The Crown Estate. A Modular Approach to Offshore Transmission Systems

The Crown Estate A Modular Approach to Offshore Transmission Systems January 2011

RESPONSIBILITY NAME DATE Author Sean Kelly 6 January 2011 Checked Chuan Zhang 10 January 2011 REVISION NUMBER DESCRIPTION DATE Draft Rev 0 Structure / text from presentation 29 November 2010 Draft Rev 1.1 Draft Rev 1.2 Issue A Draft for TCE comments Exec Summary Added Revised following TCE comments 23 December 2010 6 January 2011 16 January 2011 Front cover: the report s front cover shows the Gunfleet Sands offshore substation platform, viewed from a supply vessel. The standardisation of the design of these platforms is one of the possibilities opened up by having standardised equipment voltages and power ratings. 2

CONTENTS EXECUTIVE SUMMARY... 4 1. INTRODUCTION... 9 2. ASSUMPTIONS... 11 2.1 Technology Maturity Requirements... 11 2.2 Grid Design Rules... 11 2.3 Wind Farm Size and Transmission Rating... 12 2.4 Interconnection between Zones & Countries... 14 3. SURVEY OF DESIGNS IN SERVICE AND ON ORDER... 15 3.1 Survey of Power Modules... 15 3.2 Survey of Transmission Modules... 18 3.3 Survey Conclusions... 20 4. SURVEY OF COMPONENTS AND TECHNOLOGIES... 22 4.1 Cable and Substation Technology... 23 4.2 DC Cable and Converter Technology... 24 5. SUGGESTED STANDARD POWER MODULES... 26 5.1 Description of Modules... 26 5.2 Rationale for the Power Modules... 28 5.3 Availability of Power Modules... 30 6. SUGGESTED STANDARD TRANSMISSION MODULES... 32 6.1 Transmission Modules... 32 6.2 DC Transmission Modules... 33 6.3 Availability of Transmission Module Technologies... 35 7. ILLUSTRATIVE APPLICATION OF PROPOSED COMMON MODULES... 37 8. COST-BENEFIT ANALYSIS OF PROPOSED COMMON MODULES... 40 8.1 Qualitative Analysis... 40 8.2 Quantitative Analysis... 42 8.3 Conclusions on Cost-Benefit Analysis... 44 3

EXECUTIVE SUMMARY A modular approach promoting common modules has the potential to significantly reduce design, manufacture, installation and O&M costs over the lifetime of projects and enhance safety of construction and operation 1. As part of an initiative to help drive down offshore wind costs through economies of scale and further enhance safety via more common designs, The Crown Estate has initiated an investigation into the feasibility of establishing a voluntary modular approach featuring common settings (if possible) for the key parameters of the offshore transmission assets that will connect future offshore energy developments. Under a contract from The Crown Estate, Transmission Capital have: i) Undertaken a survey of all of the in-service, under construction and on-order transmission connections for offshore wind farms worldwide. ii) Developed a set of feasible common modules (featuring common voltage and power ratings), including technologies that are not commercially available at present but are likely to be available by the dates when orders are expected to be placed for the transmission assets associated with UK Round 3 wind farms. iii) Undertaken a high-level examination of the costs and benefits of adopting such a common modular approach. The survey of existing, under construction and on-order offshore wind farm transmission connections covered a total of 32 offshore substations ( Power Modules ) and 41 high voltage export cables ( Transmission Modules ) in 6 countries. The survey showed that 33kV is almost universally applied for wind farm array cables. Where the wind farm connects to 132kV or 150kV onshore grids the export cable will almost always be at the same voltage, but for projects connecting onshore at 400kV and/or using HVDC (as is likely to be the case in most if not all of the Round 3 projects) there is no standardisation of voltages, even among projects built by the same company. Furthermore there is no evidence of any standardisation of power capacity ratings in general offshore wind transmission projects are custom designed for the 1 Some of these benefits may only be achieved by use of standardised equipment and platform designs. Setting common sizes for transmission modules is necessary precondition for such standardisation to take place. 4

characteristics of their particular wind farms. It was noted that many projects, despite being custom-designed, have provided additional transformer capacity over the minimum required, apparently in order to reduce the impact on wind farm output of a transformer outage. This suggests that the cost of additional transformer capacity that could result from building the next standard size up may be substantially offset by improved connection availability. The survey also showed no offshore substations with capacities above 500MW: larger projects are served by multiple smaller offshore substations. On one project it has been publically stated that a pair of c. 300MW offshore substations was preferable to a single c. 600MW facility because the additional costs associated with the more extensive 33kV array cables of a 600MW block more than offset any economies of scale in substation design. The table below sets out the set of standard modules proposed by Transmission Capital. As some of these modules are based on technologies that are not yet commercially available, a date is also provided for each module. Transmission Capital believes that developers would be unwise to design their projects around new technologies if their programmes would require transmission equipment contracts to be placed prior to the dates indicated. Date of order placement Proposed Standard Transmission Modules Comments Present Power Modules (Substations): 330MW, 2x180MVA 33/132kV 500MW, 3x180MVA, 33/132kV 330MW, 2x180MVA, 33/220kV Transmission Modules: 132kV export cable, 170MW 220kV export cable, 330MW +/-320kV 1000MW HVDC-VSC link (polymeric cable) Based on: existing UK designs (132kV), modification of the Anholt design (220kV) German HVDC connections 2013/14 ( First wave Round 3 ) As above plus additional Transmission Modules: +/-500kV 1000MW HVDC-VSC link (mass impregnated cable) +/-500kV 1500MW HVDC-VSC Suitable cable designs already exist, but VSC converter designs would need to be increased to reach +/-500kV voltage. Manufacturers claim that latest VSC converter designs are modular and so an increase from +/-320kV to +/-500kV can 5

2016 ( Second Wave Round 3 ) link (mass impregnated cable) Additional Power Modules: 500MW, 3x180MVA, 33/275kV. be achieved without abnormal risk on an order in 2013/14 even though the first +/-320kV systems will not enter service until 2013. New power module design is permitted by new 275kV 3-core cable design (see below). Post 2017 Additional Transmission Modules: Single 3-core 275kV 500MW cable +/-500kV 2000MW HVDC-VSC link Power Modules of higher ratings beyond 500MW, if designs with higher array voltages are proven so that more generation can be economically connected to a single offshore substation. New 275kV 3-core cable design required: single core 275kV cable designs already exists but is unlikely to be economic compared to 3-core 220kV cable. First-of-a-kind 2000MW HVDC cable should be in service by 2015 (Hunterston- Deeside). Further stretch of VSC technology to 2000MW also required. Not possible to establish new modules until the technical and economic characteristics of any new array cable standard are known. Since only second wave projects that would not start detailed development until c. 2012 would be impacted it is possible to delay the creation of new standards based on higher array voltages until this information is available. Transmission Capital also undertook a high-level examination of the costs and benefits of adopting standard modules of the type described above. These are summarised in the tables below: Cost Factor Undiscounted present value (minus is disbenefit of standards) Explanation Standardised power module sizes need the next standard size up. () - 38m Based on sixty 500MW standard modules being built with 20% having less than 500MW of generation, but needing to use the next standard size up. Assumed that the benefit of improved availability offsets half of the additional cost. 6

Standardised power module sizes need the next standard size up. (DC) Risk of standards becoming entrenched / barrier to progress Risk UK adopts standards that are incompatible with elsewhere - 140m Nil Nil Based on four HVDC zones where the last HVDC cable in the development needs to be oversized by 250MW to meet the next standard size up. Assumed that the benefit of improved availability offsets half of the additional cost. Hard to quantify costs, and should not occur at all if standards are regularly and appropriately reviewed. Mitigated by the fact that the proposed standards are voluntary. Unlikely to occur as proposed standards are voluntary and manufacturers are unlikely to tolerate UK standards that differ from wider international standards (e.g. the work that CENELEC and Cigre have commenced) Benefit Factor Undiscounted present value (minus is disbenefit of standards) Explanation Earlier access to new technologies due to more and better R&D 150m Benefit figure is based on 2000MW HVDC VSC links become available a year earlier than would otherwise be the case due to the proposed standards allowing manufactures to develop products with a greater confidence that a market will exist. Other technology acceleration would give additional benefits, beyond what is quantified here. Acceleration of transmission system delivery and hence lowcarbon offshore wind generation 300m to 2100m A manufacturer has claimed a 6-month acceleration in delivery. Assuming that this is the critical factor limiting offshore wind delivery, and with a build out of 3GW/yr over 10 years, a 6-month acceleration will mean an additional 50 million MWhr of carbon-free generation. The value of this carbon-free generation is then assessed as per National Grid s cost-benefit analysis of its largest loss policy. Increased build-out rate may also be possible, but this potential benefit is not quantified here. Increased competition and 180m Reduced design effort means that manufactures can build more systems, and reduced bidding effort means 7

reduced design costs Reduced spares holding A few 10m that they can compete more actively to win work. Assumed that increased competition reduces 10% profit margin to 9.5% and reduced design effort cuts cost by 0.5%. Applied to 18bn capex programme. Standard equipment ratings allow for a small number of strategic spares to be shared between OFTOs. Not quantified in detail. Note that other factors (e.g. dimensions) would also need to be standardised. This high level examination of potential costs and benefits shows some very substantial benefits for the proposed standardisation, arising particularly from the environmental benefit of the increased amount of wind that can be connected as a result of shorter design-build cycles and potentially - increased manufacturer throughput. Although this is based on an assumption that it is the transmission connection of offshore wind which will determine the rate at which the UK s offshore wind build-out can progress, there is considerable anecdotal evidence suggesting that this will be the case unless suitable mitigating actions such as the introduction of these standards is taken. In undertaking discussions with manufacturers we were struck by the very positive response that we received to the concept from all of the manufacturers that we spoke to. If this concept is to be taken further it would be useful to meet with the main manufacturers to ensure that the specific standard modules we have developed are in line with what they expected when they gave their strong support to the concept. Although we have only had limited discussions with developers we have noted that some concerns were expressed about the introduction of standards, and it would certainly be a considerable change for an industry where each project has traditionally engaged in its own separate cost-benefit analysis and electrical design. We believe, therefore, that if the project is to be taken further then more comprehensive discussions with developers will be an important step which will need to be taken early in the process. 8

1. INTRODUCTION Transmission Capital has been engaged by The Crown Estate to assess the potential for promoting voluntary common modules for offshore transmission and grid connections, to consider what form such common modules might take, and to undertake an analysis of the costs and benefits of adopting such an approach. The first part of this study (presented in Section 3 of this report) was an examination of the designs already in service, under construction and on order as of December 2010. The key parameters are described for all offshore wind farms with transmission-voltage (i.e. 110kV) connections, and single line diagrams are provided for a sample of projects in a confidential annex. The second part of this study (presented in Section 4) was an examination of the components (e.g. cables, HVDC converters, transformers and switchgear) from which future designs can be assembled. A particular issue here was to identify when components of a particular rating and voltage (and particular cables and HVDC converters) would be sufficiently technically mature for their adoption as a standard to be reasonable. For ease of comparison two reference dates were established: 2013/14, when the first wave of Round 3 project might be placing orders and 2016, when the second wave might be expected to start placing orders. Based on the components discussed in the second part of this study, the third part (presented in Sections 5, 6 and 7) then proposed a number of possible voluntary standards for: i) Power Modules i.e. offshore substations and the array cables which connect them to the wind turbines. ii) Transmission Modules i.e. cables and the associated onshore substations and reactive power control. iii) DC Transmission Modules i.e. /DC converters and HVDC cables It should be noted that these voluntary standards aim only to standardise the voltages and ratings of the transmission equipment used. They do not aim to ensure component interchangeability (e.g. by standardising sizes, weights, impedances, tapping ranges, cable 9

connection locations, etc). Nor do they attempt to standardise the detailed design of offshore substation platforms to reduce design effort or operator training. Although standardisation of voltages and ratings is a precondition for equipment interoperability and standardised designs, it is not sufficient, and such an extension of the standardisation concept is outside of the scope of this study. The fourth and final part of this study (presented in Section 8) is an assessment of the cost and benefits of the proposed standards. This is undertaken using both quantitative and qualitative criteria. In view of the uncertainties involved the quantitative analysis is not put forward as a conventional cost-benefit study, but nonetheless it provides a useful indication of the relative significance of the various factors examined. Table 1.1 Report Structure Chapter Content 1 Introduction 2 Assumptions 3 Survey of designs in service and on order (as of Nov 2010) 4 Survey of Components and Technologies based on technological options available for orders placed today, for orders placed in 2013/14 and for orders placed in 2016 5 Suggested Standard Power Modules based on technological options available for orders placed today, for orders placed in 2013/14 and for orders placed in 2016 6 Suggested Standard Transmission Modules based on technological options available for orders placed today, for orders placed in 2013/14 and for orders placed in 2016 7 A demonstration of how these standard modules could be used, using the example of a fictitious set of projects 8 Cost-benefit analysis and conclusions this seeks to assess the case for voluntary industry adoption of the suggested standards 10

2. ASSUMPTIONS This section describes the assumptions that underlie the analysis. This section is divided into subsections that set out the assumptions used to determine when technologies are likely to be sufficiently mature, the assumptions regarding the rules that will be set by the Security and Quality of Supply Standards, and the relationship between wind farm size and the rating of the transmission assets. 2.1 Technology Maturity Requirements It should be noted that the technical maturity level required for adoption of a voltage/rating combination as a standard should be set higher than the requirement for first commercial use. The offshore wind industry has traditionally been rather conservative in the adoption of higher voltages (e.g. 220kV), which may partially reflect the fact that such transmission assets represent only a small part of the overall wind farm cost, thus diluting the impact of any cost saving possible though using newer technologies. We also understand that concerns regarding the availability of insurance for technologies regarded, by the insurance industry, as lacking technical maturity, may also have constrained the decisions of certain developers. It is assumed that sufficient technical maturity can be proven (to the satisfaction of developers and the insurance sector) for a new design to be adopted as a standard when the following criteria are met: i) There are a number of manufacturers able to accept orders for the component in question. ii) Installations at the same voltage and power rating are already in service elsewhere, or (if this is not the case) such installations have already been ordered for service elsewhere (the first commercial use ) at least a year previously and the change in voltage and/or power rating from installations in service is not excessive compared to historical rates of progress. iii) Type tests (possibly with enhanced tests) have already been undertaken. 2.2 Grid Design Rules Grid design rules are set out in a document called the Security and Quality of Supply 11

Standards (SQSS). All transmission companies (including OFTOs) are required to comply with the SQSS through clauses in their transmission licences. Although several projects have deviated from the requirements of SQSS through customer requested variations this would not be appropriate for a proposed standard design. In addition several of the key requirements of the SQSS, notably those relating to the single infeed loss limit (the amount of generation that can be disconnected from the system following a single transmission failure) cannot be varied at customer request since their purpose is to protect wider system security rather than to give a particular customer a reasonable quality of connection. At present (December 2010) the SQSS states that any design must ensure that no more than 1000MW of generation can be lost following the failure of a wind farm export cable. Proposed changes to the SQSS to accommodate the EPR 2 nuclear reactor design would increase this limit to 1320MW 3. 2.3 Wind Farm Size and Transmission Rating Smaller wind farms (including the smaller Round 3 projects and most of the Scottish Territorial Waters projects) are expected to be developed, permitted and built as a single unit. These smaller wind farms will vary in size, depending on the physical size of the size and the developer s decisions regarding machine size and spacing. Because of this variation in size, the introduction of standard sizes of offshore and onshore substations 4 will mean that such projects would need to select the next largest standard size, thus resulting in assets being oversized relative to a wholly customised design. However the extra cost of buying larger transformers, etc, will be offset to some extent by the benefits provided by larger units (e.g. reduced losses, increased availability should a transformer fail). Larger wind farms (i.e. the larger Round 3 projects with capacities of several GW) are being developed in stages. It should be possible to match the size of these stages to the size of the standard transmission connection designs being proposed. Although the situation may be somewhat complicated where wind farm sites are broken up by shipping lanes, etc, our 2 Areva s European Pressurised Water Reactor design 3 Nuclear reactors with gross outputs of up to 1800MW would be allowed under these rules, but wind farm export cables would be held to tighter standards. This reflects the current situation where the Sizewell B nuclear reactor is allowed to be larger than the largest wind farm export cable. 4 Standard designs for cables (the type that would be expected for smaller wind farms) already exist. This is discussed further in Chapter 6 of this report. 12

analysis suggests that the additional cost of slightly longer cables in order to connect a few detached parts of a wind farm is not significant relative to overall transmission costs. We have surveyed the major Round 3 projects and found that three of these projects (Dogger, Hornsea and Norfolk) have already announced the size of the first of their stages. Unfortunately the size of this first stage is different for each zone (1400MW for Dogger, 1200MW for Norfolk and 1000MW for Hornsea) which makes means that these projects do not give a clear lead on suitable standard sizes. Despite this, however, it should be possible to use standard size transmission assets for all of these projects as any surplus transmission capacity left after completion of a stage in the zone s development could be used in the development of the next stage 5. Thus only in the last stage of development would there be a significant cost associated with using standardised rather than customised connections. For the development of the first stage of these large zones the single infeed loss limit will apply, and will limit the useable capacity of the first circuit to the value set for this limit at the time (see section 2.2 above). For later developments, however, the HVDC converter platforms associated with each stage s cables to shore can be connected with relatively short offshore cables. Linking the HVDC platforms in this way mitigates the effect of the single infeed loss limit, since if one of the export cables fails some of its load can be taken by the other export cables. For instance: i) A single 1500MW cable can only carry 1320MW under the proposed future SQSS rules. (88% utilisation). ii) A pair of 1500MW cables can carry 2820MW (94% utilisation): if one cable fails then 1320MW of generation will be automatically shut down, with the remaining 1500MW being exported on the cable that is still in service. iii) Three 1500MW cables can carry 4320MW (96% utilisation). Although smaller cables could have 100% utilisation and would avoid the cost of links between the DC platforms, the extra utilisation is only 4% and the links between converters are relatively short. Any savings that a smaller cable design would get from these factors would be swamped by the extra cost per-mw of smaller cables due to the loss of economies of scale. 5 The economies of scale in the design of HVDC links are such that the cost of building the first stage with some surplus capacity by building cables of the next largest available size is small. For instance building a 1500MW link to accommodate a 1200MW wind farm would add only 10-12% to the transmission costs, or ~2% of the total cost of the wind farm and transmission connection. The NPV impact of the delay of a few years before the second stage is built would be on the order of 0.5%. 13

iv) A further advantage of having two or more 1500MW cables, with links between the converter platforms, is that this approach should provide a higher quality of connection. Should one cable fail auxiliary supplies to all turbines would remain and automatic shutdown of turbines will only be required where the zone is running at or near its full output. Although the example above is based on a 1500MW cable rating, the same effects would also be found for a 2000MW rating. National Grid has undertaken studies 6 which show a total saving of 2.9bn, across all zones, by using larger equipment sizes, in particular 2000MW HVDC cables in place of 1000MW. 2.4 Interconnection between Zones & Countries Arrangements have been proposed that add additional HVDC cables, either: i) Between two or more of the UK s Round 3 zones 7, or ii) Between offshore wind farms in different countries 8. Simple arrangements of this type can be achieved by using HVDC circuits with three of four HVDC converters connected to them (so called multi-ended HVDC links). More sophisticated arrangements would require HVDC hubs where several cables could connect, with facilities to ensure automatic disconnection should a failure occur. Design work by various parties (including National Grid s recent work on integrated grids and our own work for The Crown Estate on the connection of wind farms in Scottish Territorial Waters) has shown that there is little reason to expect a need for HVDC hubs even for projects placing orders in 2016. As a result it has been assumed that any HVDC standard should include a requirement for HVDC links to be able to operate in a multi-ended mode, but that HVDC hubs need not be included in any standard to be applied up to 2016. 6 See presentation Integrated Offshore Grid Solutions by National Grid to RenewableUK, 4 Aug 2010. 7 See, for instance, the Integrated Strategy option in the Offshore Development Information Statement (ODIS), National Grid, September 2010 8 See, for instance, C. Veal, C. Byrne and S. Kelly, 2007, The cost-benefit of integrating offshore wind farm connections and subsea interconnectors in the North Sea, Proc. European Offshore Wind Conference and Exhibition 2007, Berlin 14

3. SURVEY OF DESIGNS IN SERVICE AND ON ORDER Transmission Capital undertook a survey of all in-service, under construction and on-order offshore wind transmission connections worldwide 9. This covered 32 offshore substations (of 27 different designs) and 41 high voltage export circuits. The survey reflects the situation at the end of 2010 10. The survey separately examined Power Modules (i.e. the offshore substation and its associated lower-voltage connections to the turbines) and Transmission Modules (the or DC export cables, along with /DC converter stations and reactive compensation). 3.1 Survey of Power Modules Tables 3.1 to 3.3 below summarise the results of the survey of power modules. Note that: i) The voltages given are the nominal array and export-cable voltages. These nominal voltages are standardised and exclude the small variations applied by some operators. (For instance equipment with a maximum continuous voltage of 170kV is usually described as having a nominal voltage of 150kV, but is described by some operators as 155kV. The actual operating voltage is usually different from both nominal voltage figures). ii) The manufacturer name given in these tables is the manufacturer of the transformer(s) alone. In some cases this manufacturer, sometimes as part of a consortium, will have wider responsibilities for delivery of switchgear, control equipment or even the entire platform, whilst in other cases their scope is limited to the transformer itself. 9 Currently six countries (UK, Germany, Denmark, Netherlands, Belgium and Sweden) have wind farms connected at 110kV or above. Wind farms not listed here are connected to shore at below 110kV (most likely 33kV or below). 10 Since the end of 2010 a least one additional order - for the connection of Borkum Riffgat - has been placed. This involves 80km of 150kV cable, 50km offshore and 30km onshore. 15

Table number 3.1: Power Modules - Scandinavia and Benelux Module MW No of transf Date ordered Transf. Manf er Transf. voltage Transf. rating Note Horns Rev I 160MW 1 2000? Alstom 33/150kV 160MVA Horns Rev II 209MW 1 2006 Siemens 33/150kV 220/110/100MVA Three winding transformer Rodsand I (aka Nysted) 166MW 1 2002 Tironi 33/132kV 180/90/90MVA Three winding transformer Rodsand II 207MW 1 2008 Siemens 33/132kV 220/110/110MVA Three winding transformer Anholt 400MW 3 2010?? 33/220kV 3 x 140MVA Transformers have a large overload rating Lillgrund 110MW 1 2006 Siemens 33/132kV 120MVA Q7 (aka Princess Amalia) 120MW 1 2006 ABB 22/150kV 140MVA Possibly three winding? BelWind 165MW 1 2009 CG 33/150kV 200/100/100MVA Three winding transformer Thornton Bank?? 2? 2010 ABB 33/150kV?? Table number 3.2: Power Modules Great Britain Module MW No of transf. Date ordered Transf. Manf er Transf. voltage Transf. rating Note Barrow 90MW 1 2004 Alstom 33/132kV 120MVA ONAN 60MVA Robin Rigg East Robin Rigg West 90MW 90MW 1 1 2007 Alstom 33/132kV 100MVA Not connected offshore same substation onshore Gunfleet 172MW 2 2007 ABB 33/132kV 120MVA 16

Thanet 300MW 2 2008 Siemens 33/132kV 180/90/90MVA Three winding transformers Inner Gabbard 368MW 3 2008 Siemens 33/132kV 180/90/90MVA Three winding transformers. Could take up to 500MW Galloper 137MW 2 2008 Siemens 33/132kV 90MVA Walney-1 Walney-2 184MW 184MW 2 2 2007-8 2010 ABB ABB 33/132kV 33/132kV 120MVA 120MVA Not connected offshore & different substations onshore Ormonde 150MW 2 2009 ABB 33/132kV 85MVA Sherringham-1 Sherringham-2 158MW 158MW 2 2 2008 Alstom 33/132kV 33/132kV 90MVA 90MVA Not connected offshore same substation onshore Lincs (incl. LID extension) 270MW 2 2010 Siemens 33/132kV 240/120/120MVA Three winding transformer London Array -1 London Array - 2 315MW 315MW 2 2 2010 Siemens 33/150kV 33/150kV 180/90/90MVA Three winding transformers. Two platforms not connected offshore same substation onshore Gwynt y Mor 1 Gwynt y Mor 2 288MW 288MW 2 2 2010 Siemens 33/132kV 33/132kV 160MVA 160MVA Two platforms not connected offshore same substation onshore Table number 3.3: Power Modules - Germany Module MW No of transf. Date ordere d Transf. Manf er Transf. voltage Transf. rating Note Bard 1 400MW 2 2007? Siemens 33/150kV 236/118/118MVA (unverified) Three winding transformer Baltic 1 48MW 1 2009?? 33/150kV 100MVA Baltic 2 288MW 2? 2010 Alstom 33/150kV? Global Tech I 400MW? 2010 Alstom 33/150kV? 17

Alpha Ventus 60MW 1 2007 Alstom 33/110kV 75MVA Nordsee Ost 295MW 2? 2010 Siemens 33/150kV? 3.2 Survey of Transmission Modules Tables 3.4 to 3.6 below summarise the results of the survey of transmission modules. Note that: i) The voltages are nominal (as described for the power modules) ii) The ratings are approximate, and typically described the required capacity (in MW) of a circuit rather than the physical current rating (in MVA or Amps). iii) The manufacturer name given is only for the offshore cable part of the circuit. (For HVDC links the converter manufacturer is also indicated where different) Table number 3.4: Transmission Modules - Scandinavia and Benelux Module Cable voltage Approx rating Date ordered Manf er Total length Subsea length Reactive comp Notes Horns Rev I 150kV 160MW 2000? Nexans 55km 21km Shunt reactor nr. landfall 630mm2 cable Horns Rev II 150kV 209MW 2006 Nexans 98km 42km Shunt reactor nr. landfall As above, but less conservatively rated Rodsand I (aka Nysted) 132kV 166MW 2002 Prysmian 29km 11km Nearby onshore SVC? Rodsand II 132kV 207MW 2008? ABB 37km 9km? Anholt 220kV 400MW 2010 NKT 105km 25km Shunt reactor nr. landfall 1600mm2 Al conductor. Aggressively rated Lillgrund 132kV 110MW 2006?? 9km 7km None? Q7 (aka Princess Amalia) 150kV 120MW 2006 ABB 30km? 28km None? 1000mm2 Al conductor BelWind 150kV 165MW 2009 Nexans 54km? 52km Offshore shunt reactors 18

Thornton Bank 250kV 2 x??? 2010 ABB????? One circuit previously operated at 33kV Table number 3.5: Transmission Modules Great Britain Module Cable voltage Approx rating Date ordered Manf Total km Subse km Reactive comp Notes Barrow 132kV 90MW 2005 Nexans 30km 27km Fixed shunt reactor 300mm2 Robin Rigg E Robin Rigg W 132kV 90MW 90MW 2007 Prysmian 15km 13km Onshore 33kV caps (4) & shunt reactors (6) 300mm2 Gunfleet 132kV 172MW 2007 Prysmian 13km 9km None 800mm2 Thanet 132kV 2 x 150MW 2008 Prysmian 29km 26km Onshore statcom, caps, shunt reactors Gabbard 132kV 3 x 167MW 2008 Prysmian 46km 46km Onshore statcom, caps, shunt reactors Galloper 132kV 137MW 2008 Prysmian 16km 16km Offshore 33kV shunt reactors Walney-1 Walney-2 132kV 184MW 184MW 2008 2010 Prysmian 48km 49km 44km 45 km Onshore shunt reactors 800mm2 Ormonde 132kV 150MW 2009 Prysmian 46km 43km Onshore statcom, caps, shunt reactors 630mm2 Sherringham-1 Sherringham-2 132kV 158MW 158MW 2007 Nexans 43km 45km 21km 23km Onshore variable shunt reactors Lincs (incl. LID extension) 132kV 2x135 MW 2009 Nexans 62km 50km Onshore statcom, caps, shunt reactors 630mm2 (1200mm2 at shore ends) London Array -1 London Array - 2 150kV 2x158 MW 2x158 MW 2009 Nexans 54km 53km Onshore statcom, caps, shunt reactors 630mm2 (800mm2 at shore ends) Gwynt y Mor 1 Gwynt y Mor 2 132kV 2x144 MW 2x144 MW 2010 NKT 32km 21km Onshore statcom, caps, shunt reactors Table number 3.6: Transmission Modules Germany 19

Module Cable voltage Approx rating Date ordered Manf er Total length Subsea length Reactive comp Notes Alpha Ventus 110kV 60MW 2007 Prysmian 66km 59km Offshore and onshore fixed shunt reactors Baltic 1 150kV 250MW? 2009 NKT 75km 60km Offshore and onshore shunt reactors Oversized for future expansion Baltic 2 150kV 288MW? 2010 NSW 132km 120km Onshore and offshore reactors Intermediate reactive comp at Baltic 1 BorWin 1 +/- 150kV HVDC 400MW 2007 ABB 199km 122km Built in feature of VSC converter First user to be Bard Offshore 1 Borwin 2 +/- 300kV HVDC 800MW 2010 Siemens Prysmian 200km 125km Built in feature of VSC converter First users to be Global Tech 1 + Veja Mate DolWin +/- 320kV HVDC 800MW 2010 ABB 165km 75km Built in feature of VSC converter First user to be Borkum West II (Trianel) HelWin +/- 250kV HVDC 576MW 2010 Siemens Prysmian 130km 85km Built in feature of VSC converter First user to be Nordsee Ost 3.3 Survey Conclusions The survey shows a number of interesting common features across the fleet of offshore wind farm connections: i) For the array cables that connect the wind turbines to the offshore substations, 33kV is the (almost) universal standard. Indeed it is arguably the only standard voltage at present. ii) With very few exceptions, neither developers nor manufactures have standardised designs for transmission-voltage elements. For instance TenneT in Germany has different voltages for all five of the export cables it has built or ordered, while in Britain one manufacturer has 4 different transformer designs across five projects. iii) The largest power module in service or under construction is 500MW. Larger projects (e.g. London Array and Gwynt y Mor) have been broken into modules of 20

less than 500MW, apparently because of the benefit of reducing 33kV cable costs and losses 11. It is unusual for projects below 500MW to break themselves into smaller power modules and where it occurs (e.g. Walney) there generally appear to be special circumstances. iv) Transformers are frequently larger than the minimum size required to accommodate the output of the wind farm. Although minimum size transformers may have a slightly lower capital cost, this may be offset somewhat by the cost of testing new designs. In addition oversized transformers will allow for increased exports should a transformer fail and should give reduced losses and increased reliability. The widespread adoption of oversized transformer is encouraging, as it suggested that the slight over-sizing and increase in costs that may result from adopting standard ratings may well be acceptable to wind farm owners. 11 In the case of London Array the suppliers have stated that this was the reason why the project has been split into two blocks. 21

4. SURVEY OF COMPONENTS AND TECHNOLOGIES Two basic technology choices are available for the connection of offshore wind farms: alternating current () cables and high voltage direct current using voltage source converters (HVDC VSC). The key features of these systems are compared below: Table 4.1 Technology Overview Application Cost Capacity Land take onshore Electrical Losses Number of manufacturers Systems Short connections <100km (applications in the 50-100km range may require offshore reactive compensation) High per-mw.km cost, but no fixed converter costs at either end. Generally more economic for <100km Capacity per cable is low (c. 400MW max at present, possible 500MW future). 1.5-2Ha / GW (including onshore reactive power compensation) Lowest for shorter cables (no conversion losses at either ends) Five manufacturers (Prysmian, Nexans, ABB, NKT, General Cable) have built / received orders for 132kV submarine cables to offshore wind farms HVDC VSC Systems Connections >100km Low per-mw.km cost (fewer cables), but high cost associated with converters. Generally more economic for >100km Capacity per cable higher (c. 1000MW max at present, possible 2000MW future) Converter stations ~1.7Ha / GW. Lower for longer cables (higher voltage means lower per-km losses) Two manufacturers (ABB and Siemens/Prysmian) have built / received orders for HVDCVSC links for offshore wind farms Note that the table above suggests that the distance above which HVDC becomes more economic is 100km. This is based on our observation of placed orders and on our own costbenefit analysis of the two systems. The exact break-even distance is quite sensitive to input assumptions (equipment cost, loss capitalisation factors, etc) and will tend to change as technologies develop. 22

4.1 Cable and Substation Technology The table below sets out the technical options currently available for submarine cables, and their relative level of maturity and suitability. substation technology is well established at far higher voltage and power levels than those described below, and it should not be an important factor in technology selection. Level of technical maturity offshore cable technology Comments Commonplace for offshore wind farm connection In use outside of offshore wind sector Available now, first-of-a-kind application to offshore wind sector under construction Available for orders placed in 2013/14 and after 132kV and 150kV 3-core XLPE cables; up to 200MW at 150kV Single core cables up to 400kV and 1000MW (for a set of three cables) 220kV 3-core XLPE cables; up to 400MW 220kV 3-core XLPE cables; up to 400MW High-end ratings may require exploitation of non-continuous wind output characteristics and the long thermal time constants of buried cables. A set of three single core offshore cables is more expensive and higher loss than a 3- core cable. Higher voltages give additional reactive power issues. Generally unsuitable for offshore wind applications. (275kV single core is referred to in the ODIS as a possible stopgap design prior to the availability of 275kV 3-core cables, but 220kV 3-core cables are likely to be more economic). 220kV 3-core offshore cables and substations are being pioneered by Danish Anholt project (400MW). To be in service by 2012. Anholt (and the 100km Malta-Sicily link) should mean that 220kV is well proven by 2012 and suitable for adoption as a standard. Three manufacturers are offering this design already; probably more by 2013. Available for orders placed in 2016 and after 275kV 3-core XLPE cables; up to 500MW Although at least one manufacturer has suggested a 275kV 3-core design we would not be comfortable recommending this as a standard in 2013/14. By 2016 there should have been sufficient time available for this design to be 23

considered mature. National Grid 12 have stated that they believe that this design has a 5-7 year development timescale from 2010. 4.2 DC Cable and Converter Technology The table below sets out the technical options currently available for HVDC submarine cables and converter stations, and their relative level of maturity and suitability. Level of technical maturity DC cable and converter technology Comments Commonplace for offshore wind farm connection On order for offshore wind farm connection & offshoresuitable technology in other applications In use outside of offshore wind sector cables In use outside of offshore wind sector converter stations n/a Voltages of up to +/-320kV and power up to 1000MW using HVDC VSC converters and polymeric cables. Voltages of up to +/-500kV and circuit ratings up to 1500MW are possible using mass impregnated cable designs currently on order. HVDC classic (line commutated) converter. Practical voltage and power ratings limited by cable (see above) Only one HVDC wind farm connection has been energised to date, and it has yet to be fully commissioned. Six VSC links at 300-320kV are on order, three for offshore wind. Largest link on order for offshore wind is 864MW, but same technology has been ordered with 1000MW capacity for onshore application and manufacturers are presently marketing 1000MW offshore HVDC VSC links. Not used for offshore wind because VSC HVDC converters cannot currently match this rating. HVDC classic designs would need considerable modification to connect an offshore wind farm, which is likely to make the approach uneconomic. To our knowledge no HVDC equipment manufacturer is currently proposing to do this; all are offering voltage source converter (VSC) technology. 12 See the available technology section of the presentation Integrated Offshore Grid Solutions by National Grid to RenewableUK, 4 Aug 2010 24

Available for orders placed in 2013/14 and after Available for order placed in 2016 and after HVDC VSC converter stations and mass impregnated cables at up to +/-500kV, 1500MW HVDC VSC converter stations and mass impregnated (or possibly polymeric) cables at up to +/-500kV, 2000MW The cable design will be proven with 2 years of non-wind farm service by 2013. A +/-500kV converter design will be an upgrade of the +/-320kV designs currently on order, which will start to enter service in 2013. Manufacturers claim that the modular nature of VSC converters means that increasing the DC voltage 13 can be achieved with limited technical risks. National Grid expect to employ a +/-500kV 2000MW mass impregnated 14 cable for the Hunterston-Deeside HVDC circuit. By 2016 this should have been in service for one year. One manufacturer claims that +/-500kV polymeric insulation cables will be available for order by 2014, however it would seem implausible that they would be immediately be accepted as a standard for widespread offshore deployment. 2016 appears a more plausible date, and would be more consistent with the historic rate of progress of commercially available HVDC polymeric cable technology (500kV polymeric DC cables were successfully tested seven years ago, but commercially available technology is necessarily more conservatively designed). 13 By adding additional levels to a multilevel converter 14 The rating would be raised from the current limit through the use of Polypropylene Paper Laminate (PPL) insulation, a material which is currently untested in HVDC service. 25

5. SUGGESTED STANDARD POWER MODULES This section of the report sets out four power modules (i.e. offshore substations) which it is suggested could be the basis of a voluntary standard adopted by wind farm developers and/or OFTOs. Three of these modules are feasible using current technologies while the fourth relies on a technology that it is believed will be ready for widespread deployment by 2016 (i.e. the second wave round 3 projects and beyond). 5.1 Description of Modules Figure 5.1 below shows the first module: it employs a pair of 180MVA 33/132kV transformers, connected (via disconnectors/earth switches, not shown) to a pair of 132kV export cables. All of this equipment has been used already for offshore wind applications; indeed the module bears a strong resemblance to Thanet and other UK offshore substations. The design of the 33kV switchboard(s) is not shown in the drawing below, but it is suggested that a double busbar design be considered given its reliability and flexibility advantages over the single bus designs that have, to date, been universally applied in British offshore substations. Figure 5.1 Figure 5.2 shows the second module, which employs a pair of 180MVA 33/220kV transformers, connected (via a single-bus 220kV switchboard, shown as a green box) to a single 220kV export cable. A short additional standby cable (shown dashed) may be added to a nearby offshore substation of the same design if there is concern regarding the reliability of the single cable connection. Technologically this module in similar to the Anholt wind farm connection in Denmark, although the amount of generation connected has been reduced. This reduction reflects concerns expressed by manufacturers that a 400MW module might be inappropriate as a standard size since a rating this high might not be achievable in many real- 26

world situations (e.g. deep directionally drilled ducts). Figure 5.2 Figure 5.3 shows the third module, with three180mva 33/132kV transformers, connected (via disconnector/earth switches, not shown) to 132kV export cables. As with the design in Figure 5.1 this is based on widely used technology and is similar to an existing offshore substation, in this case Inner Gabbard. It should be noted that the transformer specification for both these designs is the same, potentially allowing spares to be interchangeable. Figure 5.3 27

Figure 5.4 shows the fourth design. This relies on a cable design (a 275kV 3-core cable with a 500MW rating) which is not currently available, and is not expected to be available until after the early Round 3 projects have placed orders. Apart from using three transformers in preference to two larger units this design is very similar to the 500 design that National Grid presents in the ODIS. Figure 5.4 5.2 Rationale for the Power Modules Number of transformers. All power modules have at least two transformer (as this is required by the SQSS rules, based on a cost-benefit analysis of consequences of transformer failure). No power module has four or more transformers as the cost of this would be unnecessary for module sizes of 500MW. For the 500MW module three transformers were chosen instead of two larger units, even though this is slightly more expensive 15 since the reliability of the design is improved (more energy can be exported with a transformer out of 15 Transformer manufacturers have noted that since costly special transport arrangements are needed for transformers larger than ~200MVA, the cost advantage of two larger transformers is not as strong as might be thought. 28

service, with generation only being curtailed when the wind farm is near full load). We note for this reason a three-transformer design has been applied at Anholt, in preference to the twotransformer design originally adopted. Using 3x180MVA transformers for a 500MW module also makes it possible to create a 330MW module using two transformers of the same design, as has been done with the designs in figures 5.1 and 5.2 respectively. Export cable voltage. 132kV was chosen (in preference to 150kV) because it is a standard voltage in the UK and would allow smaller power modules close to shore the option of continuing to connect to the onshore 132kV grid. As 150kV only provides 14% more capacity than 132kV it did not seem sensible to establish a further standard design for such a modest gain. 220kV and 275kV were adopted because they allow for the cheapest cable designs (i.e. 3-core cables at the highest available voltage) at present and in the near future respectively. Size of module. Discussions with developers showed that studies they have undertaken into power modules larger than 500MW found that the extra cost of the (33kV) array cables required to take power to a more distant central substation, along with the losses on these cables, more than offset the economies of scale obtainable by having a single large offshore substation rather than two smaller units. For this reason our designs contain no equivalent of the 600 design presented by National Grid in the ODIS it is better to have two 330MW power modules rather than a single module twice the size. No 220kV 2-cable design. A 500MW module with two 220kV export cables is possible but as the cables would only carry 250MW their cost per MW.km would be much higher than with the 330MW design shown in figure 5.2. With long cables this effect will make it more economic to use 3 x 330MW modules (each with one 330MW cable) than 2 x 500MW module (each with two 250MW cables). With short cables per-mw.km cable costs are less important but in this case the 132kV designs can be used. Offshore Reactive Power Compensation. The designs shown above have no reactive compensation on the offshore platform. This approach is likely to be reasonable for most Round 3 projects (where the substation will be <50km from shore or relatively close to a HVDC converter platform). Although there will be some situations where offshore compensation is necessary, these are likely to be in a minority and so it was decided that the standard designs should include no offshore reactive compensation, recognising that this may on occasion need to be varied. Array cable voltage. As noted in the survey section above, the 33kV array cable voltage is 29

virtually the only area where a single standard international voltage exists; several previous attempts have been made to explore higher voltages but to date all designers have ultimately selected 33kV, suggesting a degree of robustness to this standard. Having said that, we are aware that research into higher array voltages continues 16 (e.g. the current work by the Carbon Trust Offshore Wind Accelerator) which may establish a new standard. However it would take time for any new standard to be translated into products, and further time before these products are considered proven 17. We think it unlikely that any new array voltage could be the basis of standard modules intended to be ordered in 2013/4. It should be noted that a change in the array cable design could have a wider impact on the overall design: a higher array voltage will give lower losses and per-mw array cable costs, thus the optimum size of power module will increase. It may even be attractive to remove the offshore substation and take the array cables directly to an offshore DC converter for far offshore projects. The next section sets out how it is proposed to deal with this uncertainty. 5.3 Availability of Power Modules The table below sets out the power modules examined and when they may become available: Date of order placement Power Modules Available Comments Present 2013/4 ( First wave Round 3 ) 2016 ( Second Wave Round 3 ) 330MW, 2x180MVA 33/132kV 500MW, 3x180MVA, 33/132kV 330MW, 2x180MVA, 33/220kV As above 500MW, 3x180MVA, 33/275kV Based on existing UK designs (132kV) and modification of the Anholt design (220kV) Higher voltage export and array cable designs still at research stage. Insufficient time to have in service, proven and adopted as standards (especially given industry s previous conservatism re new voltages) The substation can be built today, but the design would not be economic until a 275kV 3-core cable design is available 16 DC array cables are a more radical approach that is also being researched. The timescales for adopting DC array cables as a standard are likely to be much longer than for higher voltages. 17 We note that generators have been quite conservative in adopting higher voltages for export cables, and presume that this will be equally true for array cables, where new compact/low-cost switchgear designs may also be required. 30