6/10/2003 – e.heineVLVnT facility-power1 EFNI H K Introduction Hypothesis Topology, technical constrains Redundancy, budget constrains Grounding Reliability Conclusion EFNI H K E. Heine, H.Z. Peek

6/10/2003 – e.heineVLVnT facility-power2 EFNI H K Hypothesis Power VLVnT structure = 10 * Antares VLVnT power < 5 * Antares = 100 kW Technology progress Structure upgrades Demands of other users? Environmental Distance shore-facility = 100 km Power cables combined with fibers Single point failures has to avoid

6/10/2003 – e.heineVLVnT facility-power3 EFNI H K Power transmission Cable behavior for AC, DC AC systems DC systems Redundancy (no single failure points) Shore VLVnT 100km mains 100kW

6/10/2003 – e.heineVLVnT facility-power4 EFNI H K Node grid Node 12 Converter Junction Converter Junction power 12

6/10/2003 – e.heineVLVnT facility-power5 EFNI H K Node grid II Node 4 35 Node Converter Junction Converter Junction power Sub node Sub node 7 Opt. station Opt. station Opt. station 10 Sub node Inst. station

6/10/2003 – e.heineVLVnT facility-power6 EFNI H K Node grid III power Node Converter Junction Node Converter Junction Node Both stations can handle full power. Installation in phases possible. How to combine with redundant optical network.

6/10/2003 – e.heineVLVnT facility-power7 EFNI H K Redundant Power distribution Node power Converter Junction Converter Junction power Converter Junction Converter Junction Two long distance failures, still 84% in charge. Installation in phases possible. To combine with optical network

6/10/2003 – e.heineVLVnT facility-power8 EFNI H K Long distance to local power AC / DC system Node DC / DC system Node AC / AC system converter junction shore power Node 2x144 el.units submarine low power conversion 2x el.units submarine high power conversion 2x el.units submarine 4 el.units on shore high power conversion

6/10/2003 – e.heineVLVnT facility-power9 EFNI H K Long distance cable behavior Higher voltage = lower current = less Cu losses = higher reactive losses AC losses (I load ²+(kU  C cable )²) R cu > DC losses (I load ²R cu ) AC charging/discharging C, DC stored energy = ½CV² DC conversions more complex then AC conversions AC cables have be partitionized to adapt reactive compensation sections. AC cable DC cable AC cable with reactive power compensation VLVnT mains R cu C cable L cable L comp. C cable L cable mains

6/10/2003 – e.heineVLVnT facility-power10 EFNI H K Redundant Node grid DC load sharing between two converter junctions R cu AC load sharing between two converter junctions Converter Junction Converter Junction Node Converter Junction Converter Junction

6/10/2003 – e.heineVLVnT facility-power11 EFNI H K Conclusions on topology AC/AC passive components initial costs losses fixed ratio by transformers extra DC conversion subm. running costs extra cabling AC/DC constant output level initial costs active components subm. running costs DC/DC constant output level cable losses running costs active components initial costs

6/10/2003 – e.heineVLVnT facility-power12 EFNI H K Node Grounding Node AC DC AC DC 48V= 380V= 5kV= Grounding is essential to relate all potentials to the environment. Prevent ground currents by use of one ground point in a circuit.

6/10/2003 – e.heineVLVnT facility-power13 EFNI H K Reliability Not something we buy, but something we make! Denson, W., “A Tutorial: PRISM”,RAC, 3Q 1999, pp. 1-2 Parts 22% Manufacturing 15% Design 9% Induced 12% No Defect 20% System management 4% Wear-Out 9% Software 9%

6/10/2003 – e.heineVLVnT facility-power14 EFNI H K General conclusions Technical issues to investigate –DC for long distance is promising-> breakeven study for costs and redundancy –node grid gives redundancy –node grid can be made of components used in railway industry –inter module grid can be made of components used in automotive industry –each board / module makes its own low voltages Organization –power committee recommend –specify the power budget (low as possible, no changes) –coordination of the grounding system before realizing –watching test reports, redundancy and reliability –try to involve a technical university for the feasibility study –coordination between power and communication infrastructures

6/10/2003 – e.heineVLVnT facility-power15 EFNI H K Industrial references HVDC, conventional HVDC, VSC-based HVAC MVAC VLVnT Sally D. Wright, Transmission options for offshore wind farms in the united states, University of Massachusetts High Voltage Direct Current Transmission, Siemens HVDC light, ABB Gemmell, B; e.a. “HVDC offers the key to untapped hydro potential”, IEEE Power Engineering Review, Volume:22 Issue: 5, May 2002 Page(s): 8-11 VLVnT Break- Even Distance DC AC Installation costs; €/MW.km Lower in time by technical evolution Expanding by technology progress

6/10/2003 – e.heineVLVnT facility-power16 EFNI H K Monopolar Cable configuration Combined with fibers for communication Redundancy powerLoad Bipolar power Load Bipolar normal operation power Load Bipolar monopolar operation power Load High Voltage Direct Current Transmission, Siemens

6/10/2003 – e.heineVLVnT facility-power17 EFNI H K Power factor corrector Classic PFC diliveredpower current voltage delivered power current voltage CC control IU ~200kHz Classic: more harmonic noise higher I²R losses PFC constant power load voltage and current in phase

6/10/2003 – e.heineVLVnT facility-power18 EFNI H K Converter types buck converter boost converter transformer isolated converter control ~200kHz input circuit switching circuit rectifying circuit output circuit control ~200kHz control ~200kHz Efficiency up to 90% Power driven

6/10/2003 – e.heineVLVnT facility-power19 EFNI H K start up sequence Local power Converters on each board or function Control of switch on/off sequences Control of Vh>Vl (by scotky diodes) Control of voltage limits; 1V8 ± 4%, 5V ± 5%, 12V ± 10% Power consistence PCB-layout (noise) efficiency 12V 3V3 3V3out 1V8out 1V8 3V3+5V 1V8+5V C1 C2 Delay C1; 290nF/ms Rise time C2; 41nF/ms 68k 15k 37k BS250 BS170 20k 4k7 2k efficiency time power Efficiency (%) voltage