20%                          Vmax 1        → Vmax  by 10%
                                                                                            2P
    → same |Imax, lines| and IF                        → |Imax, lines|  by 10% and IF 
                                      XF  Z 01
                                       2     2                                              RF


        P Dumortier et al        Slide 8           19th RF Top Conf: 3 June 2011
                  Triplet frequency response




                                                                    RF
                                                      Gmin1 
                                                                   2    2
                                                                 X F  Z01
                                                            2      2
                                                      (for RF  X F )




   Gmin1 determined by ZF and Z01
   If ideal TL 4PJ is at Vanti-node for all frequencies: Gmin2,max = 3Gmin1

        P Dumortier et al      Slide 9   19th RF Top Conf: 3 June 2011
                  Triplet frequency response


             Gmin2,max = 3
             Gmin1




                  Bandwidth function of
                   Z01/XF and Z02/Z01

                                          fopt solution of tan(β<l1>)= Z01/XF



                             fopt


   4PJ fixed in space → acts as a single frequency filter
      Maximum at fopt, when electrical junction point is at Vanti-node
      Tune response by choosing Z01, <l1>, Z02
        P Dumortier et al           Slide 10     19th RF Top Conf: 3 June 2011
                 Triplet frequency response


                   Turning
                    point


                                                        Gmin3 response curve turns
                                                        around a turning point and its
                                                        slope is determined by LSSt



                                                        fTP determined by L4PJ-SSt
                                                        (Turning point remains on Gmin2 curve)



                           fTP


   Broad-banding → Band-pass filter response in feeding line
   Gmin3 response shape determined by Z0SSt, L4PJ-SSt and LSSt

       P Dumortier et al         Slide 11   19th RF Top Conf: 3 June 2011
       Frequency response can be optimized…
   By acting on front face geometry of the antenna (↔ Gmin1)
    → Strap width, box depth, vertical septum recess, profiling…
                 RF           Partly due to external medium and partly due
    Gmin 1                   to antenna box geometry




                                                                                         Impact on Iant
               X F  Z01
                 2    2




                                                                                           Coupling
                            Mainly due to antenna box geometry, weak
                            dependence on plasma conditions

                                         VF       VF
    But |IF| ↑ when XF ↓ because IF          
                                      ZF          XF
     Trade-off between Iant,max and Vmax
    Modeling (MWS, Topica, Antiter II) + Mock-Up Phase 1               B14 – F. Louche
   By acting on the 4-port junction (↔ Gmin2)              B16 – F. Durodié




                                                                                         No impact on Iant
    → Optimal frequency solution of tan(β<l>)= Z01/XF




                                                                                          Prematching
    → Bandwidth function of Z01/XF and Z02/Z01
    → Optimize 4PJ geometry
   By acting on Service Stub: Z0SSt, L4PJ-SSt and LSSt (↔ Gmin3)
    Modeling (MWS, TL) + Mock-Up Phase 2

       P Dumortier et al      Slide 12     19th RF Top Conf: 3 June 2011
    RF properties validation using RF mock-ups
   ITER ICRH antenna - RF requirements
   Actual reference antenna
       Design choices & features
       Optimization of the antenna
       Frequency response
   Reduced-scale mock-ups
       Rationale for the use of reduced-scale mock-ups
       Phase 1: Optimization of one triplet
            Validation of antenna box optimization
       Phase 2: Validation of optimized model
            Validation of optimized front face and 4-port junction
            Broadbanding by service stub
       Phase 3: Mock-up of full antenna
            Performance evaluation
            Grounding
      Matching and Decoupling system test
   Conclusions


        P Dumortier et al      Slide 13   19th RF Top Conf: 3 June 2011
      Why using reduced-scale RF mock-ups ?
   Relatively cheap way to validate the RF simulations results
   Same impedances as full scale model if ratio between dimensions
    and vacuum wavelength kept constant (except for skin effect losses)
    → Operating frequency needs to be multiplied by reduction scale
       factor
   Realistic plasma-like load conditions obtained by putting a medium
    with a large dielectric constant, such as water, in front of the antenna
    → Load variations obtained by moving water load in front of
       antenna mock-up
   No need for large water load
    → Small concentration of salt added to water allows wave
       absorption (avoid reflections on walls leading to standing waves)




       P Dumortier et al    Slide 14   19th RF Top Conf: 3 June 2011
         Phase 1 – RF optimization validation
   Based on October 2007 design (1 strap triplet and triangular 4PJ)




   Measurements/simulations performed:
      Scan in distance mock-up – water load
      Scan in strap width and antenna box depth
             3 different strap widths
             3 different box depths
              → 9 sets of straps
        Impact of Faraday screen
        Impact of vertical septum recess
        P Dumortier et al        Slide 15   19th RF Top Conf: 3 June 2011
                    Phase 1 – Set-up




P Dumortier et al     Slide 16   19th RF Top Conf: 3 June 2011
                       Scan in load conditions
                                                                   2
                                                            Vmax
                                                 P  Gmin
                                                              2




   Good agreement with simulations (except when load against the antenna) but
        MWS: importance of BC and meshing to obtain quantitative agreement
        Measurements: importance of correct de-embedding of 20Ω-50Ω transition
   Expected frequency response - not centered in ITER band because fixed 4PJ
   Only slight frequency shift with change in loading

         P Dumortier et al     Slide 17   19th RF Top Conf: 3 June 2011
     Scan in strap width and antenna box depth
   Scan in strap width W : XF ↓ when W ↑ → shift towards higher f




   Scan in antenna box depth D : XF ↑ when D ↑ → shift towards lower f




   Good agreement with numerical simulations

       P Dumortier et al   Slide 18   19th RF Top Conf: 3 June 2011
                               Numerical optimization
       Optimization of strap width and antenna box depth




       Not very sensitive to W and D when close to optimum


           P Dumortier et al       Slide 19   19th RF Top Conf: 3 June 2011
Phase 2 – Optimized geometry and service stub
   Optimized geometry (reference June 2008)




   Measurement/simulations performed:
      Set of spacers to scan:
            4-port junction arms length
            Service stub insertion point
        Scan in 15Ω service stub length
        Scan in distance mock-up – water load
        No Faraday screen

        P Dumortier et al       Slide 20    19th RF Top Conf: 3 June 2011
      Scan in 4-port junction arms’ length




   Frequency response can be centered in band by acting on 4PJ arms’
    lengths
   But this affects Gmin2 as Gmin2,max = 3 Gmin1 (in case of ideal TL 4PJ)

     P Dumortier et al      Slide 21   19th RF Top Conf: 3 June 2011
           Scan in 4-port junction arms’ length
   Comparison Measurements – MWS and TL simulations




       P Dumortier et al   Slide 22   19th RF Top Conf: 3 June 2011
         Broad-banding by service stub




P Dumortier et al   Slide 23   19th RF Top Conf: 3 June 2011
      Comparison measurement – TL model
   Excellent representation of service stub insertion by
    Transmission Line modeling




       P Dumortier et al   Slide 24   19th RF Top Conf: 3 June 2011
  Impact of service stub insertion point




         Change LSSt → Turn around “turning point”
    Change L4PJ-SSt → Move “turning point” along Gmin curve

P Dumortier et al     Slide 25   19th RF Top Conf: 3 June 2011
                         Voltage pattern
             40MHz                                         55MHz




For Vmax in the MTL of Vmax3 = 45 kV the voltage can be higher:
      in 4PJ

      in section between 4PJ and SSt

      in SSt

  → Need to operate at Vmax3 < 45kV for some frequency ranges
     P Dumortier et al     Slide 26   19th RF Top Conf: 3 June 2011
                          Power limitation




                                                              2
                                                       Vmax
                                            P  Gmin
                                                          2




   Active power for 1 triplet and given experimental load condition
   Power from Gmin and Vmax=45kV in all regions of antenna
   Infinite extent regions

      P Dumortier et al     Slide 27   19th RF Top Conf: 3 June 2011
                          Power limitation




   But regions 1, 2 and service stub of finite extent
    → Vmax corresponding to Gmin may not be reached
    → Voltage margin
   Power constrained to Vmax=45kV reached in every region
      P Dumortier et al     Slide 28   19th RF Top Conf: 3 June 2011
                          Power limitation




                                                      = Minimum of dotted lines




   Active power for 1 triplet and given experimental load condition
   Maximum power constrained to Vmax=45kV in all regions of antenna
   Other power limitations exist (electric fields, current) B16 – F. Durodié
   Very sensitive to LSSt, less to L4PJ and rather insensitive to L4PJ-SSt
      P Dumortier et al     Slide 29   19th RF Top Conf: 3 June 2011
    Phase 3: Full antenna RF characterization
   RF characterization of full array
   Impact of Faraday Screen on coupling
   Effect of vertical septa recess
   Effect of grounding




       P Dumortier et al   Slide 30   19th RF Top Conf: 3 June 2011
    RF performance – Preliminary measurement
   Expected RF frequency response (relative)
   Skew in 0π0π response due to too low KD,water for low f     B18 – S. Champeaux
   Total radiated power for Vmax3=45kV in feeding line and fixed water load position




         P Dumortier et al     Slide 31   19th RF Top Conf: 3 June 2011
           Power distribution amongst triplets
   Array currents controlled but strong variation in active radiated power to
    straps for the different triplets due to mutual coupling
      Active radiated power can be negative for some triplets




          Crucial importance of good decoupling network
         P Dumortier et al     Slide 32    19th RF Top Conf: 3 June 2011
                  Comparison with modeling




   Preliminary analysis show fair agreement between measurements
    and modeling




        P Dumortier et al   Slide 33   19th RF Top Conf: 3 June 2011
             Impact of vertical septum recess
      Reference                 Reference + 20mm                Reference + 40mm




   Significant gain in coupling by recessing further the vertical septa
      Less gain for dipole (0π0π)
   Mutual coupling between strap triplets increased
     → check whether level is tolerable by decoupling network
     → evaluate impact on tuning elements (range, current rating, …)
   Frequency shift towards lower frequencies
     → coupling will increase further when centering in the frequency band
   Different positions of service stub for internal and external triplets
   Note: slight uncertainty on exact position on temporary set-up
   Note: full VS recess, i.e. all vertical septa recessed

         P Dumortier et al    Slide 34   19th RF Top Conf: 3 June 2011
                   Effect of Faraday Screen
   Limited decrease of coupling observed
   Slight shift of towards higher frequencies




       P Dumortier et al      Slide 35   19th RF Top Conf: 3 June 2011
Effect of grounding on RF frequency response
   Mind the gap: 20mm clearance gap between the antenna plug and the vacuum vessel
    may lead to mode excitation in the gap
                                                                          A54 – V. Kyrytsya
   Frequency response is essentially affected for monopole phasing
    → avoid monopole excitation due to unequal anti-node voltage distribution




         P Dumortier et al       Slide 36    19th RF Top Conf: 3 June 2011
             Phase 5: Matching and Decoupling
   Performed on design 2003 mock-up at present
   CT option (back-up) D. Grine – RF2009
        Adjacent poloidal triplets are connected in shunt in the circuit via a T-junction
        Matching stubs to adjust the conjugate pairs
        6 Toroidal decouplers are preset (vacuum load) capacitors
        11 feedback actuators for phase control of voltage anti-nodes (other parameters
         preset)
            Tuning stubs: 8
            Generator relative phase: 3
       Fully simulated, implemented and tested on mock-up
   Hybrid option (reference) B15 – D. Grine
       Adjacent poloidal triplets are connected to 3dB hybrid splitter
       Double stub tuning on each triplet line
       23 active feedback actuators for full antenna (other parameters preset)
            Double Stub Tuners: 8 x 2 = 16 actuators
            Poloidal decouplers: 4 actuators
            Toroidal (CD phasings) or Poloidal-Toroidal (Heating phasings) decouplers: 3 actuators
        Fully simulated, implemented and tested (CD case) on mock-up
   Decoupler tuning by voltage anti-node on adjacent lines comparison
   Starting conditions important for stability
    → in practice, starting from the vacuum conditions is OK

        P Dumortier et al         Slide 37      19th RF Top Conf: 3 June 2011
                3dB Hybrid Mock-up Implementation
                                                                   Left: 10 decouplers
                                                                   (green) between the
                                                                   ports A-H and 16
                                                                   matching stubs (red) on
                                                                   the 8 heating lines;

                                                                   Right: feedback system
                                                                   with software-based
                                                                   controller and associated
                                                                   hardware



Strap                                    Voltage anti-node ports
array                                                              Mock-up of the
                                                  Decouplers
                                                                   ITER antenna and
                                                  Double           the 3dB hybrid
        Water load                                Stub             matching circuit
        (removed)                                 Tuners

                                                  3dB Hybrids
                                                  & DST probes




          P Dumortier et al   Slide 38       19th RF Top Conf: 3 June 2011
                                    3dB Hybrid
   Impedance tuning is done via one of three developed algorithms: B15 – D. Grine
      Bang-Bang: same as CT.
      Fast Bang-Bang: improves on former by simultaneously steering the
       two tuning stubs.
      Real/Imag: steers the double stub tuner using analytically derived
       formulas based on the measured reflection coefficient at the hybrid
       outputs, both in magnitude and in phase.




         Bang-Bang                    Fast Bang-Bang                    Real/Imag

       Simulation of RA,eff excursion from 2.25Ω/m to 5Ω/m and current drive. |HO|
         for the heating lines A-H as a function of the normalized iterations n/NBB,
      where NBB is the number of iterations required for the Bang-Bang algorithm to
       Resilience study started converge

        P Dumortier et al         Slide 39    19th RF Top Conf: 3 June 2011
                            Conclusions
   Experimental measurements on mock-up validated simulation results
       Gain confidence in design optimization and expected
        performance
   Frequency response and broad-banding by service stub confirmed
   Coupling loss due to presence of Faraday screen is limited
   Coupled power very sensitive to vertical septum position
       Beneficial to recess further the vertical septum
       Need integration with decoupling and matching network
   Vital importance of decoupling network confirmed
   Grounding
       Importance of correct grounding (essentially for monopole)
   Matching and decoupling
       Suitable algorithms found and implemented
       Tested on CT and hybrid options on full array (CD case only for
        hybrid option)

       P Dumortier et al   Slide 40   19th RF Top Conf: 3 June 2011
                     Some related contributions
   R02 – R. D’Inca – Arc detection for the ICRF system on ITER
   I06 – R. Maggiora – Mitigation of parallel RF potentials by an appropriate antenna design
    using TOPICA
   I17 – E. Lerche – ICRF scenarios for ITER’s half field phase
   A54 – V. Kyrytsya – Detailed modeling of grounding solutions for the ITER ICRH antenna
   B11 – A. Mukherjee – Status of R&D activity for ITER ICRF power source
   B12 – D. Rasmussen – ITER ICH transmission line and matching system prototype
    development
   B14 – F. Louche – 3D modeling and optimization of the ITER ICRH antenna
   B15 – D. Grine – Results of the implementation on a mock-up of the full 3dB hybrid
    matching option of the ITER ICRH system
   B16 – F. Durodié – Optimization of the layout of the CYCLE ITER antenna port plug and
    its performance assessment
   B17 – M. Nightingale – Design of the ITER ICRF Antenna
   B18 – S. Champeaux – High dielectric dummy loads for ITER ICRH antenna laboratory
    testing: numerical simulation of one triplet loading by ferroelectric ceramics
   B19 – JM. Bernard – TITAN: a test bed facility for ICRH antenna and components of ITER
   B25 – D. Rathi – A simple coaxial ceramic based vacuum window for Vacuum
    transmission line of ICRF system
   B30 – A. Messiaen – Influence of the edge plasma profile and parameters on the
    coupling of an ICRH antenna. Application to ITER.
   B31 – D. Milanesio – Analysis of the impact of antenna and plasma models on RF
    potentials evaluation




        P Dumortier et al         Slide 41     19th RF Top Conf: 3 June 2011
                       Scan in load conditions
   Estimate of strap input impedance




Effective strap input resistance        Effective strap input reactance
significantly varies with load          almost insensitive to load
(distance mock-up – water load)         (distance mock-up – water load)




       P Dumortier et al   Slide 42   19th RF Top Conf: 3 June 2011
           4-port junction passive distribution
   Estimate of input strap voltage and current on the 3
    straps of one triplet




    Excellent passive power distribution operated by 4-port junction




       P Dumortier et al   Slide 43   19th RF Top Conf: 3 June 2011
                                           Conjugate-T
     Impedance tuning for the CT is done via the Bang-Bang algorithm: an
      ad-hoc (trial&error) approach using the magnitude of the reflections
      after the T and steering only one stub at a time for each CT
                          Initial wrong
                          decisions

  Matching                                     Phase
feedback on                                 feedback on
  mock-up                                     mock-up




          Simulation of load-resilience at the               Measurement of load-resilience at the
      generators for RA,eff ≈ 2.25Ω/m and Current          generators for RA,eff ≈ 2.25Ω/m and Current
                          Drive                                                Drive
              P Dumortier et al           Slide 44        19th RF Top Conf: 3 June 2011

								
To top