Design of Series-Connected Novel Large-Scale Offshore Wind Power All-DC System with Fault Blocking Capability

Abstract

The utilization of wind power all-DC systems with DC collection and transmission is an effective solution for the extensive development of wind power in deep-sea areas. However, in the event of faults occurring in wind power all-DC systems, the fault propagation speed is extremely rapid, with a wide-ranging impact, and to date, there are no complete DC engineering references available. It is crucial to research the topology and fault isolation methods applicable to large-scale offshore wind power all-DC systems in deep-sea areas. This paper proposes a novel series-connected all-DC system topology and presents corresponding fault isolation methods for internal faults in wind turbine units and faults in high-voltage DC transmission lines. The system simulation model was constructed using PSCAD/EMTDC (v4.6.3), and simulations were conducted for internal faults in the wind turbine units and DC transmission line short-circuit faults. The simulation results demonstrate that the proposed system can isolate various DC faults while maintaining stable operation, thereby validating the effectiveness of the control strategies and fault isolation methods proposed in this paper.

1. Introduction

Offshore wind power possesses characteristics such as high wind energy density, stable wind speeds, and proximity to load centers [1], leading to its rapid development in recent years. Offshore wind power systems consist of two parts: collection and transmission. Early-stage offshore wind power projects, with relatively close distances from shore, employed AC collection and transmission methods [2]. However, as offshore wind farms are being developed further from shore, issues arising from the charging current and charging power of AC cables have become prominent, thereby making flexible DC transmission the mainstream technology [3]. With the continuous increase in voltage levels and transmission capacities, the capacitive rise effect within offshore wind farms due to AC collection becomes increasingly severe [4]. Consequently, the adoption of offshore wind power full-DC systems with DC collection and transmission [5] has emerged as a new and significant development direction.

Based on the different methods of energy collection on wind farms, offshore wind power full-DC systems can be classified into parallel-connected systems and series-connected systems [6]. Typically, the output voltage level of DC wind turbine units is relatively low. After parallel collection of electrical energy, the voltage needs to be elevated to the high-voltage DC level by offshore boosting stations and then transmitted to onshore inverter stations via DC sea cables [7]. The voltage level of DC collection is generally around 50 kV, and if a voltage level of 320 kV is adopted for high-voltage DC transmission, offshore boosting stations would require DC/DC converters with a turns ratio of over 6. If direct transmission is implemented after parallel collection [8], it necessitates the construction of wind turbine units capable of directly outputting high-voltage DC. Each wind turbine unit would be equipped with a high-gain DC/DC converter on the DC side. However, achieving high gains in DC/DC converters is challenging with current technological capabilities [9]. Therefore, parallel-connected systems are not suitable for remote offshore scenarios.

In contrast to parallel-connected collection systems, series-connected collection systems serially connect DC wind turbine units to achieve high-voltage DC levels and then transmit them to onshore converter stations via DC sea cables [10]. This approach offers a simple structure, low cost, no requirement for offshore platforms, and low exit voltage of wind turbine units [11], making it a practical solution to the challenges of constructing full DC systems in remote offshore areas [12]. However, series-connected collection systems exhibit strong coupling characteristics, where each wind turbine unit in the system influences others. Fluctuations in power output or shutdown of a single wind turbine unit can cause voltage fluctuations at the ports of all wind turbine units in the system [13]. To prevent excessively high voltages in individual wind turbine units due to significant power differences among them, voltage clamping is necessary, resulting in wind power losses [14]. Furthermore, direct connection of wind turbine units to high-voltage transmission lines, considering a voltage level of 320 kV, for instance, requires the first wind turbine unit in the series array to withstand a DC bias voltage of 320 kV, posing insulation challenges to the units and restricting the increase in transmission voltage levels [15].

The research in [16] proposes a series wind power decoupling control strategy with energy storage, which reduces wind power losses. The study in [17] improves the series-connected topology by adding shunt circuits to avoid wind power losses and overvoltage phenomena, albeit at the expense of increased control complexity. The study in [18] suggests involvement of onshore converter stations in limiting wind turbine unit voltages, ensuring maximum power of all wind turbine units while limiting overvoltage. However, this approach requires communication between the wind farm and onshore converter stations, which has limitations in terms of communication delays and cable losses.

Current research on offshore wind power full DC systems primarily focuses on system topology, steady-state, and transient operation control [19]. However, the challenging environmental conditions in which these systems operate make maintenance and repair difficult [20], necessitating high reliability of the system equipment and rapid fault identification and control when faults occur. To ensure the stable operation of large-scale offshore wind power full-DC systems, research on fault-blocking methods is essential [21]. The study in [22] proposes the addition of auxiliary circuits to reduce overvoltage in series-connected wind turbine units but requires more cables and circuit breakers. The study in [23] suggests the inclusion of diodes in series arrays to limit fault currents. The study in [24] proposes the isolation of DC faults through isolation stations composed of diodes, isolation switches, and bypass switches, albeit requiring additional offshore platforms.

The current research on offshore wind power all-DC systems remains primarily theoretical. With the increasing emphasis on offshore wind energy development, there is a growing need for practical engineering solutions. To address this, the paper proposes a novel offshore wind power full-DC system with fault isolation capability in Section 2. In this system, direct current wind turbines avoid overvoltage and wind power loss phenomena. Key equipment in the system adopts topologies with direct current fault interruption capability, and the system can achieve full-power high-voltage direct current transmission. Section 3 elaborates on the system’s control strategies and fault-blocking methods, while Section 4 conducts fault condition simulations based on PSCAD/EMTDC. Finally, Section 5 provides a concise conclusion.

2. The Structure of the Novel Offshore Wind Power All-DC System with Fault Isolation Capability

2.1. The Operational Principle of the Existing Series-Connected System

The topology structure of the existing series-connected system is depicted in Figure 1. In a series-connected DC wind farm, m wind turbine units with DC output voltages are directly connected in series to achieve voltage elevation. The electrical energy generated by the wind farm is collected in series and transmitted to the shore via underwater DC cables in the form of direct current. Finally, onshore converter stations perform DC-AC conversion, invert the electrical energy, and integrate it into the grid.

In a series-connected DC wind farm, the voltage is controlled and maintained constant by the onshore converter station, while the output voltage of individual series-connected DC wind turbine units is not independently controlled. Therefore, any change in the operational state of one wind turbine unit affects other operational units within the wind farm. Let us denote the output voltage and power of the ith DC wind turbine unit (i = 1, 2, …, n) as $U_{\mathrm{wt},i}$ and $P_{\mathrm{wt},\mathrm{i}}$, respectively. The voltage, current, and power of the series-connected DC wind farm can then be represented as $U_{\mathrm{wf}}$, $I_{\mathrm{wf}}$, and $P_{\mathrm{wf}}$, respectively. Thus, the voltage and power of the series-connected DC wind farm can be expressed as:

$$U_{\mathrm{wf}}={\displaystyle \sum _{i=1}^m{U}_{\mathrm{wt},i}}$$

(1)

$$P_{\mathrm{wf}}={\displaystyle \sum _{i=1}^m{P}_{\mathrm{wt},i}}$$

(2)

Since the wind turbine units are connected in series, the current flowing through any wind turbine unit is equal to the current of the wind farm $I_{\mathrm{wf}}$. Therefore, the output power of the ith wind turbine unit satisfies:

$$P_{\mathrm{wt},i}=U_{\mathrm{wt},i}{I}_{\mathrm{wf}}$$

(3)

Furthermore, the current of the wind farm can be expressed as:

$$I_{\mathrm{wf}}=\frac{P_{\mathrm{wf}}}{U_{\mathrm{wf}}}$$

(4)

Combining Equations (3) and (4), the output voltage of the ith DC wind turbine unit $U_{\mathrm{wt},i}$ can be expressed as:

$$U_{\mathrm{wt},i}=\frac{P_{\mathrm{wt},i}}{P_{\mathrm{wf}}}{U}_{\mathrm{wf}}$$

(5)

Through Equation (5), it is evident that since $U_{\mathrm{wf}}$ is maintained constant under the control of the onshore converter station, the output voltage of the series-connected DC wind turbine units is determined jointly by their own power and the power of the wind farm. Moreover, the higher the power of a wind turbine unit, the higher its output voltage. When the operating conditions of a particular DC wind turbine unit change, it affects all wind turbine units within the wind farm, thus demonstrating strong coupling.

When a DC wind turbine unit in a series-connected DC wind farm experiences a change in power due to variations in wind speed, the dynamic process of this turbine unit is as follows:

$$(P_{\mathrm{wt},i}+\Delta P_{\mathrm{wt},i})=(U_{\mathrm{wt},i}+\Delta U_{\mathrm{wt},i})(I_{\mathrm{wf}}+\Delta I_{\mathrm{wf}})$$

(6)

In the equation, $\Delta P_{\mathrm{wt},i}$, $\Delta U_{\mathrm{wt},i}$, and $\Delta I_{\mathrm{wf}}$ represent the changes in the output power, output voltage, and wind farm current of the DC wind turbine unit, respectively.

From Equation (6), it can be observed that a change in the output power of the DC wind turbine unit will sequentially affect the output voltage of the turbine unit and the current of the wind farm. Subsequently, the change in wind farm current will further influence the output voltage of all-DC wind turbine units until the system reaches a new equilibrium point. This dynamic process is illustrated in Figure 2.

For the entire wind farm, the dynamic variation process unfolds as follows:

$${\displaystyle \sum _{i=1}^m(P_{\mathrm{wt},i}+\Delta P_{\mathrm{wt},i})}={\displaystyle \sum _{i=1}^m(U_{\mathrm{wt},i}+\Delta U_{\mathrm{wt},i})}(I_{\mathrm{wf}}+\Delta I_{\mathrm{wf}})$$

(7)

Since the voltage of the wind farm is maintained constant by the onshore converter station, the sum of the changes in output voltage of all-DC wind turbine units is zero. Therefore, Equation (7) can be expressed as:

$${\displaystyle \sum _{i=1}^m(P_{\mathrm{wt},i}+\Delta P_{\mathrm{wt},i})}=U_{\mathrm{wf}}(I_{\mathrm{wf}}+\Delta I_{\mathrm{wf}})$$

(8)

From Equation (8), it can be observed that changes in the current of the wind farm result in changes in the power of the wind farm.

If the output voltage of the DC wind turbine units is not limited, when there are significant differences in power output between wind turbine units, those with higher output power will experience higher voltages. Let us denote the reference value of the output voltage of wind turbine units as $U_{\mathrm{wt}}^{*}$.

$$U_{\mathrm{wt}}^{*}=\frac{U_{\mathrm{wf}}}{m}$$

(9)

For a wind farm with m wind turbine units, the average output power of the wind farm is given by:

$$P_{\mathrm{wt},\mathrm{avg}}=\frac{1}{m}{\displaystyle \sum _{i=1}^m{P}_{\mathrm{wt},i}}$$

(10)

Combining Equations (5), (9), and (10), the output voltage of the wind turbine unit can be expressed as:

$$U_{\mathrm{wt},i}=\frac{P_{\mathrm{wt},i}}{P_{\mathrm{wt},\mathrm{avg}}}{U}_{\mathrm{wt}}^{*}$$

(11)

From Equation (11), it is evident that when there is power imbalance between wind turbine units, the output voltage of wind turbine units with output power greater than the reference value $P_{\mathrm{wt},\mathrm{avg}}$ will exceed $U_{\mathrm{wt}}^{*}$, leading to an overvoltage operating state.

When a wind turbine unit in the wind farm stops operating due to certain reasons, even if the remaining operational wind turbine units have balanced output power, the output voltage $U_{\mathrm{wt},\mathrm{avg}}^{\prime }$ of the remaining operational wind turbine units may still exceed the reference value $U_{\mathrm{wt}}^{*}$:

$$U_{\mathrm{wt},\mathrm{avg}}^{\prime }=\frac{m}{m-k}{U}_{\mathrm{wt}}^{*}$$

(12)

where k represents the number of wind turbine units that have stopped operating, 0 < k < m.

However, for considerations of system safety and the cost of wind turbine insulation to withstand voltage, the output voltage of wind turbine units cannot infinitely increase. In actual operation, the voltage that wind turbine units can withstand will be limited. Let us denote the maximum voltage that a wind turbine unit can withstand (the clamping value) as $U_{\mathrm{wt},\mathrm{limit}}$. When the theoretical value of the output voltage of a wind turbine unit exceeds $U_{\mathrm{wt},\mathrm{limit}}$, it enters a clamping operating state. At this point, to ensure the transmission of electrical energy to the shore, the output voltage of other wind turbine units that have not entered the clamping operation state will increase to reach the transmission voltage level. According to Equation (2), when the wind turbine unit power remains unchanged, an increase in voltage will lead to a decrease in wind farm current. According to Equation (8), the power of the wind farm decreases by $U_{\mathrm{wf}}\Delta I_{\mathrm{wf}}$, resulting in wind power loss.

In summary, in the current series-connected wind power full-DC systems, wind turbine units are coupled with each other, and their output voltages influence each other. The fluctuation of wind turbine unit power directly manifests as fluctuations in output voltage. This limitation affects the tracking of the maximum power point of wind turbine units and the occurrence of overvoltage phenomena, which also impacts the stability of the system. Additionally, constrained by insulation technology development, wind farm voltages generally should not exceed 100 kV, posing limitations when applied in offshore scenarios.

2.2. The Novel Offshore Wind Power All-DC System with Fault Isolation Capability

The topology structure of the large-scale offshore DC wind power system with fault isolation capability is illustrated in Figure 3.

In the large-scale offshore DC wind power system with fault isolation capability, a DC/DC converter known as the Capacitor Energy Transfer (CET) converter is utilized based on the principle of capacitive energy transfer [25], as depicted in Figure 4.

This topology consists of three phases (j = a, b, c), with each phase having an identical structure. Each phase comprises a capacitor energy storage bridge arm and four sets of switching valves (${\mathrm{T}}_{1j}$, ${\mathrm{T}}_{2j}$, ${\mathrm{T}}_{3j}$, ${\mathrm{T}}_{4j}$). The capacitor energy storage bridge arm is composed of several half-bridge submodules (HBSMs) and a bridge arm inductor $L_{\mathrm{CET}}$ connected in series [26,27]. Additionally, a DC energy dissipating device is installed on the low-voltage side, consisting of switches, ${\mathrm{S}}_{R\mathrm{f}}$, and dissipative resistors, $R_{\mathrm{f}}$. Control of the switching valves allows for energy transfer between the two sides of the CET converter. The energy storage bridge arm serves as a carrier for energy transfer, and by controlling the conductivity of the HBSMs, connectivity between different voltage levels on both sides of the CET converter can be achieved. The CET converter possesses the capability for fault isolation, and the DC energy dissipating device on the low-voltage side can consume the electrical energy output by the wind farm in non-steady-state conditions, such as short-circuits in the DC transmission lines, to maintain system stability.

The CET converter operates in a three-phase alternating manner, and each phase operates independently. Here, we illustrate the working principle using the phase ‘a’ as an example. The schematic diagram of the two states during the operation process of the CET converter is depicted in Figure 5. The waveform diagram illustrating the working principle of the CET converter is shown in Figure 6, where $T_{\mathrm{h}}$ represents the operating period, and $f_{\mathrm{h}}$ represents the operating frequency.

The working process of the CET converter transferring energy from the low-voltage side to the high-voltage side is as follows:

• [$t_0$ − $t_1$]: Charging Stage. During this stage, the energy storage bridge arm absorbs power from the low-voltage side. The switching valve ${\mathrm{T}}_{a1}$ is in the conducting state, while the remaining switching valves are in the locked state, as shown in Figure 5. When the voltage across the bridge arm $U_{\mathrm{p}a}$ equals the low-voltage side voltage $U_{\mathrm{L}}$, the triggering signal “zero voltage” is applied to the switching valve ${\mathrm{T}}_{a1}$, causing it to conduct. As a result, a driving voltage with an amplitude of $-U_0$ is generated across the bridge arm inductor $L_{\mathrm{CET}}$, and the bridge arm current $I_{\mathrm{p}a}$ rises sinusoidally to $I_{\mathrm{L}}$. At this point, the voltage across the bridge arm is at a constant amplitude of $U_{\mathrm{L}}$, and the current is sinusoidal with a peak value of $I_{\mathrm{L}}$, achieving constant voltage and constant current charging. After conducting for 120° of electrical angle, the CET converter initiates phase shifting. At this moment, a driving voltage with an amplitude of $U_0$ is generated across the bridge arm inductor $L_{\mathrm{CET}}$, causing the bridge arm current $I_{\mathrm{p}a}$ to decrease sinusoidally to zero. The switching valve ${\mathrm{T}}_{a1}$ withstands a reverse voltage with an amplitude of $U_{\mathrm{L}}-U_{\mathrm{H}}$ and shuts off with zero current.
• [$t_1$ − $t_2$]: Discharge Stage. During this stage, power is transferred from the energy storage bridge arm to the high-voltage side. The switching valve ${\mathrm{T}}_{a2}$ is in the conducting state, while the remaining switching valves are in the locked state, as shown in Figure 5. The working principle during this stage is similar to the charging process and will not be described in detail.

By sequentially conducting the three phases of the CET converter for 120° each, a stable waveform can be outputted, eliminating the need for a filter, thereby reducing the volume and weight of the offshore boosting station. Additionally, in Figure 4, the period $T_{\mathrm{cp}}$ is referred to as the commutation stage, during which the CET converter completes the alternate conduction of adjacent phases within $T_{\mathrm{cp}}$, and $I_{\mathrm{p}a}$ is required for the sinusoidal rise/fall. $T_{\mathrm{dz}}$ is referred to as the transition stage, during which the jth phase does not participate in energy transfer. This is crucial for the switching valves to achieve soft switching.

The CET converter possesses the capability for bidirectional energy flow. When the CET converter transfers energy from the high-voltage side to the low-voltage side, the working principle is similar to that of energy transfer from the low-voltage side to the high-voltage side. The difference lies in the switching valves involved in the operation, which are ${\mathrm{T}}_{a3}$ and ${\mathrm{T}}_{a4}$, and the amplitude of the bridge arm current, which becomes $I_{\mathrm{H}}$ and $-I_{\mathrm{L}}$.

The CET converter, without the need for an AC transformer, exhibits compactness and lightweight characteristics. It features an exceptionally wide voltage operating range, enabling high-voltage DC transmission of offshore wind energy and further reducing the insulation requirements of wind farms. During the operation of the CET converter, it actively controls the bridge arm current, and the sum of the three-phase bridge arm currents represents the input/output DC current. Hence, the current of the series-connected wind farm remains constant. From Equations (1)–(3), it is evident that when the power of a particular wind turbine changes, it only affects its own DC output voltage, while the operational status of other wind turbines remains unchanged, thus avoiding wind power loss. If the CET converter controls the current on the wind farm side to its rated value, then when a wind turbine operates at full power, its output voltage precisely matches the rated value, avoiding overvoltage. Although the wind farm voltage may fluctuate with changes in wind turbine output power, the CET converter can promptly track the wind farm voltage without affecting the operation status of the high-voltage side.

The inverter of a wind turbine unit requires a wide range of voltage regulation capabilities. The Full-Bridge and Half-Bridge Submodule Mixed-Modular Multilevel Converter (FHMMC), as proposed in ref. [28], is an excellent choice. By designing the proportions of full-bridge and half-bridge submodules, the FHMMC possesses extensive DC voltage reduction capabilities.

The schematic diagram of the direct current wind turbine unit studied in this paper is illustrated in Figure 7. It mainly consists of a wind rotor, gearbox, Permanent Magnet Synchronous Generator (PMSG), isolation transformer, Wind Turbine Modular Multilevel Converter (WTMMC), isolation switches S1 and S2, and bypass switch F. The wind rotor converts wind energy into mechanical energy, while the gearbox transforms the low-speed rotational mechanical energy provided by the wind rotor into medium-speed rotational mechanical energy. The PMSG converts mechanical energy into electrical energy. The isolation transformer, based on the rated modulation ratio of the WTMMC, increases the voltage to an appropriate level and isolates the PMSG from the collection line of the series-connected direct current wind farm, reducing the insulation and withstanding the voltage requirements of the direct current wind turbine unit. When the WTMMC is operational, isolation switches S1 and S2 are closed, and the bypass switch F is open. When the WTMMC is out of operation due to faults or maintenance, isolation switches S1 and S2 are open, and bypass switch F is closed, ensuring the energy flow path of the series-connected wind farm.

The use of semi-direct drive Permanent Magnet Synchronous Generators (PMSGs) is currently the mainstream trend for high-power units [29], as they are better suited for offshore environments. If a half-bridge Modular Multilevel Converter (MMC) is used as the converter for the direct current wind turbine unit, it lacks fault isolation capability, failing to meet the high reliability requirements of offshore direct current wind turbine units. On the other hand, if a full-bridge MMC is employed, it would increase the input of power devices and enlarge the size and weight of the direct current wind turbine unit, thus reducing its economic viability. The hybrid MMC, which incorporates both half-bridge submodules (HBSMs) and full-bridge submodules (FBSMs) in a 1:1 ratio, extends the operating range of the DC output voltage to 0~1 pu. Additionally, it possesses fault isolation capability for direct current while reducing the number of Insulated Gate Bipolar Transistors (IGBTs) used by 25% compared to the full-bridge MMC.

In this paper, the Wind Turbine Modular Multilevel Converter (WTMMC) utilizes series connection for energy collection and performs the first stage of voltage boosting. The output voltage of the WTMMC is variable and only needs to output low voltage. As a result, the WTMMC does not need to operate in low modulation ratio mode.

The onshore converter station also adopts a hybrid MMC topology, fully leveraging the advantages of flexible direct current transmission technology. The capacity of the onshore Modular Multilevel Converter (onshore MMC) is larger and its voltage level is higher. Consequently, it requires more submodules, and since the alternating current side is connected to the AC grid, the control strategy differs accordingly.

3. The Control Strategy and Fault Isolation Methods of the System

3.1. The Control Strategy of the CET Converter

Due to the independent nature of the three phases of the CET converter, we will continue to illustrate the control strategy using its ‘a’ phase as an example. The control block diagram of the CET converter is depicted in Figure 8, encompassing low-side current control, energy balance control, bridge arm current control, and thyristor triggering control.

The energy balance control is responsible for maintaining the stability of the submodule capacitor voltage. It calculates the average voltage U_Cpa,avg of the submodule capacitor voltage of phase ‘a’ and compares it with the reference value $U_{\mathrm{CCET}}^{*}$. The deviation is then sent to a PI controller. Using waveform generator 1, the output is controlled to the current reference value $I_{\mathrm{H}}^{*}$ when the bridge arm is connected to the high-voltage side.

The low-voltage side current control provides the current reference value $I_{\mathrm{L}}^{*}$ when the bridge arm is connected to the low-voltage side. This reference value $I_{\mathrm{L}}^{*}$ is generated by waveform generator 2 and its magnitude is determined by the ratio of the rated power $P_{\mathrm{CETbase}}$ and the rated low-voltage $U_{\mathrm{Lbase}}$. Waveform generators 1 and 2 generate approximate trapezoidal waves with a period $T_{\mathrm{h}}$, where each trapezoid’s slopes are sinusoidal waves, and these sinusoidal waves are simultaneously tangent to both 0 and 1. The CET converter can reliably turn off the thyristors and complete the conversion of the bridge arm voltage when the outputs of waveform generator 1 and waveform generator 2 are both 0.

The reference value I_pa* for the bridge arm current of phase ‘a’ can be obtained by summing $I_{\mathrm{H}}^{*}$ and $I_{\mathrm{L}}^{*}$. Then, the actual bridge arm current $I_{\mathrm{p}a}$ sampled is subtracted from the reference value I_pa*, and the difference is sent to the PI controller of the bridge arm current control. The modulation wave U_pa^* is generated by adding the voltage $U_0$ that ensures reliable turn-off of the thyristor to the output of the PI controller.

For the commutation valves, ensuring reliable turn-off is crucial, while triggering turn-on is relatively simple. Waveform generators 3 and 4 can be used to generate trigger signals for the thyristors. Their outputs are rectangular waves with a period of $T_{\mathrm{h}}$. When the thyristors in the commutation valves receive the trigger signal, they can conduct while simultaneously bearing a forward voltage drop.

3.2. The Control Strategy of MMCWT and Onshore MMC

The controller structure of the MMCWT is illustrated in Figure 9. The MMC of the wind turbine unit employs an active outer loop controller as depicted in Figure 9, which, when combined with the CET-DCT control strategy, enables simultaneous control of the MMC output active power and wind turbine unit output voltage. The remaining components are similar to traditional MMC control and are not reiterated here. Submodule capacitance equalization control is utilized to maintain the stability of the submodule voltage in the wind turbine unit MMC, and trigger signals are generated through Carrier Phase-Shifted PWM modulation. The comprehensive control diagram of the wind turbine unit MMC is shown in Figure 9, where $P_{\mathrm{wt}}^{*}$ and $Q_{\mathrm{wt}}^{*}$ represent the active and reactive commands of the wind turbine unit MMC, respectively. $I_{d\mathrm{wt}}^{*}$ and $I_{q\mathrm{wt}}^{*}$ denote the reference commands for the current inner loop controller of the wind turbine unit MMC, while $\omega L_{\mathrm{wt}}$, $E_{d\mathrm{wt}}$ and $E_{q\mathrm{wt}}$ represent the dq-axis components of the decoupling and feedforward terms, respectively.

The control diagram of the onshore converter station MMC is illustrated in Figure 10, incorporating not only dual-loop control but also an independent DC control. Here, $U_{\mathrm{Cavg}6\mathrm{p}}$ represents the average of the capacitor voltages of the six MMC bridge arms, while $U_{\mathrm{H}}^{*}$ denotes the voltage command on the DC side of the converter station. The active outer loop controller is responsible for submodule capacitor voltage control, whereas the reactive outer loop controller remains unchanged from traditional strategies. During normal operation, the reference value for DC voltage control, $U_{\mathrm{H}}^{*}$, can be directly assigned. However, in the event of a short-circuit fault on the DC side, the DC control mode switches to DC current control, facilitating fault ride-through without tripping.

3.3. The Fault Isolation Methods of the System

When critical components such as the gearbox or the converter of the MMCWT experience failures due to internal control issues, corrosion, or other external factors, it is imperative for the system to maintain stable operation and prevent fault propagation. Upon the need for the MMCWT to be taken offline for reset or maintenance, the system should ensure continued stability. This can be achieved by correctly switching off the isolation switches and bypass switches at the DC output of the MMCWT, thereby isolating the fault and maintaining system stability. Simultaneously, locking the power devices in the converter and utilizing FBSMs to block fault currents are essential steps. The logic for handling internal faults in the MMCWT is illustrated in Figure 11, using the example of an internal fault occurring in the ith MMCWT:

When the ith MMCWT experiences an internal fault, it is imperative to promptly isolate it from the system to protect the MMCWT and prevent the fault from spreading, thus averting potentially severe consequences. In this scenario, the isolation switches S1 and S2 of the ith MMCWT are opened to isolate the fault source from the system. Simultaneously, the bypass switch F is closed to establish an energy flow path for the series-connected DC wind farm. Additionally, all power devices within the ith MMCWT should be locked to prevent further operation.

In case of a transient fault, once the ith MMCWT undergoes a reset, it can be restarted. In this scenario, the isolation switches S1 and S2 are closed, the bypass switch F is opened, and the power devices resume operation, allowing the ith MMCWT to be reintegrated into the system. However, if a permanent fault occurs, the ith MMCWT necessitates maintenance. After rectifying the fault, the isolation switches S1 and S2 are closed, the bypass switch F is opened, and the ith MMCWT is brought back into operation. As the wind turbine converter employs MMC technology, the faulty MMCWT can absorb power directly from the DC side to facilitate self-starting, eliminating the need for additional energy storage devices.

The short-circuit fault in the DC transmission line is one of the most severe faults in a wind power all-DC system. Following a line short-circuit, the DC current rapidly escalates, posing a significant threat to the power devices within the system.

Figure 12 shows the logic of the CET converter and the onshore MMC for the detection of short-circuit faults in the DC transmission line, which mainly detects the voltage and current of the DC transmission line set fault voltage coefficient $k_1$ (0 < $k_1$ < 1) and fault current coefficient $k_2$ (1 < $k_2$). When the DC voltage $U_{\mathrm{H}}$ drops below $k_1$∗$U_{\mathrm{H}}$ and the current $I_{\mathrm{H}}$ exceeds $k_2$∗$I_{\mathrm{H}}$, a short-circuit fault in the DC transmission line is considered to have occurred, with a designated fault detection time $T_{\mathrm{fault}}$.

When the CET converter detects a short-circuit fault in the DC transmission line, it can achieve fault isolation by blocking all bridge arm submodule and thyristor trigger signals. Since the CET converter operates independently in three phases, here we continue to use phase ‘a’ as an example for fault isolation analysis.

Depending on the working status of phase ‘a’ when a fault occurs, two scenarios can be distinguished:

• When the energy storage bridge arm is absorbing power from the low-voltage side, as depicted in Figure 13a. At this point, the energy storage bridge arm and the fault point are not connected, and hence there is no path for fault current to flow. Upon detecting the fault, the controller sends signals to lock the submodule and thyristor gates of the bridge arm, interrupting the charging process of the energy storage bridge arm and preventing the formation of a fault path.
• When the energy storage bridge arm is delivering power to the high-voltage side, as illustrated in Figure 13b. In this case, the energy storage bridge arm is connected to the fault point. Upon detecting the fault, the controller sends signals to lock the submodule and thyristor gates of the bridge arm. Although the submodule gates quickly close, fault current can still flow through the bridge arm inductance $L_{\mathrm{CET}}$ and the parallel diode D2 of the submodule IGBT until the fault current flowing through the inductance $L_{\mathrm{CET}}$ decays to zero. Subsequently, the thyristor can be turned off, achieving fault isolation.

While locking the submodule of the CET converter, opening the resistor switch can dissipate the electrical energy supplied by the wind farm, thus preventing damage to the switching valves and maintaining system stability.

During a short-circuit fault in the DC transmission line, the onshore MMC converter switches its DC control to the direct current control mode as depicted in Figure 10. In this mode, the direct current command is set to zero, preventing the discharge of submodule capacitors towards the fault point and controlling the DC fault current to zero to achieve isolation. Throughout this process, the submodule capacitor voltage control remains operational, preventing continuous capacitor charging. Additionally, the reactive power compensation function for the AC system remains unaffected. After the fault is cleared, the MMC can quickly resume operation.

4. Case Study

To validate the theoretical analyses presented in Section 2 and Section 3, a simulation model of a fault-tolerant series-connected large-scale offshore wind farm system was developed using PSCAD/EMTDC simulation software (v4.6.3), as depicted in Figure 14. The system comprises four MMCWT units forming the series wind farm, along with the CET converter and the onshore converter station MMC. The output voltage of the MMCWT is limited to 1.05 pu.

The parameters of the CET converter are listed in

Table 1. Main parameters of CET converter.

Items	Data
Rated power/MW	40
Rated low-voltage side voltage/kV	80
Rated high-voltage side voltage/kV	240
Rated low-voltage side current/kA	0.5
Operating frequency/Hz	150
Submodules per arm	132
Rated capacitor voltage/kV	2
Submodule capacitor/mF	1.01
Arm inductor/mH	10
Working thyristors	Tj1, Tj2

, while the parameters of the wind farm MMC and the onshore converter station MMC are provided in

Table 2. Main parameters of MMC.

Items	WTMMC	Onshore MMC
Rated power/MW	10	40
Rated DC side voltage/kV	20	240
Rated capacitor voltage/kV	1	4
Submodules per arm	20	60
Submodule capacitor/mF	9.26	0.77
Arm inductor/mH	4.8	172

.

4.1. Internal Fault of Single Wind Generator

At the initial time, the system is operating at rated conditions: all four MMCWT units have an output power of 10 MW with an output voltage of 20 kV. The CET converter operates with a low-side power of 40 MW, a low-side voltage of 80 kV, and a low-side current of 0.5 kA. The high-side power of the CET converter is also 40 MW, with a high-side voltage of 240 kV and a high-side current of 0.13 kA. At t = 1.1 s, the second MMCWT unit is removed due to an internal fault. The fault is cleared 1 s later, and at t = 2.1 s, the second MMCWT unit is reintroduced.

Figure 15 illustrates the simulation results for an internal fault of a single wind turbine. From Figure 15a,b, it is evident that the remaining three MMCWT units consistently operate under rated conditions. In Figure 15d, it can be observed that the low-side current of the CET converter remains at 0.5 kA throughout. Figure 15e shows that the high-side voltage of the CET converter consistently stays at 240 kV.

At t = 1.1 s, the second MMCWT experiences an internal fault, and after the action of the bypass switch and isolation switch, it is removed from the system, as indicated in Figure 15a,b. Consequently, the output power and output voltage of the second MMCWT decrease to 0, as shown in Figure 15c. The low-side and high-side power of the CET converter also decrease by 10 MW, as depicted in Figure 15d, and the low-side voltage of the CET converter decreases by 20 kV, as shown in Figure 15e. Additionally, the high-side current of the CET converter decreases to 0.125 kA.

When the second MMCWT is reintroduced into the system at t = 2.1 s after the action of the bypass switch and isolation switch, and the submodules are unlocked to resume operation, the system returns to its rated operating state, as depicted in Figure 15a–e.

The simulation waveforms depicted in Figure 15 indicate that the system remains stable both when an MMCWT exits operation due to an internal fault and when a faulty MMCWT is reintroduced into operation. The removal and reintegration of the faulty MMCWT do not affect the operational status of the remaining MMCWT units. Furthermore, the power generated by the wind farm continues to be fully transmitted throughout the process.

4.2. HVDC Transmission Line Fault

At the initial moment, the system is operating at rated conditions, with all four MMCWT units outputting 10 MW of power each, with an outlet voltage of 20 kV. The CET converter operates with a low-side power of 40 MW, a low-side voltage of 80 kV, and a low-side current of 0.5 kA. On the high side, the CET converter operates with a power of 40 MW, a voltage of 240 kV, and a current of 0.13 kA. At t = 1.1 s, a short-circuit fault occurs in the DC transmission line, which is cleared after 1 s. The system resumes operation at t = 2.1 s.

Figure 16 depicts the simulation results for the HVDC transmission line fault. From Figure 16a, it is observed that following the occurrence of a short-circuit fault in the DC transmission line at t = 1.1 s, both the power output from the MMC high-side and the power received by the onshore MMC drop to zero. Figure 16b illustrates that as the MMC switches to zero DC current control and the power devices of the CET converter are locked out, the DC side voltage decreases to zero. Moreover, Figure 16c shows that following the switch to zero DC current control at the onshore converter, the fault current entering the onshore MMC is rapidly suppressed, with the fault current decaying to zero within approximately 0.19 s, while the CET converter, through the locking out of power devices, causes the DC current to decay to zero. After fault clearance, the system resumes operation at rated conditions at t = 2.1 s. Notably, during the transient period of system recovery depicted in Figure 16b,c, the amplitude of the high-side current and voltage of the CET converter rises significantly. This phenomenon arises from the absence of a stable DC voltage establishment at the moment of unlocking at the onshore converter, resulting in a large deviation between the reference and actual values of the bridge arm current control in the unlocking instant. Figure 16d,e indicate that during the period of DC short-circuit fault, there are no significant deviations in the submodule capacitor voltages of both the CET converter and the onshore MMC.

5. Conclusions

The paper introduces a high-voltage, high-capacity offshore wind power full-DC system with the capability of DC fault interruption, suitable for high-voltage and large-capacity applications. Additionally, a fault isolation control method suitable for this system is proposed. For internal faults in the MMCWT, the fault current can be interrupted by locking the submodule, and the faulty MMCWT can be isolated from the system using bypass and isolation switches to ensure stable operation of the system. In the event of a high-voltage DC transmission line fault, the CET-DCT can interrupt the fault current by locking, while the onshore converter station can actively limit the fault current through independent DC loop control. During the fault process, the CET-DCT dissipates wind farm power by engaging resistive loads, thus obviating the need for MMCWT action. Simulation results using PSCAD/EMTDC validate the effectiveness of the proposed fault isolation method. However, the paper still has limitations. The transient currents and voltages during the recovery process after clearing the DC transmission line fault are too high, and the detection process of faults has not been studied, which are areas for future research focus.