A hybrid circuit breaker with fault current limiter circuit in a VSC-HVDC application

Abstract

A conventional hybrid circuit breaker (HCB) is used to protect a voltage source converter-based high voltage direct current transmission system (VSC-HVDC) from a short circuit fault. With the increased converter capacity, the DC protection equipment also requires a regular upgrade. This paper adopts a novel type of HCB with a fault current limiter circuit (FCLC), and focuses on the responses of voltage and current during DC faults, which are associated with parameter selection. PSCAD/EMTDC based simulation of a three-terminal VSC-HVDC system confirms the effectiveness and value of HCB with FCLC, by using an equivalent circuit modelling approach. Laboratory experimental tests validate the simulation results. The peak fault current is reduced according to the current limiting inductor (CLI) increase, and can be isolated more quickly. By adopting parallel metal oxide arrester (MOA) with the main branch of HCB, voltage stresses across the breaker components decrease during transient and continuous operation, and less energy needs to be dissipated by the MOA. The remnant current for all cases is transmitted to power dissipating resistor (PDR) in the final stage, and the fault current is reduced to the lowest possible value. When the current from the main branch is transferred to the FCLC branch, transient voltage spikes occur, while smaller PDR is required to absorb current in the final stage.

1 Introduction

The expansion of large-scale industry has brought economic growth for many parts of the world, but it simultaneously has caused a severe threat to environment due to the increased carbon dioxide emission. There has been increased use of renewable energy resources for power production and studies have highlighted the critical statistics of renewable energy production. However, these studies also show that there is a gap between renewable energy resources and power generation [1,2,3,4]. The apparent reason is that the enriched renewable energy sites are located far from the load centres, and this gives rise to the challenge of transmitting power over long distance. Considering all limitations, HVDC transmission becomes an appropriate technology to connect remote renewable energy with load centres.

The development of HVDC technology has spread over many decades, beginning from mercury rectifiers to VSC [5,6,7,8]. A HVDC transmission system can be divided into two major types, i.e., line-commutated converter (LCC-HVDC) and VSC-HVDC [9, 10]. A comprehensive evaluation of both technologies is given in [11]. However, along with the many benefits of VSC-HVDC technology, it has the disadvantage of a fast rising fault current when subject to DC short circuit faults. VSC-HVDC transmission systems employ circuit breakers (CBs) at both sides of the DC line to protect it from short-circuit faults. These breakers can be classified into three major types, i.e., mechanical (MCB), solid state (SSCB), and hybrid (HCB) [12,13,14,15,16,17]. For the first time, ABB presented the idea of HCB with experimental validation in [18,19,20]. Later, ALSTOM also proposed a type of HCB [21].

Various studies have highlighted the performance of HCB [22, 23]. It is concluded that, as the power handling capacity of VSC-HVDC system has increased over the time, the capacity of HCB is inadequate. To improve this situation, a FCLC is used along with a breaker to improve power handling related matters. A FCLC can be divided into two types, i.e., superconducting (SCFCLC) and non-superconducting (NSCFCLC). References [24,25,26,27] report some examples of SCFCLC, while [28,29,30,31,32] explain examples of NSCFCLC. The comparative analysis of the two types of FCLC is elaborated in [33]. A FCLC can be used as an independent circuit with a CB [34, 35], while design of the CB can also be associated with FCLC features. References [36, 37] illustrate recent cases with a modified CB. They consider FCLC features, while the limitation of an FCLC is discussed in detail, and the fault current is reduced stepwise.

To match the capacity of a HCB with a VSC-HVDC system, a FCLC can be used with HCB. Some FCLC based HCB or independent FCLC have been proposed, but they contain certain disadvantages [38,39,40]. One of the major disadvantages of existing FCLC approaches is that some FCLC components remain inside the circuit for both normal and fault conditions. Consequently, it is impossible to increase the value of these components beyond an specific point. Otherwise, it would cost extra losses and possible voltage drop at the receiving side of converter due to heavy inductance. Furthermore, the series and parallel operation of current limiting inductors (CLI) significantly increases the cost and size of the protection equipment. A detailed analysis of the mathematical formulation of fault current is reported in [41, 42]. The studies use a complicated method to estimate the fault current by considering the time-dependent variations of the voltage due to a DC-link capacitor.

A HCB with FCLC is designed by the authors [38], and Fig. 1 shows the structure of the auxiliary branch, where S₁, S₂, S₃…S_n denote the bidirectional switch (BDS) formed by IGBT, while L₁, L₂, L₃…L_n signify the CLI. In normal working condition, the current follows through the low resistive branch. When the system is in a fault situation, current flows through the auxiliary branch. After the fault current has decreased to the desired level in the auxiliary branch, the current is shifted to the fault isolation branch at the terminal stage. Thus, none of the CLI is used in the normal operating condition, so losses due to CLI in the normal condition are negligible, and the numbers and values of CLI can be increased without extra losses.

Fig. 1

Structure of the auxiliary branch of HCB with FCLC

In Fig. 2, an alternate and more straightforward scheme is developed to evaluate the performance of HCB rigorously. As shown, V_DC represents the terminal/source voltage, L_b is the breaker’s impedance, also known as current limiting reactance (CLR), L_l is the cumulative line inductance, and R_L is the load resistance. In addition, two different current flow paths are highlighted.

Fig. 2

Representation of equivalent circuit modelling approach

This paper adopts the novel HCB with FCLC proposed in [38], and focuses on the response character of voltage and current. This is associated with the parameter selection of the FCLC. PSCAD/EMTDC-based simulation authenticates the significance of HCB with FCLC by using a three-terminal VSC-HVDC model and an equivalent circuit modelling approach. Laboratory tests validate the simulation results. The main contents of the paper are as follows: Sect. 2 realizes the design and analysis for HCB with FCLC, while Sect. 3 analyses the response behavior of fault current. Section 4 shows the simulation and experimental results, and Sect. 5 contains the conclusions and general commentary on the results.

2 Design and analysis for HCB with FCLC

2.1 Conventional HCB and HCB with FCLC

Figure 3 represents the schematic diagram of conventional HCB without FCLC. This has three branches, i.e., low-resistance, fault current transmission, and fault clearing branches. The CLR, BDS constituted by IGBT, and mechanical switch UFD form the main branch and the auxiliary branch of the HCB. The third branch, known as the energy absorption branch of the HCB, contains a MOA, which is used for fault elimination.

Fig. 3

Schematic diagram of conventional HCB

The current flows through the main branch of the HCB in the normal working condition. After a DC fault and once the fault current crosses a threshold value, it is transferred to the auxiliary branch. At the final stage, the residual current is transferred to the energy absorption branch, isolated from the system, and the UFD can be operated to recover the line physically.

The adopted HCB with FCLC is shown in Fig. 4 [38]. The main and energy absorption branches are the same as the conventional HCB, while the only difference is in the auxiliary branch, as shown in Fig. 4. Similar to a conventional HCB, the current flows through the low resistance main branch during normal operating conditions. When an fault is detected, the fault current is moved to the auxiliary branch, which includes the FCLC. So in this way, the main breaker inductance L_b and FCLC inductance operate in series to suppress the fault current. This causes the fault current to be significantly reduced. Once the desired level of fault current suppression is achieved, the residual fault current is shifted to the MOA in the energy absorption branch for its complete isolation.

Fig. 4

HCB with FCLC

2.2 Operating period of HCB

The operating period for each branch of both breakers, i.e., conventional HCB and HCB with FCLC, is shown in Fig. 5, which has two parts. In the first part, various operating stages are described with the association of branches. In the second part, the current and operating period of each region are explained respectively to show the active branch of the breakers during a particular period.

Fig. 5

Operating regions of both breakers

From Fig. 5, it is apparent that from 0 to t₁, both breakers operate in the normal operating region, which is defined as R₁. The main or low resistive branch of both breakers carries the current I₁, described as the normal or rated current. A fault occurs at t₁, and the fault current begins to increase at a rapid rate from its rated value and reaches the peak point at t₂. It is important to note that during t₁ to t₂, i.e., R₂, the current remains flowing through the main branch for both breakers. This is the transient operating region, and the current during this period reaches the value of I₂. The fault is detected at t₂, and the current flowing path is transferred to the auxiliary branch of both breakers. In fact, the fault current cannot be suppressed through a conventional HCB, and will increase further to the high value during t₂ to t₃, i.e., R₃. In contrast, HCB with FCLC reduces the fault current to a considerable extent, so I₃ is much lower than I₂. The system enters R₄ after t₃, and the residual current for both breakers is switched to the energy absorption branch to be isolated completely.

3 Response behaviour of fault current

There are two most important components used in HCB with FCLC, i.e., CLR and CLI. To optimize the values of inductors used as CLR and CLI, it is crucial to understand the essence of current in various operating stages of the system.

3.1 Response behaviour aroused by short circuit fault

When the system is prone to any types of DC short circuit problem, a complete analysis of the situation considering the comprehensive information of fault current at different stages is required. For the voltage and current characteristics, both simulation and experiment need to be noted. A 200 kV voltage level is used for simulation, which is in line with various practical projects, e.g., the Trans Bay Cable project in the US [40].

Based on the parameters in [38], the simulation results are shown in Fig. 6a, b for the current and voltage, respectively. From 0 to 0.2 s, the system operates under normal conditions. The rated current during this stage is 0.5 kA, and the rated voltage is 200 kV. After 0.2 s, the system is subjected to a DC fault, and the fault current escalates at a very fast pace. Conversely, the DC capacitor begins to discharge. Eventually, the capacitor discharges completely, and the voltage and current drop to zero. As L_b and CLI are two critical components used in HCB with FCLC, it is essential to consider some critical points while selecting the appropriate values for these components.

• (a) Once the fault in the system is confirmed, the breaker must have the capacity to withstand its peak current value.

• (b) The gradual slope of the breaker current should be larger than the slope of the line current.

• (c) The fault current must be excluded before the DC-link capacitor begins to discharge.

Fig. 6

Voltage and current behavior in fault condition: a current; b voltage

These points can be analytically summarized as:

$$f\left( x \right) = \left\{ {\begin{array}{*{20}l} {\max \left( {I_{line} } \right) < \max \left( {I_{B} } \right)} \hfill \\ {\max \left( {dI_{line} /dt} \right) < \max \left( {dI_{B} /dt} \right)} \hfill \\ {t_{fct} < t_{cdt} } \hfill \\ \end{array} } \right.$$

(1)

In (1), max(I_line) indicates the peak fault current when subjected to a fault. The maximum value that the breaker can withstand is defined as max(I_B). dI_b/d_t reflects the breaker current rate of change, and dI_line/d_t denotes the line current rate of change. t_fct shows the total fault removal time, and t_cdt indicates the time taken to discharge the DC-link capacitor.

3.2 Current in different operating regions

To select CLR values in the main branch of both breakers and CLI in the auxiliary branch of HCB with FCLC, it is necessary to establish the equivalent circuits and approximate current in various operating regions. By applying KVL and Laplace transformation, the approximate value of the current can be derived. Figure 7 represents the equivalent circuits in various operating regions, such as R₁, R₂, R₃ and R₄, for both breakers.

Fig. 7

The equivalent circuits for different operating regions: a R₁ for conventional HCB and HCB with FCLC; b R₂ and R₃ for conventional HCB, and R₂ for HCB with FCLC; c R₃ for HCB with FCLC

Figure 7a represents the circuit during the steady-state operating region, which is defined as R₁ in Fig. 5. The circuit is similar for both breakers. In Fig. 7b, the circuit is effective for the transient operating region, and remains operating until fault detection. The circuit is used in the second operating region for both breakers, which is defined as R₂ in Fig. 5. It is apparent that the circuit shown in Fig. 7b is also used for operating region R₃ for conventional HCB, because no current limiting component is employed to limit the fault. Finally, the circuit in Fig. 7c is used in the operating region, which is defined as R₃ in Fig. 5 for HCB with FCLC.

Applying KVL and Laplace transformation for the circuit in Fig. 7b yields:

$$u\left( t \right) = \frac{{V_{1} \omega_{0} }}{{\omega_{n} }}e^{{ - {\Upsilon }\left( {t - t_{1} } \right)}} sin\left( {\omega_{n} \left( {t - t_{1} } \right) + \xi } \right) - \frac{{I_{1} \omega_{0} }}{{\omega_{n} }}e^{{ - {\Upsilon }\left( {t - t_{1} } \right)}} sin\left( {\omega_{n} \left( {t - t_{1} } \right)} \right)$$

(2)

$$I_{2} \left( t \right) = - \frac{{I_{1} \omega_{0} }}{{\omega_{n} }}e^{{ - {\Upsilon }\left( {t - t_{1} } \right)}} sin\left( {\omega_{n} \left( {t - t_{1} } \right) - \xi } \right) + \frac{{V_{1} \omega_{0} }}{{\omega_{n} \left( {L_{b} + L_{Line} } \right)}}e^{{ - {\Upsilon }\left( {t - t_{1} } \right)}} sin\left( {\omega_{n} \left( {t - t_{1} } \right)} \right)$$

(3)

where $u\left( {t_{1} } \right) = V_{1} ; I\left( {t_{1} } \right) = I_{1} ; \xi = R_{Line} /2\left( {L_{b} + L_{Line} } \right); \xi = R_{Line} /2\left( {L_{b} + L_{Line} } \right); \omega_{0} = \sqrt {\xi^{2} + \omega_{n}^{2} } ; \Upsilon = a\tan (\omega_{n} /\xi )$ and $\omega_{n} = \sqrt {1/(L_{b} + L_{Line} )C - (R_{Line} /2(L_{b} + L_{Line} ))^{2} }$.

This paper studies the fault current with updated parameters and develops a MATLAB algorithm for calculating the values of the CLR. Because of the complex task of estimating the current, a simple and understandable proposition is chosen. Because the voltage fluctuation caused by the DC-link capacitor is low, it is treated as a separate DC source when KVL and Laplace transformation are used for the corresponding circuits in Fig. 7. Thus, during t₁ to t₂, there are:

$$I_{1} \left( t \right) = I_{DC}$$

(4)

$$I_{2} \left( t \right) = \delta - \left( {\delta - I\left( {t_{1} } \right)} \right)e^{{ - \zeta \left( {t - t_{1} } \right)}}$$

(5)

$$I_{3} \left( t \right) = \delta - \left( {\delta - I\left( {t_{2} } \right)} \right)e^{{ - \zeta \left( {t - t_{2} } \right)}}$$

(6)

Equations (4)–(6) represent the final results of the current associated with a conventional HCB. In addition, (4) reports the current in normal working condition, which is also defined as the rated current of the system. The results in (5) and (6) are acquired using KVL and Laplace transformation for the circuit in Fig. 7b at two operating regions, R₂ and R₃. In (4), I_DC represents the rated DC current of the system. In (5) and (6), different variables are used and are defined as: δ = V₁/R_l, where V₁ represents the terminal DC voltage and R_l is the line resistance; ζ = R_l/(L_l + L_b), where L_l is the line inductance, L_b is the breaker inductance.

Equations (7)–(9) are obtained for HCB with FCLC by using a similar approach, as:

$$I_{1} \left( t \right) = I_{DC}$$

(7)

$$I_{2} \left( t \right) = \delta - \left( {\delta - I\left( {t_{1} } \right)} \right)e^{{ - \zeta \left( {t - t_{1} } \right)}}$$

(8)

$$I_{3} \left( t \right) = \delta - \left( {\delta - I\left( {t_{2} } \right)} \right)e^{{ - \xi \left( {t - t_{2} } \right)}}$$

(9)

Equations (7) and (8) are also acquired from Fig. 7a, b, respectively. For the circuit in Fig. 7c, including the CLI in FCLC, (9) is calculated using KVL and Laplace transformation. An essential point to note is that the structure of HCB with FCLC is modular in nature. As a result, according to design requirements, the number of inductors employed as CLI in FCLC, which are used in the auxiliary branch of the breaker, can be raised or lowered. This paper adopts three CLI as shown in Fig. 4. Moreover, ξ = R_l/(L_eq), where L_eq = L₁ + L₂ + L₃ + L_l. L₁, L₂, and L₃ are used as CLI in FCLC.

Based on (5) and (8), a MATLAB program is created to simulate the effect of fault current with various inductance values and fault location from the reference VSC-HVDC system terminal. The effect on I₂ is shown in Fig. 8. This effect is occasioned by inductances, distance between the fault and any reference VSC-HVDC system terminal. It can be observed that the impact of the fault is more extreme if it is closer to the converter station, and the subsequent peak current value is greater. It is clear that if the CLR value in the main branch of both breakers is higher, the rate of current change is lower, and the peak value is also reduced. This paper defines the rated current I₁ as 0.5 kA, rated terminal DC voltage as 200 kV, and 2 ms as R₂. Considering 3.5 kA/ms as the increasing rate of fault current and 100 km as the defecting distance [18], the CLR can be approximated as 38.4 mH.

Fig. 8

Impact of varying breaker impedance on current

CLI is elevated using the similar approach for CLR in the main branch. Based on (9), the influence of CLI on the current in R₃ is shown in Fig. 9. There are some important points which have been considered while implementing the algorithm.

• (a) Three CLI in FCLC have the same value to avoid mathematical complexities.

• (b) In (9), I(t₂) is the current at the previous stage. Considering the rated value of 0.5 kA current, 2 ms of the total operating time in R₂, and 3 kA/ms increasing rate of current, the maximum value of I(t₂) is 7.5 kA.

• (c) By assuming the rated terminal DC voltage of 200 kV, the value of L_b is taken as 38.5 mH.

Fig. 9

Influence of CLI on current in R₃

In Fig. 9, the results show that if larger CLI is adopted for a longer period, the fault current is reduced considerably.

4 Simulation and experimental results analysis

4.1 The simulation results

4.1.1 The three-terminal VSC-HVDC system

Figure 10 shows a three-terminal VSC-HVDC system, whose parameters are listed in Table 1 [38]. The DC link voltages for terminal 1, 2, and 3 are V_dc1, V_dc2, and V_dc3, respectively. Similarly, the AC grid voltages for terminal 1, 2, and 3 are denoted by V_ac1, V_ac2, and V_ac3. The flow of current between terminal 1 and 2 is defined as I₁₂, and I₁₃ is that between terminal 1 and 3. P₁, P₂, and P₃ represent the active power for terminal 1, 2, and 3, respectively. Finally, it is noted that the DC power is controlled by adopting a double-loop control strategy.

Fig. 10

The three-terminal VSC-HVDC system

Table 1 Parameters of the three-terminal VSC-HVDC system

4.1.2 Response of current

Figure 11 analyses the current when the system is subjected to a DC line-to-line short circuit fault. Figure 11a reports the basic response, and Fig. 11b highlights the current reaction with an explanation in different operating regions. In Fig. 11a, during 0–0.2 s, the system operates and carries the pre-defined current of 0.5 kA. The current in R₁ is denoted as I₁ in Fig. 11b. At the end of R₁, the fault is injected, and the current rises at the rate of 3.5 kA/ms for both breakers. It is considered that 2 ms is required to detect the fault, so the total span of short period R₂ is 2 ms. At the end of R₂, the current for both breakers has reached the maximum peak value, which is approximately 7.5 kA. Therefore, the current in R₂ is defined as I₂.

Fig. 11

Response of current under fault condition: a no explanation of operating region; b with explanation of operating region

We should note that the current I₁ and I₂ in R₁ and R₂ are identical for both breakers in Fig. 11b. The fault is confirmed at the R₂ boundary. For a conventional HCB, no current limiting components are used in the energy transfer/auxiliary branch, and the current I₃ in R₃ keeps rising at the rate of 3.5 kA/ms. Thus, at the end of this region, I₃ reaches almost 14 kA. On the other hand, HCB with FCLC employs three 100 mH CLI in the auxiliary branch. When the system enters R₃, the fault current drops sharply, and reaches the final value of 2 kA at the end of R₃. The residual current in both situations is transferred to the energy absorption branch once the R₃ boundary is passed, and the system enters the last function area R₄. Figure 12 reports the voltage across the breaker, and Fig. 13 depicts the energy dissipated by the MOA.

Fig. 12

The voltage across the breaker

Fig. 13

The power dissipated by MOA

Figure 12 compares the different responses of voltage for breakers without and with FCLC. The response can be divided into different segments. During 0–0.2 s, the system is in a normal operating state, and zero voltage during this period can be observed across both breakers. The fault is injected at 0.2 s, but the current continues to follow the path of the main branch, i.e., the low resistive path, until 0.202 s. When the current is switched from the main to the auxiliary branch after the current limiting operation starts, a transient voltage spike is noticed. This is caused by diverging the current flow from one branch to another. However, the voltage returns to the system rating of 200 kV after a short while. When the current is shifted from FCLC to MOA, an similar voltage spike is observed. Figure 13 shows the dissipation of the energy curve under the scope of both breakers. From the response of DC current in Fig. 11, it is evident that at the end of R₃, the peak value of current for conventional HCB is much higher than that of HCB with FCLC. Therefore, the fault current at this stage is quickly removed from the system when HCB with FCLC is adopted. Moreover, less energy is dissipated by MOA in the energy absorption branch. Figure 14 shows an additional result to observe the impact of CLI while the system is subjected to a DC short circuit fault.

Fig. 14

Impact of CLI on the system current

Four cases of C1, C2, C3, and C4, with different values of CLI of 100 mH, 150 mH, 200 mH, and 250 mH, respectively, are considered in Fig. 14. It can be seen that if the value of CLI is increased, the peak current in the fault current limiting region R₃ decreases accordingly, and the slope of current in that region also decreases. Also, the fault current can be isolated more quickly.

The impact of parallel MOA is explored and is shown in Fig. 15. Voltage stress across the breaker components decreases with parallel MOA during transient and continuous operation in Fig. 15a, although MOA raises the cost and size of the HCB. In Fig. 15b, parallel MOA is employed with the main branch of HCB, and less energy needs to be dissipated by the MOA at the final stage. In summary, it is noted that a parallel MOA has the greatest influence on voltage and power dissipation.

Fig. 15

The result of parallel MOA: a voltage, b power dissipation by MOA

4.1.3 Comparison among [18, 35, 37] and this paper

Figure 16 displays the conclusions of the comparison analysis. Cases CS1, CS2, CS3 and CS4 represent HCB in [18, 35, 37] and this paper respectively, and the equivalent circuit parameters are listed in Table 2.

Fig. 16

Comparison among [18, 35, 37] and this paper: a current, b power dissipation by MOA, c voltage

Table 2 Equivalent circuit parameters for [18, 35, 37] and this paper

All case studies from CS1, CS2, CS3 and CS4 show that the system carries the predefined normal current of 0.5 kA from 0 to 0.2 s. At 0.2 s, a fault is introduced into the system, and the current rises above the rated value of 0.5 kA. Because no fault current limiter is used in CS1, the current continues to rise at a rate of 3.5 kA/ms. After the high fault current is detected at 0.202 s for CS2, CS3, and CS4, the fault current limitation process begins. It can be seen from Fig. 16a that at the end of the fault current limitation period, the fault current for CS4 is the lowest compared to other HCBs. Consequently, at the terminal stage for CS4, the least energy is required to be absorbed by the MOA, as can be seen from Fig. 16b. Figure 16c shows the voltage profiles of all breakers. The results show that during normal operating from 0 to 0.2 s, the HCBs have no voltage drop. The transient voltage responses of the HCBs are observed at the instant when the fault current limitation process starts, and at the terminal stage when the fault current is shifted to the MOA. Based on these findings, it is clear that the proposed HCB outperforms its competitors.

4.2 Experimental results

4.2.1 Parameters for the experimental platform

Compared to the simulation, the fundamental objective of laboratory experiments is to verify the functionality of the breaker irrespective of rated voltage and current. The rated voltage and current are kept low for the laboratory experiment because of the availability of laboratory equipment and sensitivities associated with the experiment. Thus a small laboratory-scale experimental platform is developed to validate the design of the breakers, and the parameters are listed in Table 3. The circuit used to verify the conventional HCB is shown in Fig. 17.

Table 3 Parameters for the experimental platform

Fig. 17

Circuit for conventional HCB

Figure 18 highlights the circuit employed to verify the design of the HCB with FCLC experimentally, which is slightly different from the circuit shown in Fig. 3, to avoid complication and excessive use of circuit components. Line parameters for the experimental platform have been neglected because their values are too small to influence the main experimental results.

Fig. 18

Circuit for HCB with FCLC

An extra resistance R_shunt is used with the main branch switch to measure the current by observing the voltage across it. The resistance R_f is used to avoid an extreme rise of fault current, which may be aroused by the DC short circuit connection. A PDR is employed in the energy absorption branch to reduce the residual fault current to the lowest possible value. The MOSFET switches are S₁ and S₂, respectively. An inductance L_eqv is also used in Fig. 15 to represent the equivalent value of inductance in Fig. 5. R_L and V_DC represent the load resistance and power supply voltage, respectively.

4.2.2 Experimental results

This section verifies and discusses the response of current when the system is subjected to a line-to-line DC short circuit fault, with the current and voltage responses shown in Figs. 19 and 20, respectively. As shown in Fig. 19, the current response is divided into three periods. R₁ is associated with the steady-state response of the system. During this time, the system carries 50 mA rated current. When the system is subjected to a line-to-line short circuit fault, the current begins to increase from its rated value and eventually reaches twice the rated value (i.e. 100 mA) and maintains this value during R₂. The reason is that a fault resistance R_f is used, and if this resistance is not employed, the current will increase at the same incremental rate. The fault current is moved to PDR during R₃, and the current reduces dramatically.

Fig. 19

Current response of conventional HCB

Fig. 20

Voltage response of conventional HCB

In Fig. 20, the voltage response across the breaker is reported. It shows that during R₁ and R₂, there is no voltage increase across the breaker until the fault current is shifted to PDR at the end of the process. The voltage should ideally be equal to the supply voltage of 5 V. However, the results clearly indicate that the value of voltage is just above 3 V. This voltage drop is expected, because of the losses of different connecting wires and internal losses of the switches.

Meanwhile, three results are obtained for the HCB with FCLC at case C1, C2, and C3 as shown in Figs. 21 and 22. In general, the values of FCLC components are dependent on two factors, i.e., the rating of the particular VSC-HVDC project and the level of fault current suppression. Initially, for all three cases, the circuit carries 50 mA current. When the fault occurs, the fault current starts to increase and reaches twice the rated value within seconds for all cases. After that, the system enters the fault limitation stage, and the fault current falls below its peak value of 100 mA because of the CLI in FCLC.

Fig. 21

Current response of HCB with FCLC; a case C1, b case C2, c case C3

Fig. 22

Voltage response of HCB with FCLC; a case C1, b case C2, c case C3

It is interesting to note that the rate of current increase in the fault limitation stage decreases, and the peak value of fault current also decreases before shifting to PDR. This refers to cases C1, C2 and C3. This can also be observed from the results reported in Fig. 21. The remnant current for all cases is transmitted to the PDR in the final stage, and the fault current is reduced to the lowest possible value. If the value of the PDR increases, the current in the final stage can be reduced further. For the previous case when there is no FCLC, the current maintains its peak value of 100 mA before shifting to the PDR, as shown in Fig. 19. It means that the PDR of a conventional HCB needs to absorb higher current in the final stage than PDR of HCB with FCLC.

For the voltage response in the presence of HCB with FCLC shown in Fig. 22, it is observable that there are two voltage spikes. When the current is switched from one branch to another, the spikes occur. The voltage response difference between the conventional HCB and HCB with FCLC is highlighted. When the current from the main branch is transferred to the FCLC branch, the transient voltage spikes occur as shown in Fig. 22. During this stage, a small voltage drop across the breaker is also noticed. For the terminal phase, the response of voltage for both breaker designs remains the same. During this time, the voltage drop should ideally be equal to the supply voltage of 5 V. However, from all the results in Fig. 22, it can be seen that the final stage voltage drop across the breaker is 3 V. As previously described, the voltage drop is due to losses of different connecting wires and internal losses of switches.

4.3 Improvement and innovation compared to [38]

This paper focuses on the responses of voltage and current. These are associated with the parameter selection of the FCLC. The comparison between [38] and this paper is listed in Table 4, and the main improvement and innovation over [38] are as follows:

• (a) The fault peak current is reduced as CLI increases, of which the slope decreases, and can be isolated more quickly. By adopting a parallel MOA with the main branch of HCB, voltage stresses across the breaker components decrease during transient and continuous operation, and less energy needs to be dissipated by the MOA at the final stage. It is noted that parallel MOA has the greatest influence on voltage and power dissipation.

• (b) The experimental results on conventional HCB show that during R₂, fault current increases from its rated value and eventually reaches twice the rated value. It is moved to PDR and reduced dramatically during R₃. The voltage response shows that during R₁ and R₂, there is no voltage increase until the fault current is shifted to PDR at the end of the process.

• (c) The experimental results on the HCB with FCLC show that the remnant current for all cases is transmitted to the PDR in the final stage, and the fault current is reduced to the lowest possible value. If the value of PDR increases, the current in the final stage can be reduced further. The voltage response difference between the conventional HCB and HCB with FCLC is highlighted. When the current from the main branch is transferred to the FCLC branch, transient voltage spikes occur. During this stage, a small voltage drop across the breaker is also noticed. The PDR of a conventional HCB needs to absorb considerably more current in the final stage than the PDR of HCB with FCLC.

Table 4 Comparison between [38] and this paper

5 Conclusions

The paper focuses on the impact of the FCLC combined with the HCB, and presents the functional features, operating principle, and parameter selection associated with HBC with FCLC. Simulation and laboratory experiments are performed to validate effectiveness. Based on the theoretical discussion, simulation and laboratory experiments, it is concluded that the HCB with FCLC can provide a suitable scheme if large fault current suppression and interruption are required. The following are the conclusions drawn from the results of simulation and experiment:

• (a) The fault peak current is reduced as CLI increases, and can be isolated more quickly.

• (b) By adopting a parallel MOA with the main branch of the HCB, voltage stresses across the breaker components decrease during transient and continuous operation, and less energy needs to be dissipated by the MOA.

• (c) The remnant current for all cases is transmitted to the PDR in the final stage, and the fault current is reduced to the lowest possible value. When the current from the main branch is transferred to the FCLC branch, transient voltage spikes occur, while a smaller PDR is required to absorb current in the final stage.