Topology Selection for Large-Scale Offshore Wind Power HVDC Direct Transmission to Load Centers: Influencing Factors and Construction Principles

Abstract

The development and utilization of large-scale offshore wind power (OWP) are critical measures for achieving global energy transition. To address the demands of future large-scale OWP centralized development and transmission, this study systematically investigates the influencing factors and construction principles for topology selection in offshore wind power high-voltage direct current (HVDC) transmission systems delivering power to load centers. First, under the context of expanding the offshore wind power transmission scale, the necessity of transmitting OWP via HVDC overhead lines directly to load centers after landing is theoretically discussed. Five key topological influencing factors are then analyzed: offshore wind power collection schemes, multi-terminal HVDC network configurations, DC fault isolation mechanisms, offshore converter station architectures, and voltage source converter HVDC (VSC-HVDC) receiving terminal landing modes. Corresponding topology construction principles for direct HVDC transmission to load centers are proposed to guide system design. Finally, the feasibility of the proposed principles is validated through a case study of a multi-terminal HVDC system integrated into an actual regional power grid, demonstrating practical applicability.

1. Introduction

Wind energy is emerging as a crucial source of electric power for industrial and domestic applications, owing to its distinct advantages over fossil fuels [1]. Offshore wind power offers notable benefits compared to onshore installations, including zero land resource occupation, richer wind energy potential, and elimination of noise pollution [2]. According to the Global Wind Energy Council (GWEC) forecast, global offshore wind installed capacity will continue its upward trajectory over the next decade, with anticipated new installations reaching 410 gigawatts (GWs) [3]. Traditional offshore wind power development used AC transmission, but AC submarine cables’ capacitive effect limits transmission distance to around 80 km [4]. As offshore wind projects increase in transmission distance and capacity, VSC-HVDC (especially with MMC) has become mainstream for far-offshore wind power transmission due to its unrestricted distance and flexible control [5].

Globally, Europe leads in offshore wind power DC transmission. It is projected that Europe’s offshore renewable energy will reach 496 GW by 2050 [6]. ENTSO-E estimated that investment in European offshore transmission infrastructure will total approximately EUR 400 billion between 2025 and 2050, with over half allocated to offshore and onshore converter stations. In North America, particularly the United States, relevant projects are being vigorously promoted: from 2025 to 2030, efforts will be made to coordinate state-level planning of offshore transmission networks and standardize HVDC technical standards, and a national HVDC testing and certification center is planned for 2030–2040 [7]. China’s relevant policies have set requirements for offshore wind power development in terms of spatial layout and intensive transmission, with coastal provinces responding actively. China Southern Power Grid’s Yangjiang Sanshan Island Offshore Wind Power HVDC Project, adopting an integrated land–sea DC transmission scheme, is scheduled to be commissioned in October 2026, and will then become the world’s largest offshore wind power DC transmission project. Over the coming decades, research into the construction of safe and effective HVDC transmission topologies will be of substantial significance in supporting the large-scale development of global offshore wind power.

The overall configuration of an OWP grid integration system via VSC-HVDC is illustrated in Figure 1, comprising three core components: wind power collection, VSC-HVDC transmission, and onshore grid integration.

With the scaling up of OWP development, individual planned projects now reach the ten-gigawatt scale. Addressing such ultra-large-scale OWP integration, the construction of DC transmission systems faces two primary challenges:

1.

Transmission Channel Resource Constraints in Collection and Transmission:

Marine and terrestrial transmission corridors are subject to significant resource limitations. Submarine cables and overhead lines are constrained by their current-carrying capacities, requiring multiple circuits to transmit tens of gigawatts of power. However, the availability of transmission corridors is increasingly scarce, complicating infrastructure development. Converter station construction costs also pose a critical issue; for example, the Rudong OWP VSC-HVDC demonstration project in China—with a transmission capacity of 1100 MW—incurred costs of CNY 1.8 billion for the offshore converter station (40% of total project investment) and CNY 1.2 billion for the onshore station (27% of total investment).

2.

Grid Integration Challenges at the Receiving End:

The receiving-end grid faces substantial pressure for secure power absorption. Unlike dispersed onshore generators, concentrated injection of tens of gigawatts from OWP imposes strict requirements on grid infrastructure. This large-scale power influx can cause AC line congestion, approaching thermal stability limits. Expanding AC grid infrastructure to relieve congestion may increase short-circuit currents at buses, risking violations of safety thresholds and compromising grid security.

Scholars have conducted extensive research on VSC-HVDC-based OWP grid-connected systems from multiple perspectives.

In terms of offshore wind farm collection, Reference [8] proposed a topology sequence for single wind farms and clusters based on offshore distance and scale, pointing out that the selection of schemes should consider economy, reliability, and sea area intensification. Reference [9] proposed an optimization method based on a graph genetic dynamic minimum spanning tree for the collection of large-scale wind farms, taking into account both cable current-carrying capacity and cross-avoidance constraints, thus improving the method system for the topology optimization of large-scale wind farms. Reference [10] proposed a hybrid AC/DC collection topology on the basis of traditional AC collection, which reduces the cost and transmission loss of collection in large-scale wind farms.

Existing studies on VSC-HVDC transmission systems have primarily focused on lightweight DC system designs. SIMENS proposed using a full-power diode rectifier unit (DRU) for offshore converters, which offers low losses, high power density, and reduced platform size/cost. However, DRU lacks control capability, requiring grid-side converters to employ grid-forming control for autonomous AC voltage establishment [11,12]. Hybrid rectifiers combining DRU and VSC leverage DRU’s cost advantages while using auxiliary VSCs for voltage support. Reference [13] proposed a topology for offshore converter stations where DRU and MMC are connected in parallel on the AC side and in series on the DC side, and verified its control strategy and AC/DC fault ride-through capability through simulations. This topology reduces MMC DC-side voltage and IGBT requirements but cannot support wind farm black starts via reverse power flow. Reference [14] studied starting strategies for parallel DRU-MMC systems, avoiding MMC backflow through power distribution optimization, while Reference [15] investigated harmonic suppression and MMC capacity selection for reactive compensation. These schemes, despite using smaller MMCs, still require the full DC voltage withstand capability, leading to numerous sub-modules.

Regarding the grid connection of offshore wind power via the VSC-HVDC system, Reference [16] optimized the receiving-end access points of VSC-HVDC by considering wind power fluctuations, the voltage support capacity of VSC-HVDC converter stations, and investment and construction costs, aiming to ensure the voltage stability of buses in the access area. However, this study failed to consider the transmission capacity of the receiving-end AC network and the power injection of offshore wind farms. Reference [17] further factored in the wind power accommodation capability and bus vulnerability of the receiving grid, refining the grid connection point selection for offshore wind farms integrated through VSC-HVDC. Nevertheless, these studies were all limited to single-site selection for point-to-point VSC-HVDC systems, with no research conducted on the parallel connection point selection for multi-terminal VSC-HVDC systems that feature multiple converter stations at the receiving end.

A review of extant literature reveals that current research on offshore wind power VSC-HVDC transmission system construction predominantly focuses on isolated system components, lacking comprehensive analysis of integrated architectures for offshore wind power DC transmission and grid integration. This research gap impedes the formulation of systematic guidelines for future large-scale OWP development. Addressing this, the present study first evaluates the necessity of direct HVDC transmission to load centers via submarine cables following onshore grid integration through a comparative analysis of two typical connection modes for large-scale offshore wind power DC transmission systems. It then systematically discusses multi-dimensional factors influencing topology selection for offshore wind power DC transmission systems, and proposes topology construction principles tailored for large-scale OWP transmission and grid connection. Finally, the proposed principles’ feasibility is validated through a case study of offshore wind power DC transmission system topology design in a regional power grid context.

2. Necessity of Direct DC Transmission from Large-Scale Offshore Wind Farms to Load Centers

Load centers represent critical zones for absorbing large-scale OWP. When transmitting OWP via DC submarine cables to onshore landing points, two distinct grid integration approaches emerge if the AC grid load center is geographically distant from the landing point:

• Approach 1: Construct a converter station at the landing point to convert DC to AC, integrating power into the coastal AC grid. Power is subsequently transmitted to the load center via AC overhead lines.
• Approach 2: Extend the DC submarine cable into a DC overhead line for direct power delivery to the load center, where a converter station converts DC to AC for grid integration.

Currently, Approach 1 dominates China’s OWP DC transmission projects, leveraging existing AC grid infrastructure to avoid constructing new long-distance overhead lines, as required in Approach 2, thereby reducing capital expenditure. However, as OWP integration scales up, the limitations of Approach 1 become increasingly pronounced:

• Elevated System Operational Risks: Large-scale wind power transmission through AC grids induces significant power flow redistribution, leading to uneven power distribution across transmission interfaces. This increases overload risks for lines during both steady-state operation and N-1 contingency scenarios. Mitigating overloads via AC grid expansion often exacerbates short-circuit current violations at buses.
• Substantial Voltage Drop Issues: The long distance between load centers and landing points causes substantial voltage drop during AC transmission. This drop intensifies with increasing power transfer, potentially resulting in unacceptable low voltage levels at load centers.
• Higher Transmission Losses: AC overhead lines exhibit inherently greater transmission losses than DC overhead lines of equivalent voltage rating, a discrepancy that amplifies significantly over long distances.

In contrast, Approach 2 enables direct power delivery to load centers via DC overhead lines, facilitating localized OWP consumption. This configuration eliminates challenges associated with AC power flow redistribution and long-distance voltage instability while reducing transmission losses. Economically, DC transmission typically outperforms AC transmission in cost-efficiency when excluding converter station investments. Therefore, for large-scale OWP integration, direct DC transmission to load centers offers distinct advantages—especially when converter station construction at load centers is technically and financially feasible. The key characteristics of the two grid connection modes are compared in

Table 1. Comparison of characteristics between the two grid integration approaches.

Comparison Dimension	Approach 1	Approach 2
transmission path	landing point → onshore converter station → AC overhead lines → load center	landing point → DC overhead lines → onshore converter station → load center
system operational risk	higher due to power flow redistribution	lower due to localized power consumption
magnitude of voltage drop	larger	smaller
transmission loss	higher	lower
economy	short-term favorable	long-term favorable
applicable scenarios	small-scale and short-distance transmission	large-scale and long-distance transmission

.

3. Topology Construction for Direct DC Transmission from Offshore Wind to Load Centers

3.1. Cluster Collection of Offshore Wind Farms

Although DC submarine cables have no inherent transmission distance limitations, their power-carrying capacity remains constrained by physical properties. To optimize marine transmission corridor utilization, integrating multiple offshore wind farms (OWFs) into appropriately scaled clusters before high-capacity DC submarine cable transmission is essential. Key factors governing cluster configuration design include the following:

• Geographical Layout of Wind Farms:

The spatial distribution of OWFs varies significantly due to offshore wind resource heterogeneity and marine environmental constraints. Proximally located wind farms enable straightforward cluster integration, whereas widely dispersed installations necessitate long-distance AC collection systems—incurring substantial transmission losses and cable-laying costs that compromise economic viability.

2.

Installed Capacity of Wind Farms:

Individual far-sea wind farms typically range from hundreds of megawatts to gigawatt-scale, with notable capacity disparities between projects. Cluster size selection requires case-specific engineering evaluation, considering actual capacity distributions and grid integration requirements to balance technical feasibility and cost-effectiveness.

3.

Current-Carrying Capacity of DC Cables:

In OWP DC transmission systems, submarine DC cables primarily serve power transmission. Given their fixed current ratings, the total capacity of integrated wind farm clusters must not exceed the cable ampacity limits to ensure safe and efficient operation.

Table 2. Transmission capacity of HVDC submarine cables.

Cross-Section (mm2)	Power (MVA)
	±200 kV	±250 kV	±320 kV	±400 kV	±500 kV
1 × 1000	556	677	852	1057	1307
1 × 1200	608	741	932	1157	1430
1 × 1400	667	815	1023	1272	1575
1 × 1600	722	883	1112	1382	1708
1 × 1800	769	941	1187	1471	1819
1 × 2000	820	1004	1266	1570	1941
1 × 2500	-	1145	1444	1791	2215
1 × 3000	-	-	1601	1986	2455
1 × 3500	-	-	1768	2193	2712

summarizes the transmission capacities of existing DC submarine cables across standard voltage classes, providing a basis for cluster capacity planning.

3.2. Offshore Converter Station Topology

In far-sea OWP DC transmission systems, full-power MMCs are typically deployed at both sending and receiving ends. The offshore MMC employs V/f control to provide voltage support for offshore wind farms, while the onshore MMC utilizes grid-following control with constant DC voltage and constant reactive power or AC voltage regulation. In ultra-large-scale OWP centralized transmission scenarios, the prohibitive investment cost of the offshore MMC necessitates topology innovations for system lightweighting. Three alternative offshore converter configurations exist:

• Topology 1: AC-parallel DC-series connection of MMC and DRU;
• Topology 2: AC-DC-parallel connection of MMC and DRU;
• Topology 3: full-power DRU.

A comprehensive technical assessment of these topologies is presented in

Table 3. Technical comparison of different topology schemes for offshore converter station.

Topology	AC Voltage Support Capability	Black Start Capability	Reactive Power Compensation Devices	Control of Wind Generator
Full-power MMC	Offshore MMC	Onshore MMC	Not Required	Grid-following
Topology 1	Offshore MMC	Auxiliary power supply system	Not Required	Grid-following
Topology 2	Offshore MMC	Onshore MMC	Not Required	Grid-following
Topology 3	Wind power generator	Auxiliary power supply system	Required	Grid-forming

.

As indicated, Topology 2 exhibits equivalent technical performance to the full-power MMC in terms of control flexibility. Regarding offshore wind farm black start capability, both Topology 1 and Topology 3 lack power reversal functionality, necessitating additional auxiliary power supply systems to provide starting power for wind turbines. Notably, Topology 3 requires offshore wind turbine generators to adopt grid-forming control—a critical distinction from other schemes. Currently, grid-forming control technology for offshore wind turbines remains in the theoretical research stage, facing challenges such as unclear control mechanisms, complex parameter design, and the need to enhance fault ride-through capability [18,19]. Despite these hurdles, its long-term application potential is significant due to inherent advantages in system simplification.

Beyond technical disparities, converter station topologies differ notably in economic performance, particularly in investment and operational costs.

The primary cost drivers include switching devices, reactive power compensation equipment, and converter platform infrastructure. The full-power DRU utilizes only high-power diodes, significantly reducing switching device costs compared to MMCs—where IGBTs constitute the majority of expenses. Since MMCs can inherently provide reactive power and harmonic compensation for DRUs, additional filters or compensation devices are only required in the full-power DRU scheme. In terms of platform infrastructure, DRU converter stations achieve substantial size and weight reductions: compared to MMCs, a same-capacity DRU station reduces volume by 80% and weight by 67% [20], leading to proportional decreases in offshore platform construction costs.

Operational costs primarily consider the operating losses of different topological schemes. MMCs using IGBTs incur conduction, blocking, and energy storage device losses, accounting for approximately 1.43% of transmitted power. In contrast, DRUs only exhibit conduction and blocking losses, totaling around 0.417% of transmitted power [21]. This demonstrates that DRU-based converters can significantly reduce operational costs through lower energy dissipation.

3.3. Multi-Terminal DC Network Topology and DC Fault Isolation Scheme

With the large-scale development of new energy sources, the integration of renewable energy has formed a pattern featuring multiple renewable energy stations and multiple load centers, rendering the construction of multi-terminal DC systems particularly essential [22]. In the context of offshore wind power aggregation, typical multi-terminal DC topologies include star, loop, and hybrid configurations, as depicted in Figure 2.

The star topology (Figure 2a) connects all converter stations to a common DC bus via individual feeders, with each feeder rated to match the capacity of its connected converter station. This topology offers simple structure and low construction cost but exhibits inherent reliability limitations: feeder disconnection due to faults isolates the associated converter station from the grid. Offshore converter station disconnection results in partial power loss, while onshore station disconnection requires rerouting all power through remaining onshore nodes—mandating converter station capacity designs with redundant margins to accommodate such contingencies.

In contrast, the loop topology (Figure 2b) interconnects all converter stations into a closed-loop network during normal operation. Although this increases DC line investment, it significantly enhances system reliability: feeder faults only remove the affected segment from the loop without isolating any converter station, enabling full-power transmission to persist post-fault. A key tradeoff is the requirement for certain lines to handle the full system power capacity, imposing stringent design criteria on conductor size and insulation ratings.

The hybrid star–loop topology (Figure 2c) integrates advantages of both configurations: large-capacity converter stations are connected via the loop topology to ensure reliability, while small-capacity stations utilize the star topology to minimize line costs. This balanced approach optimizes both investment and operational resilience.

A critical factor in topology selection is the AC grid’s maximum tolerable power loss P_max-fail. If the worst-case power loss following a single DC fault falls within the receiving-end grid’s acceptable range, the star topology suffices; otherwise, the loop or hybrid topology with redundant paths becomes necessary. P_max-fail reflects the AC grid’s strength and frequency regulation capability—larger values indicate stronger grid resilience to power fluctuations. Thus, topology choice directly hinges on matching MTDC fault performance with the receiving system’s operational constraints.

DC line fault isolation in MTDC systems generally follows two technical approaches:

• Fault Self-Clearing Converter Valves: In MMC-based VSC-HVDC systems, this scheme utilizes full/half-bridge hybrid sub-modules or clamped dual sub-modules. Upon DC-side short-circuit faults, sub-modules detect overcurrent and immediately block operation, establishing reverse voltage through their capacitors to rapidly decay fault currents to zero. Subsequent isolation is achieved via disconnect switches. A critical limitation is that DC faults necessitate blocking all MMCs, causing system-wide voltage reduction and temporary transmission interruption—inducing significant power surges in the AC grid. This drawback is particularly pronounced in large-scale OWP-integrated MTDC systems due to their high power density and interdependent operation.
• DC Circuit Breakers: This approach enables precise fault isolation without system-wide voltage derating, offering distinct advantages for MTDC configurations. High-voltage DC circuit breakers employ numerous fully controlled devices, leading to substantial costs. Traditional designs require dual-end installation for each DC line, with the breaker count increasing linearly with line numbers—reducing economic viability as the system scale grows. In contrast, multi-port DC circuit breakers integrate breakers connected to the same DC bus, allowing shared usage across multiple lines. The assembly HVDC breaker proposed in Reference [23] features main breaking and sectionalizing components: main circuit breakers are installed at converter stations and common buses, while sectionalizing switches are configured per line. This architecture reduces the total breaker count significantly compared to traditional setups, with cost advantages amplifying as the number of terminals increases—making it a more economical choice for MTDC systems integrating large-scale offshore wind power.

3.4. Landing Modes of Receiving-End Converter Stations

The receiving-end converter stations of VSC-HVDC systems can adopt single-point landing or multi-point landing modes. Multi-point landing can be implemented via positive/negative pole differentiation or high/low voltage terminal separation. Taking a true bipolar connection system as an example, typical grid connection configurations are illustrated in Figure 3.

Landing mode selection is primarily governed by the receiving-end grid’s power acceptance capability. When single-point landing fails to ensure safe power flow diffusion, multi-point integration can be achieved by adding new outgoing lines or splitting VSC-HVDC power flows, subject to the following constraints:

• No line or transformer overload shall occur under steady-state and N-1 fault conditions;
• AC bus short-circuit currents shall not exceed the circuit breaker’s maximum interrupting capacity.

These requirements are evaluated using the following thermal stability margin and short-circuit current margin indices:

(1)

Thermal stability margin

This index quantifies the proximity of line/transformer transmitted power to its full-load capacity, defined as

$${\xi }_i={\displaystyle \frac{S_{\max ,i}-S_i}{S_{\max ,i}}}$$

(1)

where ξ_i is the thermal stability margin of branch or transformer i, S_max,i is the ultimate power that branch or transformer i can transmit, and S_i is the actual apparent power. Typically, ξ_i > 30% is required under steady state and ξ_i > 10% under N − 1 fault conditions.

(2)

Short-circuit current margin

This index characterizes the proximity of bus short-circuit current to its upper limit, defined as

$$K_{\mathrm{sc},i}={\displaystyle \frac{I_{\mathrm{b}\max ,i}-I_{\mathrm{sc},i}}{I_{\mathrm{b}\max ,i}}}$$

(2)

where K_sc,i is the short-circuit current margin index for bus i, I_bmax,i is the allowable short-circuit current upper limit (typically the circuit breaker’s interrupting capacity), and I_sc,i is the magnitude of the short-circuit current when a three-phase metallic short circuit occurs at bus i (obtained via power system simulation). For a 500 kV AC bus, I_sc,i > 63 kA is required, with K_sc,i > 0 essential for all buses.

When a fault occurs in the AC grid, the magnitude of the short-circuit current contributed by the MMC is related to its adopted low-voltage ride-through (LVRT) strategy. Generally, in strong AC grids (high bus short-circuit levels), the MMC should limit reactive current injection to reduce fault current contribution, while, in weak AC grids (low short-circuit levels), the MMC prioritizes reactive current injection for voltage support, maximizing its short-circuit current contribution.

3.5. Principles for Topology Construction of Offshore Wind Power Direct DC Transmission to Load Centers

Based on the analysis of influencing factors for each component of the OWP DC transmission and grid-integrated system, the proposed topology construction principles are as follows:

• To minimize offshore transmission corridor occupation, offshore wind farms should be clustered optimally, with cluster capacities matched to the transmission capability of existing DC submarine cables. This ensures efficient utilization of marine resources and reduces cable laying costs.
• Offshore converter station topologies shall adopt the most cost-effective configuration under technical constraints, considering comprehensive costs: wind farm black start devices, reactive power compensation systems, converter switching devices, offshore platform infrastructure, and long-term operational losses. Lightweight topologies (e.g., DRU-based hybrid configurations) are preferred for large-scale projects to balance reliability and economy.
• Ultra-large-scale OWP transmission is preferably implemented using VSC-MTDC systems. The DC network topology must ensure that the maximum power loss under single-line faults does not exceed the AC system’s tolerable threshold, P_max-fail, maintaining grid frequency stability and power quality.
• The landing modes of the receiving-end converter stations must ensure safe DC power dissipation while ensuring AC bus short-circuit currents do not exceed circuit breaker interrupting capacities. Rationality is verified by thermal stability margins (≥30% under steady state and ≥10% under N-1 faults) and short-circuit current margin indices (K_sc,i > 0) for all connected buses.

For an actual offshore wind power DC transmission project, the proposed topology construction principles can be applied in the following steps:

• Step 1: Determine a reasonable aggregation scheme based on the wind farm layout, installed capacity, and the transmission capacity of existing DC submarine cables, thereby determining the number of sending ends of the DC system.
• Step 2: Determine the number of receiving ends of the DC system in combination with the transmission capacity of existing DC overhead lines. Then, select alternative access points in the load center of the AC power grid, conduct safety checks on different access schemes, and determine the landing mode.
• Step 3: After determining the number of sending and receiving ends, determine the multi-terminal DC network topology in combination with the maximum tolerable loss power of the AC power grid.
• Step 4: Compare the costs of different offshore converter station topologies considering DC fault handling, and determine the final offshore converter station topology.

4. Case Study

This section presents a case study of a specific regional power grid. In this scenario, nine OWFs with a total installed capacity of 5600 MW require centralized transmission. The onshore AC grid load center is situated 220 km inland from the shoreline.

Table 4. Installed capacity of each wind farm.

Wind Farms	Installed Capacity (MW)
WF1	600
WF2	500
WF3	400
WF4	400
WF5	900
WF6	500
WF7	500
WF8	800
WF9	1000

lists the installed capacity of each wind farm, and Figure 4 depicts the geographic layout of the OWFs and the AC grid connection configuration. Specifically, the load center area features 3876 MW of local power generation, 17,406 MW of active load demand, 13,530 MW of power imported from external regions, and a system-wide maximum tolerable power loss P_max-fail = 4200 MW.

4.1. Wind Farm Clustering

According to the transmission capacity of existing DC submarine cables listed in Table 2, to achieve centralized transmission of 5600 MW offshore wind power while minimizing the number of DC submarine cables, three circuits of DC submarine cables with a cross-sectional area of 2000 mm² at the ±500 kV voltage level are sufficient. Considering the geographic layout and installed capacity of the wind farms, the formed offshore wind farm clustering scheme is shown in

Table 5. Collection schemes for offshore wind farms.

Cluster	Clustering Wind Farms	Group Capacity (MW)
Cluster 1	WF1, WF2, WF3, WF4	1900
Cluster 2	WF5, WF6, WF7	1900
Cluster 3	WF8, WF9	1800

.

4.2. Grid Integration and Landing Modes

Given the 220-km distance between the load center and the offshore wind power onshore landing point, connecting OWFs to the nearby coastal AC grid would incur significant transmission losses and voltage drop; thus, DC overhead lines are adopted for direct power delivery to the load center. Considering the transmission capacity of ±500 kV DC overhead lines (3000–5000 MW per circuit), two parallel circuits are deployed to meet the 5600-MW transmission requirement, necessitating two receiving-end converter stations with a rated capacity of 2800 MW each.

Figure 5 illustrates two alternative landing configurations for the converter stations: Configuration 1 (Bus 9 + Bus 12) and Configuration 2 (Bus 10 + Bus 12). Safety verification using thermal stability and short-circuit current margins for single-point grid connection shows no AC line thermal overload risks in either option. However, when applying the 63-kA upper limit for 500 kV bus short-circuit currents, Configuration 2 results in Bus 15 exceeding this threshold, as detailed in

Table 6. Short-circuit current at Bus 15 under different schemes.

Scheme	Short-Circuit Current of Bus 15 (kA)
Original System	62.30
Scheme 1	63.63
Scheme 2	62.76

for Bus 15 under different setups. Through comprehensive safety assessment, the landing arrangement selecting Bus 10 and Bus 12 is validated to satisfy grid connection requirements without violating thermal or short-circuit current constraints.

4.3. Multi-Terminal DC System Topology

Given the DC system’s configuration, which includes three offshore terminals, two onshore terminals, and all converter station capacities below the AC grid’s P_max-fail = 4200 MW, a star topology is selected for the DC network. The designed multi-terminal DC system is depicted in Figure 5, where HSSdenotes the high-speed switch of the DC line. Offshore wind farms are aggregated via three DC submarine cables to an onshore switch station, which converts the power to two DC overhead lines for transmission to the load center. These lines terminate at Bus 10 and Bus 12, respectively, forming a classic star structure: all converter stations connect to the common DC bus through individual feeders, ensuring compliance with the AC grid’s fault tolerance requirements. This topology efficiently balances cost-effectiveness and operational reliability, with each terminal’s capacity within the grid’s tolerable power loss threshold.

4.4. Economic Comparison of Converter Station Topologies Considering DC Fault Handling

The three lightweight offshore converter topologies introduced in Section 3.2 lack inherent DC fault self-clearing capability, necessitating the integration of DC circuit breakers. This section incorporates the assembly HVDC breaker proposed in Reference [23], which comprises IGBTs and thyristors, into the switching device cost calculations for different topology configurations. Four comparative schemes are analyzed:

• Scheme 1: Full/half-bridge hybrid sub-modules are adopted for both sending and receiving converter stations, with a sub-module quantity ratio of full-bridge to half-bridge set at 1:1.
• Scheme 2: Offshore converter stations employ the DRU series-connected half-bridge MMC with a 2:1 steady-state power transmission ratio, where the MMC provides reactive power compensation for the DRU. Onshore stations use the half-bridge MMC, and assembly HVDC breakers isolate DC faults.
• Scheme 3: Offshore converter stations utilize the DRU parallel-connected half-bridge MMC with a 2:1 power transmission ratio, with the MMC offering reactive power support to the DRU. Onshore stations adopt the half-bridge MMC, and assembly HVDC breakers handle fault isolation.
• Scheme 4: Offshore converter stations use full-power DRU, paired with the onshore half-bridge MMC. The assembly HVDC breakers manage fault isolation, and offshore wind turbines employ grid-forming control.

The comparative results of investment and operation costs for different schemes are summarized in

Table 7. Cost comparison of different schemes.

Scheme *	Switching Device	Black Start Source	Reactive Power Compensation	Converter Platform Infrastructure	Operational Losses	Total Cost
Scheme 1	0.656	0	0	14.192	0.702	15.55
Scheme 2	0.344	1.001	0	6.923	0.284	8.552
Scheme 3	0.331	0	0	6.923	0.284	7.538
Scheme 4	0.266	1.001	0.343	3.930	0.143	5.683

* Cost unit: CNY 1 billion.

.

As shown in Table 7, DRU-based offshore converter topologies (Scheme 2, Scheme 3 and Scheme 4) achieve cost savings of 45.00%, 51.52%, and 63.45%, respectively, compared to Scheme 1. The most substantial reductions occur in converter platform infrastructure costs, driven by the compact design of DRU systems. In terms of switching device expenses, Scheme 2, Scheme 3 and Scheme 4, which utilize the assembly HVDC breakers, reduce costs by 47.56%, 49.54%, and 59.45% relative to Scheme 1’s full/half-bridge hybrid sub-modules. This reflects the economic advantage of diode-based components over IGBT-intensive configurations.

A comparative analysis between Scheme 2 and Scheme 3 reveals that Scheme 2’s lack of power reversal capability necessitates additional black start power supplies for offshore wind farms, increasing investment by approximately 12.13% of the total cost. Other cost discrepancies between the two schemes remain negligible. Although Scheme 4 requires investments in offshore black start systems and reactive power compensation equipment due to its full-power DRU design, its overall cost remains the lowest among all configurations. This makes Scheme 4 the most economical choice, particularly when considering the long-term benefits of simplified converter architecture.

From a technological readiness perspective, if grid-forming control for offshore wind turbines reaches maturity, full-power DRU-based offshore converter stations will exhibit unparalleled economic advantages. For grid-following offshore wind turbine systems, where power reversal capability is not a strict requirement, DRU-parallel MMC topologies (Scheme 3) offer an optimal balance between cost efficiency and technical feasibility. These topologies avoid the need for auxiliary black start equipment while maintaining acceptable operational losses.

5. Conclusions

This paper systematically analyzes the key influencing factors and proposes topology construction principles for large-scale offshore wind power HVDC transmission and grid-integrated systems. The primary conclusions are as follows:

• Direct HVDC transmission of large-scale OWP to load centers outperforms near-shore AC grid connection in operational efficiency, avoiding significant transmission losses and voltage drops inherent in long-distance AC transmission to ensure more reliable power delivery to inland load centers.
• On the premise of mature technology, grid-following offshore wind turbines are suitable for DRU-parallel MMC topologies in offshore converter stations, while grid-forming offshore wind turbines exhibit the greatest economic advantage when transmitted via full-power DRU.
• The proposed construction principles cover the three links of offshore wind power aggregation, transmission, and grid connection, addressing offshore wind farm clustering, offshore converter topologies, multi-terminal DC network topologies, and receiving-end converter connection modes. These principles are applicable to the topology construction of multi-farm, ultra-large-scale offshore wind power centralized transmission systems and provide guidance for the planning of deep-sea large-scale offshore wind power transmission systems.

This paper has made certain achievements in researching the construction of offshore wind power HVDC system topologies; however, there are still some limitations. Firstly, the cost analysis is not sufficiently comprehensive. Secondly, there is a lack of research on the impact of different DC system topologies on system stability. In future studies, environmental and social costs incurred by different topologies could be taken into consideration, and further research could be conducted on how different topologies affect system stability. Additionally, this paper primarily focuses on research regarding offshore wind power HVDC systems at the planning stage, and future research may explore the application of artificial intelligence technology to optimize the operation and maintenance of such systems.