Pilot Protection Based on Zero-Sequence Current Resistance-Capacitance Component for Large-Scale Inverter-Interfaced Power Stations

Abstract

At present, zero-sequence current protection is generally used as the main protection for single-phase ground faults in resistance-grounded inverter power stations. However, limited by the principle, it is difficult for current protection to take into account selectivity and rapid action when the neutral point resistance is large, so there is protection mismatch with the inverter-type power supply for low voltage ride through leads to the risk of large-scale disconnection of non-fault lines. Aiming at the above problems, firstly, a fault analysis model of the inverter power station considering the capacitance to ground is established to study the distribution characteristics of resistive and capacitive zero-sequence currents in the collection system, when single-phase ground short circuit occurs on different types of lines. Then, based on the characteristic difference between the resistance and capacitance components of the zero-sequence current flowing through the two ends of the tie line in case of internal and external faults, a zero-sequence pilot protection algorithm is formed. Compared with the traditional zero-sequence current differential protection, the proposed protection algorithm only transmits logic information without synchronous sampling, and has significant economy. Finally, the feasibility and effectiveness of the proposed protection algorithm are verified by an engineering simulation example.

1. Introduction

Building a new power system with new energy as the main body is an effective way to achieve the goal of “carbon peaking and carbon neutrality”. Taking the carbon peaking and carbon neutrality goals into account, by 2060, it is predicted that the combined installed capacity of wind power and photovoltaic power generation in China will account for about 80%, and the combined power generation will account for about 70% [1]. In the future, large-scale new energy grid connection will bring many challenges to the new power system, among which the safe and stable operation of large-scale inverter power grid connection is one of the major challenges [2,3,4,5].

The large-scale centralized grid-connected inverter-interfaced generator power stations (IGPS), adopt decentralized inverter and centralized grid-connected system structure [6]. Hundreds of inverter power generation units are connected by segmented cables into collector lines, and several collector lines are connected to the collector bus in parallel, and are then sent to the boost transformers through the tie lines, before finally forming a multi-level radial collection system network. The station collection system of IGPS uses cable lines widely. In order to effectively reduce the overvoltage level of single-phase grounding fault, at present, the operation mode of neutral with resistance grounded has been widely used. The grid specification stipulates that IGPS should have protective measures to rapidly remove the single-phase grounding fault of the station collection systems [7].

A large number of domestic and foreign literatures have carried out research on relay protection with the inverter-interfaced Distributed Generator (IIDG) connected to the distribution network [8,9,10,11,12,13,14,15]. Some scholars have analyzed different control strategies for IIDG under various circumstances, as well as the fault characteristics of IIDG under different control strategies [8,9]. Literature [8] analyzes the short-circuit current characteristics of photovoltaic system under different fault conditions through simulation, and points out that the output characteristics of inverter power source are different from that of traditional one. Literature [9] qualitatively gives the detailed characteristics of the IIDG model and inverter power source in the case of fault, but without the output characteristics. Literature [10,11] combined different fault characteristics of IIDG to analyze the influence of IIDG on protection elements of distribution network and distance protection, but did not carry out detailed analysis and setting calculation for current protection. Literature [12,13] discusses the influence of distributed power source on current protection, distance protection, reclosing, and distribution network voltage after it is connected to the distribution network, but no improvement measures are given. In literature [14], the positive sequence voltage at the installation of protection is introduced to modify the fixed value of the positive sequence current velocity protection adaptively. In literature [15], according to IIDG operation control parameters, the fixed value of the protection of the adaptive current velocity break is calculated online by the iterative solution method, so as to meet the protection requirements under different IIDG outputs.

The above studies effectively solve the sensitivity problem of current protection by reducing or eliminating the influence of IIDG on fault current. However, they do not consider the coordination relationship between protection and the Low Voltage Ride Through (LVRT) of power grid specifications. The topological characteristics of the centralized IGPS collection system require that the relay protection must be able to cooperate with the LVRT of the generating unit. Otherwise, the voltage fall behind of the collector bus may not be recovered timely, due to the mismatch between the protection and LVRT, leading to the problem of large area off-grid of the generating unit in the non-fault section [16,17]. Literature [18] transformed the coordination relationship between large-scale IGPS protection and LVRT into specific requirements for protection quickness and selectivity, and then proposed a zero-sequence current protection scheme considering LVRT coordination. However, the selectivity of current protection is based on the sufficient discrimination at the head and end of the current curve. With the gradual increase of the neutral point resistance, when the current curve is too gentle to ensure the discrimination, the selectivity can only be guaranteed by increasing the action time limit, which is obviously contradictory to the quick action. Pilot protection has well action characteristics, and absolute selectivity, so it can be used as a protection scheme with both rapidity and selectivity. Literature [19] proposed a current differential protection scheme suitable for a large-scale photovoltaic power source connecting to the distribution network. However, the existing communication facilities of distribution network cannot meet the requirement of high synchronization of fault information in differential protection [20]. In literature [21], the transmission of current scalar weakens the strict requirement of differential protection for synchronization error. The node branch auxiliary criterion solves the influence of unmeasured load branch on protection and eliminates the “dead zone” of the main criterion. However, in the case of heavy load and high resistance fault, the protection may refuse to work. Literature [22] proposed a new pilot protection scheme based on positive sequence voltage difference, which improves the access capacity of IIDG and reduces the requirement of multi-terminal protection on pilot channel. By combining voltage information, literature [23] proposed to realize active distribution network protection by using the comparison of phase between positive sequence component of bus voltage and the fault component of positive sequence current of each feeder before the fault occurs. All the literatures above require voltage information, which is difficult to achieve in engineering practice [24]. The influence of unmeasured load branch on protection is also not considered. In literature [25], the inverse time current differential protection scheme is proposed without voltage information, and the influence of unmeasurable load branches is discussed. However, the influence is only analyzed when the fault occurs outside of the station. It is still restricted by the current communication channel and data synchronization technology of distribution network, so it is difficult to obtain the current synchronization phase information, which may lead to protection misoperation.

In view of the above deficiencies, this paper deeply analyzes the distribution characteristics of zero-sequence current resistance-capacitance components in the large-scale IGPS collection system, when single-phase grounding short-circuit occurs in different types of lines, and proposes a pilot protection principle based on zero-sequence current resistance-capacitance components. Compared with the traditional zero-sequence current differential protection, this principle does not require synchronous sampling to reduce the strict requirements of the pilot protection for synchronization accuracy, and only transmits logical information, reducing the requirements for communication channels, thereby significantly improving the economy of the protection, which is conducive to large-scale promotion and use. Finally, an engineering simulation example is established by using the MATLAB. Through the quantitative protection setting calculation and verification process, it is verified that the proposed protection algorithm can accurately judge the internal and external faults within the full resistance range of the neutral point resistance.

2. Fault Characteristic Analysis of Large-Scale Centralized IGPS

In order to make fault analysis universal, three fault points K͵ K’ and K’’ are added to the centralized IGPS general topology, as shown in Figure 1, located on the tie line and the two collector lines, respectively.

As shown in Figure 1, a collection system of large-scale IGPS contains several medium-sized IGPS, which can be divided into two types according to topological structure. The first type is IGPS1 with collector lines directly connected to the booster station, and the second is called IGPS2-IGPSm, connecting to the booster station through a 35 kV tie line.

2.1. Fault Characteristic Analysis of Tie Line

When a single-phase (A-phase) grounding fault occurs at point K on the tie line MNm, the equivalent fault zero-sequence network is shown in Figure 2.

In Figure 2, since the booster transformer T1 is Y/Δ connection, the zero-sequence path on the system side is provided by a grounded transformer and a neutral resistance, Z_0R is the equivalent zero-sequence impedance of the neutral resistance branch, X_Ci is the reactance of capacitance to the ground of cable collector lines of IGPSi (i =1, 2,…, m), X_Ci = 1/jωC_0i, Where C_0i is capacitance to the ground of the collector line, and the line impedance of the collector line is ignored. Z_0Li is the zero-sequence impedance of Mni. When Z_0L1 = 0, the tie line is mostly overhead line and capacitive current is only 1/50 to 1/30 of the cable line, excluding the capacitance to the ground of the tie line; Z_0MK is the zero-sequence impedance between the fault point K and the bus M, Z_0KNm is the zero-sequence impedance between the point K and the bus Nm, and R_f is the transition resistance at the fault point. I_0R is the zero-sequence current flowing through the resistance branch, I_0K is the zero-sequence current of the fault point, I_0i is the zero-sequence current flowing through the capacitance branch, equal to the zero-sequence capacitive current to the ground of cable line in the IGPSi; U_f is the virtual power source at the point of failure, U_0M is the zero-sequence voltage of bus M, and U_0K is the zero-sequence voltage at the fault point K. (throughout this work, bold italics indicates a phasor).

According to the series-parallel relationship of each branch in Figure 2, the equivalent zero-sequence impedance $Z_{\sum 0}$ at the fault point K can be obtained as follows:

$$Z_{\sum 0}=\frac{1}{\frac{1}{Z_{0\mathrm{M}}+Z_{0\mathrm{M}\mathrm{K}}}+\frac{1}{X_{\mathrm{C}\mathrm{M}}+Z_{0\mathrm{K}\mathrm{N}\mathrm{M}}}}$$

(1)

where Z_0M is the equivalent zero-sequence impedance of capacitance to ground branch and the neutral resistance branch in parallel form at bus M at all non-fault IGPSs, and the expression is as follows:

$$Z_{0\mathrm{M}}=\frac{1}{\frac{1}{Z_{0R}}+{\displaystyle \sum _{j=1}^{\mathrm{m}-1}\frac{1}{X_{\mathrm{C}j}+Z_{0\mathrm{L}j}}}}$$

(2)

According to the principle of voltage division, the zero-sequence voltage U_0K at the fault point and the zero-sequence voltage U_0M of the system bus M can be obtained as follows:

$${\mathit{U}}_{0\mathrm{K}}={\mathit{U}}_{\mathrm{f}}\frac{Z_{\sum 0}}{Z_{\sum 0}+3R_{\mathrm{f}}}$$

(3)

$${\mathit{U}}_{0\mathrm{M}}={\mathit{U}}_{0\mathrm{K}}\frac{Z_{0\mathrm{M}}}{Z_{0\mathrm{M}}+Z_{0\mathrm{M}\mathrm{K}}}$$

(4)

The zero-sequence current I_0R flowing through the neutral resistance branch is:

$${\mathit{I}}_{0\mathrm{R}}={\mathit{U}}_{0\mathrm{M}}/Z_{0\mathrm{R}}$$

(5)

The zero-sequence current flowing through capacitance to ground branch of the non-fault IGPSj (j =1, 2, …, m − 1) is:

$${\mathit{I}}_{0j}={\mathit{U}}_{0\mathrm{M}}/(X_{\mathrm{C}j}+Z_{0\mathrm{L}j})$$

(6)

The zero-sequence current I_0M flowing through side M of the fault tie line is:

$${\mathit{I}}_{0\mathrm{M}}={\mathit{I}}_{0\mathrm{R}}+{\displaystyle \sum _{j=1}^{\mathrm{m}-1}{\mathit{I}}_{0j}}$$

(7)

The zero-sequence current flowing through the N_m side of the fault tie line, namely, the zero-sequence current I_0m of capacitance to ground branch of IGPSm is:

$${\mathit{I}}_{0\mathrm{M}}={\mathit{U}}_{0\mathrm{K}}/(X_{\mathrm{C}\mathrm{M}}+Z_{0\mathrm{K}\mathrm{N}\mathrm{M}})$$

(8)

The zero-sequence current I_0K flowing through the fault point:

$${\mathit{I}}_{0\mathrm{K}}={\mathit{I}}_{0\mathrm{M}}+{\mathit{I}}_{0\mathrm{M}}={\mathit{I}}_{0\mathrm{R}}+{\mathit{I}}_{0\mathrm{C}\sum }$$

(9)

where ${\mathit{I}}_{0\mathrm{C}\sum }$ is the sum of zero-sequence capacitive current in the collection system.

In general, Z_0Li $\ll$ X_Ci, i.e., the line impedance of each capacitance to ground branch is far less than the capacitive reactance, so the line impedance of each capacitance branch in series can be ignored. The zero-sequence impedance of grounding transformer is usually small. When the value of neutral point resistance R_g is large, such as Z_0R $\approx$ 3R_g, then X_Ci = 1/jωC_0i substituted into Equation (2), can be obtained:

$$Z_{0\mathrm{M}}=\frac{3R_{\mathrm{g}}}{1+3\mathrm{j}\omega R_{\mathrm{g}}({\mathrm{C}}_{0\sum }-{\mathrm{C}}_{0\mathrm{m}})}$$

(10)

where ${\mathrm{C}}_{0\sum }={\mathrm{C}}_{01}+{\mathrm{C}}_{02}+\cdots +{\mathrm{C}}_{0\mathrm{M}}$ is the sum of capacitance to ground of the cable line in the collection system.

In addition, in order to limit the arc overvoltage level, the ground capacitance of the collection system is at least twice that of Z_0R, so the resistance value of R_g mainly determines the size of Z_0M.

When R_g is large, Z_0M $\gg$ Z_0MK, in order to simplify the analysis process, usually ignore Z_0MK:

$$Z_{\sum 0}=\frac{3R_{\mathrm{g}}}{1+3\mathrm{j}\omega R_{\mathrm{g}}{\mathrm{C}}_{0\sum }}$$

(11)

Equation (11) is the expression of equivalent zero-sequence impedance of short-circuit point with ground capacitance in low resistance grounding system commonly found in existing literature. However, it should be noted that since Z_0MK actually reflects the change of fault location on the line, it must be taken Z_0MK into account when analyzing the characteristics of short-circuit current related to the fault location.

By substituting Equation (11) into Equation (3), the zero-sequence voltage at the fault point is:

$${\mathit{U}}_{0\mathrm{K}}={\mathit{U}}_{\mathrm{f}}\frac{R_{\mathrm{g}}}{R_{\mathrm{g}}+R_{\mathrm{f}}(1+3\mathrm{j}\omega R_{\mathrm{g}}{\mathrm{C}}_{0\sum })}$$

(12)

Equation (12) indicates that the zero-sequence voltage at the fault point is closely related to the neutral resistance R_g and the transition resistance R_f. Once R_g is determined and fixed after the photovoltaic power station is built, the zero-sequence voltage at the fault point will be mainly affected by the transition resistance R_f in grounding short circuit. Obviously, the zero-sequence voltage at the fault point reaches the maximum value when Rf = 0, that is, gold grounding short-circuit occurs. The short-circuit current flowing through each branch in the system also reaches the maximum value.

Based on the above analysis, the fault characteristics of single-phase (A-phase) grounding short-circuit in MN_m of the tie line can be as follows:

• Zero-sequence current I0M flowing through the fault tie line MNm and the M side of the system bus is the vector sum of the current I0R of the zero-sequence resistance branch and zero-sequence capacitive current to ground of cable lines in all non-fault IGPSs, as shown in Equation (7). The zero-sequence current flowing through the Nm side of the photovoltaic bus is the zero-sequence capacitive current I_0m of IGPSm.
• The zero-sequence current flowing through the non-fault tie line MNj (j = 1, 2,…, m–1) is the zero-sequence capacitance current I0j of IGPSj.
• The zero-sequence current flowing through the installation of protection on collector lines is the zero-sequence capacitive current to the ground of its own line.

2.2. Fault Characteristic Analysis of Collector Line

2.2.1. Fault of Collector Line Directly Connected to Booster Station

When the single-phase grounding fault occurs at the point K “on the collector lines 1-N₁ in IGPS1, there is the fault equivalent zero-sequence network, as shown in Figure 3. Due to X_Ci $\gg$ Z_0Li and Z_0M $\gg$ Z_0MK’’, the series impedance Z_0Li of the capacitive branch and the line impedance Z_0MK” between the fault point K” and the bus M are ignored. Because the ground reactance X_Ci is and R_g is large, they have little influence on the short-circuit current of the non-fault tie line.

The meaning of X_C2 to X_Cm in Figure 2 is exactly the same, which is the reactance of the ground capacitance from non-fault IGPS₂ to IGPS_m. Since the fault occurs on the collector line in IGPS₁, the ground capacitance of the fault collector line and the non-fault collector line are equivalent, respectively. X_C1 is the ground reactance of all the non-fault cable collector line in IGPS₁, X_Cu is the ground capacitance of the fault cable collector line 1-N₁ upstream of the fault point K”. I_01u is the zero-sequence capacitive current flowing through X_Cu, X_Cd is the ground capacitance of the downstream part of the fault point K”, I_01d is the zero-sequence capacitive current flowing through X_Cd. I_01-n1 is the zero-sequence current flowing through the first end of fault collector line 1-N₁ (side M of system bus). In order to facilitate comparative analysis, MNm is marked separately in the figure.

The equivalent zero-sequence impedance of the collection system at the fault point K” is as follows:

$$Z_{\sum 0}^{″}=\frac{3R_g}{1+3\mathrm{j}\omega R_g{\mathrm{C}}_{0\sum }}$$

(13)

It can be seen from Equation (13) that although the fault location changes the circuit of the ground-to-ground capacitance in the zero-sequence equivalent network, the reactance of all capacitive branches in parallel is exactly the same as that the fault of collector line occurs, so the zero-sequence voltage U_0K“ at the fault point is also exactly the same as the expression in Equation (12).

Therefore, the zero-sequence current I_01-N1 flowing through the bus side M of the fault collector line 1-N₁ (side M of system bus):

$${\mathit{I}}_{01-\mathrm{N}1}={\mathit{I}}_{0R}+{\displaystyle \sum _{j=1}^{\mathrm{m}}{\mathit{I}}_{0j}}$$

(14)

where I_0R is the current flowing through the resistance branch of the neutral point, and I_0j is the zero-sequence capacitive current flowing through non-fault grounded capacitive branch. The zero-sequence current I_0K” flowing through the fault point is:

$${\mathit{I}}_{0{\mathrm{K}}^{″}}={\mathit{I}}_{01-\mathrm{N}1}+{\mathit{I}}_{01\mathit{U}}+{\mathit{I}}_{01\mathrm{d}}={\mathit{I}}_{0R}+{\mathit{I}}_{0\mathrm{C}\sum }$$

(15)

It can be seen from Equation (15) that when the fault of collector line 1-N₁ occur, the zero-sequence current flowing through the fault point I_0K” is the same as that in the case of fault tie line, and both are the vector sum of resistance current I_0R and the total zero-sequence capacitive current in the system.

According to the above analysis, the fault characteristics of single-phase (A-phase) grounding short-circuit on collector line 1-N₁ are as follows:

• Zero-sequence current I_01-N1 flowing through fault collector line 1-N1 (side M of system bus) is the vector sum of resistance current I_0R and zero-sequence capacitive current to the ground of all non-fault collector lines in the system (including all collector lines from IGPS2 to IGPSm and non-fault collector lines from IGPS1).
• The zero-sequence current flowing through the installation of non-fault collector line protection is the zero-sequence capacitive current to the ground of this line.
• The zero-sequence current flowing through the non-fault tie line MNi is the zero-sequence capacitive current of its subordinate IGPSi, especially that the zero-sequence current flowing through MNm is the zero-sequence capacitive current I_0m of IGPSm.

2.2.2. Fault of Collector Line at the Lower Level of the Tie Line

When the single phase (A phase) grounding fault occurs at K’ point on the collector line m-n_m in the IGPSm below the tie line MN_m, the fault equivalent zero-sequence network is shown in Figure 4, and the impedance of the non-fault capacitive branch is ignored.

X_C1 to X_CM1 in Figure 4 are the reactance of the ground capacitance of the non-fault IGPSi (i =1, 2,…, m − 1). The ground capacitance of the fault collector line and the non-fault collector line are equivalent in IGPSm, respectively. X_Cm is the ground reactance of all non-fault collector line in IGPSm, X_Cu is the ground capacitance of the fault collector line upstream of the fault point K’, I_01u is the zero-sequence capacitive current flowing through XCu, XCd is the ground capacitive current flowing downstream of the fault point K’, I_01d is the zero-sequence capacitive current flowing through X_Cd. I_0M is the zero-sequence current flowing through the whole MNm tie line, and I_0m-nm is the zero-sequence current flowing through the m-nm first end of the fault collector line.

The equivalent zero-sequence impedance of the collection system at the fault point K’ is as follows:

$$Z_{\sum 0}^{\prime }=\frac{1}{\frac{1}{Z_{0\mathrm{M}}+Z_{0\mathrm{L}\mathrm{M}}}+\frac{1}{X_{\mathrm{C}\mathrm{M}}}+\frac{1}{X_{\mathrm{C}\mathit{U}}}+\frac{1}{X_{\mathrm{C}\mathrm{d}}}}$$

(16)

If the value of R_g is large and Z_0M $\gg$ Z_0Lm, Z_0Lm is ignored:

$$Z_{\sum 0}^{\prime }=\frac{3R_g}{1+3\mathrm{j}\omega R_g{\mathrm{C}}_{0\sum }}$$

(17)

I_0m-nm is zero-sequence current flowing through the Nm side of the bus of the fault collector line:

$${\mathit{I}}_{0\mathrm{M}-\mathrm{N}\mathrm{M}}={\mathit{I}}_{0R}+{\displaystyle \sum _{i=1}^{\mathrm{m}-1}{\mathit{I}}_{0i}}+{\mathit{I}}_{0\mathrm{m}}$$

(18)

where I_0m is the zero-sequence capacitive current of all non-fault cable collector line, I_0j is the zero-sequence current flowing through non-fault ground capacitive branch X_Cj (j =1,…, m − 1).

I_0M is the zero-sequence current flowing through whole tie line MN_m, without I_0m:

$${\mathit{I}}_{0\mathrm{M}}={\mathit{I}}_{0R}+{\displaystyle \sum _{i=1}^{\mathrm{m}-1}{\mathit{I}}_{0i}}$$

(19)

I_0K’ is the zero-sequence current flowing through the fault point:

$${\mathit{I}}_{0{\mathrm{K}}^{\prime }}={\mathit{I}}_{0\mathrm{M}-\mathrm{N}\mathrm{M}}+{\mathit{I}}_{01\mathit{U}}+{\mathit{I}}_{01\mathrm{d}}={\mathit{I}}_{0R}+{\mathit{I}}_{0\mathrm{C}\sum }$$

(20)

It can be seen from Equation (20) that the zero-sequence current I_0K’ flowing through the fault point when the collector line M-nm fault occurs has the same composition as the above two kinds of line faults, and I_0K’ is the vector sum of the resistance current I_0R and the total zero-sequence capacitive current in the system.

The fault characteristics of single-phase (A-phase) grounding short-circuit on collector line m-n_m are as follows:

• Zero-sequence current I0m-nm through the first end of the fault collector line is the vector sum of the resistance current I0R and the zero-sequence capacitive current of all the non-fault collector line in the collection system, including all the collector line of the non-fault photovoltaic station and IGPSm).
• The zero-sequence current flowing through the installation of non-fault collector line protection is the zero-sequence capacitive current to the ground of this line.
• The zero-sequence current flowing through the whole tie line MNi connected to non-fault IGPSi (i =1, 2,…, m − 1) is the zero-sequence capacitive current I0i of the cable in the station.
• Zero-sequence current I0M flowing through the whole superior tie line MNm of fault collector line M-NM is the sum of resistance current I_0R and zero-sequence capacitive current of all non-fault collector line of IGPSs, as shown in Equation (19).

3. Blocking Zero-Sequence Pilot Protection

3.1. Protection Algorithm

Figure 5 shows the installation position diagram of zero-sequence pilot protection for the tie line MNm in large-scale IGPS. P_M and P_N are installed on the M side of the booster station bus and the Nm side of the photovoltaic bus, respectively.

In the application circumstances of large-scale IGPS, based on the fault characteristic analysis in the previous section, a blocking zero-sequence pilot protection algorithm is designed for the tie line MNm, as shown in Figure 6.

Figure 6a,b shows the main components and action logic of the half blocking zero-sequence pilot protection P_M on the M side of the tie line and P_N on the Nm side, respectively. T1 is the instantaneous action delay return element, t2 is the delayed action instantaneous return element.

I₁ is the low constant starting element, and I₂ is the high constant starting element, which are set as follows:

$$I_1=3\left|{\mathit{I}}_{0\mathrm{m}}\right|/{\mathrm{K}}_{\mathrm{sen}}$$

(21)

$$I_2=3\left|{\mathit{I}}_{0\mathrm{R}}\right|/{\mathrm{K}}_{\mathrm{sen}}$$

(22)

where ${\mathrm{K}}_{\mathrm{sen}}$ is sensitivity coefficient, 1.5~2.

Equations (22) and (23) indicate that the low constant starting element I₁ is set to ensure sufficient sensitivity to the capacitive current $3\left|{\mathit{I}}_{0\mathrm{m}}\right|$ of IGPSm at the lower level of the tie line, while the high constant starting element I₂ is set to ensure sufficient sensitivity to the resistance current $3\left|{\mathit{I}}_{0\mathrm{R}}\right|$. Only qualitative analysis is made here. In the next section, quantitative protection setting calculation and verification will be carried out through specific examples, and detailed steps are given.

In particular, at the design stage for large-scale IGPS, in order to secure the security of primary side equipment, over-voltage of arc light level is limited. In the selection of neutral point resistance, it has been guaranteed that the resistance current is at least 2 times of the capacitive current of the collection system, that is, $\left|{\mathit{I}}_{0\mathrm{R}}\right|\ge 2\left|{\mathit{I}}_{0\mathrm{C}\sum }\right|>2\left|{\mathit{I}}_{0\mathrm{m}}\right|$, I₂ must be greater than I₁, and the high and low fixed values can ensure a reliable degree of differentiation.

The following are detailed analysis of the working process of the proposed protection scheme under different fault circumstances.

• When the power station is in normal operation, there is no zero-sequence current in the collection system, I₁ and I₂ are not started, then the four and-gates of P_M and P_N on both sides have no output, and the protection does not start.
• When the ground fault occurs at any point K” of the adjacent collector line outside the area, the zero-sequence current flowing through the whole tie line MNm is the zero-sequence capacitive current I_0m of its subordinate IGPSm. So I₁ start, I₂ do not. “and-gate 1” on P_M side has instantaneous action, start the transmitter, send locking signal, and lock protection on both sides. I₂ on the P_N side does not start and outputs 1 through the not gate, but it is sent to “the and gate 3” after t2 delay. Therefore, “the and gate 3” also outputs 1 before t2, and the starting transmitter sends a locking signal to close the trip circuit of “the and gate 4” on the local side. After t2, “the and gate 3” returns and stops sending signal, but the receiver has received the locking signal from P_M side, locking “the and gate 4” tripping circuit. To sum up, when there is a fault on the adjacent collector line outside the area, both sides of the protection can be reliably locked, will not miss-operate.
• When the fault occurs on other tie lines outside the zone, the zero-sequence current flowing through the whole tie line MNm of the non-fault tie line is still the zero-sequence capacitive current I_0m, and the subsequent protection work is the same as that of (2).
• When the ground fault occurs at any point K” of the subordinate collector line outside the area, the zero-sequence current flowing through the whole tie line MNm is shown as Equation (20), which is the vector sum of the zero-sequence resistance current I0R and the zero-sequence capacitive current of all non-fault capacitive branches. Therefore, the starting elements I₁ and I₂ on both sides of the protection are in action. “the and gates 1 and 3” are instantaneously started by I₁ before t2, and the transmitter sends a blocking signal. After t2, “the and gate 1” returns, and the transmitter on P_M side stops sending. However, after “the and gate 3” receives the 1 output by I₂ through the nor gate, the P_N side transmitter continues to send the message and sends the locking signal to lock the protection on both sides. To sum up, when there is a fault on the outside subordinate collector line, both sides of the protection can be reliably locked, will not miss-operate.

When the out-of-area faults of (2) and (3) are removed, the P_M side protection does not stop sending locking signal, “the and gate 4” element of P_N protection does not operate, and P_N protection does not trip. When the out-of-zone fault is removed, the starting element I₁ of P_M protection returns, but “the and gate 1” stops sending after t1 (generally 100 ms) delay, which can ensure that even if the P_N protection I₁ returns slowly, “the and gate 4” element can be reliably locked when the external fault is removed.

When the out-of-area fault of (4) is removed, P_N protection does not stop sending locking signal, P_M protection “the and gate 2” components do not operate, P_M protection does not trip. When the out-of-zone fault is removed, the P_N protection starting element I₁ returns, but “the and gate 3” stops sending after t1 delay, which can ensure that even if the P_M protection I₂ returns slowly, the “the and gate 2” element can be reliably locked when the external fault is removed.

5.

When the grounding fault occurs at the point K on the fault tie line MNm in the area, the zero-sequence current flowing through the M side of the fault line is the vector sum of the zero-sequence resistance current I0R and the zero-sequence capacitive current, as shown in Equation (7). Therefore, both I₁ and I₂ of P_M protection operate, and the zero-sequence current flowing through the Nm side of the fault tie line is zero-sequence capacitive current I_0m. PN protection only operates I₁ and I₂ does not start. Both “the and gates 1 and 3” are instantaneously started by I₁ before t2, and the transmitter sends a locking signal. After t2, P_M’s I₂ outputs 1, starts the tripping circuit of “and 2”, “the and gate 1” returns, and the transmitter stops. PN’s I₂ does not start and output 1 through the nor gate, start “the and gate 4” trip circuit, “the and gate 3” returns and the transmitter stops, both sides of the protection action trip.

In conclusion, it shows that the proposed protection scheme can accurately distinguish inside and outside fault based on the two sides protection of zero-sequence current logical judgment. The delay element is the transmission time waiting for the latching signal in the channel. As the length of the tie line is generally less than 20 km, the delay time is very short, generally less than 10 ms, which can fully meet the requirements of protection quickness and LVRT coordination.

3.2. Comparison with Traditional Zero-Sequence Current Differential Protection

Traditional zero-sequence current differential protection (hereinafter referred to as traditional protection) does not consider the distinction between resistance current and capacitance current, but uses the phasor sum of zero-sequence current on both sides of the line as the action criterion. Taking the interconnection line MN as an example, the zero-sequence current phasors at both ends of MN are I_0M and I_0N, respectively. The amplitude of the sum of the two phasors |I_0M + I_0N| is large when an internal ground fault occurs, but almost zero when an external fault or normal operation occurs. This difference in fault characteristics constitutes the traditional protection.

Traditional protection can meet the protection requirements of rapid operation of the whole interconnection line, but it requires real-time transmission of analog quantities in the channel, strict synchronous sampling at both ends of the line, and real-time transmission of sampling data through the communication channel. This requires high requirements for communication channels and protection equipment, which is difficult to apply in large scale at this stage.

By analyzing the characteristics of capacitive current and resistive current, the proposed protection has designed a low setting starting element I₁ corresponding to capacitive current and a high setting starting element I₂ corresponding to resistive current, so that the proposed protection does not need synchronous sampling, thus avoiding the input of sampling synchronization equipment, and the protection on both sides only requires interactive logic information, and has lower requirements on communication channels, which is more suitable for wide range applications.

The comparison between the traditional zero-sequence current differential protection and the proposed protection is shown in

Table 1. Comparison between the traditional protection and the proposed protection.

Requirement	The Traditional Zero-Sequence Current Differential Protection	The Proposed Zero-Sequence Current Differential Protection
Synchronous sampling	Necessary	Unnecessary
Real time transmission	Necessary	Unnecessary
Communication channel	High	Low
Investment	High	Low

.

4. The Example Analysis

The large-scale IGPS network is taken as an example for analysis and the Matlab is used for simulation analysis of the system, which are shown in Figure 7. Photovoltaic power source adopts the positive sequence voltage controlled current source model mentioned in literature [15]. See Appendix A for specific network parameters.

4.1. Fault Characteristic Analysis of Tie Line

When a single-phase (A-phase) metallic grounding fault occurs at point K on the tie line MN₂, IGPSs is equivalent to a large-capacity photovoltaic VCCSs, and the positive sequence augmented network with the fault tie line MN₂ is shown in Figure 8.

In the Figure 8, ${\mathit{I}}_{\mathrm{PV}i}^{+}$ and ${\mathit{U}}_{\mathrm{PV}i}^{+}$ (i = 1, 2) are positive sequence short-circuit current and terminal voltage output by photovoltaic VCCSs respectively. Z_S is the system impedance (positive and negative sequence equal). $Z_{\Delta }$ is the additional impedance. In the case of single-phase grounding short-circuit, $Z_{\Delta }$ is the series of $Z_{\sum 2}$ and $Z_{\sum 0}$, $Z_{\Delta }=Z_{\sum 2}+Z_{\sum 0}$. The zero-sequence equivalent impedance of the fault point can be obtained from Equation (1), where m = 2.

In order to clarify the change rule of resistive current and capacitive current in fault zero-sequence current, the calculation results of zero-sequence resistive current I_0R flowing through neutral resistive branch and zero-sequence capacitive current I₀₁ and I₀₂ flowing through ground capacitive branch are given, as shown in

Table 2. Zero-sequence resistance currents at different fault location.

Rg (Ω)	Zero-Sequence Resistance Current I0R.S When Point K Is at the Head (A)	Zero-Sequence Resistance Current I0R.m When Point K Is at the Midpoint (A)	Zero-Sequence Resistance Current I0R.E When Point K Is at the End (A)
10	740.74∠−5.66°	595.50∠−25.54°	477.72∠−36.82°
15	496.41∠−2.57°	443.46∠−17.68°	387.44∠−28.29°
20	372.92∠−1.00°	350.25∠−13.02°	321.75∠−22.45°
50	149.35∠1.83°	151.81∠−3.33°	152.94∠−8.25°
100	74.67∠2.77°	77.55∠0.26°	80.20∠−2.30°
150	49.78∠3.09°	52.03∠1.49°	54.22∠−0.18°
200	37.33∠3.25°	39.14∠2.11°	40.94∠0.90°

,

Table 3. Zero-sequence capacitive currents of XC1 branch at different fault location.

Rg (Ω)	Zero-Sequence Capacitive Current I01.S When Point K Is at the Head (A)	Zero-Sequence Capacitive Current I01.m When Point K Is at the Mmidpoint (A)	Zero-Sequence Capacitive Current I01.E When Point K Is at the End (A)
10	16.33∠85.48°	13.13∠65.61°	10.53∠54.32°
15	16.41∠88.20°	14.66∠73.08°	12.81∠62.48°
20	16.44∠89.57°	15.44∠77.55°	14.19∠68.12°
50	16.46∠92.06°	16.73∠86.90°	16.86∠81.98°
100	16.46∠92.89°	17.09∠90.37°	17.68∠87.82°
150	16.46∠93.16°	17.20∠91.57°	17.93∠89.89°
200	16.46∠93.30°	17.26∠92.17°	18.05∠90.96°

and

Table 4. Zero-sequence capacitive currents of XC2 branch at different fault location.

Rg (Ω)	Zero-Sequence Capacitive Current I02.S When Point K Is at the Head (A)	Zero-Sequence Capacitive Current I02.m When Point K Is at the Midpoint (A)	Zero-Sequence Capacitive Current I02.E When Point K Is at the End (A)
10	6.53∠85.49°	5.97∠80.38°	5.60∠80.14°
15	6.57∠88.20°	6.33∠83.45°	6.10∠81.54°
20	6.58∠89.57°	6.52∠85.53°	6.40∠83.17°
50	6.58∠92.06°	6.80∠90.29°	6.97∠88.65°
100	6.58∠92.89°	6.87∠92.15°	7.14∠91.35°
150	6.58∠93.16°	6.88∠92.80°	7.18∠92.35°
200	6.58∠93.30°	6.89∠93.13°	7.20∠92.87°

.

In order to show the current component more clearly, not only the amplitude of the zero-sequence current, but also the phase of the current, are given in the tables. The phase of system voltage source ES is used as the reference phase. Subscripts S, m, and E indicate that point K is located at the head, midpoint, and end of the tie line, respectively.

By comparing Table 2 with Table 4, it can be found that when R_g is small, the zero-sequence resistive current I_0R is larger than the fault point zero-sequence current I_0K flowing through fault point. I_0R is not a purely resistive current, but a current flowing through the neutral resistance circuit.

Since the resistive circuit contains the grounding variable zero-sequence impedance is mainly inductive, so when R_g is small, the inductive current component in I_0R is high compared to R_g when it is large, and the phase angle between I_0R and the capacitive current ${\mathit{I}}_{0\mathrm{C}\sum }$ (${\mathit{I}}_{0\mathrm{C}\sum }={\mathit{I}}_{01}+{\mathit{I}}_{02}$) is greater than 90 degrees, so after taking the vector sum, some of them will cancel each other out. Finally, I_0K is slightly smaller than that of I_0R. This is more obvious after the resistive circuit is added to the line impedance, because the inductive resistance is larger in the tie line.

Looking at the capacitive currents in Table 2 and Table 3, there are two overall trends in the vertical and horizontal directions.

(1)

Longitudinal trend: As the neutral point resistance R_g gradually increases, the capacitive current gradually increases.

I₀₂ is flowing through the X_C2 capacitor, because I₀₂ = U_0K/X_C2, X_C2 is unchanged, causing I₀₂ change is due to the U_0K.

According to the voltage and current relationship in the sequence network of Figure 8, the expression of the zero-sequence voltage U_0K at the fault point can be obtained as follows:

$${\mathit{U}}_{0\mathrm{K}}={\mathit{E}}_{\sum }\frac{Z_{\sum 0}}{Z_{\sum 1}+Z_{\sum 2}+Z_{\sum 0}}$$

(23)

where ${\mathit{E}}_{\sum }$ is the equivalent integrated power source, which is derived from the combined action of the system power source and the PV inverter power sources ${\mathit{I}}_{\mathrm{PV}1}^{+}$ and ${\mathit{I}}_{\mathrm{PV}2}^{+}$, defined as: ${\mathit{E}}_{\sum }=Z_{\mathrm{S}}{\mathit{I}}_{\mathrm{S}}+(Z_{\mathrm{S}}+Z_{\mathrm{T}1}){\mathit{I}}_{\mathrm{PV}1}^{+}+(Z_{\mathrm{S}}+Z_{\mathrm{T}1}+Z_{\mathrm{M}\mathrm{K}}){\mathit{I}}_{\mathrm{PV}2}^{+}$.

By the Formula (23) to analyze the change of U_0K, integrated power source ${\mathit{E}}_{\sum }$ change is very small, do not consider. When R_g gradually increases, the proportion of $\left|Z_{\sum 0}\right|$ in $\left|Z_{\sum 1}+Z_{\sum 2}+Z_{\sum 0}\right|$ gradually increases, making $\left|{\mathit{U}}_{0\mathrm{K}}\right|$ gradually increase, so the trend of $\left|{\mathit{I}}_{02}\right|$ also gradually increases with the increase of Rg.

I₀₁ is flowing through the X_C1 capacitor, there is I₀₁ = U_0M/X_C1 and U_0M = (U_0KZ_0M)/(Z_0M + Z_0MK), as R_g gradually increases, both Z_0M and $\left|{\mathit{U}}_{0\mathrm{K}}\right|$ gradually increase, making $\left|{\mathit{U}}_{0\mathrm{M}}\right|$ also gradually increase, so the trend of $\left|{\mathit{I}}_{01}\right|$ also gradually increases as R_g increases.

(2)

Cross-sectional trend: as the point K moves from M to N, both $\left|{\mathit{I}}_{01}\right|$ and $\left|{\mathit{I}}_{02}\right|$ gradually decrease when R_g is small, and both $\left|{\mathit{I}}_{01}\right|$ and $\left|{\mathit{I}}_{02}\right|$ gradually increase when R_g is large. As discussed separately below.

(a)

when R_g is small:

First taking I₀₂ of U_0K into consideration, when R_g is small, ignoring the capacitance to ground, $Z_{\sum 0}\approx$ 3R_g + 3Z_MK, and $Z_{\sum 1}=Z_{\sum 2}=Z_{\mathrm{S}}+Z_{\mathrm{T}1}+Z_{\mathrm{MK}}$, substituting into (23), have the following expressions:

$${\mathit{U}}_{0\mathrm{K}}={\mathit{E}}_{\sum }\frac{3R_{\mathrm{g}}+3Z_{\mathrm{MK}}}{2(Z_{\mathrm{S}}+Z_{\mathrm{T}1})+3R_{\mathrm{g}}+5Z_{\mathrm{MK}}}$$

(24)

When the fault K point is located at the M end, there is $Z_{\mathrm{MK}}=0$, due to $(Z_{\mathrm{S}}+Z_{\mathrm{T}1})\ll$3R_g, so ${\mathit{U}}_{0\mathrm{K}}\approx {\mathit{E}}_{\sum }$. As K moves from M to N, Z_MK gradually increases, making U_0K gradually decreases, then $\left|{\mathit{I}}_{02}\right|$ also gradually decreases.

Look again at U_0M, U_0M = (U_0KZ_0M)/(Z_0M + 3Z_MK), R_g is smaller, X_C1 can be ignored, then there is $Z_{0\mathrm{M}}\approx$ 3R_g is also smaller, as K moves from M to N, Z_MK gradually increases, U_0K gradually decreases, which inevitably leads to U_0M gradually decreases, then $\left|{\mathit{I}}_{01}\right|$ also gradually decreases.

• (b)

  When R_g is large:

Still first look at U_0K, R_g larger, must take into account the capacitance to ground, when $Z_{\sum 0}$ is resistive characteristics. Where the resistive component by R_g decision, the capacitive component by X_C decision. In R_g take a certain value, $Z_{\sum 0}$ in both components remain unchanged. As the fault point K moves from M to N, the inductive line impedance Z_MK gradually increases, which will offset part of the capacitive component of $Z_{\sum 0}$. Because the denominator of Equation (23) $Z_{\sum 1}+Z_{\sum 2}$ also contains Z_MK, which can obviously offset more capacitive components, so $\left|Z_{\sum 0}\right|/\left|Z_{\sum 1}+Z_{\sum 2}+Z_{\sum 0}\right|$ will gradually increase with the increase of Z_MK, making U_0K gradually increase, $\left|{\mathit{I}}_{02}\right|$ also gradually increase.

Looking at U_0M, U_0M = (U_0KZ_0M)/(Z_0M +3Z_MK), Z_0M is resistive-capacitance characteristic, inductive impedance Z_MK increase will offset part of the capacitive resistance component in Z_0M, which leads to the denominator Z_0M +3Z_MK decrease, accompanied by U_0K increase, so U_0M must gradually increase with Z_MK increase, then $\left|{\mathit{I}}_{01}\right|$ also gradually increase.

In order to facilitate the subsequent protection calibration and verification, the zero-sequence current amplitude flowing through the M side of the system bus is given $\left|{\mathit{I}}_{0\mathrm{M}}\right|$, as shown in

Table 5. Zero-sequence currents at M side of tie line with different fault locations.

Rg (Ω)	Zero-Sequence Current $\left|{\mathit{I}}_{0\mathbf{M}.\mathbf{S}}\right| \mathbf{When} \mathbf{Point} \mathbf{K}$ Is at the Head (A)	Zero-Sequence Current $\left|{\mathit{I}}_{0\mathbf{M}.\mathbf{m}}\right| \mathbf{When} \mathbf{Point} \mathbf{K}$ Is at the Midpoint (A)	Zero-Sequence Current $\left|{\mathit{I}}_{0\mathbf{M}.\mathbf{E}}\right| \mathbf{When} \mathbf{Point} \mathbf{K}$ Is at the End (A)
10	740.59	595.37	477.62
15	496.46	443.50	387.48
20	373.11	350.43	321.91
30	249.29	245.19	237.35
50	150.19	152.65	153.80
100	76.43	79.37	82.09
150	52.41	54.78	57.09
200	40.78	42.76	44.72

.

4.2. Fault Characteristic Analysis of Outside Tie Line

4.2.1. The Fault of Adjacent Outside Collector Line

When a single-phase (A-phase) grounding fault occurs on collector line 1-5 in IGPS1, the positive-sequence of fault augmented network can be obtained, as shown in Figure 9.

In Figure 9, $Z_{1\mathrm{d}}=Z_{\mathrm{Cf}2}+Z_{\mathrm{d}}$ indicates the positive-sequence impedance between the downstream PV power source of the fault point and the fault point. The zero-sequence equivalent impedance ${\displaystyle Z_{\sum 0}^{\prime }}$ of the short-circuit point in Figure 9 can be obtained from Equation (13), where m = 2.

As special attention is paid to the calculation and verification of the protection setting for of the tie line MN2, only the zero-sequence current flowing on the tie line MN2 is given below. According to the above analysis, when the fault occurs on the collector line 1-5, the amplitude of the zero-sequence current $\left|{\mathit{I}}_{02}\right|$ flowing through the whole the tie line MN2 is shown in

Table 6. Zero-sequence currents of X_C2 branch at different fault location on line 1-5.

Rg (Ω)	Zero-Sequence Current $\left|{\mathit{I}}_{02.\mathbf{S}}\right| \mathbf{When} \mathbf{Point} {\mathbf{K}}^{″}$ Is at the Head (A)	Zero-Sequence Current $\left|{\mathit{I}}_{02.\mathbf{m}}\right|\mathbf{When} \mathbf{Point} {\mathbf{K}}^{″}$ Is at the Midpoint (A)	Zero-Sequence Current $\left|{\mathit{I}}_{02.\mathbf{E}}\right| \mathbf{When} \mathbf{Point} {\mathbf{K}}^{″}$ Is at the End (A)
10	6.44	6.03	5.66
15	6.52	6.24	5.98
20	6.55	6.34	6.14
30	6.58	6.44	6.31
50	6.60	6.52	6.45
100	6.61	6.58	6.55
150	6.62	6.60	6.58
200	6.62	6.61	6.60

.

4.2.2. The Fault of Subordinate Outside Collector Line

When a single-phase (A-phase) grounding fault occurs on subordinate collector line 2-2 in IGPS2, the positive-sequence of fault augmented network can be obtained, as shown in Figure 10.

In Figure 10, $Z_{2\mathrm{d}}=Z_{\mathrm{Cf}2}+Z_{\mathrm{d}}$ indicates the positive-sequence impedance between the downstream PV power source of the fault point and the fault point. The zero-sequence equivalent impedance ${Z^{\prime }}_{\sum 0}$ can be obtained from Equation (16), where m = 2.

According to the above analysis, when the fault occurs on the collector line 2-2, the amplitude of the zero-sequence current flowing through the whole the tie line MN₂ is the vector sum of resistance current I_0R and the zero-sequence capacitive current I₀₁ to ground of non-fault IGPS1, $\left|{\mathit{I}}_{0\mathrm{M}}\right|=\left|{\mathit{I}}_{0\mathrm{R}}+{\mathit{I}}_{01}\right|$.When the fault occurs on the different location of collector line 2-2, the amplitude of the zero-sequence current $\left|{\mathit{I}}_{0\mathrm{M}}\right|$ flowing through the whole the tie line MN2 is shown in

Table 7. Zero-sequence currents of MN2 at different fault location on line 2-2.

Rg (Ω)	Zero-Sequence Current $\left|{\mathit{I}}_{0\mathbf{M}.\mathbf{S}}\right|\mathbf{When} \mathbf{Point} {\mathbf{K}}^{\prime }$ Is at the Head (A)	Zero-Sequence Current $\left|{\mathit{I}}_{0\mathbf{M}.\mathbf{m}}\right|\mathbf{When} \mathbf{Point} {\mathbf{K}}^{\prime }$ Is at the Midpoint (A)	Zero-Sequence Current $\left|{\mathit{I}}_{0\mathbf{M}.\mathbf{E}}\right|\mathbf{When} \mathbf{Point} {\mathbf{K}}^{\prime }$ Is at the Rnd (A)
10	475.22	472.59	469.32
15	385.40	384.33	382.27
20	321.22	319.90	318.57
30	237.04	236.44	235.84
50	153.72	153.57	153.41
100	82.10	82.12	82.14
150	57.10	57.14	57.18
200	44.74	44.78	44.82

.

4.3. Blocking Zero-Sequence Pilot Protection Setting and Verification of Tie Line

Based on the short-circuit calculation results in Section 3.1and Section 3.2, this section carries out specific setting calculation and action behavior verification for the proposed protection scheme. By substituting the capacitive current in Table 3 and the resistance current in Table 1 into the setting calculation Equations (21) and (22), respectively, the zero-sequence pilot protection starting values I₁ and I₂ under different values of neutral point R_g can be obtained. It should be noted that since the setting principle is to ensure that I₁ and I₂ have sufficient sensitivity to start, the minimum value of short-circuit current in different fault locations should be taken for setting calculation.

1.

When R_g = 20 Ω and the fault of the end tie line MN2 occurs, the resistance current $\left|{\mathit{I}}_{0\mathrm{R}.\mathrm{E}}\right|$ and capacitive current $\left|{\mathit{I}}_{02.\mathrm{E}}\right|$ are taken to calculate the protection fixed value as follows:

$$I_1=3\left|{\mathit{I}}_{02.\mathrm{E}}\right|/{\mathrm{K}}_{\mathrm{sen}}=3\times 6.4/1.5=12.8\hspace{0.17em}\mathrm{A}$$

(25)

$$I_2=3\left|{\mathit{I}}_{0\mathrm{R}.\mathrm{E}}\right|/{\mathrm{K}}_{\mathrm{sen}}=3\times 321.75/1.5=643.5\hspace{0.17em}\mathrm{A}$$

(26)

When single-phase ground fault occurs on three different lines in the collection system (R_g = 20 Ω), the zero-sequence current flowing through the collection bus M and N₂ is shown in Figure 11.

Figure 11a–c shows the zero-sequence current before and after fault on tie line MN2, collector line 1-5 and collector line 2-2, respectively. In order to observe and determine the action of starting element under different fault conditions, the threshold values of high constant value starting element I₁ and low constant value starting element I₂ are also added in the current diagram.

Table 8. The action results of protection element for an internal fault on tie line MN₂ (R_g = 20 Ω).

Current Location	Minimum Current Amplitude (A)	Start Setting of I1 (A)	Start Setting of I2 (A)	Element Action Result
PM	965.73	12.80	643.50	I1, “1” I2, “1”
PN	19.20	12.80	643.50	I1, “1” I2, “0”

,

Table 9. The action results of protection element for an external fault on collector line 1-5 (R_g = 20 Ω).

Current Location	Minimum Current Amplitude (A)	Start Setting of I1 (A)	Start Setting of I2 (A)	Element Action Result
PM	18.42	12.80	643.50	I1, “1” I2, “0”
PN	18.42	12.80	643.50	I1, “1” I2, “0”

and

Table 10. The action results of protection element for an external fault on collector line 2-2 (R_g = 20 Ω).

Current Location	Minimum Current Amplitude (A)	Start Setting of I1 (A)	Start Setting of I2 (A)	Element Action Result
PM	955.71	12.80	643.50	I1, “1” I2, “1”
PN	955.71	12.80	643.50	I1, “1” I2, “1”

, respectively, show the judgment results of the starting elements of P_M and P_N protection on both sides of the tie line when there is a fault on the tie line MN₂ in the protection area, the adjacent collector 1-5 and the lower collector 2-2 outside the protection area, where “1” indicates that the element starts, and “0” indicates that the element does not start.

2.

When R_g = 150 Ω and the fault of the head of tie line MN₂ occurs, the resistance current $\left|{\mathit{I}}_{0\mathrm{R}.\mathrm{S}}\right|$ and capacitive current $\left|{\mathit{I}}_{02.\mathrm{S}}\right|$ are taken to calculate the protection fixed value as follows:

$$I_1=3\left|{\mathit{I}}_{02.\mathrm{S}}\right|/{\mathrm{K}}_{\mathrm{sen}}=3\times 6.58/1.5=13.16\hspace{0.17em}\mathrm{A}$$

(27)

$$I_2=3\left|{\mathit{I}}_{0\mathrm{R}.\mathrm{S}}\right|/{\mathrm{K}}_{\mathrm{sen}}=3\times 49.78/1.5=99.56\hspace{0.17em}\mathrm{A}$$

(28)

When single-phase ground fault occurs on three different lines in the collection system (R_g = 150 Ω), the zero-sequence current flowing through the collection bus M and N₂ is shown in Figure 12.

Table 11. The action results of protection element for an internal fault on tie line MN₂ (R_g = 150 Ω).

Current Location	Minimum Current Amplitude (A)	Start Setting of I1 (A)	Start Setting of I2 (A)	Element Action Result
PM	157.23	13.16	99.56	I1, “1” I2, “1”
PN	19.74	13.16	99.56	I1, “1” I2, “0”

,

Table 12. The action results of protection element for an external fault on collector line 1-5 (R_g = 150 Ω).

Current Location	Minimum Current Amplitude (A)	Start Setting of I1 (A)	Start Setting of I2 (A)	Element Action Result
PM	19.74	13.16	99.56	I1, “1” I2, “0”
PN	19.74	13.16	99.56	I1, “1” I2, “0”

and

Table 13. The action results of protection element for an external fault on collector line 2-2 (R_g = 150 Ω).

Current Location	Minimum Current Amplitude (A)	Start Setting of I1 (A)	Start Setting of I2 (A)	Element Action Result
PM	171.30	13.16	99.56	I1, “1” I2, “1”
PN	171.30	13.16	99.56	I1, “1” I2, “1”

, respectively, show the judgment results of the starting elements of P_M and P_N protection on both sides of the tie line when there is a fault on the tie line MN₂ in the protection area, the adjacent collector 1-5 and the lower collector 2-2 outside the protection area.

The action results of the protection elements, in Table 8, Table 9, Table 10, Table 11, Table 12 and Table 13, are in complete agreement with the design of the protection scheme in the previous section, indicating that the proposed blocking zero-sequence pilot protection can reliably judge the in-zone and out-zone faults when R_g = 20 Ω and R_g = 150 Ω, and has selectivity and quickness.

5. Conclusions

In this paper, a fault analysis model of large-scale centralized IGPS collection system with ground capacitance is established. Based on the fault characteristic analysis, a blocking zero-sequence pilot protection scheme with economic characteristics is proposed. Finally, the effectiveness of the proposed protection scheme is verified by an example analysis. The main conclusions are as follows:

• We established a fault analysis model of large-scale centralized IGPS collection system with ground capacitance. By analyzing the characteristics of single-phase grounding fault on different lines, we proposed a blocking zero-sequence pilot protection scheme with economic characteristics. The protection scheme only uses the current to constitute the protection criterion without the measurement of voltage, and the protection on both sides does not need synchronous sampling data, only needs to transmit logic information, which can save the investment cost of engineering construction.
• The influence of neutral resistance on different components of zero-sequence current is analyzed. When the neutral resistance is small, the zero-sequence equivalent impedance of short-circuit point is mainly resistive, and the zero-sequence current is mainly resistive and inductive, and the influence of ground capacitance on fault current is small. When the resistance of the neutral point is large, the zero-sequence equivalent impedance of the short-circuit point is mainly characterized by resistance and capacitance, and the zero-sequence current is mainly characterized by resistance and capacitance.
• Through the quantitative protection setting calculation and verification process, it is verified that the proposed protection scheme can reliably judge the in-zone and out-zone faults, and can give consideration to the protection selectivity and rapidity when the value of neutral point resistance is low or high.