1 Introduction

Transmission lines are occassionally tapped to provide intermediate connections to loads or reinforce the underlying lower voltage network through a transformer. Such a configuration is known as multi-terminal transmission lines. For strengthening the power system, MTLs are frequently designed as a temporary and inexpensive measure. However, they can cause problems in the protective system [1].

As a part of a continuous endeavor to eliminate the problems caused by MTLs and enhance the reliability of the protective system, many protection methodologies have been developed. A few of them are discussed here. Abe et al. [2] developed asynchronous measurements based protection methodology for fault location in MTLs. The MTLs have been transformed into two terminal lines to achieve fault location accurately. Nagasawa et al. [3] have proposed an algorithm for protection of parallel MTLs using asynchronous differential currents at each terminal. Though the algorithms proposed by the authors [2, 3] performed well, their accuracy may be affected by unbalance in the line parameters when different fault conditions occur. Funbashi et al. [4] have proposed methods to identify the fault point in double circuit MTL using measurements from capacitor voltage transformer (CCVT) and current transformer (CT). However, for accurate fault location, measurements are required from all the terminals. In [5], Qiu et al. have developed a multi-agent algorithm for protection of MTLs. It consists of organization agent, coordination agent and executive agent. They exchange the information among themselves regarding trip information. However, lack of global synchronous measurements acquired from different agents may lead to mal-operation of the relay. Gajic et al. [6] have proposed differential protection with innovative charging current compensation algorithm for MTL protection. However, the reliability of the algorithm depends on the availability of the current channels.

Forford et al. [7] have designed differential current algorithm for protection of MTLs. The proposed algorithm can differentiate internal fault, external fault and normal load conditions using electric mid-point (EMP). Arbes [8] has developed differential line protection scheme for the protection of double lines, tapped lines and short lines. However, the performance of the proposed scheme [8] depends on the local voltage and current measurements. Al-Fakhri [9] has proposed differential protection methodology against internal and external faults using asynchronous measurements. Hussain et al. [10] have proposed a fault location scheme for MTLs using positive sequence voltage and current measurements. However, the synchronization process may be affected due to metering errors. For reliable operation of proposed methods [9, 10], the precise time synchronization of analog information between the line ends must be required for the differential calculation to be accurate.

In addition to voltage and current based methodologies [2,3,4,5,6,7,8,9,10], traveling wave-based protection schemes have been proposed for MTL protection [11, 12]. Authors of [11] have used single traveling wave and fundamental measurements for fault location in MTLs. However, the performance of the methodology will be affected by arcing faults and variation in fault impedance. In [12], Zhu et al. developed a current traveling wave based algorithm. Fault detection and location functions are accomplished using arrival time of current waves at a terminal.

Technological developments in measurements, communication, control and monitoring of power grids have brought a paradigm shift in the protection philosophy of transmission lines. The reliability of power system has been enhanced by early detection of wide-area disturbances and optimal utilization of assets. Some of the protection methodologies based on Synchrophasor measurements are discussed here. Lin et al. [13] demonstrated the performance of PMU based fault location algorithm on MTLs. Faults are identified and located using synchronized positive sequence voltage and current phasors. Further, for accurate fault detection and location in MTLs, Brahma [14] has employed time-stamped voltage and current phasors obtained from all the terminals. Ting Wu et al. [15] have formed a novel fault location technique for multi-section non-homogeneous transmission lines. However, the accuracy of [14, 15] may be lost in case of medium and long MTLs. For decades, the distance protection is widely employed for the protection of transmission lines as it is simple and fast. The distance protection can protect most of the protected line, and it is virtually independent of the source impedance. However, the performance and the reliability of the distance protection are influenced by infeed and outfeed currents in MTLs [16,17,18].

In this paper, PMU based adaptive zone settings of distance relays (PAZSD) methodology has been proposed to improve the performance and the reliability of distance protection by adjusting zone settings adaptively. The PAZSD methodology employs current coefficients to adjust the zone settings of the distance relays during infeed situations. These coefficients are calculated in phasor data concentrator (PDC) at system protection center (SPC) using the magnitude of current phasors obtained from PMUs. The functioning of the distance relays during infeed condition with and without the proposed methodology has been demonstrated through different fault case studies carried out on a four-bus model in PSCAD/EMTDC environment. Further, a laboratory prototype of EHV MTLs is considered to validate the performance of distance relays during infeed condition. For phasor estimation, PMUs are realized in real-time using NI cRIO-9063 chassis embedded with data acquisition sensors (NI-9225 & NI-9227) and Global Positioning System (GPS) synchronisation module (NI-9467) in conjunction with LabVIEW software. The results indicate that the proposed PAZSD methodology can improve the performance and the reliability of conventional distance protection during infeed situations under different fault conditions.

2 Synchrophasor Technology

The cutting-edge Synchrophasor technology entails estimation of time-stamped phasor measurements on GPS time reference. It has been used to provide accurate information regarding the state of the power system for implementing immediate corrective actions. The process of phasor estimation starts with a sampling of an analog signal (x(t)) at a sampling frequency fs (= Nfo). With this sampling frequency, N number of samples per cycle are obtained. The time-stamped fundamental phasors of three-phase voltage and current signals per cycle are estimated using the Discrete Fourier Transform (DFT) [19]. In general, the kth estimation of the original signal is given by Eq. 1.

$$ {X}_k=\frac{\sqrt{2}}{N}\sum \limits_{n=0}^{N-1}x\left(n\Delta T\right){e}^{-j\frac{2\pi kn}{N}} $$
(1)

where x(n∆T) is sampled version of x(t) (voltage or current analog signals),

∆T is sampling time in seconds,

fo is nominal frequency (Hz),

T is time period in seconds,

N is number of samples per cycle,

n is sample number starting from n = 0 to N-1,

For k = 1, X k gives the fundamental frequency phasor.

A concise description of the infeed problem encountered by distance protection in MTLs and proposed solution (PAZSD methodology) under different fault conditions are discussed in the following section.

3 Methods

Fig. 1 shows a part of an interconnected power system for illustrating the effect of infeed on the performance of distance relays. Let the distance relays protecting the lines i-l and l-j are R il & R li , and R lj & R jl respectively. Likewise, the distance relays of the teed terminal l-k are R lk and R kl . Assume PMUs are installed at all the buses which communicate to the PDC at SPC through a modem and fiber optical cables. Assuming the currents Ip and Iq are in phase. For illustration purpose, the PMUs at Bus i and k are considered.

Fig. 1
figure 1

Proposed PAZSD methodology for multi-terminal transmission lines (MTL)

3.1 Infeed effect

In order to explain the infeed effect, only relays are assumed to be present in the above power system network (No PMUs, PDC and SPC are present). Assume that a fault has occurred on the line l-j at a point D as shown in Fig. 1. The resultant currents are indicated in Fig. 1. Under such conditions, the impedance observed by the relay R il at Bus i is obtained using KVL:

$$ {V}_i={I}_p\left({Z}_{il}+{Z}_{lD}\right)+{I}_q{Z}_{lD} $$
(2)
$$ \frac{V_i}{I_p}=\left({Z}_{il}+{Z}_{lD}\right)+\frac{I_q}{I_p}{Z}_{lD} $$
(3)
$$ \mathrm{Let}\kern1.25em {Z}_{il}+{Z}_{lD}={Z}_{iD} $$
$$ \frac{V_i}{I_p}={Z}_{iD}+\kern0.5em \frac{I_q}{I_p}{Z}_{lD} $$
(4)

where.

V i is voltage at a Bus i,

Z il is the impedance of transmission line i-l,

Z lD is impedance of transmission line l-j from Bus l to the fault point D,

I p is current flowing from Bus i to l,

I q is current flowing from Bus k to l,

From Eq. (4), it is observed that the impedance seen by the relay R il is more than the impedance observed (Z iD ) when there is infeed. Therefore, the relay R il under reaches during fault condition. The main cause for such phenomenon is that the relay R il cannot sense the current (I q ) flowing from Bus k to l. The amount of under reach depends on the magnitude of current I q .

To address the above infeed issue and to ensure reliable operation of the relay R il , the subsequent section proposes the PAZSD methodology. This proposed methodology guides the relay to change the zone settings according to the infeed conditions using Synchrophasor technology. The PAZSD methodology which is executed in the PDC sends the new zone settings to the corresponding relay to ensure reliable operation during the infeed condition.

3.2 Flowchart of proposed PAZSD methodology

The sequence of execution of the proposed PAZSD methodology, as shown in Fig. 1, is illustrated in step by step manner to eliminate the infeed problem as discussed in the previous section.

Step 1: Synchronized time-stamped voltage and current phasor data are estimated in the PMU at Bus i and k, and transmitted to PDC at SPC.

Step 2: In PDC, three-phase current coefficients (K1, K2 and K3) are calculated from the magnitudes of the current phasors using Eqs. (5) to (7)

$$ {K}_1=\frac{\mid {I}_{Ri}\mid +\mid {I}_{Rk}\mid }{\mid {I}_{Ri}\mid } $$
(5)
$$ {K}_2=\frac{\mid {I}_{Yi}\mid +\mid {I}_{Yk}\mid }{\mid {I}_{Yi}\mid } $$
(6)
$$ {K}_3=\frac{\mid {I}_{Bi}\mid +\mid {I}_{Bk}\mid }{\mid {I}_{Bi}\mid } $$
(7)

Step 3: If K1 ≈ K2 ≈ K3, adjust the reach settings of the relay R il using Eq. (8).

$$ {Z}_{\mathrm{set}\hbox{-} \mathrm{new}}={K_1}^{\ast }\ {Z}_{\mathrm{set}\hbox{-} \mathrm{old}} $$
(8)

Else if K1> (K2 & K3), adjust the reach settings of the relay R il using Eq. (9).

$$ {Z}_{\mathrm{set}\hbox{-} \mathrm{new}}={K_1}^{\ast }\ {Z}_{\mathrm{set}\hbox{-} \mathrm{old}} $$
(9)

Else if K2> (K1 & K3), adjust the reach settings of relay R il as given in Eq. (10).

$$ {Z}_{\mathrm{set}\hbox{-} \mathrm{new}}={K_2}^{\ast }\ {Z}_{\mathrm{set}\hbox{-} \mathrm{old}} $$
(10)

Else adjust the reach settings of the relay R il as given in Eq. (11).

$$ {Z}_{\mathrm{set}\hbox{-} \mathrm{new}}={K_3}^{\ast }\ {Z}_{\mathrm{set}\hbox{-} \mathrm{old}} $$
(11)

where.

K1K2K3 are current cofficients for infeed condition,

I Ri , I Yi & I Bi are three-phase current phasors flowing from Bus i to l,

I Rk , I Yk & I Bk are three-phase current phasors flowing from Bus k to l,

Zset-old are old three zone reach settings of the relay R il ,

Zset-new are new three-zone reach settings of the relay R il with infeed line (between Bus k and l).

The following section describes the implementation of the proposed PAZSD methodology on a four-bus system to eliminate the infeed problems as discussed in section 3.1.

3.3 Case studies

A four-bus model shown in Fig. 2 is considered and implemented in PSCAD/EMTDC software. Various case studies (Case 1, Case 2 and Case 3) are conducted to illustrate the functioning of distance relays for infeed condition. The base MVA and kV of the system are 100 and 400 (line to line) respectively. The positive sequence resistance, inductive and capacitive reactance of the transmission line are 0.0234 Ω/km, 0.298 Ω/km, and 256.7 kΩ*km respectively. The values of negative sequence parameters are same as that of the positive sequence parameters. Similarly, the values of zero sequence resistance, inductive and capacitive reactance of the transmission lines are 0.388 Ω/km, 1.02 Ω/km, and 376.6 kΩ*km respectively. The length of each transmission line is 350 km. The zone settings of the distance relays are given in Table 1. Assume PMUs are installed at all buses.

Fig. 2
figure 2

A four-bus model implemented in PSCAD/EMTDC software to illustrate the functioning of the distance relays during infeed condition

Table 1 Zone settings of the distance relays

3.3.1 Performance of distance relays without PAZSD methodology during infeed condition

Three case studies (Case 1, Case 2 and Case 3), as shown in Fig. 2, are considered to illustrate the performance of the distance relay R12 without PAZSD methodology. Assuming that no PMU, PDC and SPC technology are present in the case studies.

Case 1

Assume that a triple line fault (RYB) occurred at a distance of 10 km from Bus-1. In other words, in Zone-1 of the relay R12 and Zone-2 of the relay R21. For such condition, the impedance trajectory of the distance relays, R12, R21, and R23, is portrayed in Fig. 3.

Fig. 3
figure 3

Performance of (a) R12, (b) R21 & (c) R23 for LLL (RYB) fault at 10 km from Bus-1 (Zone-1 of R12 and Zone-2 of R21)

From Fig. 3, the relays R12 and R21 have observed the impedance in Zone-1 and Zone-2 respectively. Whereas, the relay R23 has not observed the impedance in any of its zones. Hence, it is clear that none of the relays are affected by the infeed condition and the respective relays have correctly detected a fault condition.

Case 2

A double line to ground fault (RYG) is created at a distance of 500 km from Bus-1 (i.e., 150 km from Bus-2). This indicates the relays R12 and R23 should detect the fault in Zone-2 and Zone-1 respectively. The corresponding impedance trajectory of the relays R12, R21 and R23 is shown in Fig. 4. From Fig. 4, it is observed that the relay R12 has observed the trajectory in Zone-3 whereas the relay R23 has observed in Zone-1. The relay R21 has not observed the trajectory in any of its zone due to its inherent directional property. Therefore, it is understood that the infeed at Bus-2 has caused the relay R12 to mal-operate.

Fig. 4
figure 4

Performance of (a) R12, (b) R21 & (c) R23 for LLG (RYG) fault at 500 km from Bus-1 (i.e. 150 km from Bus-2)

Case 3

A double line fault (RY) is created at a distance of 200 km from Bus-2 which lies in Zone-3 of the relay R12 and Zone-1 of the relay R23. For this event, impedance trajectory observed by the relays, R12, R21 and R23 are shown in Fig. 5. From figure, it is concluded that the relays R12 and R21 have not seen the impedance in any of their zones. Whereas, the relay R23 has seen the impedance in Zone-1. Therefore, it is clearly known that the infeed at Bus-2 has influenced the relay R12 to mal-operate.

Fig. 5
figure 5

Performance of (a) R12, (b) R21 & (c) R23 for LL (RY) fault at 550 km from Bus-1 (i.e. 200 km from Bus-2)

From the above three case studies, it is clear that the performance of the relay is affected by the infeed at Bus-2 when a fault occurs on the line 2–3.

3.3.2 Performance of distance relays with proposed PAZSD methodology during infeed condition

The case studies discussed in the previous subsection are reconsidered with the implementation of the proposed PAZSD methodology using PMUs and PDC. The same four-bus system is considered, and the proposed PAZSD methodology is implemented in PDC at SPC with the data acquired from each PMU. Once the current coefficients (K1, K2, and K3) are estimated in PDC, the new zone settings are calculated and communicated back to the corresponding relay. The following case studies prove the advantages of the proposed methodology to eliminate the infeed problem discussed in the previous section.

Case 1

A triple line fault (RYB) is created with the same fault conditions as discussed in section 3.3.1.1. The relay zone settings (Zset-old) are shown in Table 2. The current coefficients (K1, K2, and K3) and new zone settings estimated for the above fault condition are tabulated in Table 2 (according to the methodology proposed in section 3.2). These new zone settings are updated in the respective relays, and the relay operate as per the new settings. The impedance trajectory of the distance relays, R12, R21, and R23 is portrayed in Fig. 6. From Fig. 6, the relays R12 and R21 have observed the impedance in Zone-1 and Zone-2 respectively. Whereas, the relay R23 has not observed the impedance in any of its zones. Hence, it is clear that none of the relays are affected by the infeed condition and the respective relays have properly detected the fault condition with new zone settings.

Table 2 Zone settings of the distance relays, R12, R21 & R23 using PAZSD methodology
Fig. 6
figure 6

Performance of (a) R12, (b) R21 & (c) R23 for LLL (RYB) fault at 10 km from Bus-1 (Zone-1 of R12 and Zone-2 of R21) with the proposed methodology

Case 2

A double line to ground fault (RYG) is considered with same fault conditions as discussed in section 3.3.1.2. The estimated current coefficients (K1, K2, and K3) and new zone settings and tabulated in Table 2. The values of current coefficients K1, K2 and K3 are 1.82, 1.79 and 1.9 respectively. Since K3 > (K1 & K2), as per the proposed methodology the new zone settings of the relay R12 (Relayset-new = K3*Relayset-old) are 12.449 + j158.536, 23.342+ j297.255 and 32.678+ j416.157. The impedance trajectory of relays R12, R21 and R23 are shown in Fig. 7, and it is clear that the relay R12 has detected the fault in Zone-2. From these case studies (3.3.1.2 & 3.3.2.2) it is observed that because of implementation of the proposed methodology, the relay R12 could detect the fault condition in Zone-2 (instead of Zone-3), which averts the mal-operation of the relay.

Fig. 7
figure 7

Performance of (a) R12, (b) R21 & (c) R23 for LLG (RYG) fault at 500 km from Bus-1 (i.e. 150 km from Bus-2) with the proposed methodology

Case 3

A double line fault (RY) is considered with same fault conditions as discussed in section 3.3.1.3. The current coefficients (K1, K2, and K3) and new zone settings estimated for the above fault condition are tabulated in Table 2. The values of the current coefficients K1, K2 and K3 are 1.8, 1.78 and 1.89 respectively. Since K3 > (K1 & K2), as per the proposed methodology the zone settings of the relay R12 (Relayset-new = K3*Relayset-old) are 12.383 + j157.702, 23.219 + j295.691 and 32.506 + j413.967. The impedance trajectory is shown in Fig. 8 and it is clear that the relay R12 has detected the fault in Zone-3. From these case studies (3.3.1.3 & 3.3.2.3) it is observed that because of implementation of the proposed methodology the relay R12 could detect the fault condition properly in Zone-3, which averts the mal-operation of the relay.

Fig. 8
figure 8

Performance of (a) R12, (b) R21 & (c) R23 for LL (RY) fault at 550 km from Bus-1 (i.e. 200 km from Bus-2) with the proposed methodology

From the above case studies, it is understood that the performance of the relay R12 is satisfactory with new zone settings when a fault occurs on the line 1–2. However, the performance of the relay R12 has been corrected by the proposed PAZSD methodology when a fault occurs on the line 2–3. Therefore, the performance of distance relay (R12) has been enhanced during the infeed condition with the help of the proposed methodology.

In the subsequent section, the efficacy of the proposed PAZSD methodology in enhancing the performance and reliability of the distance relay is validated in real-time on a laboratory prototype model of EHV MTL.

4 Results and discussion

A scale down laboratory prototype model of EHV MTL is shown in Fig. 9. As shown in figure, two three-phase 440 V, 50 Hz power supplies are connected to Bus B1 and B4 through autotransformers. The autotransformer steps down the supply voltage from 440 V to 110 V at 50 Hz. A three-phase variable load of 3.75 kW is connected at the receiving end (Bus B3). PMUs are connected at all buses.

Fig. 9
figure 9

Single line diagram of the scale down laboratory model of EHV MTL system for infeed condition

The length of the transmission lines 1–2 and 2–3 is 200 km each with the Π-model transmission line. Each 200 km transmission line is divided into four 50 km Π-sections connected in series. The parameters of the transmission line per 50 km are considered with resistance 1.8 Ω, inductance 10.07 mH and capacitance 2.2 μF. PMUs are implemented using NI cRIO-9063 chassis embedded with NI-9225 Voltage, NI-9227 Current and NI-9476 GPS modules programmed in LabVIEW FPGA software. As shown in Fig. 9, the three-phase voltage and current phasors are acquired from PMUs and communicated to the PDC. The sampling frequency of NI cRIO-9063 considered for phasor estimation is 2 kHz. To evaluate the performance of distance relay (RA), numerous faults with different fault impedances (0.2 Ω, 1.7 Ω, and 4.9 Ω) are simulated and discussed in the following section.

4.1 Real-time performance analysis of the conventional distance relay without PAZSD methodology

The zone settings of the relay RA are calculated for 400 km and tabulated in Table 3.

Table 3 Zone Settings of the distance relay RA

The following conditions are used to detect the zone of fault point.

$$ \mathrm{Zone}\hbox{-} 1:\kern1em \mathrm{If}\mid {\mathrm{Z}}_{\mathrm{Calculate}\hbox{-} 1}\mid <11.647 $$
(1)

where |ZCalculate-1| is the distance from the center of Zone-1 circle to the fault point.

$$ \mathrm{Zone}\hbox{-} 2:\kern1em \mathrm{If}\ 11.647<\mid {\mathrm{Z}}_{\mathrm{Calculated}\hbox{-} 2}\mid <21.8388 $$
(2)

where |ZCalculate-2| is the distance from the center of Zone-2 circle to the fault point.

The performance of the distance relay RA without the proposed methodology is evaluated for various faults with different fault impedances (0.2 Ω, 1.7 Ω and 4.9 Ω) and tabulated in Tables 4, 5 and 6.

Table 4 Performance of the distance relay RA without PAZSD methodology for different faults with FR = 0.2 Ω at different distances
Table 5 Performance of distance relay RA without PAZSD methodology for different faults with FR = 1.7 Ω at different distances
Table 6 Performance of distance relay RA without PAZSD methodology for different faults with FR = 4.9 Ω at different distances

From Table 4, for example, consider an LG fault occurred at 50 km from Bus B1. The corresponding impedance observed by the relay RA is 3.9612∠65.2140. The relay RA has seen the impedance in Zone-1 since the magnitude of the impedance is less than 11.647 (Condition 1). Therefore, the relay operates as per its settings. Figure 10 displays the LabVIEW front panel for relay RA in PDC at SPC (without the proposed methodology). For the case study, LabVIEW front panel displays a glowing LED for Zone-1 fault. Figure 10 also shows the voltage and current phasor data acquired from PMUs at B1 and B4, and the calculated impedance. A similar explanation holds good for LL fault at 50 km, LLG & LLL faults at 150 km from Bus B1 as given in Table 4.

Fig. 10
figure 10

LabVIEW front panel display of PDC at SPC showing the performance of distance relay RA without the proposed PAZSD methodology for LG fault (FR = 0.2 Ω) at 50 km from B1

Further, consider double line fault (LL) at 250 km from Bus B1 as given in Table 4. The impedance observed by the relay RA is 32.092∠84.440. The relay RA has seen the impedance in Zone-2 since the magnitude of the calculated impedance (|ZCalculated-2|) is less than 21.8388 (Condition 2). Therefore, the relay operates in Zone-2 rather than in Zone-1. Thus, the infeed at Bus B2 has caused the relay RA to mal-operate. A similar explanation holds good for LG, LLG & LLL faults at 250 km from Bus B1 as given in Table 4.

Furthermore, consider the double line to ground fault (LLG) at 300 km from Bus B1 as given in Table 4. The impedance observed by the relay RA is 48.136∠80.370. The relay RA has seen the impedance neither in Zone-1 nor in Zone-2 since the magnitude of the calculated impedance (|Z Calculated|) is higher than 21.8388 (Condition 2). Therefore, the relay does not operate since the observed impedance has fallen out of its zone settings. Thus, the infeed at Bus B2 has caused the relay RA to mal-operate. A similar explanation holds good for LG, LL & LLL faults at 350 km from Bus B1 as given in Table 4. The fault conditions for all the cases are shown in Table 4 considering a fault resistance of 0.2 Ω at the fault point.

Tables 5 and 6 show similar case studies as discussed in Table 4 but with different FR of 1.7 Ω and 4.9 Ω respectively. From tables, it is clear that with the change in FR the relay RA malfunctions for many cases because of infeed condition at Bus B2. Few cases are discussed below.

Consider double line fault (LL) at 250 km from Bus B1 as given in Table 5. The impedance observed by the relay RA is 32.092∠84.440. The relay RA has seen the impedance in Zone-2 since the magnitude of the calculated impedance (|Z Calculated-2|) is less than 21.8388 (Condition 2). Therefore, the relay operates in Zone-2 rather than in Zone-1. Thus, the infeed at Bus B2 has caused the relay RA to mal-operate. Figure 11 displays the LabVIEW front panel for relay RA in PDC at SPC (without the proposed methodology). The LabVIEW front panel displays glowing LED for Zone-2 fault. Figure 11 also shows the voltage and current phasor data acquired from PMUs at B1 and B4 and the calculated impedance.

Fig. 11
figure 11

LabVIEW front panel display of PDC at SPC showing the performance of distance relay RA without the proposed PAZSD methodology when an LL fault (FR = 1.7 Ω) occurs at 250 km from B1

Consider a double line to ground fault (LLG) with FR of 4.9 Ω at 350 km from Bus B1 as given in Table 6. The impedance observed by the relay RA is 80.335∠68.630.

The relay RA malfunctions because the impedance observed by the relay is greater than 21.8388 (distance from the center of the Zone-2 circle to the fault point). A LabVIEW front panel display for this case study is shown in Fig. 12. Figure 12 also displays the LabVIEW front panel for relay RA in PDC (without the proposed methodology). The voltage and current phasor data acquired from PMUs at B1 and B4, and the calculated impedances are also shown in Fig. 12.

Fig. 12
figure 12

LabVIEW front panel display of PDC at SPC showing the performance of distance relay RA without the proposed PAZSD methodology for an LLG fault (FR = 4.9 Ω) at 350 km from B1

The subsequent section presents the performance of the distance relay RA when the proposed methodology has been implemented at SPC.

4.2 Real-time performance analysis of the conventional distance relay with PAZSD methodology

The following conditions are used to detect the zone of fault point using the proposed PAZSD methodology.

$$ \mathrm{Zone}\hbox{-} 1:\kern1em \mathrm{If}\mid {\mathrm{Z}}_{\mathrm{Calculate}\hbox{-} 1}\mid <\left(|{\mathrm{Z}}_{\mathrm{new}}{\_}_{\mathrm{Z}\mathrm{one}1}|/2\right) $$
(3)
$$ \mathrm{Zone}\hbox{-} 2:\kern1em \mathrm{If}\ \left(|{\mathrm{Z}}_{\mathrm{new}}{\_}_{\mathrm{Z}\mathrm{one}1}|/2\right)<\mid {\mathrm{Z}}_{\mathrm{Calculated}\hbox{-} 2}\mid <\left(|{\mathrm{Z}}_{\mathrm{new}}{\_}_{\mathrm{Z}\mathrm{one}2}/2\right) $$
(4)

In this subsection, the enhanced functioning of the distance relay RA with PAZSD methodology for the same case studies (studied in subsection 4.1) is discussed. The new zone settings of the relay RA using the current coefficients (K1, K2 & K3) for different faults with different fault conditions are tabulated in Tables 7, 8 and 9.

Table 7 Performance of the distance relay RA with PAZSD methodology for different faults with FR = 0.2 Ω at different distances
Table 8 Performance of distance relay RA with PAZSD methodology for different faults with FR = 1.7 Ω at different distances
Table 9 Performance of distance relay RA with PAZSD methodology for different faults with FR = 4.9 Ω at different distances

From Table 7, for the LG with the same fault conditions as discussed in subsection 4.1, the current coefficients K1, K2 & K3 estimated by the proposed methodology are 1.2656, 1.6712 & 1.5703 respectively. Since K2 > (K1 & K3), the new zone settings of the relay RA as per the proposed methodology are 38.930∠60.36° and 72.994∠60.36°. For this condition, the impedance observed by the relay RA is 4.011∠63.5990. The zone of fault detection is Zone-1 since the magnitude of the observed value is less than 18.29 (Condition 3). Therefore, the operation of the relay RA with new zone settings is same as with the old zone settings. Figure 13 displays the LabVIEW front panel for relay RA in PDC at SPC (with the proposed methodology). The LabVIEW front panel displays glowing LED against Zone-1 fault. Figure 13 also shows the voltage and current phasors acquired from PMUs at B1 and B4 and the calculated impedance. A similar explanation holds good for LL fault at 50 km, and LLG & LLL faults at 150 km from bus B1 as given in Table 7.

Fig. 13
figure 13

LabVIEW front panel display of PDC at SPC showing the performance of distance relay RA with the proposed PAZSD methodology for an LG fault (FR = 0.2 Ω) at 50 km from B1

Similarly, for double line fault (LL) with the same fault conditions as discussed in subsection 4.1, the current coefficients K1, K2 & K3 estimated by the proposed methodology are 4.3359, 5.1328 and 2.0313 respectively. The new zone settings of the relay RA as per the proposed methodology are 119.567∠60.36° and 224.188∠60.36° as K2 > (K1 & K3). For this condition, the impedance observed by the relay RA is 32.092∠84.440. The relay RA has seen the impedance in Zone-1 since the magnitude of the calculated impedance (|ZCalculated-1|) is less than 59.7835 (Condition 3). Therefore, the relay operates correctly, i.e. in Zone-1 whereas without the proposed methodology the relay RA operates in Zone-2. Thus, the effect of the infeed on the relay RA performance has been eliminated by the proposed methodology. A similar explanation holds good for LG, LL, LLG & LLL faults at 250 km from bus B1 as given in Table 4. Likewise, for double line to ground fault (LLG) with the same fault conditions as discussed in subsection 4.1, the current coefficients K1, K2 & K3 estimated by the proposed methodology are 2.4688, 4.0469 and 4.375 respectively. Since K3 > (K2 & K1), the new zone settings of the relay RA using 4.375 are 101.914∠60.36° and 191.090∠60.36°. The impedance observed by the relay RA is 48.136∠80.370.

The relay RA has seen the impedance in Zone-1 since the magnitude of the calculated impedance (|Z Calculated-1|) is less than 50.957 (Condition 3). Therefore, the relay does operate correctly, i.e. in Zone-1 whereas without the proposed methodology the relay RA does not operate. Thus, the infeed at Bus B2 has not affected the performance of the relay RA. A similar explanation holds good for LG, LL & LLL faults at 350 km from Bus B1 as given in Table 7. The fault conditions for all the cases are shown in Table 7 considering a FR of 0.2 Ω at the fault point.

Tables 8 and 9 show similar case studies as considered in Table 7 but with FR of 1.7 Ω and 4.9 Ω respectively. From tables, it is clear that with a change in FR the relay RA with the proposed methodology functions correctly for all the cases regardless of the infeed condition at Bus B2. Few cases are discussed below for better understanding.

From Table 8, consider double line fault (LL) with the same fault conditions as discussed in subsection 4.1 (Table 5). The current coefficients K1, K2 & K3 estimated by the proposed methodology are 4.3359, 5.1328 and 2.0312 respectively. The new zone settings of the relay RA as per the proposed methodology are 119.558∠60.36° and 224.191∠60.36° as K2 > (K1 & K3). The impedance observed by the relay RA is 31.75∠ − 275.120. The relay RA has seen the impedance in Zone-1 because the magnitude of the calculated impedance (|Z Calculated-1|) is less than 59.77 (Condition 3). Therefore, the relay RA with the proposed methodology does operate in the right zone, i.e. Zone-1 whereas without the proposed methodology the relay RA operates in Zone-2. Thus, the mal-operation of the relay RA is averted. Figure 14 displays the LabVIEW front panel for relay RA in PDC at SPC (with the proposed methodology). The LabVIEW front panel displays glowing LED for Zone-1 fault. Figure 14 also shows the voltage and current phasor data acquired from PMUs at B1 and B4 and the calculated impedance. The fault conditions for all the cases are shown in Table 8 considering a FR of 1.7 Ω at the fault point.

Fig. 14
figure 14

LabVIEW front panel display of PDC at SPC showing the performance of distance relay RA with the proposed PAZSD methodology for an LL fault (FR = 1.7 Ω) at 250 km from B1

Consider LLG fault with FR of 4.9 Ω at 350 km from Bus B1 as given in Table 9. The current coefficients K1, K2 & K3 estimated by the proposed methodology are 4.3359, 4.75 and 5.0547 respectively. The new zone settings of the relay RA as per the proposed methodology are 117.748∠60.36° and 220.777∠60.36° as K3 > (K1 & K2). The impedance observed by the relay RA is 80.614∠68.660. The relay RA has seen the impedance in Zone-2 since the magnitude of the calculated impedance (|Z Calculated-2|) is less than 110.389 (Condition 4). Therefore, the relay RA with the proposed methodology does operate in the right zone, i.e., Zone-2 whereas without the proposed methodology the relay RA does not operate. The LabVIEW front panel display for LLG fault with FR of 4.9 Ω at 350 km from Bus B1 is shown in Fig. 15. Figure 15 also displays the LabVIEW front panel for relay RA in PDC at SPC (with the proposed methodology). The voltage and current phasor data acquired from PMUs at B1 and B4, and the calculated impedances are also shown in Fig. 15. The fault conditions for all the cases are shown in Table 9 considering an FR of 4.9 Ω at the fault point.

Fig. 15
figure 15

LabVIEW front panel display of PDC at SPC showing the performance of distance relay RA with the proposed PAZSD methodology for an LLG fault (FR = 4.9 Ω) at 350 km from B1

Thus, from the above elaborated discussion, it is clear that the proposed PAZSD methodology has improved the performance of the conventional distance relay.

4.3 Reliability analysis of the conventional distance protection without and with the proposed methodology

Despite the simple and dependable performance of the conventional distance protective system, the reliability of distance protection is affected when infeed condition exists in MTLs. In-feed conditions jeopardize security in the power system due to the non-adaptive property of distance protection system and provide an obscure view of the system conditions. Further, the function of the conventional distance protection may not be accurate for faults with different fault impedances.

The reliability attribute of the conventional distance protection with and without the proposed PAZSD methodology has been analyzed in this section as per the definition of reliability [19].

From Table 4, for instance, when an LL fault occurred at 50 km from Bus B1, the relay RA (without the proposed PAZSD methodology) has observed the fault point in Zone-1 which shows that the relay RA operates correctly as per zone settings. Thus, the reliability attribute of the relay has not been influenced by the infeed at Bus B2.

Similarly, the reliability of the relay had not influenced by the infeed when LG at 50 km, LLG & LLL faults at 150 km from Bus B1 are separately created as given in Table 4. However, when LG fault occurred at 250 km from Bus B1, the relay RA has observed the fault in Zone-2, even though the fault is in Zone-1. Thus, the FR has influenced the reliability of the relay RA. A similar explanation holds good for LG, LL, LLG & LLL faults at 250 km from Bus B1 as given in Table 4. Likewise, when the double line to ground fault (LLG) occurred at 300 km from Bus B1, the relay RA has seen the impedance neither in Zone-1 nor Zone-2. Thus, the infeed at Bus B2 and FR has influenced the reliability of the relay RA. A similar explanation holds good for LG, LL & LLL faults at 350 km from Bus B1 as given in Table 4.

Similarly, Tables 5 and 6 show similar case studies as discussed in Table 4 but with FR of 1.7 Ω and 4.9 Ω respectively. From tables, it is clear that with the change in FR, the reliability of the relay RA has been influenced by many cases because of infeed at Bus B2. However, from Table 7, for the same fault conditions as discussed in Table 4, the reliability of the relay RA has been improved by the proposed PAZSD methodology by changing the zone settings adaptively as per the requirement. Likewise, the reliability of the relay RA has been improved by the proposed methodology for all the case studies as tabulated in Tables 8 to 9.

The above concise discussion underlines the importance of the proposed methodology in improving the reliability of conventional distance protection during infeed condition and impedance faults in MTLs.

5 Conclusions

This paper proposed a PMU based methodology for adaptive zone settings of distance relays to improve the performance and reliability of distance protection. The operation of distance relays during infeed condition with and without the proposed methodology has been demonstrated through a four-bus model implemented in PSCAD/EMTDC environment. Further, a laboratory prototype of EHV MTLs is considered to validate the performance of distance relays during infeed condition. The PAZSD methodology employs current coefficients to adjust the zone settings of the relays during infeed situations. The results strongly convey that the proposed PAZSD methodology is effective in improving the performance and reliability of distance protection during infeed condition.