Power System - Fault Clearance with Limited Fault Current A study of fault detection and clearance in presence of converter interfaced generators Master of Science Thesis ANUP ARYAL Department of Electrical Engineering Division of Electrical Power Engineering CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2024 www.chalmers.se http://www.chalmers.se/ MASTERS THESIS 2024 Power System - Fault Clearance with Limited Fault Current A study of fault detection and clearance in presence of converter interfaced generators ANUP ARYAL Department of Electrical Engineering Division of Electrical Power Engineering CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2024 Power System - Fault Clearance with Limited Fault Current A study of fault detection and clearance in presence of converter interfaced generators © Anup Aryal, 2024 Supervisor: Adjunct Professor Daniel Karlsson, Electrical Engineering Department Examiner: Associate Professor Peiyuan Chen, Electrical Engineering Department Master’s Thesis 2024 Department of Electrical Engineering Division of Electric Power Engineering Chalmers University of Technology SE-412 96 Gothenburg Telephone +46 31 772 1000 Typeset in Microsoft Word Printed by Chalmers Reproservice Gothenburg, Sweden 2024 Power System - Fault Clearance with Limited Fault Current ANUP ARYAL Department of Electrical Engineering Division of Electric Power Engineering Chalmers University of Technology Abstract The integration of renewable energy sources like wind and solar power into the grid, typically through converter interfaces, has introduced significant technical challenges, particularly in fault current management. This thesis explores the evolving challenges and solutions in power system protection as the share of converter interfaced generators continues to grow in response to global ambitions for reducing greenhouse gas emissions. Traditional synchronous generators provide predictable and substantial fault currents that are essential for the reliable operation of existing protection schemes. However, converter interfaced generators, governed by converter controllers, produce fault currents that are not only lower in magnitude but also less predictable, creating a need to reassess and adapt current protection systems. This research investigates the impact of high converter interfaced generators penetration on short-circuit capacity and the preparedness of transmission system operators to manage these changes. The study highlights the limitations of conventional protection schemes, such as overcurrent and distance protections, where they exhibit under-reaching of distance relays and delayed tripping of overcurrent relays. These issues arise when fault currents are significantly reduced and exhibit varying characteristics, including diminished negative sequence currents and a wider range of phase angles. The research also examines the potential of new protection technologies and contingency strategies, such as grid forming converters, synchronous condensers, adaptive protections, and advanced methods like traveling wave-based protection, in addressing these challenges. Despite promising developments, the thesis identifies several obstacles to the widespread implementation of these new technologies, including high costs, complex performance evaluations, communication and cybersecurity concerns, and a lack of standardization, which hinders interoperability among different equipment vendors. Furthermore, evolving grid codes and new requirements for fault ride through and reactive power control demonstrate the ongoing efforts to mandate better fault management characteristics for converter interfaced generators. The reliability of overcurrent protection in distribution system and distance protection in transmission system can be significantly compromised in the evolving grid. There has been significant progress in adapting protection systems to accommodate more renewable resources, but further research and development are crucial to overcoming the remaining challenges and ensuring reliable fault detection and clearing in the modern power systems. Keywords: Protection, Relay, Converter, Wind turbine, Fault, Short circuit, Grid code Kraftsystem - Felbortkoppling med begränsad felström ANUP ARYAL Institutionen för Elektroteknik Avdelningen för Elektrisk Krafteknik Chalmers Tekniska Högskola Sammanfattning Integrationen av förnybara energikällor som vind- och solkraft i elnätet, ofta genom omriktargränssnitt, har skapat betydande tekniska utmaningar, särskilt när det gäller hantering av felström. Denna avhandling undersöker de föränderliga utmaningarna och lösningarna inom kraftsystemskydd i takt med att andelen generatorer med omriktargränssnitt fortsätter att öka som svar på globala ambitioner att minska växthusgasutsläpp. Traditionella synkrona generatorer ger förutsägbara och betydande felströmmar som är avgörande för tillförlitlig drift av befintliga skyddssystem. Omriktarstyrda generatorer producerar dock felströmmar som inte bara är lägre i storlek utan också mindre förutsägbara, vilket skapar ett behov av att ompröva och anpassa nuvarande skyddssystem. Denna forskning undersöker påverkan av hög penetrering av generatorer med omriktargränssnitt på kortslutningseffekt och beredskapen hos transmissionssystemoperatörer att hantera dessa förändringar. Studien lyfter fram begränsningarna i konventionella skyddssystem, såsom överströmsskydd och distansskydd, när felströmmarna är betydligt reducerade och uppvisar varierande egenskaper, inklusive minskade minusföljdsströmmar och ett bredare spektrum av fasvinklar. Forskningen undersöker också potentialen hos nya skyddsteknologier och beredskapsstrategier, såsom nätstyrande omriktare, synkronkompensatorer, adaptiva skydd och avancerade metoder som skydd baserat på vandringsvågor, för att hantera dessa utmaningar. Trots lovande utveckling identifierar avhandlingen flera hinder för den utbredda implementeringen av dessa nya teknologier, inklusive höga kostnader, komplexa prestandautvärderingar, kommunikations- och cybersäkerhetsproblem samt brist på standardisering, vilket hindrar interoperabilitet mellan olika utrustningsleverantörer. Dessutom visar nätkoder och nya krav för felridning och reaktiv effektreglering på de pågående ansträngningarna att ställa högre krav på felhanteringsegenskaper för generatorer med omriktargränssnitt. Sammanfattningsvis, även om det har gjorts betydande framsteg i att anpassa skyddssystem för att rymma mer förnybara resurser, är ytterligare forskning och utveckling avgörande för att övervinna de återstående utmaningarna och säkerställa tillförlitlig feldetektering och felbortkoppling i moderna kraftsystem. Acknowledgements This thesis was written to fulfill the requirements of the master’s degree program in Sustainable Electric Power Engineering and Electromobility at Chalmers University of Technology. This work was produced during scholarship period funded by the Swedish Institute. I am deeply grateful for all their support which has helped me in my academic and personal growth. I extend my heartfelt gratitude to my supervisor, Daniel Karlsson, Adjunct Professor at Department of Electrical Engineering, Chalmers University of Technology, for his unwavering guidance, expert insights and support throughout the thesis work. His guidance and mentorship has helped me navigate through research work and shape the trajectory and depth of this thesis work. I would also like to express my sincerest appreciation to my examiner, Peiyuan Chen, Associate Professor at the Department of Electrical Engineering, Chalmers University of Technology, for his thoughtful guidance, constructive feedback and help with PSS/E. His guidance and assessment has greatly enhanced the quality of this work. My heartfelt thanks go out to Abdin Mohamed and Abhishek Subramanyam for their invaluable assistance with both the simulations and the report. Their expertise and guidance greatly eased the process, making it much more manageable for me to complete this work. To my family, whose unwavering support and encouragement has been the foundation of my academic journey – this accomplishment would not have been possible without them. Their belief in me, even during the most challenging times, has been a constant source of strength and motivation. And to my friends, who have inspired, guided, and uplifted me: those who encouraged me to pursue a master’s degree, those who convinced me to take the leap and move to a new country, and those who supported me every day over the past two years. I am truly fortunate to have friends who make each day a little brighter. Thank you all for being a part of this journey. Anup Aryal Gothenburg, October 2024 List of Acronyms Below is the list of acronyms that have been used throughout this thesis listed in alphabetical order: CIG Converter Connected Generator CT Current Transformer DFIG Double Fed Induction Generator DG Distributed Generation ENTSO-E European Network of Transmission Service Operators for Electricity FCL Fault Current Limiters FRT Fault Ride Through GFC Grid Forming Converters IDE Intelligent Electronic Device IEC International Electrotechnical Commission POTT Permissive Overreach Transfer Trip PPM Power Park Module P.U. Per Unit PUTT Permissive underreach Transfer Trip RES Renewable Energy Source SC Synchronous Condenser SCC Short Circuit Capacity SG Synchronous Generator TSO Transmission System Operator TW Travelling Wave WTG Wind Turbine Generator Contents 1 Introduction ........................................................................................................................ 1 1.1 Background ................................................................................................................ 1 1.2 Objective .................................................................................................................... 2 1.3 Scope .......................................................................................................................... 2 1.4 Structure of the report ................................................................................................ 2 2 Power system faults and protections schemes ................................................................... 3 2.1 Overcurrent protection ............................................................................................... 4 2.2 Distance protection .................................................................................................... 4 2.3 Line differential protection ........................................................................................ 6 3 Grid strength and fault response from converter interfaced generators ............................. 8 4 Standard for fault current calculation .............................................................................. 11 5 Fault current study on a power system model ................................................................. 15 5.1 Simulation tool ......................................................................................................... 15 5.2 Model description .................................................................................................... 16 5.3 Methodology ............................................................................................................ 18 5.4 Results of simulation................................................................................................ 19 5.5 Interpretation of simulation results .......................................................................... 21 6 Impact of converter interfaced generators on existing protection scheme ...................... 23 6.1 Impact on protection of distribution network .......................................................... 23 6.2 Impact on protection of transmission network......................................................... 25 6.2.1 Distance protection .......................................................................................... 25 6.2.2 Overcurrent protection ..................................................................................... 29 6.2.3 Line differential protection .............................................................................. 29 6.2.4 Other protection schemes ................................................................................. 30 7 Commercially available techniques ................................................................................. 31 7.1 Current reversal protection ...................................................................................... 31 7.2 Weak end infeed protection...................................................................................... 32 7.3 Instantaneous residual overcurrent protection ......................................................... 32 7.4 Four step residual overcurrent protection ................................................................ 34 7.5 Sensitive directional residual protection .................................................................. 35 7.6 Line differential protection ...................................................................................... 36 8 Power grid codes .............................................................................................................. 39 9 Promising technologies and required support .................................................................. 43 9.1 Fault current limiting technologies .......................................................................... 43 9.1.1 Voltage monitoring based protection ............................................................... 43 9.1.2 Fault current limiters ........................................................................................ 44 9.2 Distance protection in distribution system ............................................................... 44 9.3 Adaptive protection .................................................................................................. 45 9.4 Permissive overreach ............................................................................................... 47 9.5 Travelling wave protection ...................................................................................... 48 9.6 Line differential protection ...................................................................................... 49 9.7 Grid forming converters ........................................................................................... 50 9.8 Synchronous condensers .......................................................................................... 51 10 Discussion ........................................................................................................................ 54 11 Conclusions and recommendations.................................................................................. 59 11.1 Conclusions .............................................................................................................. 59 11.2 Recommendations .................................................................................................... 60 Bibliogaphy .............................................................................................................................. 61 1 1 Introduction This section aims to serve as an introduction providing a brief overview of the thesis project, its relevance, and interesting aspects. In addition, it also provides a problem description, objectives, scope, and limitation of the project work. This section ends with a description of the structure of the report. 1.1 Background The ambition of cutting greenhouse gas emissions and increasing energy efficiency to tackle climate change has created a shift in how electricity is produced and consumed. The share of renewable energy sources (RES) in the power system is increasing every year. This energy transition has brought many challenges to the power system. Wind and solar power are connected to the grid through a converter interface. Converters are electric circuits that have semiconductor switches used to alter amplitude, frequency, and phase angles of the electrical parameters. These devices are used in AC/DC conversion in rectifiers or inverters. Flexible AC transmission system (FACTS) and high voltage direct current (HVDC) links also use power electronic converters. HVDC is particularly popular in integration of offshore wind farms whereas FACTS devices are used in transmission systems to maintain voltage and angle stability. These devices are capable of absorbing and injecting reactive power in the transmission system at different points. While the converters offer significant advantages like flexibility and control, it has also introduced several challenges like harmonic distortion, voltage and frequency instability, fault current limitation, etc. Among these, limited fault current from converter connected generators has emerged as a major challenge in power system protection. The fault response from converter interfaced generators (CIG) are governed by the converter controller and can provide the fault current in the same range as rated current (1-1.5 p.u.). Whereas the traditional synchronous generators (SGs) can provide fault current of around five times its rated operational current. Fault response from the traditional generators is predictable and can be used during design of protection system for power system. Fault current from CIGs on the other hand depends on the type of controller and is not well defined. Protection schemes are designed and implemented to detect abnormal conditions, locate the abnormal area, and isolate faulty parts from the rest of the system. It helps in keeping equipment and people safe from electrical hazard and maintains the safe operation of the rest of the system. Fault current is important for relays to detect faults and execute protection schemes. Overcurrent and impedance-based protection schemes like distance protection schemes use fault current amplitude to detect fault in the system. The power system protection schemes in use today are based on the fault current contribution from conventional generators. With the increasing penetration of CIGs, the protection system needs to be investigated for adequacy and preparedness. In Swedish power system where wind and solar power are increasing, new hydropower are not being built and nuclear plants might get shut 2 down, a situation where CIGs contribute most of the fault current doesn’t seem so distant. This thesis aims to study short circuit capacity in CIG dominated grid, investigate how prepared the Transmission system operators (TSO) are, and study the changing role of existing protection schemes and promising new technologies. 1.2 Objective The main objective of this thesis is to study, describe and sort different principles used in fault clearing when converter-based generation is dominant in power system. A power system model is used to study how the short circuit capacity of the Swedish grid will change if the nuclear plants are discontinued and are replaced by wind power. The technology and principles available today are studied and analysed on whether they will be helpful in new network topology. Suitability and implementation of some promising technologies and principles are also analysed. With the help of literatures, computer simulation, equipment manuals, grid codes, and standards, sufficiency of existing and new technologies, technical challenges and required support are presented in this report. 1.3 Scope The focus of this report is on fault detection and clearance with limited fault current. Detection of low fault current and contribution from generators for fault current support are part of the study. The primary focus will be on the transmission network. However, a few effects of CIG domination on distribution system are also studied. Wind turbine CIG is the main focus of the study and solar will not be studied in detail. Effects of compensation devices, HVDC links, DC-DC converter, DC power system are not included. 1.4 Structure of the report This report has 11 chapters which serve to fulfill the objectives. Chapter 2 introduces the power system faults and some protection schemes that are in practice. Chapter 3 provides insights into the grid strength and characteristics of fault current response from CIGs. Chapter 4 evaluates fault current calculation standard. Chapter 5 studies the impact on fault current level on the grid when CIGs replace SGs, with the help of a power system model. Chapter 6 describes the impact of integration of CIGs in transmission and distribution level on protection coordination as well as the common protection scheme. Chapter 7 examines some of the functionalities present in the available protection devices and their appropriability for the CIG dominated grid. Chapter 8 investigates some old and new grid code provisions for the integration of CIG and. Some available and implemented technology and some promising technologies and principles are analyzed in chapter 9. Chapter 10 aims to tie together the findings of the thesis work and provide discussion on the previous chapters. Lastly, conclusions and recommendations for future work are presented in chapter 11. 3 2 Power system faults and protections schemes Events like lightning strikes on overhead lines, insulation breakdown, trees falling on a transmission line, open circuit due to broken conductors, mechanical damage of components or mis-operation of breakers create faults in power system. Only the faults that connect live phases to ground or other phases, also known as shunt faults, are within the scope of this thesis. Such power system faults are classified as balanced/symmetrical and unbalanced/unsymmetrical faults depending on number of phases involved in the fault. Three phase to ground is a balanced fault that is least frequent yet the most severe type. The most common type of fault is single line to ground (SLG), an unbalanced fault. Line to line (LL), and line to line to ground (LLG) are less common unbalanced faults. Fault studies are essential for selecting the appropriate ratings of protective switchgears and coordination of relays. The balanced faults can be solved using the per phase approach. But during the unbalanced faults, per phase impedances are not identical and it results in unbalanced current and voltages. Then it’s not possible to use the per phase approach as in balanced system. Solution using three phase approach is much more numerically complicated. So, the unbalanced faults are analyzed using the symmetrical components method as it is simpler to solve three different single phase circuits than solving a one single phase circuit in one set of equation. The unsymmetrical phase components are linearly transformed into a set of symmetrical components i.e., the positive, negative and zero sequence components. The unbalanced components can be decomposed into three sets of balanced components. Positive sequence components consists of a set of balanced three phase components with phase sequence ABC, where A, B and C are three phases with a phase angel of 120, and negative sequence components are in ACB phase sequence. Zero sequence components consist of three single phase components that are equal in magnitude and have the same phase angles. The balanced sequence networks, formed by the sequence components, are connected to the point of unbalance where a fault has occurred. Then, the three symmetrical circuits are solved individually to obtain fault current [1]. Power system faults are usually accompanied by a resistance known as fault resistance. Such resistance can be an arc resistance, resistance of object connecting live wire to the ground. Fault resistance affects the magnitude of fault current and could hinder fault detection. The protection system is implemented to detect faults and isolate the faulty part from the rest of the system. Effective protection systems will not let the abnormality propagate and disturb the stability and safety of the rest of the system. The protection system consists of a measurement section consisting of current and voltage transformers. Which means that the input to the protection system are voltage and current. Input data is fed to a relay, electromechanical or digital, where the data is interpreted and compared to a predefined value. Any abnormal situation is detected in this process and a trip signal is created. This signal is used to operate a breaker and disconnect all three phases, regardless of how many phases are affected by the fault. Relays are equipped with functions that can be set to serve desired protection scheme. Protection system design criteria include speed of operation, 4 selectivity of fault, sensitivity to weak signals and reliability when required. The most used protection systems are explained in the following section. 2.1 Overcurrent protection Overcurrent protection scheme discriminates between fault current and load current by the magnitude of current. During a short circuit, current magnitude can rise to a much higher value than load current and easily to detectable. It is commonly used in the medium voltage (MV) radial networks and distribution networks. A time delayed overcurrent protection is often used in the case of a radial system with multiple sections. Overcurrent relays are coordinated in such a way that the device closest to the fault trips in the shortest time. A directional relay feature is added to the protection scheme when there is presence of distributed generators and fault can be fed from multiple sides. Hence, it is a reliable measure for detecting a fault in a radial distribution system. However, overcurrent protection has several challenges. Fault resistance and neutral ground impedance can limit the fault current magnitude. A weak source, a long line or presence of high impedance components like transformer line will reduce the fault current magnitude. Since overcurrent relays are based on inverse time characteristics, fault clearing time also increases if the fault current is low. Coordinating a directional overcurrent relay is a difficult challenge in meshed system. Hence, its use is limited and in a meshed network, distance relay is preferred. Overcurrent protection can be used to protect the system against ground faults by measuring residual currents. A residual overcurrent protection scheme measures the vector sum of the currents in all three phases and neutral, if present, and detects the residual current during ground fault. Residual current is the imbalance current in the three-phase system where the sum of three phase currents should ideally be zero. The residual current has a permitted limit, and the relay operates when the threshold is exceeded. This protection system is designed to detect ground faults and earth leakage currents. It is used in protection of devices like generators and transformers where detection of ground fault is critical. 2.2 Distance protection Distance protection uses impedance measurements to detect faults. It takes advantage of the fact that fault not only increases the current level in the network but also decreases voltage level. Current and voltage, whose ratio gives the impedance, are measured in the line end and apparent impedance is computed. This computed value is compared to the known impedance of the transmission line. In normal operating conditions, the calculated impedance is much higher than the set amount. Whereas, during fault current increases rapidly and voltage decreases resulting in low impedance calculation. A comparison between the two is made and a fault is declared. The measured apparent impedance value gives the fault location as well. Distance protection is assigned to a portion of the line, known as reach, that the relay is supposed to protect. With fixed reach and measurement of current and voltage, distance 5 relays are more sensitive and precise than overcurrent relays. A directional unit can be added to define the reach in forward direction only. Distance protection scheme is usually divided into instantaneous zone and one or more time delayed zones. The instantaneous zone usually covers 80-90% of the immediate protected line. Zone two usually covers 120% of the line impedance. More zones can be added which serve as a backup protection with a time delay. Other zones can be set depending on the section of network with further time delay. Figure 1 shows a section of a transmission line where R1 and R2 are distance relays on two ends, and I represent direction of current flow. Zone 1 and Zone 2 are also marked in the figure. Figure 1. Distance protection scheme An impedance plane, known as R-X diagram, is used to represent distance relay characteristics and impedance of the line seen by the relay. An R-X diagram, a tool used to visualize the operation of distance relay, is presented in figure 2. A relay either has a circular or quadrilateral characteristics, selected based on fault characteristics they are supposed to clear. The region inside the circle or quadrilateral is called trip region and the outside is called block region. The line “Z” in the figure represents line impedance locus between the relay and points along the line. A solid fault will lie on this line but if there is a change in impedance due to the fault resistance, the fault will lie outside of the line. If the impedance seen by the relay is within the block region, normal operation condition is assumed. If the fault is in the trip zone, trip signal can be issued. Distance relay with quadrilateral characteristics is more popular because the resistive reach can be extended independently of reactive reach. This is one advantage it possesses over circular distance relays or mho relays by having enough reach for high resistive faults. 6 Figure 2. XR diagram with quadrilateral and circular characteristics Distance protection is equipped with communication schemes where relays at ends of protected parts communicate with each other to reduce fault clearing time. The relays in two ends, R1 and R2 in figure 1, are referred to explain the schemes. Direct transfer trip (DDT) is one basic type of communication scheme that is often used in radial lines. In this scheme, when a fault is detected by relay R1, it sends a trip signal to the relay R2 at the other end of the line and both relays trip immediately. This scheme, however, is not widely used due to being prone to malfunctioning. Some more advanced types of communication scheme are more popular and discussed in the following section. Permissive underreach transfer trip (PUTT) scheme uses two-way communication for fast fault clearance. When relay R1 detects a fault in its zone 1, it sends permissive signal to remote end relay R2 to trip. R2 receives the signal and detects the fault in its zone 2 and issues a trip command. R2 can trip without time delay even for the fault in its zone two. In case there is a failure in protection scheme, the relays can revert to the original scheme where each relay trips for faults in a zone it sees. In permissive overreach transfer trip (POTT) scheme, relays are set to detect faults beyond the protected line, extending into the next line section allowing faults to be detected near the zone boundaries. When a fault occurs somewhere within the overreaching zone, relays at both ends detect the fault and sends permissive signals to each other. Breakers near both relays trip simultaneously. 2.3 Line differential protection Differential protection is very common in generators, transformers, and busbars protection. It’s also used for protection of short transmission lines and cables. Differential protection is 7 based on Kirchoff’s current law. Current is measured at each end of the line and compared. If there is any difference in the measured value, it violates Kirchoff’s law and fault is declared. Communication between two measurement location to the relay is very important in differential protection. Current transformers will also not always read same current values in two ends. External faults in the networks can cause errors in phase angle, magnitude measurements and saturations of equipment. This creates a need to set overcurrent relay above a maximum error current during external faults. A percentage differential relay is used to fix this problem. The differential current must exceed a fixed percentage of the average of the primary and the secondary current of the protected element for the percentage differential relay to operate. Primary and secondary current here meaning the current measurements in the two ends of the protected element and is input to the relay. This average is known as through current or restraining current or bias current. The bias current helps distinguish between actual faults within the protected zone and external conditions that might otherwise cause unintended operation of the relay. The maximum differential current that is allowed without declaring fault depends on the current level where permitted differential current is larger for higher fault currents. Figure 3 shows operating characteristics of percentage differential relay through a differential current vs restraining current or biased current graph. The slope of the relay determines the trip zone. The restraining zone encompasses margin for safety and error like CT errors and current ratio mismatches [2]. Figure 3. Percentage differential relay operation characteristics. 8 3 Grid strength and fault response from converter interfaced generators Grid strength is the measure of capacity and robustness of power grid to reliably supply electricity without interruptions. A strong grid can maintain voltage and frequency stability of the network during disturbances and maintain uninterrupted power supply. The short circuit capacity (SCC) of a grid is a critical parameter and is a measurement of maximum short circuit current that can flow in the grid during a bolted three phase fault and is a strong indicator of the grid strength. Knowledge of SSC is important in the selection of equipment with the correct ratings and specifications. Components like transformers, switchgear, generators, cables, etc., must be selected to withstand the maximum fault current. SSC is essential for selection and coordination of protection devices such as circuit breakers and relays. Proper protection coordination ensures that the devices nearest to the fault operate quickly and stops the fault from spreading into the system. Traditionally, protective devices are operated based on detection of rapid rise of current in a short interval as discussed in previous section. Hence, coordination among protection devices is naturally easy in grid with high SCC. Conventional power generators or SGs have rotational mass that is coupled with the generators. It allows SGs to help during abnormal situations with inertial and frequency support. This feature helps restore stability in the system during frequency and power angle instabilities [1], [3]. But one of the most important characteristics of SG in maintaining grid strength is its response during a fault. SGs can inject fault current between 3-6 per unit (p.u.). Fault current injection from SGs during fault is predictable and usually easy to distinguish between normal load current and fault current. The short circuit current form SGs is proportional to the back EMF and the terminal voltages in a SG can be maintained by control of generator excitation current [3]. Although the traditional power system has faced issues like voltage instability and collapse, above mentioned inherent nature of SGs and ample SCC, has helped to mitigate the issues. CIGs, however, inject just 1-1.5 p.u. fault during a fault. Fault response from CIGs is determined by control scheme of power electronic devices and are not universal like SGs. The fault current injection of such power electronic interfaced generators is limited because of possible damage to the converters due to overcurrent and overheating of switches [4]. Other factors like fault ride through, reactive current support, etc., guided by the grid codes, also determine how a fault current is contributed [5]. Global push for net carbon zero by 2050 and the subsequent energy transition has created a steep rise in investment in RES. Figure 4 shows the recent trend in increase in wind and solar generation capacity according to International Renewable Energy Agency [6]. Wind and solar are trusted as the leaders in the energy transition due to low climate impact and recent technological advancements. According to [7], if all the countries abide by their pledges to the carbon neutralization target, RES will rise from about 30% of the total share of electricity in 2020 to nearly 70% in 2050. Solar and wind together is expected to supply half of the total 9 electricity demand. Since both sources have converter interface with the grid, a significant reduction in SCC can be expected in the future. Figure 4. Global wind and solar capacity 2010-2023 [6] Two commonly used types of wind turbine generators (WTGs) are Double-Fed Induction Generators (DFIG or Type III) and Full-Scale Converters (Type IV). In DFIG, the stator is directly connected to the grid, but the rotor is connected through rectifier- inverter pair. Full- scale converters are fully connected via rectifier-inverter pair. Type IV WTG is fully decoupled from the grid and shows a different fault response than Type III WTG. Grid connected solar plants also have a DC-AC interface and are decoupled from the grid [8]. So, Type III WTG operates differently during a fault compared to other CIGs. A crowbar is used in Type III WTG to disconnect rotor from converter during fault to protect the converter from thermal damage. Crowbar is kept in connection during mild faults and disconnected during sever faults. Mild faults are disturbances like minor voltage dips that do not cause extreme changes in the system. Severe faults are more extreme such as large voltage drops and prolonged faults. The rotor of the WTG faces induction of large voltage during severe fault, and this can cause high currents and stress on the power converter. The use of crowbar makes fault response from Type III WTG more complicated [8]. Since solar PVs are not in the scope of this thesis, it will not be explained in detail, but they are also completely decoupled from the grid with the converter interface and their fault response is similar to type IV WTG. Magnitude of the continuous fault current from CIG has a nonlinear relationship with the terminal voltage, and the current remains low due to converter constraints. Hence, phase angle of the fault current changes dynamically depending on the control and the terminal voltage [5]. The fault current from CIGs can be resistive, inductive, or capacitive unlike a SG which normally produces inductive fault current. The control scheme of a CIG impacts the VI characteristics during a fault near CIG. The change in angular relationship may lead to mis operation of protection function, like directional elements, that depend on phase angle [5], [8]. 0 200 400 600 800 1000 1200 1400 1600 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 T o ta l C ap ac it y ( G W ) Year Global wind and solar capacity 2010-2023 Solar Wind 10 CIGs generally contribute to positive sequence fault current only. A CIG fault contribution does not generate a zero-sequence component unless a transformer with a high-side wye- grounded winding is used to establish a ground connection [5]. The amplitude of negative sequence component is also suppressed depending on type of controller. The lack of sequence components may lead to mis-operation of protection equipment [5]. All these characteristics create the necessity to give each type of CIG individual consideration while studying SCC of the system and protection coordination. Wind and Solar are prominent sources of renewable energy among the alternative energy sources and their integration to grid is taking place at rapid pace. But its limited fault current capacity is not the only thing that’s making the grid weaker. The internal impedance of SGs affects the network impedance from its terminal to the point of fault. Since they are connected in parallel to the network, fewer number of SGs mean the equivalent impedance seen from the location of fault increases. A higher equivalent impedance results in reduced the fault current level [3]. The increasing number of CIGs in the grid means that SCC of the grid is going down and more weak areas in the grid are being created. Generally, RESs are integrated to the grid at distribution level as a concept of distributed generation. Wind and solar are a common source in microgrids as well. However, When the generation is in bulk like a solar plant or an offshore wind farm, they may be integrated at transmission or sub transmission level [9]. Increasing penetration of CIGs can cause fault current limitation which can lead to undetected fault in distribution network [10]. Dominance of limited fault feeding source will make transmission network weak. Such CIGs pose questions on reliability of protection system and there is a need to rethink the whole protection system. 11 4 Standard for fault current calculation International Electrotechnical Commission (IEC) has prepared and published a standard, IEC 60909: Short-circuit currents in three phase AC system [11], to harmonize international standards for short circuit calculations. This standard is important in the design, calculations of parameters, operation, and protection of electrical components, ensuring the equipment can successfully handle short circuit current. It provides guidelines to represent the network and its components to calculate short circuit currents accurately. The standard provides manual correction methods, but it is commonly used in power system analysis software for fault analysis on complex and large networks. It addresses the contribution from different sources but the short circuit current calculation from wind power station with doubly fed asynchronous generator and full-size converter is focused on this section. • Short circuit impedances: Short circuit impedances (Zk) are defined as the impedance of the system viewed from the location of short circuit. In this standard, generators of wind power stations and their unit transformers are combined into one unit for calculations of short circuit currents. The short circuit current at the terminal of DFIG or full-size converter connected WTGs are not dealt with in this standard. Many grid codes dictate that WTGs should feed mostly the reactive current during short circuit and during the interval, in which duration, station acts as current source [11]. These are the basis for impedance calculations. The total positive sequence short circuit impedance (ZWD) of DFIG WTG is given by (5). ZWD= √2∗KWD∗UrTHV √3∗iWDmax (5) Where, KWD is factor for calculation of peak short circuit current referred to HV side of transformer. This value is given by the manufacturer and depends on converter protection equipment like crowbar resistance. UrTHV is the rated voltage of the unit transformer at the HV side. iWdmax is the highest instantaneous value for three phase short circuit. In the case of unbalanced short circuits, the design and control strategy influences negative sequence impedance, while the type of transformer and earthing scheme determines zero sequence impedance. These values can be varied among different manufacturers. Similarly, full-scale converter WTGs are modelled as current sources in the positive- sequence system. The source current varies across different WTGs and is influenced by the type of short circuit. The positive sequence impedance is assumed to be infinite. The zero sequence impedance is also infinite. For unbalanced short circuits, the controller determines the negative sequence impedance, which is provided by WTG manufacturer. 12 • Calculation of short circuit current: Knowledge of maximum short circuit current is necessary for mechanical and thermal stress withstanding capacity of an equipment while minimum short-circuit current is important for selection of protection system. IEC 60909 uses voltage factor “c” to calculate maximum and minimum short circuit currents when the national standard for grid is not available. The voltage factor represents voltage variation at the bus. The maximum and minimum values of factor varies from 0.9 to 1.1 depending on the voltage level. The “c” factor, maximum and minimum, is multiplied with the voltage available at the fault location in the network during normal operation to get maximum and minimum voltage. The voltage values are used with most likely impedance for each component. In this method, an equivalent source voltage is taken which provides short circuit current towards the fault impedance and all other voltages are set to zero. IEC 60909 places particular emphasis on the initial symmetrical short circuit current (IK”), peak short circuit current (ip), symmetrical short circuit breaking current, and steady state short circuit current (IK). These values are illustrated in the figure 5. The initial short circuit current is the RMS value of the AC symmetrical component the moment the short circuit occurs. Symmetrical short circuit breaking current is the RMS value taken over a full cycle of the symmetrical AC component of the short-circuit current at the instant when the breaker operates. The peak short circuit current represents the highest instantaneous value of the available short circuit current. The key difference between the initial symmetrical short circuit current and the symmetrical short circuit breaking current is that the former decreases over time, particularly near generators. The steady state short circuit current is the RMS value of the short circuit current once transient currents have decayed. IEC60909 provides the formula for calculation of initial, peak, and symmetrical breaking short circuit current from full size converter interfaced power generation units. There is no formula for calculation of steady state short circuit current provided by DFIG and full-size converter. This data is based on design and test and is provided by the manufacturer. 13 Figure 5. Fault current according to IEC 60909 [11] The IEC 60909 steady state fault current calculation provides a good estimation in the SG dominated power system [12]. However, when fault current from a CIG, which is of different characteristics than that from SG, is injected, the standard might not be as efficient. In the past the fault currents from CIGs were usually ignored in steady-state fault current calculations as fault currents from such generators were limited and it didn’t affect the accuracy of estimation. But with increasing penetration of CIGs and grid codes adopting FRT requirements, the steady state fault calculation from such devices can’t be ignored anymore [13]. In recent version of IEC 60909, CIGs are recommended to be considered as a current source model according to its maximum overrating capability if its fault current is more than 5% of the total fault current [13]. In a current source model, a decoupled CIG is modeled as a current source, and it injects fault current according to a predefined value. But reference [12] shows that efficiency of this method might be dependent on location of the fault. This model was found more efficient when the fault location was near to the source. Reference [13] proposes a method to improve the accuracy of steady state fault current calculations of IEC 60909 utilizing full converters. The method proposes reactive current injection from CIG in accordance with FRT requirements. The study is based on modeling CIG as a current source model with infinite impedance. The model ignores the impedance of the fault path and assumes a balanced fault. The fault current calculation procedure is two steps where fault current is calculated separately without considering CIGs and considering only CIGs. In the first step, equivalent impedance is calculated at the faulty point using Thevenin equivalent method. Then, faulty point voltage, nominal pre-fault phase voltage is divided by the calculated impedance to get initial symmetrical current. For the second step where only CIGs are considered, the value of current source is obtained from the manufacturer of the generator. The transfer impedance is calculated between faulty buses where CIGs are connected. Then the fault current contribution from CIG is calculated by multiplying the maximum current from the current source with the transfer impedance and dividing the result by equivalent impedance calculated without considering CIGs. Two fault 14 currents from two steps are added to get total fault current. The proposed method for improving fault current calculations involves adjusting the value of the current source. Rather than using the fixed value supplied by the manufacturer, a variable value is assigned to CIGs based on the voltage dip and FRT and reactive current injection grid code requirements. The proposed steps are listed below: • The fault current calculated without considering RES and the impedance matrix is used to calculate the voltage distribution at buses during the fault. • The characteristic proportional gain or k factor and threshold voltage is obtained from grid code. It is also a voltage at which FRT control becomes active. The maximum combined current, which is the value of active and reactive current injection during a fault, is obtained from the converter manufacturer. • Then the current from each CIG is calculated using the calculated post fault voltage for the respective bus in the first step. Here, the maximum combined current must be higher than the injected reactive current that was calculated. • Then the total fault contribution from CIG can be calculated using calculated current value from the previous step rather than the one provided by the manufacturer. The rest of this step is as per the IEC 60909 standard. These steps give different fault current fed from the CIG depending on the voltage and reactive current injection and is used to calculate total fault current. The paper then implements the proposed method in power system models and compares the obtained result with the ones obtained from IEC 60909 standard. The IEEE 14 bus power system model is used for the study with the proportional gain value of 2 and 10. The study finds that the IEC standard calculates erroneous fault current calculations regardless of the fault location. However, the calculation results are acceptable if the fault is near the CIGs. The study concludes that this method can reduce the error in fault current calculation in high CIG penetration scenarios by 50%. The proper estimation of fault current calculation is pivotal for a proper design and coordination of protection system. Overestimation of fault current can lead to selection of oversized and more expensive equipment. The risk of fault going unnoticed cannot be ignored either. Underestimation can also compromise the reliability and safety of the power system. With the increasing share of renewables and CIGs in the grid, the fault current contribution from them must be integrated in compliance with the grid code requirements. Accurate fault current calculations are even more crucial for protection system design of CIG dominated grid. 15 5 Fault current study on a power system model This section is a study of how short circuit current level is affected by a CIGs using a power system model. The fault current available in transmission and distribution systems is studied and compared between the system rich in SGs and a system where some of the SGs have been replaced by CIGs. It gives a better perspective of whether the lack of short circuit current in the system could become a concern in the future. 5.1 Simulation tool A power system model with transmission, sub-transmission and distribution network has been used for the study of fault current with different levels of penetration of renewable energy. A model based on the Nordic44 power system model has been designed for the study. The Nordic44 power system model is a sophisticated framework used for analysis and management of power systems in the Nordic region. The Nordic44 model has been used by market operators to acquire key data on cost generation and transmission capabilities for energy trading and by the power system developers to study accommodation of new components including renewable energy. A simplified Nordic44 model overlapped on the map of Nordic region is presented in figure 6. The model used in this project has been developed by The Norwegian University of Science and Technology and is available in Power System Simulation for Engineers (PSS/E) [14]. PSS/E is a high-performance transmission simulation, planning and analysis software. It is a comprehensive set of programs for analysis of power transmission networks for steady state and dynamic simulations. Steady state short circuit current calculations are used for design of power system, selection of equipment and insulation, and protection coordination. Hence, in this thesis project, steady state simulation has been done in PSS/E to study fault current on different parts of the power system model. 16 Figure 6. The simplified Nordic44 model overlapped on the map of Nordic region [15]. 5.2 Model description The Nordic44 model covers the grid in Norway, Sweden, and Finland. Each country is divided into different regions and the model has 44 buses. But for sake of this project, the Norwegian and Finnish part has been removed and the study has been conducted in the Swedish transmission system only. The Nordic44 model is vast comprehensive representation of Nordic power system and the fault current contribution from Swedish generators should be enough in to understand short circuit capacity. The network reduction keeps the calculations straight forward. The model has 420 kV transmission and eleven 420 kV buses on the HV transmission level. The model has been divided into four regions, i.e., SE1, SE2, SE3 and SE4. The short circuit 17 current study in this project will be focused on SE3. The nuclear reactors which are connected to the 420 kV bus bar are in this region. Oskarshamn bus bar, marked bus number 3300 in the figure 7, has been expanded to include a transmission and distribution network. The distribution network consists of seven buses of 135 kV, 50 kV and 11 kV voltage level. Bus 13501 and 13502 are 135 kV, 5001 and 5002 are 50 kV busses, 1101, 1102 and 1103 are 11 kV busses. The expanded network has step-down transformers and medium voltage transmission lines. The distribution network is radial and does not have any distributed generators. The loads have been left as they were in the existing transmission and Five loads of 20MW and 5 MVAR each have been added to the distribution network. Figure 7. Added distribution network. The power lines are modelled as pi-sections with series resistance, reactance, and shunt capacitive line charging. The inter-country connection lines have been removed from the Nordic44 model. The impedance values of the added distribution line are presented in table 1. Table 1. Power line impedances Line voltage level Resistance (pu) Reactance (pu) Charging (pu) 135 0.001 0.0040 0.4 50 0.005 0.0015 0.02 The transmission network has one 420/135 kV transformer. Seven stepdown transformers have been added to the distribution network. Regular two winding transformers without tap changers used. Since our concern is only short circuit current available at the location of fault, tap changing is not a concern. Ratings and parameters of the transformer are listed in table 2. 18 Table 2. Transformer ratings Voltage Capacity (MVA) Resistance (pu) Reactance (pu) 420/135 1000 0.0005 0.120 135/50 63 0.0010 0.120 135/11 30 0.0010 0.100 50/11 30 0.0010 0.100 The model has several conventional generators. All the sources in Sweden including hydropower, nuclear, wind, CHP, etc. have been modeled as conventional source and connected to the transmission network. Since the goal of this simulation is to study fault current level in presence of CIGs, this model is perfect. The generators have been provided with sub transient reactance. Since the simulation is for balanced faults, the sequence data was not necessary for any components. The SGs that would be replaced by a converter connected generators are listed below in table 3. Table 3. Synchronous generators in the busses under study Connected bus Number of units Base MVA Sub-transient reactance Oskarshamn 3 1100 0.1600 Forsmark 3 1300 0.2250 Ringhals 6 1350 0.1937 5.3 Methodology Apart from removing the Norwegian and Finnish networks and adding a distribution network, the Nordic44 model was used as made available. In other words, the generators that were switched off were left as they were in the base case as well as the whole short circuit study. The only change made was in the 12 generators on Ringhals (3359), Oskarshamn (3300), and Forsmark (3000) buses. The generators that were feeding 4-6 p.u. fault current was modified to mimic replacement of nuclear and thermal plants connected to these busses with a wind farm. The sub transient reactance of the generators were set to a value that would only contribute fault current of 1.3 p.u. All the fault current calculations were done using the IEC 60909 standard methodology in PSS/E. Three phase faults were applied to different buses in the model to study the short circuit current level on them. Multiple assumptions were made to simplify the calculations which are discussed here: • Only three phase faults have been calculated in the bus bar. In symmetrical faults, calculations are simplified due to absence of sequence components and transformer grounding has less impact. Three phase faults are rare in the system but usually results in high fault current. 19 • The fault current calculated with IEC 60909 are initial symmetrical short circuit current (I”k) which is the initial current that occurs immediately after the inception of fault. • The tap ratios and phase shift angles have been left unchanged. Adjusting the tap setting changes the impedance of the transformer as seen by the fault and can change the fault current calculation. When the tap ratios and phase shift are unchanged, the transformer ratios are kept to nominal settings and retain the phase shift angles. It is essential to represent transformers phase relationships accurately. • Line charging, shunt and load have been left unchanged. These elements are parts of the fault current paths and influence reactive power and voltage profile during fault conditions. Including them rather than setting them to zero provides more accurate modeling of power system and can result in more precise calculation of fault currents magnitude and phase angle. • Voltage factor ‘C’ has been set to maximum (1 p.u.). This will simulate the worst-case scenario where the voltage is at upper limit at the inception of the fault resulting in highest possible fault current. In the first step, faults were applied individually to buses 3000, 3300, and 3359. The total fault current and the fault current contribution from each generator and the transmission lines connected to each bus were calculated. The same procedure was repeated after modifying the conventional generators to contribute a maximum of 1.3 p.u., simulating the fault current contribution from CIGs. For the distribution network, faults were applied to all buses when all generators were synchronous. In the next step, faults were applied again, but this time the generators on buses 3000, 3300, and 3359 were set to contribute only 1.3 p.u. fault current. 5.4 Results of simulation In the 420kV network, a significant drop was observed in all three buses under consideration when the fault currents were limited. In Oskarshamn bus, fault current reduced from initial value of 59.57 kA to 37.25 kA when the SGs in the bus were replaced and 31.97 kA when SGs on the three buses were replaced. The numbers represent about 62% and 53% of the fault current originally available respectively. Similarly, in Forsmark bus, fault current reduced from 61.10 kA to 44.39 kA when the SGs on the bus were replaced and 37.37 kA when SGs were replaced on all three buses under consideration. This represents about 72% and 61% of the fault current available with SGs respectively. But the largest drop in fault current level was observed on Ringhals bus where loss of up to 59% fault current was observed. The fault current dropped down from 72.71 kA to 30.05 kA and 29.26 kA. The results are presented in figure 8 and table 4 below. All the current values are in kilo amperes. 20 Figure 8. Short circuit currents on different 420 kV busses under different level of CIG penetration Table 4. Summary of short circuit currents(kA) on 420kV buses Oskarshamn Forsmark Ringhals Synchronous generator 59.57 61.10 72.71 CIG on respective bus 37.25 44.39 30.05 CIG on all busses 31.97 37.37 29.26 Percentage drop with CIG on respective bus 37.5% 27.4% 58.7% Percentage drop with CIG on all buses 46.3% 38.8% 59.8% In the radial medium voltage network, a similar trend was observed. In the 135 kV buses, bus 13501 and bus 13502, a drop of about 3% of fault current was observed. The fault current drop on the lower voltages were even less significant. The results are presented the figure 9 and listed on table 5. All the current values are in kilo amperes. Oskarshamn Forsmark Ringhals 0 10 20 30 40 50 60 70 80 420 kV buses S h o rt c ir cu it c u rr en ts ( k A ) Short circuit Currents Synchronous generator CIG on respective bus CIG on all three busses under study 21 Figure 9. Short circuit currents with SG and when all SG are replaced by CIGs in all three transmission level busses under study Table 5. Short circuit currents (kA) on distribution network Bus With SG With CIG Percentage decrease 3300 58.88 31.15 47.10% 13501 55.54 54.1 2.59% 13502 53.6 52.26 2.50% 5001 62.54 61.87 1.07% 5002 62.09 61.43 1.06% 1101 42.87 42.18 1.61% 1102 18.97 18.84 0.69% 1103 41.94 41.28 1.57% 5.5 Interpretation of simulation results On the transmission level, the drop in short circuit current level on buses close to the generators was as expected. But the significant amount of drop could affect the protection scheme that is based on the detection of high current magnitude. The reduction in fault current level reflects on fundamental change in how power system responds to faults. Miscoordination can be expected where circuit breakers and relays may not operate correctly or in proper sequence. On transmission level, where stability is of utmost priority, reduced short circuit capacity can affect voltage stability. At the same time, reduced inertia means lower critical clearing time pointing to the need to clear faults in an even shorter time. Transient stability of individual generators will experience larger voltage dips during faults and have to ride through more severe disturbances. Low short circuit capacity can also lead to 0 10 20 30 40 50 60 70 3300 13501 13502 5001 5002 1101 1102 1103 Short circuit current (kA) Bu ss es Short circuit current With CIG With Sycnhronous generator 22 lower system kinetic inertia leading to faster frequency deviations and lower resilience to disturbances. This change necessitates a re-evaluation of protection schemes, potential modifications to the grid infrastructure, and an understanding of the broader implications for grid stability and resilience. In the 135kV, 50kV and 11kV bus bars, the difference in fault current when switching from SG to CIG was less significant. This can be attributed to the fact that at 400kV, the impedance seen from the fault point to the source is lower and contribution of the source is directly reflected in the fault current response. In the lower voltage levels, the impedance between the fault point and the generator increases due to the presence of high impedance elements like transformers. This increased impedance tends to dominate the fault current calculation, reducing the impact of the source's contribution on the total fault current. As a result, the difference in fault current when switching from a SG to a WTG becomes less pronounced at lower voltage levels. In essence, the lower impedance at higher voltage levels means the generator's fault current characteristics has a greater impact, while at lower voltage levels, the network and step-down transformer impedance these differences less significant. On the other hand, a radial network-like figure 7 is not very common and some generators that add impedance and contribute fault current can be expected to be present at 135kV level and the lower voltage level as well. In any case, the replacement of SGs with CIGs at transmission, sub transmission or distribution can be expected to cause issues in fault detection and protection coordination. 23 6 Impact of converter interfaced generators on existing protection scheme The contrasting short circuit capacity in SG and CIG dominated grids have been established in the previous chapter. The existing protection schemes are designed based on the fault current characteristics of SGs. This section explores the impact of fault current contribution from CIGs in the protection scheme and their coordination. 6.1 Impact on protection of distribution network Existing protection systems in distribution networks are designed on the assumption that the power flows unidirectionally in the radial network. Hence, they are protected using current sensing devices like overcurrent relays, fuses and reclosures. These devices continuously monitor the current and generate a trip signal when the current level exceeds a predefined value. In the case of distributed generators or loop distribution networks, directional overcurrent relays (DOR) are used to avoid sympathetic tripping where a relay operates incorrectly due to fault in adjacent circuit. But with higher penetration of converter interfaced intermittent sources, changes in short circuit capacity and DORs with fixed settings cannot be relied upon [9]. A few issues introduced by the variable nature and weak infeed of fault current from the CIGs in protection coordination of distribution network with distributed generation are discussed below. • Blinding of protection: The CIGs can supply up to 1.5 p.u. fault current. As shown in figure 10, fault current sensed by relay R1, IfG, will be less than the total fault current due to the fault current, IfDG, from the distribution generator. Relays have a certain pickup value for breaker operation. Fault downstream of R2 will be supplied from grid and the DG. Fault current form DG will vary depending on its rating and impedance. Due to this varying fault current and change in impedance, R1 will face under reach issue. This phenomenon where sensitivity of R1 is compromised is known as blinding of protection [9], [16]. 24 Figure 10. Blinding of a protection in a distribution network • False tripping or sympathetic tripping: False, sympathetic, or spurious tripping occurs when a DG is connected close to a substation, supplies fault current to an adjacent feeder with common bus and the magnitude of the fault current exceeds the pickup value of the relay [16]. Relay R1 in figure 11 could trip and isolate the feeder from the grid due to fault in another feeder. The consequence of bidirectional power flow is visible in this example. In a meshed distribution system, some relays experience higher fault current than pickup values and trip before the primary relay and isolate unwanted feeder [9]. Coordination of overcurrent relays can be more complicated with penetration of more DGs. Figure 11. False tripping in distribution network in presence of DG. • Islanding: In the network in figure 11, if R1 senses enough fault current to trip, then it will isolate the DG from the grid. The DG will continue to supply the local load in an islanding operation. The islanding may face unbalanced operation with power imbalance, voltage and frequency fluctuations [17]. 25 6.2 Impact on protection of transmission network At the transmission level, overcurrent relays, distance relays, and differential protections are commonly used. Distance protection, with multiple zones, is the primary protection scheme for transmission lines. Several other backup protections are also provided in case the distance relay fails or does not detect high resistance faults. Directional overcurrent relays are a good back up option. It can be coordinated with adjacent line relays and can be set to detect fault in particular direction. The availability of communication channels makes it easier to use other backup protection like differential protection and overreaching and underreaching pilot protection schemes. Line differential protection is commonly used in short and medium length transmission lines. In longer transmission lines, pilot protection schemes provide faster backup protection with communication signals between line ends. Differential protection schemes are primary protection for generators and transformers. They are fast, sensitive and provide good selectivity. Restricted earth fault protection also provides high sensitivity for ground faults near neutral points of transformers. Over current relays can also be employed near line terminations for transformer protection. Generators have overcurrent protection for backup for overloads and external faults, and stator ground fault protection for ground faults in stator windings. Most of these protection schemes rely directly on measuring current amplitude, and with reduced fault current, their reliability can be compromised [9]. The reduced fault current magnitude, along with changes in phase angle and the nature (capacitive, reactive) of the fault current from CIGs, can significantly impact above mentioned protection schemes. The following section discusses the impact on some major protection schemes. 6.2.1 Distance protection Impedance-based protection systems with quadrilateral relay characteristics are more common. However, due to the low infeed and variable nature of CIGs, these impedance- based protections can cause underreach and overreach of relays, potentially resulting in faults going unnoticed within the relay's designated zone. Distance relays are also used in coordination with other types of relays, but low fault current can disrupt the cooperation and coordination among different relays [9], [18]. Figure 12. Distance protection scheme in presence of DG 26 In figure 12, the impedance ZAP for the fault location seen by the distance relay R1 is given by (1). ZAP = ZA + ZB + ( 𝐼𝐷𝐺 𝐼𝐺 ) ZB (1) As infeed current IDG depends on the input availability of the DG, the factor IDG/ IG is different for different instances. If the variable current infeed from the CIG is not considered while coordinating distance relay, the impedance seen by relay during fault will be reduced due to CIG current creating under reach. And if the infeed from the CIG is considered during coordination and fault occurs when the infeed is low, the distance relay may over reach [9]. Increasing penetration of CIGs may impact the operation of distance protection mostly due to reduced reach accuracy. Misoperation caused by low supervising current, misfunction of directional elements, incorrect selection of faulty phase are other issues that reduce the reliability of distance protection in presence of CIGs. The issues in operation of distance relay and its cooperation with other protection scheme are explained below. 6.2.1.1 Impedance and reach In a CIG dominated system, the impedance can change dynamically. This will cause complications in distinguishing faults in certain zones of protection. The intermittent nature of the source plays an important role in determining the reach of distance relay. SGs offer predictability, control, inertia, and voltage regulation, which stabilize the grid and reduce variations in impedance seen by distance protection relays. But fluctuating wind speed causes variations in voltage level which in turn changes the apparent impedance seen by the distance protection relay. Varying measurement of impedances causes changes in the reach of the distance relay settings [19], [20]. Type III WTGs are widely used with wind power due to its ability to work on unity power factor with power electronic converters, adaptability to variable wind speeds and reduced converter size. [21], [22]. In addition to the limited current capability of converters, other characteristics of CIGs, such as DFIGs, reduce fault current level in the grid. With the increasing number of RESs in the power system, new grid codes related to fault ride through (FRT) have been established, requiring DFIGs to remain connected to the grid during faults. This requirement leads to high current flow in the generator, risking damage to its rotor. To prevent such damage, a crowbar circuit is typically connected to the rotor. The crowbar circuit provides a path for the high short-circuit current, resulting in different fault current than normal operating conditions. During a three-phase fault, the crowbar resistance and rotor winding resistance combine to make the fault impedance very high. This change in impedance affects the reach of transmission line protection schemes, making it challenging for relays to detect faults accurately [21]. Impact of change of source impedance on reach of a distance relay has been discussed in [23]. Memory-polarized mho distance relays are common in transmission line protection. 27 These relays measure impedance based on pre-fault conditions. They use pre-fault voltage as a reference to accurately measure the impedance and detect faults along the line. The expansion of the mho distance circle is influenced by memory polarization, which depends on the amplitude and phase angle of the impedance of the source that lies behind the relay. When the source is a CIG with a varying internal impedance, this causes inconsistent initial expansion of mho circle [5], [23]. A larger ratio of source impedance to the line impedance causes greater expansion of the mho circle relay [23]. When a CIG is connected radially to a grid, it creates a weak end. A weak end is the remote end of a transmission line where source or generation is limited or nonexistent. Such an end has low voltage stability and fault current strength. When a CIG is supplying a fault current in line with weak end and there is variation in phase angle, the fault impedance has significant influence on the apparent impedance measured by the distance relay [24]. In the figure 13, Relay R1 measures voltage VA, and current IA and calculates the apparent impedance ZAP. ZA is the line impedance from the relay to the point of fault. If IDG is the fault current from the DG and RF is the fault impedance, the apparent impedance seen by the relay is given by (2). ZAP = 𝑉𝐴 𝐼𝐴 =ZA + ( 𝐼𝐷𝐺 𝐼𝐴 ) RF (2) The factor IDG/IA is dependent on the fault current source. The apparent impedance seen by the relay is significantly different from the actual impedance in case of a CIG. If there is a significant difference in phase angle between load current and fault current, a reactive part is added to the apparent impedance and it deviates towards X in the R-X plane [24]. In a SG dominated system, the pre fault and post fault current are almost identical. However, in CIG, some converter controllers can cause voltage phase jumps. All these characteristics complicates relay coordination regarding zone settings. Figure 13. A CIG connected radially to a grid. 6.2.1.2 Fault identification logic Different identification algorithms are used by distance protection relays to identify the type of fault and the faulted phase. In general fault identification logic, the faulted phase loop is identified using the mathematical relationship between zero and negative sequence current. According to [5], this mathematical relationship holds true in a SG dominated system because of the inductive nature of negative sequence network and SG impedances [5], [25]. 28 Traditionally, this was an effective and logical algorithm as negative sequence current was sufficiently supplied by the SGs. However, CIGs may not inject any negative-sequence current. Hence, in CIG dominated system, the mathematical relation might behave differently. This phenomenon has been studied in [5] where current based fault identification logic has been simulated in networks supplied by only SGs, Type IV CIG with no negative sequence current control and Type IV CIG with negative sequence current control according to German grid code. It was found that for a single line to ground fault in a SG system, negative sequence current lagged positive sequence current by -4 and the fault identification logic could declare the fault on the right phase. A similar relationship was observed for Type IV CIG with negative sequence current control. However, in Type IV CIG with no negative sequence current control, negative sequence current lead zero sequence current by -140. This caused fault declaration on the wrong phase i.e., fault was applied in phase A but declared on phase C due to the angel difference. In situations where negative and zero sequence current are not in sufficient amount, fault is identified using voltage components [25]. 6.2.1.3 Direction identification Directional elements use negative sequence current and negative sequence voltage to identify direction of the fault. This approach is effective in traditional power systems dominated by SG, where the negative sequence components are significant and consistent due to the inductive nature of the generators and the network. However, in systems with domination of CIGs, the fault current often lacks or has very limited negative-sequence components. Consequently, the reliability of directional elements is compromised in CIG-dominated systems, as they cannot accurately determine fault direction without the expected negative sequence components. This leads to potential mis operation and reduced effectiveness of directional elements. Operation of directional negative sequence overcurrent element used in distance relay to determine direction of fault is discussed in references [5], [25]. It operates by comparing the phase angle of negative sequence current and negative sequence voltage. This element operates on the assumption that the grid impedance, source impedance and, consequently, the negative sequence network is inductive in nature. The logic is that, ideally in forward faults, the negative sequence current leads to the negative sequence voltage by 90°. In case of a reverse fault, the negative sequence current lags the negative sequence voltage by 90°. This phenomenon is observed in a SG dominated network but, in a CIG, dominated network, where fault current can be resistive or capacitive as well, this assumption is invalid. 6.2.1.4 Communication schemes In the POTT scheme, relays at each terminal trip when they receive local and remote trip commands. Directional elements detect sensitive unbalanced faults, potentially supplementing or replacing ground distance elements in communication-based protection schemes [5]. However, if the directional negative sequence element misoperates, it can cause a cascading failure in communication-assisted protection schemes. When this element malfunctions, the distance relay will not send a permissive trip signal to the remote relay. This will cause the POTT scheme to mis-operate for faults within the protected zone [25]. 29 To illustrate this misoperation, a SLG fault was applied to a transmission line with POTT protection scheme in reference [5], [25]. The relays on the line end had ground distance relay and directional negative sequence overcurrent relay. In a scenario where the fault is supplied by a SG dominated system, The POTT operated successfully where ground distance relay and directional negative sequence overcurrent relay picked up in both end and successfully sent the permissive trip command to the remote relays. However, when the SGs were replaced by a Type IV CIG with no negative sequence current control, the directional negative sequence overcurrent relay of one end generated transient trip signals and could not send assertive trip command to remote end relay. The POTT scheme could not trip remote end relay for an in- zone fault due to change in phase angle of negative sequence current injected from Type IV CIG. When the sources were replaced with a Type IV CIG with negative sequence current control the POTT operated as expected. The negative sequence current control allows reactive negative sequence current injection and the relays can successfully send permissive trip signal to the remote relays [25]. 6.2.2 Overcurrent protection Overcurrent protection scheme is based on detection of current magnitude and is best to understand the impact of CIG on fault detection. Several factors like fault resistance and neutral ground impedance, a long line or presence of high impedance components like transformer can limit the fault current magnitude even in a conventional SG dominated network. Limited fault current from the CIG makes this scheme even less reliable. As overcurrent relays are based on inverse time characteristics, fault clearing time increases with low fault current. It creates an increased risk of equipment and personal damage. 6.2.3 Line differential protection Line differential protection is a communication assisted protection scheme popular in short length transmission lines where security is critical and use of other protection scheme is not enough [5]. Traditionally, line differential protection is activated when sum of amplitude of current entering and leaving the protected region of a transmission line are different than user defined value. However, other algorithms, like current-ratio plane, are also used by some vendors. Different types of CIGs inject fault current of different characteristics. If this difference is significant, fault current may fall in the wrong region of defined differential characteristics. In [5], a simulation study has been carried out to see how traditional and alpha plane line differential operate differently in presence of SG and CIG. In traditional differential protection scheme, the fault current from SG and CIGs don’t affect line differential protection if the communication channels are intact. In current ratio plane (alpha plane) differential protection however, protection was found to act differently for SG and CIG fault current. The This differential technology uses a complex ratio of terminal currents to determine differential value. The location of this complex ratio in the alpha plane determines the action 30 of the relay. When the Type IV WTG with no negative sequence current control was supplying the fault current, the alpha plane line differential scheme failed to detect the internal fault. This mis operation was due to low magnitude of fault current together with change in phase angle of negative sequence current from type IV CIG. Further simulation with negative sequence control suggests that this issue could be fixed when the negative sequence current is almost in phase. 6.2.4 Other protection schemes Asynchronously connected CIGs reduce the system inertia, and it impacts the reliability of system designed to operate in inertia rich system. Some protection schemes affected by loss of inertia in CIG dominated system are explained below. • Rate of change of frequency (ROCOF) ROCOF protection, a protection scheme used to avoid unintentional island formation in the grid, is one of the affected schemes with reduced inertia. A rapid rate of change of frequency suggests that an island is forming on the grid and the protection system should operate to isolate the island and protect the rest of the grid. The ROCOF settings has a wide range varying from 0.1 Hz/s to 2Hz/s depending on the standard defined on grid codes [5]. ROCOF is inversely proportional to system inertia. When a generator is lost in a system with low inertia, ROCOF may exceed the predefined settings. This could be interpreted as an islanding event by ROCOF relays and could trigger unnecessary tripping of CIGs. This single mis operation could create destabilization of the network [5]. • Power swing protection The variation of power flow induced by disturbances like switching of transmission line or transformers, load or generator throw off etc. are known as power swings. Such disturbances cause changes in voltages and currents. Under/over voltage protection and overcurrent relays may interpret these even as fault and trigger unnecessary tripping. Power swing protection is deployed to differentiate such disturbances from a fault and block unnecessary tripping. The rate of change of impedance is considerably slower during a power swing compared to a fault. This information used to differentiate between power swing and fault in the protection scheme known as power swing blocking [5]. Since CIGs in its current setting do not provide any inertia, the rate of change of impedance is likely to increase during power swing. This creates a case for increased CIGs reducing reliability of power swing protection. Power swing protection scheme also monitors the stability of the swing. In presence of CIGs, some stable power swings could be mis-interpreted as unstable and mis operation of the protection devices could occur [5]. 31 7 Commercially available techniques A range of relay models from different manufacturers are utilized by TSOs in power grid to protect and control their transmission and distribution networks. The choice of specific relay model depends on the application requirements, voltage level, technical specifications like capacity, switching time, size, etc., compatibility with existing equipment, Compliance with industry standards and certifications, reliability, and quality. Some popular manufacturers and their commonly used relay models are listed below. ABB: • ABB RED670: Numerical protection relay Siemens: • Siemens SIPROTEC4, SIPROTEC5: Numerical relays with distance, differential and overcurrent protection, etc. GE • GE Multilin UR series • GE Multilin D60 Line distance relay. Schneider • Schneider Electric MiCOM P Series Schweitzer Engineering Laboratories • SEL Series In the following section, the provisions for detection and fault clearing with low fault current is discussed. The discussion is based on the functions available on some of the relay models presented above. 7.1 Current reversal protection Fault current reversal protection (logic) is used when parallel lines are connected to common buses between a weak and a strong source. Overreaching permissive communication scheme can trip healthy line even when fault has been cleared on the faulty line. Such unwanted trip can be explained based on figure 14. Figure shows parallel transmission lines L1 and L2, connected between a strong source and a weak source. Transmission lines are equipped with relays R1, R2, R3 and R4 and breakers. For a fault in line L2 close to R4, as shown in the figure, fault current from the strong source will pass through R1, R2 and R3. Fault current from weak source will pass through R4 only. R1 and R2 will detect fault in forward direction and R3 will detect the fault in reverse direction. R4 will operate instantaneously, irrespective 32 of communication signals. R2 detects fault in its zone 2 but waits for permissive signal from R4 to trip. R1 also detects fault in its zone 2 and issues a permissive signal with a time delay to R3. Figure 14. Parallel transmission line between two sources After the opening of breaker near R4, the fault currents are redistributed. When R4 opens, the fault current that was passing through R3 will reverse its direction and the fault current from weak source will flow through L1 and to the fault in L2. R2 is still connected due to time delay. During this time if the communication signal from R1 to R3 is not reset, R3 will keep reading fault in forward direction and will trip unnecessarily. The false trip on unfaulted line due to current reversal can be prevented using a logic to detect current reversal. 7.2 Weak end infeed protection Operation of permissive communication scheme depends on whether the relay can detect sufficient minimum fault current. An open breaker on one end or a weak source can make fault current too low for a relay to detect. Fault current is initially distributed, causing low fault current in one end. Later when the fault is cleared in the strong end, fault current increases in the weak end. Then, the relay detects fault in zone 1 of the weak end and issues a trip command. Relays use weak end infeed (WEI) echo logic to tackle this issue. When the distance protection elements fail to detect the fault in weak end, WEI function of relays sends back the received signal to the stronger end. 7.3 Instantaneous residual overcurrent protection Instantaneous earth fault protection provides good selectivity and speed of tripping when the fault current is restricted to a specified level by impedance of some object [26]. In relays like RED670, this setting is dedicated to operating for the faults in the protected object only, ensuring selectivity. The residual currents that are used by the protection device are calculated with the help of SLG and LLG fault current on a line in a meshed system. The fault current calculation can be explained by the figure 15 and 16. Here, a transmission line with impedance ZL is shown in between two sources. ZA and ZB are source impedances on the 33 sides A and B respectively. R1 is a relay on the home end. Fault current is calculated for the fault at end B, a remote end, and end A. Figure 15. Fault current from bus A to bus B Figure 16. Fault current from bus B to bus A For fault current IfB at end B, ZA in higher source impedance operational state and ZB in lower source impedance operational state is used. Conversely, for fault current IfA at end A, the setup uses high source impedance operation state for ZB and low source impedance ZA. The function operates for a defined range of current. The minimum theoretical current setting (Imin) is given by (3). Imin ≥ Max(IfA, IfA) (3) However, in practice, the current setting is 30% above the theoretical value accounting for safety margin against static inaccuracy, possible transient overreach and inaccuracy of instrument transformers under transients and system data. A base current value is taken based on the capacity of a protected object and is used as reference for current setting. The percentage of base current is used for protection setting. 34 7.4 Four step residual overcurrent protection The four step residual overcurrent protection scheme, available on RED670, is used for sensitive earth fault protection of transmission lines. It is more sensitive for resistive earth faults compared to distance protection. This function offers up to four individually settable steps of flexible current operating levels and time delays. It can function in either directional or non-directional modes. A non-directional mode is used when a protected object cannot feed the fault. When selectivity and fast fault clearance is required, especially in a meshed and effectively earthed transmission system, directional mode is preferred. The directional function commonly uses voltage polarizing quantity which is decided in the settings. Selectivity between different overcurrent protection schemes is achieved through coordination of their operating times. This function in the relays also offer various time characteristics but for optimal coordination, all residual overcurrent protection schemes should use the same time characteristic. Therefore, multiple standardized inverse time characteristics, such as IEC and ANSI standards, are used. Various reset characteristics are also available in this protection scheme. Four step residual overcurrent protection should reset once fault current level is below operation level unless a delayed reset is desired. Current operating level can be also changed for some time using a setting of a multiplication factor. This function also has the setting of second harmonic restrain. This setting can be useful when there is a power transformer getting energized and a large inrush current flow in. Such inrush current can have residual current component accompanied by significant second harmonic component that create the risk of unwanted operation. A restrain signal can be created to prevent unwanted trip. The four step residual overcurrent protection can be used to protect a transmission line in a meshed and effectively earthed system. Relays measure residual current and uses a directional function for the protection of the line. Residual voltage is internally generated from three phase set of potential transformers and used as polarizing quantity. The four-step approach of this protection scheme helps in detecting low earth fault current by providing a structured and hierarchical response to varying fault conditions, including those with low current levels. The four steps are described below: Step 1: Directional Instantaneous function The relay calculates the fault on the remote busbar and sets the trip threshold above the calculated value. The relay adjusts for the condition if a line connected to the remote end is disconnected or a power transformer is disconnected. If a parallel line is present with mutual zero sequence impedance and a fault occurs on that line, the residual current can be greater than it would be for an earth fault on a distant busbar. The highest calculated residual current is taken for the current setting. This function is highly sensitive to immediate faults, preventing overreaching the protected lines. Step 2: Directional function with short time delay This function incorporates a time delay of about 0.4 second and detects earth fault on the line not covered by step 1. This function ensures that the entire line is covered and does not operate for faults on other lines connected to remote busbar. The trip threshold is set 35 considering minimal earth-fault current and ensuring selectivity with adjacent protection. By ensuring the entire line is covered, step 2 enhances the detection of faults with lower current levels that may not trigger an instantaneous response. Step 3: Directional function with longer time delay This function is usually set with a slightly longer time delay, often 0.8 second. It ensures selective tripping for earth faults with certain fault resistance and prevents step 2 from being triggered. The trip threshold is selected based on the settings in step 2 on the faulted line and maintains selectivity with other earth-fault protection in the network. This step ensures the low earth fault currents are selectively detected. Step 4: Non-directional function with long time delay This function detects and trips for high resistance earth and series faults. The current threshold is typically set around 100A. A predefined time delay of 1.2 - 2 second or current dependent inverse time delay characteristics is used for higher selectivity. The four-step approach in this protection, especially steps 3 and 4 are tailored to detect low fault currents. The directional function is useful in detecting direction of fault currents in case of distributed generation. The adjustable settings with current levels and time delays allow it to be tuned to the specific characteristics of CIGs. This flexibility ensures that the protection settings can be optimized for lower fault currents, enhancing the sensitivity and reliability of protection scheme. The standardized inverse time characteristics helps in coordination with other protection devices in the system, easing the process of prot