D5.1 Selection of electrical system architecture

This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 SCARLET Superconducting CAbles foR sustainabLe Energy Transition 1.09.2022 – 28.02.2027 Call identifier: HORIZON-CL5-2021-D3-02 D5.1: Selection of electrical system architecture Lead partner: University of Bologna Authors: Christophe Creusot, SuperGrid Institute Pierre-Baptiste Steckler, SuperGrid Institute Alberto Bertinato, SuperGrid Institute Antonio Morandi, University of Bologna Submission date: 31/08/2023 Dissemination level PU Public X SEN Sensitive, limited under the conditions of the Grant Agreement This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 Document history Issue date Version Changes made / Reason for this issue 16/08/2023 00 Initial version 31/08/2023 01 Final version This document only reflects the authors’ view. The program authorities are not responsible for any use that may be made of the information contained therein. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 3 INDEX INDEX 3 EXECUTIVE SUMMARY 5 1. INTRODUCTION 6 2. GENERAL PRINCIPLE 8 2.1 UNIDIRECTIONAL CASES 9 2.1.1 PRODUCTION CONFIGURATION 9 2.1.1.1 Wind energy transfer 9 2.1.1.2 Solar energy transfer 9 2.1.2 CONSUMPTION CONFIGURATION 10 2.2 BIDIRECTIONAL CONFIGURATION 10 2.3 SELECTED STUDY CASES 10 3. 1 GW CONVERTER FEASIBLITY 11 3.1 TRANSMISSION ARCHITECTURE 11 3.1.1 STATE OF THE ART 11 3.1.2 PROPOSED ARCHITECTURE 12 3.2 MMC CONVERTER ARCHITECTURE 13 3.2.1 MODULAR MULTILEVEL CONVERTERS BACKGROUND 13 3.2.2 SELECTED CONVERTER FEEDER RATINGS 16 3.3 AC FEEDER SIDE 17 3.3.1 GENERALITIES 17 3.3.2 PARALLELIZATION 19 3.4 RESULTING +/- 50 KVDC ARCHITECTURE 20 3.5 RESULTING +/- 25 KVDC ARCHITECTURE 21 3.6 EXPECTED RIPPLE CURRENT CHARACTERISTICS 22 4. OFFSHORE WIND PARK 24 4.1 OFFSHORE WINDMILL CONVERSION CHAIN 24 4.1.1 CURRENT TYPICAL CONVERSION CHAIN 24 4.1.2 PROPOSED CONVERSION CHAIN TO ACHIEVE THE EXPORT DC VOLTAGE 26 D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 4 4.2 STRING CABLE 28 4.3 COLLECTOR BUS BAR 30 5. SC CABLE 31 5.1 SC CABLE STRUCTURE 31 5.2 CRYOGENIC MACHINE AUXILIARY POWER SUPPLY 34 5.2.1 GENERAL CONSIDERATIONS 34 5.2.2 CONSIDERED ARCHITECTURES 35 5.2.3 TAPPING CONVERTER 36 5.2.4 FEEDER CABLE 37 6. PROTECTION STRATEGY 39 6.1 DC CIRCUIT BREAKER 40 6.2 DC SERIES REACTOR 43 6.3 RSFCL 44 6.4 DISCONNECTORS AND EARTHING SWITCHES 46 7. CONCLUSIONS 49 BIBLIOGRAPHY 50 D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 5 EXECUTIVE SUMMARY The purpose of this deliverable is to describe the electrical system that could allow for the use of a 1 GW superconducting (SC) cable in different application cases using MVDC (< 100 kVdc) rather than HVDC voltage. More precisely, two MVDC system voltages are considered, namely +/- 50 kVdc and +/- 25 kVdc corresponding to the HTS and MgB2 cable technologies, respectively. Currently, no converter exists for these voltage and power levels. This deliverable analyzes the state of the art of power electronics and high-power conversion solutions, and then proposes what is believed to be an industrially feasible solution. In the particular case of offshore wind power export, a new conversion chain fitting inside the windmill allows for the power output right at the export voltage. Combined with the feasibility of the new onshore converter architecture, this makes it possible to suppress the costly offshore conversion platform. However, unlike resistive cables, an SC cable needs active auxiliary devices to enable its outstanding conducting properties. In this respect, designing the offshore power supply for the intermediate cooling machines of long SC cables is challenging. Different concepts are evaluated in this deliverable, keeping in mind the objective of limiting the cost of the small offshore platforms hosting those machines. The change of paradigm induced by the use of superconducting cables (i.e., use of medium voltage associated with very high - respectively 10 kA and 20 kA - nominal current) has consequences on the requirements of protection and switching functions. Taking into account the new opportunities offered by the proposed conversion architecture, and the particulars of superconducting versus resistive cables, the architecture of a protection strategy is described. This strategy combines the advantages of Resistive Superconducting Fault Current Limiter and DC circuit breakers. Last but not least, the high nominal current associated with medium voltage requires an optimization of the system architecture in light of the availability or feasibility of classical electrotechnical devices like transformers or disconnectors. Their technical limitations as well as design flexibility are considered in the proposed system architecture. Though the focus is on the offshore wind export case, the proposed converter architecture and protection solutions are pertinent for other cases, as well: not only bulk production export, but also large industrial area feeding and point-to-point exchange. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 6 1. INTRODUCTION SCARLET addresses connections on the order of 1 GW, at medium voltage up to 100 kVdc and using a superconducting cable. Point-to-point connections are considered, discarding for the time being Multi Terminal DC Grids at this voltage level. To date, at this power level, point-to-point connections are built using two AC/DC converters at 320 kVdc or 525 kVdc. They either connect one offshore wind farm to onshore HVAC transmission grid or connect two HVAC transmission grids. The two HVDC converters occupy a very large surface on the order of magnitude of a football playground each. The limitation of the permanent current capability of power electronics, switches and conventional cables pushes to increase the transmission voltage when gigawatts of renewable power are to be exported. SCARLET proposes to use the properties of superconducting materials to export the GW power at medium rather than high voltage. Considering the export of offshore or onshore wind energy, the idea comes to match the windmill output voltage with the transportation voltage to suppress one DC conversion platform. Switching from HVDC transmission using two converters to MVDC transmission using only one converter (see Figure 1), and switching from around 2000 Adc nominal current to 10 000 Adc or 20 000 Adc nominal current requires not only switching from a conventional to a superconducting cable technology but also rethinking the full conversion chain along with the switching and protection devices. Figure 1 : Basic incentive of switching from HVDC transmission to MVDC transmission applied to offshore wind energy export (original figure -top leftfrom GE Grid Solutions DOLWIN 3 website communication) Suppressing one converter (the offshore converter in Figure 1), allows for a very high simplification of the remaining offshore platform (see Figure 2) where the valve halls, transformer bays, Air Insulated Switchgear (AIS) switching bays could be suppressed, leaving DC busbars and MVDC switches to ensure disconnection for maintenance work as an example. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 7 Figure 2 : Illustration of a 2 GW 525 kVdc offshore platform from the website of TENNET. The size of the platform is estimated with a footprint of ~ 8 000 m2 and a volume of ~ 325 000 m3 not considering the supporting structure. Assumptions of offshore platform costs can be expressed in k€ /m2 or k€/m3 depending on the sources. Based on [1], a cost of 42 k€/m2 is used as a hypothesis for an offshore platform (this model would need to be improved especially for small platforms, where fixed costs would not be negligible with respect to variable costs). In this respect, a 50% reduction of the footprint leads to a gain of 168 M€, and a 75% reduction leads to a gain of 252 M€. This significant gain needs to be used positively for the other functions so that the overall cost is reduced compared to the HVDC export solution. To enable MVDC superconducting cables as a means of power transfer, the architecture and design of the electrical system have to be reviewed. The MVDC converter design, the windmill voltage forming chain, the protection devices and protection strategy have to be modified so as to ensure the highest protection of the converter and superconducting cable and allow for a high service continuity upon a certain number of events. Therefore, to establish the targeted MVDC export link, the impacted functions can be summarized as follows, keeping in mind that their cost should have a significant margin within the gain due to the offshore platform simplification: - Switching from a HVDC XLPE cable to an MVDC superconducting cable and its auxiliary functions D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 8 - Transferring the AC to DC conversion to each individual windmill - Converting offshore side AC switching functions into DC - Adaptation of DC busbars - Adaptation of onshore Conversion from HVDC to MVDC architecture - Adaptation of the associated AC switchyard - Implementation of DC protection devices As a reminder, an objective of Work Package 1 of SCARLET is to evaluate more deeply this economic impact and quantify the potential gain of this concept. 2. GENERAL PRINCIPLE Different application cases are described in the following paragraphs considering two families: - Unidirectional family: the power always flows in the same direction. - Bidirectional family: the power can flow in one or the other direction. The behaviour and dimensioning of the electrical system and especially the SC cable will depend on these application cases. If a SC cable is thought to be used in any of these applications, the worst case of transient events has to be considered. As an illustration (Figure 3), let’s consider a fault in the situation of power flow in one direction and power flow in the other direction, with the same absolute nominal current. The critical current of the SC cable is a factor (>1) of the nominal current. In the example of Fig. 3 this factor is 1.4. In case of fault, as the nominal current is high, the time for the current in the SC cable to reach its critical current will significantly depend on the initial power flow direction. In the illustration, the prospective current is evaluated without considering RSFCL action and DC CB action (protection strategy is described within section 6). In the example, depending on the power flow direction, the time to reach the critical current (Ic) is either in the range of 250 µs or in the range of 2.7 ms. This duration competes with the DC breaker breaking time and the RSFCL quenching time resulting in a possibly significant difference in the stress to sustain by the SC cable. Figure 3 : Illustration of the impact of the power flow direction (the right picture is a zoom on the initial phase of the fault) in a +/- 50 kVdc system. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 9 2.1 Unidirectional cases Unidirectional cases are further distinguished between renewable energy export / transportation (later called Production configuration) case and feeding very large consumption centers (later called Consumption configuration). In these two cases, power always flows in the same direction when the system operates in normal conditions. 2.1.1 Production configuration 2.1.1.1 Wind energy transfer Wind energy production can be onshore or offshore, distance from wind park to shore can be in the order of 100 km. Power flows always from the wind park to the AC grid. Figure 4 : Production case – Large bulk wind energy transfer 2.1.1.2 Solar energy transfer Also in the case of a very large solar farms of the order of magnitude of 1 GW, it could be imagined to transfer the power at medium DC voltage to the AC grid. Power flows always from the solar park to the AC grid. Figure 5 : Production case – Large bulk solar energy transfer D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 16 Figure 13 : MM7 converter towers from GE Power Conversion [7]. 3.2.2 Selected converter feeder ratings Given the information presented in the previous section, regarding both the semiconductor devices available and the existing medium voltage converter developments, a maximum converter rated current of 3000 𝐴 DC was considered feasible. This results in at least 4 parallel feeders in ±50 𝑘𝑉 (10 𝑘𝐴), and at least 7 of them in ±25 𝑘𝑉 (20 𝑘𝐴). This is compatible with the ratings of the corresponding DC circuit breakers, which are assumed to be available up to 3000 𝐴 as well. A rated current of 2500 𝐴 was selected, according to the following observations: - It does not increase the number of converter feeders in ±50 𝑘𝑉 (4 are needed anyways). While it does increase the minimum feeder count from 7 to 8 in ±25 𝑘𝑉, this will be shadowed by the practical requirement of an even feeder count, which will be stated in Section 3.5. - This lower rated current increases the number of devices commercially available for a physical realization: not only 3000 𝐴 press-pack devices are guaranteed to fit with a comfortable margin, but so is a parallel connection of two 1200 𝐴 − 1500 𝐴 flat-pack modules, and probably even 2000 𝐴 press-pack devices with a careful thermal design. Most (if not all) MMC applications are based on power semiconductors with blocking voltage ranging from 3300 𝑉 to 6500 𝑉, most using 4500 𝑉 devices. Conveniently, this blocking voltage is also the one of most of the high-current press-pack component offer, such that it appears as a relevant choice for SCARLET applications. For their convenience (no bulky turn-on 𝑑𝐼/𝑑𝑡 limiting circuit [8]) and higher number of manufacturers, IGBTs 4 are chosen over IGCTs, and a blocking voltage of 4500V is selected. The nominal capacitor voltage of each submodule is, according to the capabilities of these devices, selected at 2500 𝑉, which is a safe choice. Thus, not accounting for redundancy, the required number of submodules in each arm is the ratio between the DC link voltage and the submodule voltage: 2 × 50 𝑘𝑉 ÷2500 𝑉 = 40 in ±50 𝑘𝑉, and 2 × 25 𝑘𝑉 ÷2500 𝑉 = 20 per arm in ±25 𝑘𝑉. It is worth noting that in both cases, a total submodule 4 IEGTs are close enough to IGBTs to be included in the same term at this stage, for reasons of conciseness. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 17 count of 4 × 6 × 40 = 8 × 6 × 20 =960 is needed in a conversion substation, which is much lower than what is typically encountered in 1 𝐺𝑊 HVDC applications (e.g. 1536 with 4.5 𝑘𝑉 devices in ±320 𝑘𝑉, not including redundancy), which is due to the comparatively higher current rating of each submodule. The converter footprint may be lower because of this fact, associated to the much lower insulation voltage level, motivating an optimized AC yard architecture in order not to lose this lead because of the parallel connection constraints. One last aspect regarding the number of submodules is the redundancy requirement. MMC converters are designed such that the failure of one submodule is transparent and does not jeopardize the outer power flow. Different solutions exist, some using dedicated fast mechanical bypass switches, and some based on control actions and/or the fail-to-short 5 property of press-pack modules, but in any case, faulty submodules are eliminated by shorting their outer terminals. Hence, the target reliability is obtained by including more submodules than necessary, such that some of them can fail without causing any power flow interruption, and they are replaced during the next maintenance operation. No detailed analysis was performed for SCARLET, but according to known examples in HVDC on the one hand, and SuperGrid Institute’s experience on the matter on the other hand, a redundant submodule count of 10% was selected. Consequently, 44 submodules per arm will be used in ±50 𝑘𝑉, and 22 submodules per arm in ±25 𝑘𝑉. Additionally, the number of parallel converters play a role in this choice, since a consumption of all the redundancy installed in an arm would, at worst, cause the loss of either 25% (1/4) or 12.5% (1/8) of the power capacity, and not 100% as if no paralleling was employed. 3.3 AC feeder side 3.3.1 Generalities A typical HVDC AC/DC substation schematic is presented in Figure 14. 5 Press-pack devices are designed such that they fail (whatever the cause) in a stable short-circuit mode. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 18 Figure 14 : Example of a simplified AC/DC substation schematic. Its main elements are listed below, corresponding to the numbers in the figure: 1) A main AC circuit breaker, in charge of protecting all the substation against faults. 2) A transformer group, generally made of three single-phase transformers in HVDC 6 , which adapts the incoming AC grid voltage to the converter voltage (which is related to the DC link voltage) 3) A primary-side earthing impedance, to comply with the grid code request with respect to zerosequence impedance. A star coupling is assumed for the grid-side winding. 4) A converter-side earthing device, which ensures the balance of both poles voltages across ground while limiting a possible third-harmonic current circulation through the cable screens. It is pictured for a YNyn transformer, hence it is simply an impedance connecting the converter-side star point to the ground. Starpoint reactors are typically used when a YND transformer is used instead. 5) A precharge resistor bank, limiting the inrush current of the converter during energization. The DC cable is precharged at the same time. 6) A bypass switch, which eliminates this resistance after the precharge phase. It is normally a disconnector in high voltage, but load switches or circuit breakers can be used in medium voltage, providing more flexibility. 7) Arm reactors, which are needed in the MMC, and additionally mitigate the ripple currents due to the switching actions of the converter and limit its short-circuit current. 8) Finally, the arms itself, which are stacks of submodules as described previously. 6 On the one hand, the three units are easier to transport and install than a single three-phase transformer; on the other hand, it reduces the expenditure associated with redundancy, as the spare unit is a single-phase transformer with only 1/3 of the rated power. Practically, 4 single-phase transformers are installed instead of 3, which is less costly than installing 2 three-phase transformers instead of 1. Valve Valve Valve Valve Valve Valve 13 4 25 6 7 8 D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 19 Many elements are not shown for reasons of conciseness: on the one hand, disconnectors and earthing switches for safety and maintenance; on the other hand, voltage and current transformers for control and protection purposes. 3.3.2 Parallelization As discussed previously, the required parallelization brings an opportunity for redundancy and for partial power flow capability, but in order to be usable, it must come with an adapted substation architecture. To this end, three notable abilities are desired: - The elimination of a faulty converter feeder must be allowed by dedicated switchgear. - Its maintenance must be possible while the other feeders remain active. - After maintenance, its reinsertion must be possible without de-energizing the full system. As stated before, adverse effects of the large number of parallel converter feeders shall be mitigated in order to keep a low footprint, and maximizing these abilities could lead to an increased number of elements (no mutualization). Consequently, a tradeoff is desirable. To this end, a particular attention was given to the connections, most of them being normally air insulated busbars. Indeed, to the author’s knowledge, most onshore HVDC systems use Air Insulated Switchgear (AIS) equipment. First, it is proposed to use three-phase transformer units, which require no external coupling busbars connections. This is allowed by the converter feeder power which is, at most, 250 MW (±50 kV case). In order not to have eight 125 MW transformers, it is decided to group the ±25 kV converters by pairs on a three-windings transformer. Then, the precharge system is moved to the primary side: although it will require more insulation (the energy being comparable), it will be located with the mandatory grid (common busbar) side switchgear, and may be integrated well in its footprint. This will bring a drastic reduction of the transformer inrush current as a side benefit. Using this approach, a direct cable connection from the transformer (oil/cable termination) to the valve hall (cable/air termination) is possible, allowing for a minimal footprint and a maximum layout flexibility. Next to this cable termination, an earthing switch is located for the maintenance of the valve hall. With this approach, the only remaining element on the transformer secondary side is the earthing impedance: it can either be cable-connected to the neutral point and located outside, or preferably directly integrated into the transformer oil tank. Finally, integrating the current measurements inside the transformer terminations is common practice, and like the earthing impedance, voltage transformers/dividers could be integrated in the oil tank too, especially at this relatively low voltage level (around 66 kV RMS line-line in the most stringent ± 50 kV case). This type of complex transformer design is a common practice on high power, medium voltage diode rectifiers for variable speed motor drives, which require several secondary windings with different vector groups. Thanks to the common oil tank, integrated sensors and direct oil-to-cable terminations toward the converters, the global footprint remains limited. It is considered that, for D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 20 SCARLET, the converter-side AC voltage is moderate enough (66kV or 33kV) to adopt a comparable design. This kind of integration has also been proposed for HVDC [9]. On the DC side, the SCARLET project adds DC circuit breakers and their associated 𝑑𝐼/𝑑𝑡 limiting reactors, one per pole and per converter feeder (refer to section 6). Disconnectors and earthing switches are required too, and it is expected that their marginal footprint increase is negligible with respect to the breaker itself and its reactor. They connect each converter feeder to the common DC busbar, which goes to the superconducting fault current limiter (SCFCL), then to the superconducting cable termination. 3.4 Resulting +/- 50 kVdc architecture Following these guidelines, the simplified single line diagram (SLD) of the ±50 kV substation is built and presented in Figure 15. The disconnectors, earthing switches, current and voltage transformers are not represented (different options of busbar topologies are possible as described in 4.3). The lefthand-side of the AC busbar, where two incoming AC feeders are drawn, is more related to the transmission grid substation than to the AC/DC conversion station itself: the only assumption is the presence of a circuit breaker on each incoming feeder, which will serve the protection strategy as depicted in section 6. The MMC converters are not detailed, but they follow strictly the schematic presented in Figure 9, with six submodule stacks and six arm reactors connected on the DC-grid side. The ratings are as presented in Section 3.2.2: 44 submodules per arm, a nominal capacitor voltage of 2500 𝑉 and a DC current rating of 2500 𝐴. Figure 15 : Simplified Single Line Diagram of the proposed ±50 kV AC/DC substation architecture. AC busbar 400 kV DC busbar ± 50 kV Toward SCFCL + DC cables 400/60kV 50 kV 2500 A 400/60kV 50 kV 2500 A 400/60kV 50 kV 2500 A 400/60kV 50 kV 2500 A From transmission grid MMC MMC MMC MMC D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 21 As mentioned previously, four 250 𝑀𝑊 three-phase, two-windings 7 transformers are used. Assuming a reactive power requirement of 40% of the active power (that is, 100 𝑀𝑉𝐴𝑅 per converter), 270 𝑀𝑉𝐴 transformers are selected. Although it may be fine-tuned during the refined converter sizing step, a secondary-side voltage of 60 𝑘𝑉 is selected, adapted to the 100 𝑘𝑉 DC each MMC is providing. Finally, the DC circuit breakers are rated for 50 𝑘𝑉 −2500 𝐴 (each pole), with breaking capabilities to be defined. 3.5 Resulting +/- 25 kVdc architecture Following the same full parallelization approach presented in the last section, eight three-phase transformers with an individual rating of 135 𝑀𝑉𝐴 (125 𝑀𝑊) would be required. However, it was decided to use only four 3-winding transformers, with a primary rated for 270 𝑀𝑉𝐴 and two 135 𝑀𝑉𝐴 secondaries, each one feeding a separate MMC converter. Keeping the pre-charge system on the primary-side, it forces both converters of a pair to be energized at the same time, so a fault can cause a permanent loss of 1/8 of the nominal power capacity, but a temporary loss of 1/4 could be necessary to re-energize the faulty feeder after its maintenance. As a consequence, equivalent or better service continuity is obtained compared to the ±50 𝑘𝑉 case, and better in any case than a monolithic HVDC converter (no parallelization), which was considered sufficient. This architecture results in the same high voltage AC yard (breakers and precharge system), with a marginal complexity increase in the medium voltage AC yard. The resulting SLD is illustrated in Figure 16. Due to the lower DC voltage, the converter-side AC voltage has also been reduced: keeping the same ratio, a 30 𝑘𝑉 line-to-line RMS voltage is adopted on each winding. 7 Not mentioning possible auxiliary windings. An auxiliary tertiary winding is a common practice in HVDC, providing one source of power for the control system (among others, like the local distribution grid). D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 22 Figure 16 : Simplified Single Line Diagram of the proposed ±25 kV AC/DC substation architecture The MMC converters feature less submodules (22 per arm), keeping the total count constant in the converter station, but are otherwise identical. Finally, the DC circuit breakers have a two times lower nominal voltage, with the same nominal current. 3.6 Expected ripple current characteristics In HVDC, the harmonic currents caused by the switching actions of a MMC are normally negligible from the AC and DC PCC, which is a consequence of the high count of submodules (high number of voltage levels), and of the relative insensitivity of conventional cables to these low amplitude harmonic currents. However, medium voltage superconducting transmission raises two concerns with this regard: - First, the lower DC link voltage allows a vast reduction in the number of submodules, which increases the voltage harmonic distortion mutatis mutandis; - Second, ac components, superimposed on the DC component, induce AC losses in the superconducting tape or filament. Those losses are, at first approach, proportional to the product of the ac component amplitude by its frequency. The sensitivity to these losses depends case by case on the selected superconducting material and cable design. MMC MMC MMC MMC MMC MMC MMC MMC AC busbar 400 kV From transmission grid 25 kV 2500 A 25 kV 2500 A 25 kV 2500 A 25 kV 2500 A 25 kV 2500 A 25 kV 2500 A 25 kV 2500 A 25 kV 2500 A DC busbar ± 25 kV Toward SCFCL + DC cables D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 23 Consequently, the harmonic current emission from the DC link terminals shall be assessed carefully. Due to the complexity of the MMC converter itself and its control and modulation scheme, this task is not trivial, and will be performed during the project. At this stage, only guidelines can be provided. The MMC can be represented as an equivalent Thevenin generator with an inductive impedance, which is a function of the arm, equivalent AC network and DC network inductances. The order of magnitude of this value is the mH. The cable link being inductive itself, with an estimated inductance of several hundreds of mH for 100 km, it provides a drastic decrease of the harmonic current emission of the converter. This is illustrated in Figure 17. Figure 17 : Equivalent circuit showing the influence of the cable link inductance on the DC current ripple produced by the MMC. The Thevenin parameters have yet to be determined but are estimated to a few kV at several kHz for the former, and a few mH for the latter. The cable inductance is estimated to be one or two orders of magnitude higher. The Thevenin voltage, on the other hand, depends on the converter control and/or modulation scheme. Although the number of levels is reduced with respect to HVDC (e.g., from 250 to 20 in the most stringent ±25 kVdc), it is still significant. As submodules switch one after the other, voltage steps of only 2500 V are generated (one submodule capacitor voltage). Owing to the low harmonic emission of the MMC, and the positive contribution of the cable link itself, the first analysis provided a good confidence in the cable-converter compatibility in the case of the SCARLET project. The total harmonic distortion of the DC cable ripple current is expected to be 1% at most, implying RMS ripple currents below 100-200 A. Its bandwidth should be limited to 10 kHz, with a nondeterministic spectrum, as the switching frequency is typically not chosen at an integer multiple of the fundamental frequency [10]. A finer analysis will be performed in future work once the low-level control scheme is determined. Vth 0V Load-sideConverter-side Lth ΔL D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 24 4. OFFSHORE WIND PARK 4.1 Offshore windmill conversion chain As stated in section 1, one main interest of a superconducting transmission link for offshore windfarm export is the suppression of the bulky offshore AC/DC conversion platform. Hence, the collection grid needs to be in DC and at the transmission voltage level, that is, either ±50 𝑘𝑉 or ±25 𝑘𝑉. In this section, a typical AC-collected offshore wind farm electrical architecture will be described as a reference, then the proposed philosophy and architecture for the MVDC collection network will be presented. Fully-fed (type IV), high power, medium-voltage wind turbines are considered, with a power rating of 10 𝑀𝑊 per turbine. 4.1.1 Current typical conversion chain Given the power level of the wind farm, a 66 𝑘𝑉 AC collector network is considered as a reference. An illustration of the wind turbine conversion chain is provided in Figure 18. The Rotor-Side Converter (RSC), the Grid-Side Converter (GSC) and the DC capacitor bank in-between are typically in the same enclosure. Some additional elements, like input and output LC filters and a DC bus chopper, are not presented. Figure 18 : Block diagram of a wind turbine made for AC collection. The conversion chain is highlighted in green. As an example, the schematic of PCS6000 wind turbine converter from ABB [11] is presented in Figure 19, and its internal layout is shown in Figure 20. This converter is then connected, through medium voltage cables, to the turbine transformer increasing the voltage up to 66 𝑘𝑉. The primary winding of the transformer is connected to the collection network through 72.5 𝑘𝑉 Gas Insulated Switchgear (GIS). D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 25 Figure 19 : PCS6000 wind turbine converter [11], based on three-level NPC inverters with RC-IGCT semiconductors. The nominal ratings are 3.3kV (generator and grid) and 7 MW (8.5 MVA). A braking chopper is included on the DC bus. Figure 20 : PCS6000 wind turbine converter internal layout [11], corresponding to Figure 19. All the medium voltage equipment (converters, filters and switchgear) is included inside the cabinet, excepted the braking resistor. This architecture is illustrated in Figure 21. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 32 Figure 28 Two monopoles HTS cable as designed within Work Package 2 Figure 29. Layout of one MgB2 monopole cable as designed within Work Package 4 The cable is placed inside a double wall stainless steel cryostat. The two walls are thermally insulated with vacuum and spacers, while liquid nitrogen or liquid hydrogen flows inside the inner tube to cool down the cable. The internal and external pipes are grounded at the cable extremities together with the screen. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 33 The two monopole cables have the same structure, one is used for the cooling fluid one way circulation, the other is used for the cooling fluid return circulation. The two cables can be separated by the distance “d”, as it shown in Figure 30. The impact of this distance will be on: - Installation logistics, cost and right of way: the simplest is to lay the two monopoles together at the same time in the same trench. - Apparent magnetic field: adjacent cables will not exhibit an external permanent DC magnetic field, each of the cables screening the other. - Mutual Inductance characteristics that have possibly an impact on short circuit current for fault at the load side of offshore side of the cable. The mutual inductance might also have an influence on transient current circulating in the different conductive layers of the cable. Cable shield grounding also has an influence on the transient currents in the different conductive layers. Depending on the cable characteristics, it might be necessary to implement a grounding resistance of several hundreds of mOhms. Detailed studies will be necessary to give recommendations to the manufacturer. Figure 30 2 monopole cables layout To evaluate the cable behaviour (more precisely the thermo-electrical behaviour of the conductive layers) in different shield grounding conditions and transient events (faults and loading), WP5 will establish the detailed electrical model of the cables. As the HTS cable differs from the MgB2 cable, two specific models will be built. This electrical model will have to consider the capacitive and inductive components as well as the temperature and current dependent resistive components of the conductors. The inductive components will be built in the form of a direct and mutual inductance matrix, where all the mutuals between all the conductors will be evaluated: for instance, if we consider a cable structured with former + superconductive layer + shield + inner tube + outer tube, a 10 x 10 inductance matrix will finally be built so as to consider the coupling between all layers of the two monopoles laid at the given separation distance “d”. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 34 5.2 Cryogenic machine auxiliary power supply 5.2.1 General considerations The SC Cable needs machinery to maintain the coolant temperature and pressure at the right operational values all along the cable length. This is described in more detail in work packages 2 and 4. However, in work package 5, we are looking at the feasibility of the feeding of such machinery through a dedicated auxiliary power supply, especially in the case of an offshore cable. We will focus on the HTS cable, operating with LN2 as a coolant, to establish the requirements of the auxiliary power supply. Results and conclusion of this study could then be transposed to the MgB2 cable solution. operating with LH2. Note that, as regards as LH2 is concerned, the flow conditions are not governed by the MgB2 cable cooling requirements but by the business case of bulk LH2 transportation, contrary to the HTS cable case where LN2 flow conditions are exclusively determined by the HTS cable cooling requirements. The main hypothesis we consider are the following: - 100 km link. - Power supply requirements o low voltage supply 3Ph - 400 Vac o Cryogenic machines: 2 MW every 25 km o Pumps o Chiller o Platform air conditioning o Life base It results in a total power supply estimation of 3 MW every 25 km. - Consider that the first machinery is onshore and does not need to be powered by this dedicated line. - Consider that we need a switching platform at the SC cable extremity, therefore it could advantageously be used to house cooling machinery. Different distributions of the cryogenic intermediate platforms can be considered. A conceptually simple approach is shown in Figure 31, where the first cable section is cooled by the onshore substation, and each of the following sections is cooled by its shore-side extremity, through a dedicated offshore platform. Figure 31 : A possible layout of the cooling substations, with single-sided cooling. Off-shore switching platform 25 km 25 km 25 km 25 km Cooling platform On-shore station with cooling Cooling platform Cooling platform Coolant input/output D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 35 Using this architecture, the last intermediate platform to supply is one section closer to shore than link end; that is, 75 km for the considered case study (4 sections of 25 km). However, it requires three intermediate platforms. It is assumed that the number of the latter has a significant impact on the overall cost, pleading in favor of larger but less numerous platforms. Following this rationale, more intricate but possibly more cost-effective schemes are possible, such as shown in Figure 32. Figure 32 : An alternative layout of the cooling substations, featuring double-sided cooling. The hydraulic length is kept unchanged, while the total number of platforms (terminal included) is halved. In this schematic, the hydraulic connectors at 25 km and 75 km loop the coolant flow from one cable to the other, isolating both sides. Thus, each cooling unit has the same burden as previously, but only one intermediate platform is necessary instead of three, in addition to the cable extremity switching platform. The existing terminal platform is now also used as a cryocooling station, and the intermediate platform contains two of the latter. It should be mentioned that the second option is more demanding with respect to the auxiliary power supply: besides the first quarter, that is directly consumed on shore, half the total power is transmitted to mid-length (50 km), and the remaining quarter is fed to the transmission system extremity (100 km). This is not, however, expected to change the conclusion of the following part. 5.2.2 Considered architectures Two architectures of auxiliary power supply are introduced and discussed: - Feeding of the auxiliary power through the SC cable itself, using the SC cable former and pick the power via a 50 kVdc– 1 kVdc converter and then a standard commercial 1 kVdc – 3 Ph 400 V inverter - Supply with an additional three phase AC cable from shore As detailed above, the platform load has a complex nature, including both direct-on-line motors (e.g., pumps, fans), variable speed motors (e.g., chillers) and auxiliary loads (e.g., life base, control and monitoring system), most of them being expected to operate with a three-phase AC supply at 400 V, with currently available options. Hence, for both options, it is proposed to focus on this voltage level and to assume a lumped load, even though this solution is probably suboptimal regarding the losses (number of cascaded conversion stages). Indeed, given the power level (3 MW), if the cryo-machines could be adapted, an intermediate medium voltage may be interesting for the largest rotating Off-shore switching platform with cooling 50 km 50 km Cooling platform On-shore station with cooling Coolant input/output Hydraulic separation D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 36 machines, which is then stepped down again for the remaining lower power motors and auxiliary loads. This implies, however, the development of specific cooling systems, which is out of scope of this project. 5.2.3 Tapping converter In this option, the power required for the cryogenic machines to operate is fed by the DC cable itself, through an additional converter. Assuming a DC link voltage of ±50 𝑘𝑉 (with no loss of generality), this converter transforms 100 kVdc into 3x400 Vac. For practical reasons, it is proposed to separate the latter in two conversion stages: one being a DC/DC converter stepping down the 100 kVdc into a low intermediate voltage (typ. 600-1000 V), which is then fed to a conventional three-phase, lowvoltage inverter (e.g., 750 Vdc to 3x400 Vac). This architecture is presented in Figure 33. Figure 33 : Auxiliary supply principle with a tapping converter. All the platform load (in blue) is assumed to be fed in 3x400 V, which is supplied by the grid-forming auxiliary converter (in red) from the superconducting cable itself. At first, the tapping converter appears to be a very attractive option: first, it makes the cryogenic platforms operate independently, with no electrical link besides the SC cables. Secondly, as the power is not taken from the shore but regularly on the main link, no additional transmission distance constraints are introduced by the auxiliary power system. Finally, no change is required on the superconducting cable system, keeping its manufacturing and installation cost as low as possible. However, the first analysis brought attention on several challenges and limitations: - The 100 kVdc – 750 Vdc tapping converter is not commercially available and may require complex and expensive developments. Indeed, its complexity is not much inferior to that of the main AC/DC onshore converter, even though its current rating is very limited, due to its high voltage constraints. Indeed, due to the lack of a market, no commercially available AC/DC converter hardware is adapted to this voltage and current levels 11 . 11 This affirmation may look to be conflicting with the proposed solution for the wind turbines DC/DC converters in this report, which also has to manage this high voltage at comparable currents. What massively reduces the 750V 400V M 3~ AC DC DC AC Aux. M 3~ DC DC DC AC ±50kV SC cable SC cable D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 37 - This DC-DC converter is likely to be significantly oversized due to the lack of appropriated power semiconductors (high voltage, low current), with adverse effects on its efficiency and footprint, impacting the footprint of the intermediate platforms. This increase is critical, as it must not outweigh the footprint reduction of the offshore conversion platform, which is one of the main expected benefits of superconducting transmission. - The tapping converters introduce, by nature, a coupling between the main and auxiliary power systems. As a consequence, their possible contribution to faults, the stress they receive from the latter and their ability to quickly restore the power flow after faults must be carefully assessed. - Instead of a simple hydraulic connection to the pipes for the coolant inlet/outlet, which is at ground potential, tapping the auxiliary power from the main line requires an access to the inner superconducting cables. The termination is thus more complicated, with electrical insulation, and may therefore require a specific development. - Finally, before power transmission, the whole system must be cooled from ~300 K to ~77 K. Due to thermal constraint gradients of the cable accessories, the expected duration of this cooling phase is not less than 10 hours, during which all the cryogenic machines operate at full power. In other terms, the auxiliary power is maximum while the main cable is not superconducting yet, implying that its copper core must be able to carry it in steady state like conditions without any degradation and without adding significant heat load. 5.2.4 Feeder cable Due to these limits, the more conservative option of an auxiliary cable system is also considered. In this case, a three-phase AC cable would be installed along with the main superconducting cable system, providing all the auxiliary power from the onshore AC grid. This architecture is presented in Figure 34. WT converter complexity is its unidirectionality in the increasing voltage direction, which allows for the use of a passive rectifier bridge on the high-voltage side. On the other hand, the auxiliary converter is unidirectional but step-down, such that its high voltage bridge must be active. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 38 Figure 34 : Auxiliary supply principle with a feeder cable. All the platform load (in blue) is assumed to be fed in 3x400 V, which is supplied by a MVAC feeder cable through a local three-phase transformer (in red), from the shore power. Echoing what has been done for the tapping solution, the following benefits are expected: - Using only a power transformer at the onshore connection point, and a transformer on each intermediate platform, this solution is more simple, commercially available, and more reliable. - As no correlation is forced between the main and auxiliary voltages, the latter can be tailored to the project needs regarding distance and auxiliary power consumption. - Besides a possible magnetic coupling due to their proximity, the main and auxiliary power systems are electrically decoupled, which simplifies the fault management. - As a corollary, the energization and cooling phase is vastly simplified, as the auxiliary cables are not superconducting and do not require to be heavily cooled down to be operational. However, using an auxiliary cable also introduces drawbacks, among which are the following: - This solution is likely to be unsuitable to very long links, because of the voltage drop and charging current of the cables. Mitigation solutions exist to a certain extent. - For the ease and cost of installation, the auxiliary cables are likely to be bundled with the superconducting cable system or at least to be installed at the same time in the same trench. Many design parameters are available, both regarding the electrical design and the ease of installation, such as the nominal voltage of the auxiliary supply, the cross-section and insulation thickness of each cable 12 , and the use of reactive power compensation systems in the intermediate platforms. Although these choices remain to be decided, the initial assessment performed for the SCARLET project did not 12 Using a higher voltage class cable (e.g., 66kV cable in 36kV RMS) reduces the capacitive charging current. 400V M 3~ AC DC DC AC Aux. M 3~ ±50kV 36 kV AC SC cable Aux cable SC cable Aux cable D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 39 raise feasibility issues for its length, configuration, and power requirements, therefore the use of an auxiliary cable is currently the preferred solution. 6. PROTECTION STRATEGY To protect the electrical network, a selective fault clearing strategy combining DC circuit breakers and RSFCL is proposed that can be applied on production cases, consumption cases as well as the point-topoint cases. Combining the two types of protection devices relaxes the stresses on the superconducting cable, and, at the same time, the DC breaker reduces the burden on the RSFCL and also allows for continuity of service in case of loss of one of the parallel converters. This last advantage is a natural consequence of the parallel onshore converter architecture, as seen in sections 3.4 and 3.5. The faulty converter branch can be isolated within a few milliseconds thanks to the DC CB, and power can still flow with the remaining modules. Moreover, each of the AC feeders has an AC circuit breaker that will naturally be tripped in case of fault on this feeder or its transformer. But important as well, it will be used as a backup breaker in case of failure of the DC breaker of the same converter branch. In such an event, due to the much longer opening time of the AC breaker, the current limitation property of the RSFCL will be of very high impact on the SC cable protection. In this strategy, and to design properly the protection devices, a mapping (Figure 35) of the possible fault is necessary as well as defining the protection sequence in each case. To simplify, we will distinguish pole to ground faults and pole to pole faults. We will also distinguish faults on DC bus bar and SC cable from faults on one converter branch or one offshore string. Pole to ground faults will have very small currents limited by the AC transformer neutral grounding impedances. The counterpart is the voltage shifting of the non-faulty pole that will have to be managed by all the components. That should be the case because this constraint is considered in the existing testing guides of HVDC equipment that will be used as guidelines or standards to qualify the electrical system components. Pole to pole faults are the worst cases in terms of prospective currents and they will be a designing factor both for the protection devices and for the design of the ±50 kV and the ±25 kV cables. Depending on the location, the prospective fault current can reach several tens of kiloamperes (as illustrated in Figure 3). Detailed simulations based on this fault mapping will be performed in Work Package 5 and presented in Deliverable D5.2. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 40 Figure 35: Mapping of fault cases 6.1 DC circuit breaker DC circuit breakers differ from AC Circuit breakers in the sense that the fault current does not exhibit any natural current zero as illustrated in Figure 36. The device must generate its own interruption opportunities with these constraints: - Interruption has to be faster than in AC within a few milliseconds instead of the typical value of 3 cycles of the power frequency. - The DC CB has to absorb magnetic energy in order to force the current decaying to zero. Figure 36: representation of a fault clearance in AC and in DC Different technologies are proposed by manufacturers and research organisations as described in [17], and range from full electronic devices to mechanical devices with counter current injection, through hybrid devices combining both advantages of power electronics and mechanical devices. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 41 - All the technologies will have a limitation in permanent current capability in the order of 2500 A – 4000 Adc due to different reasons. - Electronic or hybrid breakers use the benefit of the switching capability and rapidity of components like IGBTs. - Mechanical breakers use the benefit of their simplicity but naturally have slower reaction times than electronic components (milli-seconds against micro-seconds). To minimize their opening time, the moving mass (therefore the mass of necessary copper to mitigate the temperature rise under nominal current condition) must be very low, impairing the permanent current capability. - For the reason of permanent current, a DC CB cannot be inserted in series of the superconducting cable (where we target 10 kA or 20 kA nominal current depending on the SC cable technology) unless many modules are assembled in parallel which makes difficult the control, the synchronism, and the current sharing of the parallel modules. - The parallel topology of the converter gives the opportunity to install the protection DC CBs in each of the parallel converter branches, that has a permanent current compatible with the CB, as depicted in Figure 15 and Figure 16. It also provides the advantage of selectively isolate any of faulty converter branch. - The proposed DC CB for the SCARLET application is based on one module of a mechanical HVDC CB [18] as illustrated in Figure 37. - One module is of 35 kVdc nominal voltage and can interrupt up to 20 kA in less than 5 ms. - For the 50 kVdc application, two modules are used in series to reach 50 kVdc nominal voltage. - For the 25 kVdc application, only one module can be used. Figure 37: Full 525 mechanical DC breaker representation D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 48 Figure 44: Example of a Generator Circuit Breaker disconnector D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 49 7. CONCLUSIONS In this deliverable, we showed that transporting 1 GW in direct current at medium voltage using a superconducting cable is possible provided several modifications of the classical HVDC schemes: • the onshore conversion architecture is made of lower voltage MMC Converters connected in parallel implying less series connected submodules. • these modules are feasible based on actual power electronics technology and industrial converters offered by different companies. • a 1 GW +/- 25 kVdc or +/-50 kVdc can be of symmetrical monopole scheme that offers higher simplicity compared to bipolar scheme. • the windmill conversion chain has to be modified in order to get directly the right transport output DC voltage instead of 3 phase ac voltage. This can be done at the simplest by keeping the actual wind turbine converter but adding in series a compact 3-winding transformer and 12-pulse rectifier. • the converter parallel structure allows to use DC breakers that can advantageously perform selective protection. • to achieve maximal protection of the SC cable, a RSFCL is proposed in coordination with the DC CBs. • high nominal current switches are available on the market or could be upgraded by manufacturers to meet the high current ratings in the converter branches or windmill strings on one hand and on the common busbars on the other hand. The next activity in Work Package 5 is the specification of stresses seen by the different components of this electrical network. They can be evaluated by implementing the proper component models, especially those of the superconducting cable and resistive fault current limiter. These models are developed to be implemented in standard Electro Magnetic Transient Programs that will be used to simulate the whole electrical system. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 50 BIBLIOGRAPHY [1] SuperGrid Institute, «NSWPH Validation Technical Requirements,» North Sea Wind Power Hub, 2022. [2] R. M. Anton Lesnicar, «An Innovative Modular Multilevel Converter Topology Suitable for a Wide Power Range,» de Power Tech Conference, Bologna, Italy, 2003. [3] Siemens Energy, «HVDC Plus - the decisive steap ahead,» [En línea]. Available: https://assets.siemens-energy.com/siemens/assets/api/uuid:83dc80ed-cbd6-4027-9e4883a87eab8e0b/263_160390_ws_hvdcplususlowres.pdf. [4] S. Dennetière, «Modeling of Modular Multilevel Converters for the France-Spain link,» de International Conference on Power Systems Transients, Vancouver, Canada, 2013. [5] ABB, «PCS 8000 AC Excitation - AC Excitation for hydro pymp energy storage plants,» [En línea]. Available: https://library.e.abb.com/public/2a266c635cf88923c12576c500311a95/PCS%208000%20AC %20Excitation.pdf. [6] Hitachi, «More power for railway lines,» [En línea]. Available: https://search.abb.com/library/Download.aspx?DocumentID=8DAN000001&LanguageCode=e n&DocumentPartId=&Action=Launch. [7] GE Power Conversion, «MM7 Modular multilevel converter (MMC) for various applications,» [En línea]. Available: https://www.gepowerconversion.com/sites/default/files/202111/GEA34256_Modular%20Multilevel%20Converter_FLY_en_20210322.pdf. [8] ABB, «Applying IGCTs,» [En línea]. Available: https://library.e.abb.com/public/c3c6e3a514494584a6ea77a67d0cb9d0/Applying%20IGCTs_ 5SYA%202032-04-16-06-2016.pdf?xsign=PrY+Tm5QsMhDjIM+CRYmLsxEyb9k8Qa2WuhuhRJE5WT2lZEHaSgREVrCQfid1tW5. [9] S. Seman, «Modelling of the Diode-Rectifier Based HVDC Transmission Solution for Large Offshore Wind Power Plants Grid Access,» de Transmission Networks for Offshore Wind Power Plants, Vienna, Austria, 2016. [10] C. B4-110, «VSC-HVSC Transmission with Cascaded Two-Level Converters,» 2010. [11] ABB, «ABB wind turbine converters - PCS6000, full power converter, up to 15MW,» [En línea]. Available: D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 51 https://search.abb.com/library/Download.aspx?DocumentID=3BHS351272&LanguageCode=e n&DocumentPartId=&Action=Launch. [12] Alex Alefragkis; Sertkan Kabu, «Next Generation Offshore Grid Connection Systems: TenneT’s 2 GW Standard,» ELECTRA n° 321, 2022. [13] TenneT, «T01 - 2 GW HVDC Grid Design and Reliability & Availability,» TenneT position paper, 2019. [14] datasheet ABB , «Gas-insulated Switchgear ELK-04». [15] Energinet, «OFFSHORE ENERGY HUBS - High level system description and requirements,» 2023. [16] M. Ahmadi, A. Bertinato and I. Boussaad, «Reliability analysis of the bus-bar systems in multiterminal HVDC systems,» 23rd European Conference on Power Electronics and Applications (EPE'21 ECCE Europe), Ghent, Belgium, 2021. [17] J. B4/A3.80, «Design, Test and application of HVDC circuit breakers,» CIGRE, 2022. [18] W. Grieshaber; D.L. Penache; Y. Yang; F. Robert; I. Aït Brahim, «Adjustable Active Current Injection DC circuit breaker for MV and HV,» IEEE-PES-GM, 2022. [19] M. Stemmle, F. Merschel, M. Noe, A. Hobl, «Ampacity_project__Worldwide_first_superconducting_cable_and_fault_current_limiter_ins tallation_in_a_German_city_center,» CIRED2013. [20] M.J. Kim, S.J. Kim, G.H. Lee, C.Y. Bae, Y-G. Kim, «Development of a 22,9 kV/2000 1 Compact RSFCL,» CIGRE General Session A3_PS1_10626. [21] M. Moyzych, P. Ustyuzhanin, E. Magommedov, A. Telnova, L. Sabirov, K. Baburin, S. Samoilenkov, «Results of two-year operation of 220 kV pilot high temperature superconducting fault current limiter (SFCL) in Moscow power grid,» CIGRE General Session 2022 - A3_PS1_10357. [22] M. Song, S. Dai, C. Sheng, L. Zhong, X. Duan, G. Yan, Y. Huang, C. Shen, L. Li, L. Cai, T. Ma, «Design and Performance Tests of a 160 kV/1.0 kA DC Superconducting Fault Current Limiter,» Physica C: Superconductivity and its Applications. [23] Datasheet COELME - EGIC, Available: https://www.coelmeegic.com/fr/attachment/827_(DS)_CBD_36-52kV_FR_(2020-05).pdf. [24] Datasheet SDCEM, Available: https://www.sdcem.com/products/sbe-sectionneur-a-coupureverticale-dexterieur/. D5.1: Selection of electrical system architecture This project has received funding from the European Union’s Horizon Europe research and innovation program under grant agreement No 101075602 52 [25] Datasheet HITACHI, Available: https://www.hitachienergy.com/products-andsolutions/generator-circuit-breakers. [26] Datasheet GE Grid Solutions, Available: https://www.gegridsolutions.com/hvmv_equipment/catalog/generator_circuitbreakers.htm.