Design and Development of Superimposed Directional Comparison Protection Scheme

To achieve the UK Net Zero future by 2050, National Grid needs to integrate significantly more renewable generation into the power grid. This increases the level of harmonics, reduces system inertia and adversely affects the fault level and the performance of existing protection relays. One solution to the protection problem is the use of new types of protection that use the change in the voltage and current caused by the fault, often referred to as a superimposed or incremental based protection technique. This paper describes how a superimposed directional comparison protection scheme performed when applied to a reduced section of the full UK National Grid network and relates this to the operating performance of traditional protection. Tests are performed using the simulators DIgSILENT and RelaySimTest configured with different source levels, fault types, fault locations and fault resistances. Results show the superimposed based protection scheme achieves faster fault detection and tripping than conventional protection and is capable of detecting higher resistive faults on networks where the source capacities vary from strong to weak.


Introduction
To achieve Net Zero by 2050, the UK needs to integrate significant intermittent renewable generation into the Transmission and Distribution Power Grids and stop using fossil fuel based power stations. In recent years, the UK decommissioned most of its coal stations, but Combined Cycle Gas Turbines still dominate the production of electrical energy and these need to be phased out over the next 10 to 20 years. The UK has limited hydro resources and consequently the future relies on the integration of power electronic interfaced wind and solar PV resources [1,2]. This increases the harmonics seen on the grid and reduces the power system inertia and fault levels, which have a negative impact on existing protection relays [3][4][5][6]. To address these issues, new types of protection are required that operate faster, but are also more sensitive and dependable, and provides the security that ensures the "correct" relays operate for all faults within their protective zone. In addition, relays must not falsely operate for non-fault events or operate when a fault can be cleared by the protection located closer or adjacent to the faulted item of plant. The operating performance of traditional distance and overcurrent protection is adversely affected by the replacement of high-inertia fossil fuelled generators with low-inertia intermittent renewable sources. This led to the development of new types of protection which use alternative techniques to detect and locate the fault, including those based on travelling waves and/or superimposed or incremental components. Travelling wave techniques are based on information deduced from the surges caused by the sudden collapse in voltage at the fault point when a fault occurs. These propagate at nearly the speed of light and travel away from the fault in both directions along the overhead transmission lines towards the relays at each end of the line. To detect these surges, travelling wave relays require a high sampling rate and generally incorporate complex signal processing algorithms [7]. Superimposed directional relays have a response time of a few to several milliseconds, which is achieved by subtracting from the instantaneous values of the voltage and current signals their corresponding values acquired exactly one cycle earlier. The resulting voltage and current changes or incremental signals are then used to determine if the fault is "infront-of" or "behind" the relay. This technique allows the use of a simple communication channel that delivers messages corresponding to the directional decisions associated with each relay (forward or reverse) [8]. Superimposed techniques have been used in many different protection schemes, including ones that: determine the directionality of conventional relays; predict fault levels; select faulted phases; monitor power swings; and evaluate short circuit levels [8].
This paper analyses a superimposed directional comparison protection scheme and compares the performance with a traditional protection scheme. The evaluation process involves different source levels, fault types, fault locations and fault resistances. All the relays are applied to a double circuit line operating in a reduced section of the full UK National Grid network model. Figure 1 shows how to extract superimposed signal from faulted signal. Any faulted network can be separated into a superimposed network and a pre-fault network using the Thevenin theorem and the principle of superimposed [9]. The faulted network contains the fault-generated voltage and current and the load voltage and current, whilst a pre-fault network only contains the load voltage and current. The faulted and prefault voltage and current at any given time and location can be obtained from voltage and current transformers and the superimposed signals are the difference between the faulted signal and the pre-fault signal [9]. For example, the superimposed voltage can be calculated using (1).

Superimposed Based Function Operating Principle
where: ΔV is the superimposed voltage. V is the fault voltage.
Vpre is the prefault current. The superimposed directional decision requires the evaluation of the superimposed voltage signals and the superimposed replica current (ΔIz). Figures 1 and 2 show this superimposed replica current is directly proportional to the superimposed voltage (ΔV) measured at the relay location. The directional decision involves the comparison between the polarity of the superimposed voltage and superimposed replica current signals. For forward faults the polarities are opposite and for reverse faults they are the same [10]. For example, Eq. (1) can be obtained from Fig. 1. The Eq. (2) means Relay 1 in Fig. 1 detects a forward direction and Eq. (3) means Relay 1 in Fig. 2 detects a reverse direction [11,12].
(1) The superimposed directional elements TD32 incorporated within the superimposed based feeder protection relay is studied in this paper, all the other functions, incremental distance elements (TD21), incremental non-directional overcurrent supervision (TD50), incremental directional overcurrent supervision (TD67), travelling wave directional elements (TW32) and travelling wave differential elements (TW87) are disabled [13]. The TD32 element is used to determine the direction to the fault (forward or reverse) and the "local" relay will trip when it detects the forward fault and receive a permissive signal from the relay at the remote end of the feeder [14].

Simplified National Grid Model
simplified model, which are then connected into the model used for short circuit calculations and EMT simulations. Table 1 lists all the transmission line parameters in the reduced model. Sources are changed to fit the different scenarios discussed in Sect. 5. The A731(1) (L4) line from Bolney and Ninfield is protected using Relay 1 on the left side and Relay 2 on the right side.

Lab Tests Setup
The superimposed directional elements in time domain feeder protection scheme installed at both ends of the protected line (L4) were tested using the signals generated by the power system simulators. As shown in Fig. 6, The Ethernet network interface is used to connect between the power system simulators and the test sets. The superimposed directional elements in time domain feeder protection relays were installed at both end of the protected line in the RelaySimTest using two relay test sets. For comparison purposes, two traditional relays, DIgSILENT model of generic overcurrent relays and generic distance relays were setup and tested in DIgSILENT. Relay models for both these relays are available in DIgSILENT, which means it was not necessary to physically test these relays. All the network and relay parameters use the same values within both DIgSILENT and RelaySimTest [15,16]. Figure 7 shows the arrangement for the simplest equivalent sources at both end of the protected line, A731(1) L4. These are changed to different ohmic values to understand the sensitivity of the superimposed directional elements. Source strengths are described by the SIR (Source Impedance Ratio), and those selected for S1 and S2 in the three scenarios are listed in Table 2 for a strong source, SIR is 1.59 and for a weak source, SIR is 30. Table 3 lists the different fault types, locations, resistance values and inception angles analysed under the three source scenarios. Table 4 details the relay trip time results obtained under the "strong-strong" source scenario. The operating times were monitored under conditions involving various fault types, different fault locations and a-phase fault inception angles (FIA) of 0° and 90°.

Relay Tests for 'Strong-Strong' Source
The results in Row 1-9 were drawn as bar-charts 1(a), 1(b), 3(a) and 3(b) in Fig. 8. The charts show the results for different fault locations and inception angles when 0Ω single phase to ground fault and three phase fault occurs respectively.
Results for 100Ω and 73Ω fault resistances are shown in 2(a) and 2(b), 4(a) and 4(b) and 5(a) and 5(b). The x axis for all the charts is the percentage fault location from Bolney to Ninfield and the y axis is the relay tripping time in milliseconds.
The results in Table 4 and Fig. 8 demonstrate both time domain feeder protection relays tripped correctly and in an appropriate time when an internal 0Ω single phase to ground or three phase fault occurs. This is because the superimposed elements are detected and calculated correctly, and each relay receives the correct permissive signal from the relay at the opposite end of the protected line. However, some problems are experienced when the fault resistance is increased to 100Ω. For example, Fig. 8 2(a) shows relay1 cannot initiate a trip response when a 100Ω single phase to ground fault with 0° fault inception angle occurs at the far end of the line. This is because the fault resistance is high and a fault occurring at a zero crossing in the voltage signal results in very small superimposed current and voltage signals, and these cannot be detected by the relay. Recognising the fault resistance is extremely high, the process was repeated with fault resistances set to lower values. Both relays were found to trip reliably for all fault types, all locations and all fault inception angles, if the fault resistance was less than 73 Ω.    trip when a 100Ω fault with 90° inception angle occurs close to Relay 1. This is because the superimposed signals resulting from the fault are small and cannot be detected correctly.

Relay Tests for "Weak-Weak" Source
Comparing the results in Fig. 8, charts 2(a), 2(b), 4(a) and 4(b) with Fig. 9, charts 2(a), 2(b), 4(a) and 4(b), indicates the protection performs better with two weak sources (both SIR = 30) rather than two strong sources (both SIR = 1.59). This illustrates the superimposed voltage signals are larger and easier to detect if the source is weak. indicates the relays cannot trip correctly for a high resistance (100Ω) fault and especially one that involves only one phase and ground. This is because the single phase-ground fault current is smaller than the three phase current and the superimposed signals cannot be detected.

Comparison with Traditional Relays and Analysis
The simulated overcurrent relay and the distance relays were evaluated and their operating performance compared with the results obtained under similar conditions when using the superimposed elements available in the superimposed directional elements in time domain feeder protection relay. Figure 11 shows the results for overcurrent relays tests under strong-strong condition. The overcurrent relays operate correctly for all internal faults with 0 Ω fault resistance. Both overcurrent relays tripped correctly when the fault resistance was below 10 Ω. However, for a high fault resistance (100 Ω), both relays did not trip, because the fault current was less than the relay operating current. Table 3 All Case Studies for Superimposed Based Relay S1:Strong-S2:Strong S1:Weak-S2:Weak S1:Strong-S2:Weak   Figure 12 shows the results obtained when testing a distance relay under a 'strong-strong' source scenario. The distance relays correctly initiated a trip signal for all internal faults with 0Ω fault, but could not operate when a single phase to ground fault occurs with 10 Ω fault resistance. This was because the relays did not detect the high resistance fault within their protected zones.

Conclusions
A reduced section of the full UK network with double circuit transmission line was modelled in DIgSILENT and Relay-SimTest to test the performance of a superimposed directional comparison protection scheme and traditional protection scheme. Tests are performed with different source level, fault types, fault locations and fault resistances.
Results show the superimposed based protection scheme achieves faster fault detection and tripping than conventional  Testing results for the "weak-weak" source scenario Fig. 10 Testing results for the "strong-weak" source scenario protection and it is capable of detecting higher resistive faults on networks where the source capacities vary from strong to weak. For a "strong-strong" sources (both SIR = 1.59) scenario, the superimposed based relay can trip when the faut resistance is up to 73 Ω for all fault types and locations, and it can trip higher resistances single phase to ground faults than three phase faults. For the system with weaker source (SIR is increased to 30), the superimposed based protection scheme performs better than strong source because superimposed voltage signals are larger and easier to detect if the source is weak. This shows there is significant future potential for superimposed based relays, especially when utilised in a low inertia renewable dominated power network.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Ziye Guo is a PhD student in electrical power system protection from the University of Manchester. Her research is about wide area protection using superimposed directional agents.