Remote Sea Hearth 20 Fibers Cable M Sedita

Remote Sea Hearth 20 Fibers Cable M. Sedita, WPs Meeting - Paris 23/24 February 2009

ODI Four types of ODI connectors are utilised on Nemo § NRH Hybrid optical and electrical connectors at the output interface § High Power For the electrical connection between the CTA and MVC § Nautilus Diver Mate For the electrical connection between the MVC and Junction box § Rolling seal optical Replaces the Dry Mate DMO • In response to INFN concerns about the DMO connector in a previous application For the optical connection between the CTA and Junction box M. Sedita, WPs Meeting - Paris 23/24 February 2009

ODI High Power § Substantial re-design of HP connector to improve HV performance Includes new pin, pin seal and mechanical arrangement § Qualification underway includes Composite connector testing complete • 10 Kpsi, 15 k. Vdc • 1. 000 hours of testing at 60 °C successfully complete • 1. 000 hours of testing at 5 °C complete Pin long term testing complete • 1. 000 hours of unmated testing complete • Subsequent 40 k. V DC 4 hour withstand passed • A supplementary 1. 000 hours of mated connector testing has also been completed. Mate/De-mate complete Mechanical • Thermal cycles and thermal shock complete, • B&V complete • Pressure cycles and hold completed AC Breakdown testing Complete AC and DC high current testing completed DC breakdown testing Complete • Results show very high breakdown voltages • ODIs 10 k. V 25 year survival probability based on these results is 0. 9998 M. Sedita, WPs Meeting - Paris 23/24 February 2009

ODI Pin Long term testing (45 °C, 15 kpsi) • Some pins in the 2. 000 hour group showed cracks Composite pin testing (restarted after previous test set up problem) • 10 pins, 15 kpsi, 45 °C, 20 k. Vdc, 2% strain • Pin breakdowns started to occur in this group at around 1. 200 hours ODI believe that both of these failure are induced by the test pressure which was 3 times the operating pressure. ODI have conducted a detailed investigation of this problem which shows: • that radial cracks propagate form the centre of the pin towards the outside • these cracks tends to propagate along the knit line of the moulding • Test of the knit line demonstrate that this is the weakest pint • but is well above the material specification. • There are some very small radial micro cracks around the centre of all pins which result directly from the residual stress from the moulding process. • It is likely that these act as mechanical stress raisers but they are not themselves of a sufficient size to cause a problem in their own right. M. Sedita, WPs Meeting - Paris 23/24 February 2009

ODI have commenced an assessment program at pressures closer to the normal operating conditions. • Pins subjected to 2. 000 hr and 3. 000 hr 5 kpsi 5 °C have been HV tested with HV step stress results above 100 k. V • Some more pins are continuing on this test and will be HV tested and sectioned at 4. 000 hours • One of the 3000 hour pins was sectioned and shows some signs of crack propagation • A new composite pin test has been started at 7. 5 kpsi 45 °C • 1 pin broke down at around 1900 hour • 5 pins broke down dung the subsequent 1 week 30 k. V voltage step stress • 4 pins broke down during the subsequent 1 week 40 k. V voltage step stress • Sectioning of the first break down showed some crack propagation in areas other than the failure point • Sectioning of pins from the 10 kpsi composite connector test has shown no indication of crack propagation. • A further 2000 hour 10 kpsi pin pressure test group has completed. • These pins showed significant degradation of HV performance compared to the 2. 000 hour 5 Kpsi pins and sections showed some crack propagation ODI remain confident that they will be able to demonstrate the performance of these pins and the reliability of the connector. They expect to have completed their work and presented the data to us by the 2 nd week of February. M. Sedita, WPs Meeting - Paris 23/24 February 2009

ODI are following 2 back up plans • Modifications to the moulding process in an attempt to minimise the internal stresses in the moulding • The 1 st moulding trials are complete • Analysis of these pins show significant improvement in the residual stress though the micro cracks are not completely eliminated. • A further test program would be required in order to demonstrate this is sufficient. • Assessment of alternative pin moulding materials • Material coupon trials are underway • ODI have selected the material they plan to use • And are progressing this in parallel with the other options • We are working with ODI to finalise the exact timescales for this option M. Sedita, WPs Meeting - Paris 23/24 February 2009

ODI § Penetrator Pin wiring During the MV qualification test unit assembly we identified that some of the wires attached to the penetrator pins had bird cadging under the rubber boot. ODI have agreed to rework these wires to eliminate this and inspect them by X-ray. All Nemo HP harnesses are being returned to ODI for this rework Re-delivery from ODI is due by the 6 th March if re-moulded pins are required Subject to the existing pin qualification being acceptable we will endeavour to improve this date with ODI. M. Sedita, WPs Meeting - Paris 23/24 February 2009

HV connection Current configuration Low Voltage divermate Connector High Voltage wetmateable Connector MVC Sea earth CTA Harness assembly Splice Canister ROV Connector Panel M. Sedita, WPs Meeting - Paris 23/24 February 2009

HV connection Alternative solutions § Other supplier wet-mateable connector: No existing solution Despite double check with : Tronic, Diamould, Deutsch, … § Dry joint in a separate housing with ODI penetrator to penetrator harness both side: Not a workable way through : penetrator exhibit similar behavior as connector Pin M. Sedita, WPs Meeting - Paris 23/24 February 2009

HV connection Alternative solution ASN pigtails ASN cable joints Sea earth return (2 wires) MVC CTA Harness assembly Splice Canister ROV Connector Panel M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC Single Board 2 subconverters Stack 8 subconverters 1 Control Board Assembled Converter Housed Unit Internal Unit M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC P 2 C Status § Model used for extensive integration program with no problem from April to September § August 2008 Nemo Tower Power System tested § Final test of input short-circuit performed December 2008 M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC P 2 C Status § Input short-circuit results V PFE = 10. 000 V. Distance MVC to fault = 50 km – Peak current = 28 A ; MVC restart OK V PFE = 7. 500 V, . Distance MVC to fault = 100 m – Peak current = 80 A ; MVC restart OK V PFE = 10. 000 V, MVC to fault = 100 m – Peak current over 100 A ; – MVC Failed Ø Check un-powered MVC OK Ø Re apply 7. 5 k. V not started OK Ø MVC failed into Hv breakdown after attempt to restart M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC Status § Input short-circuit 10 k. V >100 A Filter capacitor discharge Short-circuit M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC Status § Input short-circuit >100 A 10 k. V Filter capacitor voltage reverse ; current flows into diodes Short-circuit M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC Status § Input short-circuit on P 2 C : Fault analysis – Input low impedance = All input stages of converters likely have failed or are clamped by input relays. – Failure cause : reverse current - created by resonance of the capacitor with input inductance-flowing through by-pass diodes of the input stage MOS-FETs Solutions – Insert a 10 m. H inductor (equivalent to 50 km of cable) at the input. Ø Simulation tests confirm the effectiveness of this solution but some side effects on dynamic performances : Simulation done with 8 m. H: no change on dynamic performances Ø Verification on sub-converter (scaled test 1/48) : tests done on “mini-unit (1/3 scale) with 2. 9 m. H coil model : start and step load not affected. Input short-circuit test planed after firmware validation. Ø 8 m. H model build : 10 k. V surge tested (coil only, no MVC) Ø OK to install the coil in P 5 – Replace input stage diodes (shall also prevents blowing diodes when input clamps) Ø Existing diode not robust enough Ø New diode validated All sub-converter for Nemo reworked M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC improvements Sub converter input short circuit test With inductance (8 m. H/ 4) With no inductance Peak diode current = 70 A Peak diode current = 7 A Peak Mos. Fet current = 53 A Peak Mos. Fet current = 6 A M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC improvements § Input diodes replacement Existing diodes characteristics (Si. C) = 1 A average / 32 A; 10 µs peak New diodes = 6 A average / 60 A; 8. 6 ms (half sine) peak Validation tests : – Reworked boards validated – Minimum penalty on efficiency All Subconverters will be reworked with new diodes M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC improvements § Design improvement Hardware / Power – New input diodes selected with much higher over-current capability. Slight penalty on efficiency : Total added power dissipation = 45 W – Input coil designed to prevent blowing up MVC when HV input short-circuit close to the node Control / Firmware – New setting on Control loops to make all PWM signals within 10% = smooth transient whatever the voting logic (disturbed or not). Penalty = about ±. 25% variation of 375 V DC output – Simplified voting logic (no more control with 10 MHz clock) but overall reliability unchanged – Change in Clock select process to prevent changing clock when MVC working : preferable to shut-down and restart rather that having a gap in control – Clock monitoring (shutdown criteria when clock fails) more accurate but done on six pulses (less likely to shut-down following Ground fault disturbance) – Improved reliability of maximum duty cycle limit setting (to prevent the risk of transformer saturation) Penalty = 100 V higher minimum input voltage at full load. M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC Qualification Plan of Work § Short POW : P 4 investigation : to be completed in week 53 done – mini unit (Ctrl from P 4 + 2 SC)*3 built week 51: done – Firmware modification s week 51, validated week 52, 53 specification done Built and test of a new subgroup : – Modified Sub-Converter board with new diodes : done – New Firmware validated week 904 : Ø new “voting logic” CPLD firmware developed and validated on mini unit Ø New “clock select” CPLD firmware partly validated on mini unit. – Other hardware changes on control board (duty cycle) frozen and validated New Qualification unit (P 5) to be built from modified production sub-assemblies : – Internal unit assembly and test : wk 906 -907 – Housing wk 908 confirmed – Testing wk 909 Confirmed Re-start of P 5 qual program in week 910 up to 914. No change in the plan M. Sedita, WPs Meeting - Paris 23/24 February 2009

MVC Qualification and production units Qualification program § New Model (P 5) for Qualification to be built and tested § Program that should be completed by wk 914 no change § Operational Test including : confirmed § – Long term operation – Short circuit Bump tests & Vibration tests week 13 -14 Temperature cycling (storage conditions) After qualification : Life test (2 to 4 weeks) confirmed M. Sedita, WPs Meeting - Paris 23/24 February 2009

Target Plan of Work – current ODI designs / MVC issues resolved § Manufacturing MVC FAT (#1) week 13 2009 MVC (#1) Arrival at Catania MVC FAT (#2) week 16 2009 week TBA MVC (#2) Arrival at Catania CTA FAT week TBA (to be available for load) week 11 CTA Tailing and delivery to Catania week 16 § Frame Assembly & Backbone monitoring Assemble MV / CTA into frame Ready To Load week 17 - week 19 week 20 § Installation Test recovered lay cable week 20 Joint frame on board, test & deploy Lay complete week 20 week 21 § Testing System Testing week 22 – week 23 RFPA approx 5 th June 2009 M. Sedita, WPs Meeting - Paris 23/24 February 2009

Target Plan of Work – Back up scenarios § Key Decision Review / Milestone March 5 - 09 Review of the ODI Qualification results of the current design Review of Alternative Scenarios and planning (ODI and ASN) § Alternative scenarios: Modifications to ODI moulding process – This programme would be approx 6 weeks later than the Target POW. Feasibility study for the use of ASN technology nearly completed – This programme would be approx 11 weeks later than the Target POW. M. Sedita, WPs Meeting - Paris 23/24 February 2009