Evaluation of offshore HVDC grid configuration options

Transcription

1 Evaluation of offshore HVDC grid configuration options Keith Bell and Callum MacIver Dept. of Electronic and Electrical Engineering University of Strathclyde, UK This work has benefited from support by: Centre for Doctoral Training on Wind Energy Systems (EP/G037728/1) Transformation of the Top and Tail of Energy Networks (EP/I031707/1) Presented by Keith Bell, ScottishPower Professor of Smart Grids and a co-director of the UK Energy Research Centre

2 The news yesterday Graphic: The Guardian

3 Offshore wind in Europe

4 Offshore wind in Europe High capacity factors Plentiful locations Wind farm costs coming down e.g. excluding costs of connections to shore, 54.50/MWh for Borssele 3 and 4, 49.90/MWh for Krieger s Flak Developments becoming larger and further offshore

5 Reducing /MWh of network AC or HVDC connection? Developers have greater confidence in AC than in HVDC but what is the cross-over distance at which HVDC is cheaper? If shunt compensation deployed rationally, cross-over distance is further than previously thought Figure: Elliott et al, A comparison of AC and HVDC options for the connection of offshore wind generation in Great Britain, 2015 However, System benefits of HVDC option? Transient over-voltages and harmonic resonance for AC option Incentives on developer to minimise losses? SO-TO Code requirements at network owner interface? Can an AC option be extended to become part of an offshore network?

6 Reducing /MWh of network Maximise utilisation by making connections part of a network Between two synchronous areas Embedded within a single synchronous area

7 Is an offshore network like an onshore network, just with its feet wet? 1. Undersea cables are much more expensive than onshore overhead lines. 2. Offshore substations are very much more expensive than those onshore since they depend on purpose built platforms and marinised equipment. 3. For HVDC branches at a certain voltage, a connection to the AC system must use a power converter that will incur a certain minimum cost that is very large. Minimise the number of converters; use multi-terminal HVDC grid 4. High-voltage DC circuit breakers (DCCB) have not yet been demonstrated in such a way as to fully establish commercial viability are likely to be both large and expensive relative to AC circuit breakers at comparable voltages. 5. Aside from oil and gas production platforms, there is no demand connected offshore continuity of supply is not so important The above can lead to different design conventions for offshore

8 The need for DC breakers Suppose that area 1 has a loss of infeed limit of 1.3GW Without protection to detect and DC breakers to clear HVDC grid faults, 2GW would be lost for a fault anywhere on the HVDC grid Add DC breakers and protection Or limit offshore wind production Lost value from offshore energy and increase in the effective levelised cost If we are not to exceed the loss of infeed limit, a fault between OSC1 and point A should be cleared by breakers opening at A

9 Another option? Split the network pre-fault; re-configure post-fault to restore generation Pre-fault split: maximum loss of infeed in area 1 is 1GW See Bell, Xu and Houghton, Considerations in Design of an Offshore Grid, CIGRE Science & Engineering, 2015 Fault between OSC1 and A Fault cleared from AC side De-energised DC grid re-configured Disconnectors at A opened Disconnectors at B closed Re-energise offshore grid

10 What form should an offshore HVDC grid take? Several options available to developers of offshore grids Technology type of VSC converter? DC Breakers? Topology radial vs meshed, monopole vs bipole Protection strategies influenced by above choices Few studies have considered reliability performance of networks extensively How faults are managed under different network options Now some work in area, e.g. CIGRE WG B4.60 Key open questions What is the value of redundancy? Radial vs Multi-terminal vs Meshed Should we emulate onshore style protection? Requires DC breakers throughout offshore grid? Cost, availability? 10

11 What is the value of different network designs? How much of the available offshore energy cannot be sold because of network faults? How long does an unplanned outage last? Depends on the repair time which depends on vessel availability and the next sufficiently long weather window Weather window depends on the season How much offshore wind energy was undelivered? Depends on weather conditions which depends on season Repair times calculated with reference to wave heights Wave heights and wind speeds are correlated Major Offshore (long repair) Minor Offshore (short repair) Major Onshore Minor Onshore Components cable, transformer converter, DC breaker transformer converter Weather window continuous non - continuous - - Procurement delay fixed - fixed -

12 Reliability assessment Sequential Monte Carlo simulation with correct temporal correlations Sample faults and next sufficiently long weather window Offshore grid component reliability data is fairly sparse Three data cases developed based on the spread of data that is available and industry expert opinion Best, Central and Worst case set of failure rate assumptions Central Reliability Scenario Inputs Components Mean time to failure (Hours*) * Transmission cable - Hours/100km Minimum time to repair (Hours) Fixed Delay Repair time Onshore Converter 6480 (10 months) - 6 Offshore Converter 6480 (10 months) - 6 Onshore Transformer (50 years) 2160 (3 months) 72 Offshore Transformer (40 years) 2880 (4 months) 120 HVDC Transmission Cable (25 years) 2160 (3 months) 144 DC Circuit Breaker (25 years) - 6

13 Case study Assume 4 700MW clusters of wind farms connected somewhere like Dogger Bank What is the value of network redundancy and DC breakers? What is the undelivered wind energy for different network designs? How much income is lost?

14 Initial HVDC connection options Each of 4 initial options uses half bridge MMC VSC converters symmetrical monopole configuration: 2 cables at ±320 kv Option 1: Radial Option 2: Radial+ Fewer converters Reduced cable km Option 3: Multi-terminal Option 4: Meshed Redundant paths added

15 Capital expenditure Assumes DC Breaker costs = 1/6th of VSC converter station Radial option has very high cable costs Number and cost of DCCBs rises sharply as grid options become more interconnected Cost of meshed grid highest despite cable savings compared with radial option

16 Reliability performance Expected annual undelivered energy ~1/3 rd reduction in undelivered energy with alternative transmission paths Additional benefit of meshed grid relatively small Big discrepancy between best case/worst case reliability scenarios

17 Overall value of grid options 25 year NPV of grid options Value of Energy: 150/MWh Discount Rate: 6% Radial: worst option in all scenarios high CAPEX poor reliability performance Radial+: favourable when component reliability is good low CAPEX Multi-terminal: favourable under central + worst case reliability performance Meshed: unfavourable due to high CAPEX with only marginally improved reliability performance

18 Further options Option 5: Minimum Breaker 200km 1400MW 200km 1400MW 15km 700MW 20km 1400MW 15km 700MW Full bridge AA-MMC converters used to block reverse fault current into DC Grid alongside reduced number of DC breakers DC Breakers used to isolate healthy grid section from faulted section allowing continued operation

19 Further options Option 6: AC Protected 200km 1400MW 200km 1400MW 15km 700MW 20km 1400MW 15km 700MW Half bridge MMC converters used Link between two radial+ grid sections switched out under normal conditions AC side protection used in event of fault and, if required, link switched in after delay and power re-routed

20 Further options Option 7: Bipole transmission links 200km 1400MW 200km 1400MW 15km 700MW 20km 1400MW 15km 700MW Half bridge MMC converters used with DC Breakers Main transmission links use bipole configuration with extra LV return conductor allowing partial power transfer under certain cable/converter fault conditions

21 Capital expenditure Assumes DC Breaker costs = 1/6th of VSC converter station Min breaker and AC protected options show significant savings compared with multiterminal solution Bipole grid option has high capital costs due to extra cabling requirement

22 Reliability performance Expected annual undelivered energy Min Breaker and AC protected options suffer no reliability penalty and actually marginally improve performance vs multiterminal option Bipole grid option has significantly improved reliability performance

23 Overall value of grid options 25 year NPV of grid options Value of Energy: 150/MWh Discount Rate: 6% Min Breaker: Higher losses in AA-MMC but lower CAPEX and undelivered energy so better value than multi-terminal AC Protected: Lowest CAPEX and good reliability performance make it best value option in best and central case scenarios Bipole: High CAPEX but very good reliability performance mean most favourable in worst case scenario

24 Conclusions Redundant transmission paths reduce undelivered energy Reliability significantly improves: Radial Multi-terminal Bipole Case for meshed grids less apparent Alternative methods to HVDC circuit breakers viable No reliability penalty and financially favourable New operating conditions must be managed Reliability performance a key design choice alongside capital cost Trade-off between reliability and capital cost, compromises required Overall reliability highly sensitive to fail/repair rate assumptions More certainty required or least regret approach may be required See MacIver, Bell, and Nedic, A Reliability Evaluation of Offshore HVDC Grid Configuration Options, IEEE Trans on Power Delivery, April 2016.

25 Major relevant European projects TWENTIES Work Package 5, BestPaths, October 2014 September 2018 PROMOTioN, Further work on Fault detection Fault clearance Converter and wind turbine inter-operability Diode rectifier units See

26 PROMOTioN WP1 review of past projects System planning & grid topologies Main Findings A range of studies have identified different roadmaps often different assumptions and methodologies so hard to compare directly but: multiple regional grids more likely than a single European super grid Case for complex grids dependant on level of offshore wind deployment Case for complex grids depends on detailed optimisation Main Gaps Need to identify how where when and why complex offshore topologies might develop Need to understand chronological evolutions towards multi-terminal or meshed grids

27 PROMOTioN WP1 review of past projects Operation of Converters in DC Grids Main Findings A range of studies have investigated steady state operation & control as well as dynamic stability in DC grids: Solutions identified and tested to various degrees that suggest DC grids can be operated effectively Various droop control schemes proposed (V-I, P-V etc.) Main Gaps Need to develop more extensive modelling (steady state, RMS, EMT) and testing of control strategies based on MMC and Diode Rectifier Unit (DRU) converter technology Need to test interoperability of different converter technology/vendors (especially MMC and DRU) and between converters and wind turbines Relevant Grid Codes still under development

28 PROMOTioN WP1 review of past projects DC Grid Protection Systems Main Findings Numerous protection schemes proposed by academia/industry Non-Unit (local measurements): fast Unit (data communication between 2 ends): greater selectivity Protection philosophy dependent on grid size, cost of protection, size of acceptable loss Main Gaps No single solution fulfils all requirements so no final consensus Combination of basic principles required Need to identify and develop solutions to be deployment ready Multi-vendor compatibility required

29 PROMOTioN WP1 review of past projects DC Circuit Breakers Main Findings 3 main design options: Resonant/mechanical breakers (low loss, low cost, 5-10ms) Solid state breakers (high loss, high cost, <1ms) Hybrid breakers (low loss, high cost, <5ms) Several large scale test on hybrid and resonant designs Main Gaps Development of sufficient models for detailed and system level studies Development of sufficient test/demonstration procedures to mimic full stresses experienced in DC grid application Continued development of design concepts to enhance speed, losses and cost performance