Stefan Fink: MATEFU Insulation co-ordination and high voltage testing of fusion magnets Le Chateau CEA Cadarache, France April 7th, 29 Insulation co-ordination Some principle considerations of HV testing Testing of ITER TF Model Coil ITER TF
Insulation co-ordination Insulation co-ordination is the selection of test voltage(s) in relation to the operating voltages and overvoltages which can appear on the system. System analysis Voltage value, waveform Representative voltages and overvoltages Test voltages Multiplying with factors Example in conventional HV engineering: waveform for a standard lightning impulse Voltage value, waveform, test time
System analysis: RL discharge on a TF coil Current with initial value I L (t = ) = I = 5 ka must be decreased to A in case of a quench L 1 2 R U kv 1 U I 5 I ka 1H.1 8 4 6 3 4 2 U = R * I I = I * e -t / τ = I * e-t / (L / R) => U =.1 Ω * 5 ka = 5 kv High voltage (HV)! 2 1 2 3 4 t 5 s 1 Increase of the voltage is in range of few ms or faster => TF coil is a high voltage impulse coil => Testing of coil and coil components only with a DC test is not sufficient
Representative voltages for TF coil discharge Difficult to make a single HV test which is relevant for all voltages (and overvoltages) which may appear on the coil => A set of tests with different waveform is used Direct voltage ("DC") Winding 1 2 Case Alternating voltage ("AC") Winding 1 2 Case Impulse Winding 1 2 Case Representative for fall Simple, cheap Low destructive Representative for increase if arc chute breakers are used Non destructive insulation diagnostic possible (e. g. partial discharge (PD)) Representative for fast excitations (fast switching, faults) Most representative Stresses all types of insulation
General aspects of HV testing of large devices Large devices may have internal overvoltages if they are subjected to fast excitations => calculation of transient behaviour: Non linear voltage distribution? Oscillations? Non destructive test methods => Partial discharge measurement 2 kv transformer of a 5 ka power supply
Special aspects of HV testing of Paschen tight apparatus A Paschen tight device can be operated independently of the surrounding dielectric properties (e. g. during vacuum breakdown). The ITER TFMC was designed with solid insulation covering completely the HV areas. The insulation is covered with conductive paint. This paint is grounded. Verification if a coil is Paschen tight is performed by HV DC testing with the transition of the Paschen curve of the surrounding air in the cryostat at room temperature. Current Lead Paschen tight apparatus Insulated test sample covered with conductive paint Undefined gas or vacuum
Some special aspects of HV testing of cryogenic apparatus Fault detection under cryogenic conditions is expensive and time consuming => make pre-tests at room temperature The dielectric strength of the cooling material may be a weak point under room temperature testing of cryogenic apparatus => increase the pressure or replace He by N 2 or SF 6 in cooling channels with insulation breaks
ITER Toroidal Field Model Coil (TFMC) Coil parameters: Rated current 8 ka Rated voltage +5 kv / -5 kv Double pancakes 5 Turns per pancake 1 (or 9 for outermost) Coil Case Winding Pack 3 different insulation types: Conductor insulation Radial plate insulation Ground insulation Cross Section Design of ITER TFMC
FEM and network model for ITER TFMC 2D-FEM model of ITER TFMC as basis for calculation of the lumped elements of network model University Karlsruhe Network model of ITER TFMC University Karlsruhe
Results of transient calculation for TFMC G(f) 3,5 3, 2,5 2, 1,5 1,,5, 1 2 3 4 5 Frequency [khz] Transfer function at node 1 of the ITER TFMC network model with symmetric voltage excitation ±5 kv Case 1 Case 2 Case 3 First resonance frequency appears at 29 khz for the relevant cases 2 and 3. (This was later conformed by low voltage / high frequency measurement on ITER TFMC.) The selected configuration with connection of the radial plate by 1.2 MΩ resistors and a symmetrical grounding gives no relevant overvoltages for rise times above 2 µs => No high overvoltages expected for all prepared HV tests
Typical HV tests for ITER TFMC DC test on ground insulation Impulse test DC test on ground insulation DC and AC test on ground, radial plate and conductor insulation without room temperature instrumentation cables DC test on ground insulation Radial plate insulation Conductor insulation Ground insulation Conducto Radial plate Grounded case DC test voltage value for ground insulation was 1 kv (test voltages for other insulation types and waveforms had been lower) Tests were performed at room and cryogenic temperature AC tests included partial discharge measurement
Results of HV tests of ITER TFMC at room temperature All tests under ambient conditions were passed successfully During Paschen test it was found that TFMC is not Paschen tight 2 potential fault locations were found, Tedlar tapes were forgotten to remove during manufacturing at one location Fault location at helium inlet tubes
Results of HV tests and HV discharge on ITER TFMC at cryogenic temperature Breakdown strength for AC and esp. impulse testing under cryogenic conditions does not fulfil the specification High current discharge with I = 8 ka and U < 1 kv was possible High voltage discharge was reduced from +5 kv / -5 kv to / 4.4 kv => ITER TFMC does not fulfil the HV specification U kv 5 4 3 2 1 SC1116.QDA U plus terminal 1 2 3 t µs Breakdown during an impulse test with 5 kv at the plus terminal
ITER TF Coil design parameters: Rated current 68 ka Voltage @ fast discharge 3.5 kv Number of coils 18 Double pancakes / coil 7 Number of turns / pancake 11 (outer DP: 3, 9) ITER TF coils Cross section of an ITER TF coi
Detailed network model of ITER TF INCLUDE: Kopplungen_1kHz.txt C_P1_GE C_P2_GE C_P3_GE C_P4_GE C_P5_GE C_P6_GE C_P7_GE Lumped elements of the coil(s) are calculated with 2D-FEM for different frequencies PARAMETERS: C_Lage1 = 8.9nF C_Lage2 = 81.8nF C_Lage3 = 82.7nF C_Lage4 = 83.6nF C_Lage5 = 84.5nF C_Lage6 = 85.3nF C_Lage7 = 86.2nF C_Lage8 = 87.nF C_Lage9 = 87.9nF C_Lage1 = 88.8nF C_Lage11 = 89.6nF PARAMETERS: L_Lage1 = 1.4485uH L_Lage2 = 1.543uH L_Lage3 = 1.5527uH L_Lage4 = 1.644uH L_Lage5 = 1.6513uH L_Lage6 = 1.6977uH L_Lage7 = 1.7421uH L_Lage8 = 1.7865uH L_Lage9 = 1.8247uH L_Lage1 = 1.8716uH L_Lage11 = 1.914uH PARAMETERS: R_Lage1 = 797.4u R_Lage2 = 85.8u R_Lage3 = 814.32u R_Lage4 = 822.96u R_Lage5 = 831.6u R_Lage6 = 839.16u R_Lage7 = 847.8u R_Lage8 = 856.44u R_Lage9 = 865.8u R_Lage1 = 873.72u R_Lage11 = 882.36u PARAMETERS: R_Anbindung = 1k {C_Lage9} {C_Lage8} 132.97n 22.5n 22.5n 22.5n 22.5n 22.5n C_P1_P2 C_P2_P3 C_P3_P4 C_P4_P5 C_P5_P6 C_P6_P7 147.24n 147.24n 147.24n 147.24n 147.24n 147.24n R23 R34 R45 R56 R67 R78 R89 R1 R111 R122 C23 C34 C45 C56 C67 C78 C89 C1 C111 C122 {R_Lage11} {C_Lage11} L23 L34 L45 L56 L67 L78 L89 L1 L111 L122 {L_Lage11} R22 R33 R44 R55 R66 R77 R88 R99 R11 R121 {R_Lage1} C22 C33 C44 C55 C66 C77 C88 C99 C11 C121 {C_Lage1} L22 L33 L44 L55 L66 L77 L88 L99 L11 L121 {L_Lage1} R12 R21 R32 R43 R54 R65 R76 R87 R98 R19 R12 R131 C12 C21 C32 C43 C54 C65 C76 C87 C98 C19C12 C131 {R_Lage9} L12 L21 L32 L43 L54 L65 L76 L87 L98 L19 L12 L131 {L_Lage9} R2 R42 R64 R86 R18 R13 R11 R31 R53 R75 R97 R119 C11 {R_Lage8} C2 C31 C42 C53 C64 C75 C86 C97 C18 C119 C13 132.97n Detailed network models in were established for different frequencies PARAMETERS: C_Messkable = 24nF {C_Lage7} {C_Lage6} L11 L2 {L_Lage8} R1 R19 C1 {R_Lage7} C19 L1 L19 {L_Lage7} R9 R18 C9 C18 {R_Lage6} L9 L18 {L_Lage6} R17 R8 C8 {R_Lage5} C17 L31 R3 C3 L3 R29 C29 L29 R28 C28 L42 L53 R41 R52 C41 C52 L41 L52 R4 R51 C4 C51 L4 L51 R39 R5 C39 C5 L64 R63 C63 L63 R62 C62 L62 R61 C61 L75 R74 C74 L74 R73 C73 L73 R72 C72 L86 L97 R85 R96 C85 C96 L85 L96 R84 R95 C84 C95 L84 L95 R83 R94 C83 C94 L18 L119 R17 R118 C17 C118 L17 L118 R16 R117 C16 C117 L16 L117 R15 R116 C15 C116 L13 R129 C129 L129 R128 C128 L128 R127 C127 L8 L17 L28 L39 L5 L61 L72 L83 L94 L15 L116 L127 {L_Lage5} R7 R16 R27 R38 R49 R6 R71 R82 R93 R14 R115 R126 C7 {R_Lage4} C16 C27 C38 C49 C6 C71 C82 C93 C14 C115 C126 Implementation = Utf_L8_1 L7 L16 L27 L38 L49 L6 L71 L82 L93 L14 L115 L126 Implementation = Utf_L8_2 {L_Lage4} S Vterminal1 R3 C3 C6 R6 R15 C15 C26 R26 R37 C37 C48 R48 R59 C59 C7 R7 R81 C81 C92 R92 R13 C13 R114 C114 R125 C125 C134 R134 {R_Lage3} S Vterminal2 L3 L6 L15 L26 L37 L48 L59 L7 L81 L92 L13 L114 L125 L134 {L_Lage3} The mutual inductances are "invisible" included in: include Kopplungen_1kHz.txt {L_Lage2} {L_Lage1} R2 C2 L2 R1 C1 L1 R5 C5 L5 R4 c4 L4 R14 R36 R25 R47 C14 C25 C36 C47 L14 L25 L36 L47 R13 R24 R35 R46 C13 C24 C35 C46 L46 L13 L35 R58 R69 C58 C69 L58 L69 R57 R68 C57 C68 L68 L57 R8 C8 L8 R79 C79 L79 R12 R124 R91 R113 R133 C91 C12 C113 C124 C133 {R_Lage2} L91 L12 L113 L124 L133 R123 R9 R11 R112 R132 C9 C11 C112 C123 C132 {R_Lage1} L9 L112 L132 L123 L11 R_P1_L C_MessKable1 C_MessKable2 C_MessKable3 C_MessKable4 C_MessKable5 C_MessKable6 R_P2_L R_P3_L R_P4_L R_P5_L R_P6_L R_P7_L C_MessKable7 Network model of the ITER TF single coil for a frequency of 1 khz (established by University of Karlsruhe, IEH)
Resonance frequencies of ITER TF The resonance frequency of a single ITER TF coil is calculated to be 5 khz 35 3 U terminal2 U HeIn7 U HeIn6 25 U HeIn5 U HeIn4 2 U HeIn3 U HeIn2 15 1 U HeIn1 U R134:2 - RP7 U R131:2 - RP7 5 5 1 15 2 U = f(f) on the 5 khz model for an excitation with 1 V. First resonance occurs at 5 khz => natural frequency is calculated to be 5 khz
ITER TF discharge circuit FDU FDU FDU 18 TF coils 9 fast discharge units (FDUs) Soft grounding TF Coil Grounding resistor Fast discharge unit => A model is required to calculate terminal voltages FDU FDU FDU TF discharge circuit (simplified)
Network model of 18 ITER TF coils Output: maximum terminal to ground voltage maximum terminal to terminal voltage S3 I1 V3 + + FDU1 FDU2 - - S L1 L2 L3 L4 ein aus ein aus.349h.349h.349h.349h C1 C28 C3 C3 C4 C31 C5 C32 C33 C2 C29 R2 C6 R3 R4 R2 L19 R1 R22 L2 L21 5 TF_FDU R27 L22 R29 L23 L24 5 R25 5 5 R31 TF_FDU FDU3 FDU4 L5 L6 L7 L8 ein aus ein aus.349h.349h.349h.349h C7 C34 C35 C9 C36 R6 C1 C37 C11 C38 C12 C39 V+ V V- C8 R7 R8 R5 5 R32 L25 R34 L26 R36 L27 TF_FDU R38 L28 5 R4 L29 R42 L3 TF_FDU 5 5 FDU5 FDU6 L9 L1 L11 L12 ein aus ein aus.349h.349h.349h.349h C13 C4 C41 C42 C16 C43 C17 C44 C45 C14 R1 C15 C18 R11 R12 R44 L31 R9 R46 L32 5 R48 L33 TF_FDU R5 L34 R52 L35 L36 TF_FDU 5 5 R54 5 FDU7 FDU8 C19 R56 L13 L14 L15 L16 ein aus ein aus.349h.349h.349h.349h C46 C47 C21 C48 C22 C49 C5 C51 C23 C24 C2 L37 R13 R14 R6 L39 TF_FDU R62 L4 R64 L45 R16 R66 L41 TF_FDU 5 R58 R15 5 5 5 FDU9 L17 L18 ein aus C25 R69 C52 L44 R17 C26 5 R71 C53 R18 L43 5 C27 R73 C54 L42 TF_FDU R68 5 ITER TF system with 18 simplified superconducting coils (established by University of Karlsruhe, IEH)
Detailed network model of ITER TF INCLUDE: Kopplungen_1kHz.txt C_P1_GE C_P2_GE C_P3_GE C_P4_GE C_P5_GE C_P6_GE C_P7_GE 132.97n 22.5n 22.5n 22.5n 22.5n 22.5n 132.97n Maximum voltages (to ground or terminal to terminal) are used to excite two detailed models (1 khz and 5 khz). Maximum internal voltages are identified and located. PARAMETERS: C_Lage1 = 8.9nF C_Lage2 = 81.8nF C_Lage3 = 82.7nF C_Lage4 = 83.6nF C_Lage5 = 84.5nF C_Lage6 = 85.3nF C_Lage7 = 86.2nF C_Lage8 = 87.nF C_Lage9 = 87.9nF C_Lage1 = 88.8nF C_Lage11 = 89.6nF PARAMETERS: L_Lage1 = 1.4485uH L_Lage2 = 1.543uH L_Lage3 = 1.5527uH L_Lage4 = 1.644uH L_Lage5 = 1.6513uH L_Lage6 = 1.6977uH L_Lage7 = 1.7421uH L_Lage8 = 1.7865uH L_Lage9 = 1.8247uH L_Lage1 = 1.8716uH L_Lage11 = 1.914uH PARAMETERS: R_Lage1 = 797.4u R_Lage2 = 85.8u R_Lage3 = 814.32u R_Lage4 = 822.96u R_Lage5 = 831.6u R_Lage6 = 839.16u R_Lage7 = 847.8u R_Lage8 = 856.44u R_Lage9 = 865.8u R_Lage1 = 873.72u R_Lage11 = 882.36u PARAMETERS: R_Anbindung = 1k PARAMETERS: C_Messkable = 24nF {C_Lage9} {C_Lage8} {C_Lage7} {C_Lage6} C_P1_P2 C_P2_P3 147.24n 147.24n R23 R34 C23 C34 {C_Lage11} L23 L34 {L_Lage11} R22 R33 {R_Lage1} C22 C33 {C_Lage1} L22 L33 {L_Lage1} R12 R21 R32 C12 C21 C32 {R_Lage9} L12 L21 L32 {L_Lage9} R2 R11 R31 C11 {R_Lage8} C2 C31 L11 L2 L31 {L_Lage8} R1 R19 R3 C1 {R_Lage7} C19 C3 L1 L19 L3 {L_Lage7} R9 R18 R29 C9 C18 C29 {R_Lage6} L9 L18 L29 {L_Lage6} R17 R8 R28 C_P3_P4 147.24n R45 R56 C45 C56 L45 L56 R44 R55 C44 C55 L44 L55 R43 R54 C43 C54 L43 L54 R42 R53 C42 C53 L42 L53 R41 R52 C41 C52 L41 L52 R4 R51 C4 C51 L4 L51 R39 R5 C_P4_P5 147.24n R67 R78 R89 C67 C78 C89 C1 L67 L78 L89 R66 R77 C66 C77 L66 L77 R65 R76 C65 C76 L65 R64 C64 L64 R63 C63 L63 R62 C62 L62 R61 L76 R75 C75 L75 R74 C74 L74 R88 C88 C99 L88 R87 C87 C98 L87 R86 C86 C97 L86 R85 C85 C96 L85 R73 R84 C73 C84 C95 L73 L84 R83 R72 C_P5_P6 C_P6_P7 147.24n 147.24n R1 R111 R122 C111 C122 {R_Lage11} L1 L111 L122 R99 R11 R121 C11 C121 L99 L11 L121 R98 R19 R12 R131 C19C12 C131 L98 L19 L12 L131 R18 R13 R97 R119 C18 C119 C13 L97 L18 L119 L13 R96 R17 R118 R129 C17 C118 C129 L96 L17 L118 L129 R95 R16 R117 R128 C16 C117 C128 L95 L16 L117 L128 R15 R127 R94 R116 C8 {R_Lage5} C17 C28 C39 C5 C61 C72 C83 C94 C15 C116 C127 L8 L17 L28 L39 L5 L61 L72 L83 L94 L15 L116 L127 {L_Lage5} R7 R16 R27 R38 R49 R6 R71 R82 R93 R14 R115 R126 C7 {R_Lage4} C16 C27 C38 C49 C6 C71 C82 C93 C14 C115 C126 Implementation = Utf_L8_1 L7 L16 L27 L38 L49 L6 L71 L82 L93 L14 L115 L126 Implementation = Utf_L8_2 {L_Lage4} S Vterminal1 R3 C3 C6 R6 R15 C15 C26 R26 R37 C37 C48 R48 R59 C59 C7 R7 R81 C81 C92 R92 R13 C13 R114 C114 R125 C125 C134 R134 {R_Lage3} S Vterminal2 L3 L6 L15 L26 L37 L48 L59 L7 L81 L92 L13 L114 L125 L134 {L_Lage3} The mutual inductances are "invisible" included in: include Kopplungen_1kHz.txt {L_Lage2} {L_Lage1} R2 C2 L2 R1 C1 L1 R5 C5 L5 R4 c4 L4 R14 R36 R25 R47 C14 C25 C36 C47 L14 L25 L36 L47 R13 R24 R35 R46 C13 C24 C35 C46 L46 L13 L35 R58 R69 C58 C69 L58 L69 R57 R68 C57 C68 L68 L57 R8 C8 L8 R79 C79 L79 R12 R124 R91 R113 R133 C91 C12 C113 C124 C133 {R_Lage2} L91 L12 L113 L124 L133 R123 R9 R11 R112 R132 C9 C11 C112 C123 C132 {R_Lage1} L9 L112 L132 L123 L11 R_P1_L C_MessKable1 C_MessKable2 C_MessKable3 C_MessKable4 C_MessKable5 C_MessKable6 R_P2_L R_P3_L R_P4_L R_P5_L R_P6_L R_P7_L C_MessKable7 Network model of the ITER TF single coil for a frequency of 1 khz (established by University of Karlsruhe, IEH)
Calculated voltages in time domain 4 3 2 1-1 FD without fault - L8 U L8 terminal1 U L8 terminal2-2 5. 5.2 5.4 5.6 5.8 5.1 For an ideal fast discharge all coils have the same maximum voltage of 3.5 kv to ground and between both terminals of each coil. The calculated terminal voltages are in good agreement with some ITER DDDs. But non linear internal voltage distribution was found already during fast discharge without fault which was not in agreement with the simple calculations of the ITER DDDs (where only linear internal voltage distribution is assumed). => HV tests are required to confirm proposed test voltages are compatible with ITER design
Long term testing on ITER TFMC Insulation: conductor and radial plate insulation (ground insulation has fault) Maximum voltage test value derived from calculation of transient behaviour: 11 kv peak (factor compared to TFMC acceptance tests: 4 for DC and 8 for AC) Voltage waveform: DC and AC Duration of 3 voltage steps each: 1 h No voltage breakdown appeared during DC test (U DC, max = 11 kv) Breakdown appeared after 9 h 39 min of 7.78 kv rms on ground insulation during conductor insulation test on known fault location (increase of PD activity 15 min before breakdown) => Proposed test values for conductor and radial plate insulation would be OK ITER TFMC outside the cryostat
Burn out of fault location on ITER TFMC The burn out confirms the assumption of the fault location Flashes around the helium tubes during burn out
Conclusion for ITER Calculation of terminal voltages and assuming only linear voltage distribution is not enough for prediction of internal voltages A Paschen Test is indispensable to prove high voltage strength during vacuum breakdown A cold test is recommended to verify reliable HV operation at cryogenic temperature Conductor and radial plate insulation can withstand the proposed test voltages derived from calculation of transient behaviour of ITER TF in special fault case for 1 h without breakdown. => 1 working day (8 h) Paschen Test with permanently applied high voltage would be possible
End
3D FEM model for ITER TFMC 3D-FEM model of ITER TFMC for direct voltage calculation (University of Karlsruhe)
Terminal voltages in time domain (TF-7) FD without fault - L8 failure of FDU 2 and 3 + earth fault 3-1 4 2 U terminal 2:1 3 15 U terminal 2:2 U terminal 8:1 2 U L8 terminal1 U L8 terminal2 U terminal 8:2 1 1 5-1 -2 5. 5.2 5.4 5.6 5.8 5.1-5 5. 5.2 5.4 5.6 5.8 5.1 For an ideal fast discharge all coils have the same maximum voltage of 3.47 kv to ground and between both terminals of each coil. Rise time: t r = 1.6 ms. Maximum voltage to ground in fault case 2 is 16.35 kv (t = 5.877 s, t r = 3.5 ms, terminal L8:2)
Frequency measurements on ITER TFMC G FRS calculated Case 2 ±5 kv Calculated (network) and measured resonance frequency show good agreement for the relevant cases G(f) 2,5 2 1,5 FRS.QDA Damping directly in resonance case and above was calculated with poor accuracy sometimes too low and sometimes too high 1,5 1 2 3 4 5 f khz Comparison of the transfer functions on outermost inner pancake joints for radial plates connected over resistors and symmetric excitation.