* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Fault Location Algorithms for Electrical Power Transmission Lines
Survey
Document related concepts
Telecommunications engineering wikipedia , lookup
Voltage optimisation wikipedia , lookup
Mains electricity wikipedia , lookup
Transmission line loudspeaker wikipedia , lookup
Power engineering wikipedia , lookup
Three-phase electric power wikipedia , lookup
Two-port network wikipedia , lookup
Stray voltage wikipedia , lookup
Ground (electricity) wikipedia , lookup
Immunity-aware programming wikipedia , lookup
Protective relay wikipedia , lookup
Amtrak's 25 Hz traction power system wikipedia , lookup
Alternating current wikipedia , lookup
History of electric power transmission wikipedia , lookup
Electrical substation wikipedia , lookup
Transcript
Fault Location Algorithms for Electrical Power Transmission Lines Methodology, Design and Testing Master of Science Thesis Shreya Parmar Intelligent Electrical Power Grids - EWI Fault Location Algorithms for Electrical Power Transmission Lines Methodology, Design and Testing Master of Science Thesis For the degree of Master of Science in Electrical Power Engineering at Delft University of Technology Shreya Parmar July 8th, 2015 Faculty of Electrical Engineering, Mathematics and Computer Science (EWI) · Delft University of Technology Delft University of Technology Department of Electrical Engineering The undersigned hereby certify that they have read and recommend to the Faculty of Electrical Engineering, Mathematics and Computer Science (EWI) for acceptance of a thesis entitled Fault Location Algorithms for Electrical Power Transmission Lines by Shreya Parmar in partial fulfillment of the requirements for the degree of Master of Science Electrical Power Engineering Dated: July 8th, 2015 Thesis Commitee : Dr. ir Marjan Popov Dr. ir A. Rodrigo Mor Dr. ir D. Jeltsema Dr. ir Gert Rietveld c Electrical Engineering Copyright All rights reserved. Abstract Electric power transmission lines are the veins which pump life into the modern day world, delivering electricity to consumers at their homes, offices and industries. It is important to ensure a smooth operation of transmission lines to deliver a minimally interrupted power supply making necessary for reliable operation of electrical power lines. This need has given rise to Fault Location detection techniques so that the economic impact of the fault situations can be mitigated and their correction can be rendered simpler and precise. Over the years, much research has been done on the various techniques for accurate fault location detection. Many methods use line data from one or more terminals for determining the fault locations. Providing a certain degree of accuracy, these methods lose out on the ease of detection as they may use line parameters extensively which are sometimes difficult to assess. To counter these problems, some techniques of fault location detection were developed on the basis of the available voltage and current measurements across the terminals of the faulted lines. These methods minimize the effects of parameters due to varying load and weather conditions. This report presents a comparison between two different approaches towards fault location detection with and without applying transmission line parameters. Firstly, an impedance based parameter dependent algorithm, derived by using the Modal Transformation theory and Discrete Fourier Transform is presented. The methodology has the ability to locate the fault whether it is on an overhead line or an underground power cable. The second algorithm is an improved parameter-independent fault location method that only uses time synchronized measurements. The unknown fault location will be determined from voltage and current phasors, measured at both line end terminals. This approach of fault detection renders the prior knowledge of line parameters as obsolete, which is of great assistance to grid technicians and engineers. The thesis report presents the results of the algorithms tested through the use of EMTP/ATPDraw and RTDS/RSCAD v2.025. The results of the line parameter-independent algorithm will be compared with those provided by the parameter-dependent algorithm. Master of Science Thesis Shreya Parmar ii Both algorithms are tested for the most common to occur asymmetrical faults with varying fault resistances. The results of the tests are then related to gauge the effectiveness and accuracy of these algorithms. Shreya Parmar Master of Science Thesis Table of Contents Preface ix Acknowledgments xi 1 Introduction 1-1 Importance of Fault Location . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2 Thesis Overview and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . 2 Fault Location Techniques 2-1 Fault Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2 Fault locators and Relays . . . . . . . . . . . . . . . . . . . . . . . 2-2-1 Fault Locators . . . . . . . . . . . . . . . . . . . . . . . . . 2-2-2 Fault Locators vs Relays . . . . . . . . . . . . . . . . . . . . 2-2-3 Distance Relays . . . . . . . . . . . . . . . . . . . . . . . . 2-3 Fault Location Algorithms . . . . . . . . . . . . . . . . . . . . . . . 2-3-1 Impedance based Algorithms . . . . . . . . . . . . . . . . . 2-3-2 Travelling Wave based Fault Location Algorithms . . . . . . . 2-3-3 Knowledge based Fault Location Algorithms . . . . . . . . . 2-3-4 High Frequency component based Fault Location Algorithms . 2-3-5 Synchronous Sampling in Fault Location Algorithms . . . . . 2-3-6 Time and Frequency Domain Analysis . . . . . . . . . . . . . 1 2 2 . . . . . . . . . . . . 5 5 9 9 9 10 11 12 12 13 14 15 16 3 Transmission Line Fault Location Algorithms 3-1 Line Parameter Dependent Algorithm . . . . . . . . . . . . . . . . . . . . . . 3-2 Line Parameter Independent or Parameterless Algorithm . . . . . . . . . . . . 19 19 22 Master of Science Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shreya Parmar iv Table of Contents 4 Modelling Methodologies 4-1 Network Modelling . . . . . 4-2 Source Modelling . . . . . . 4-3 Transmission Line Modelling 4-4 Fault Control Modelling . . 4-5 FLA Block . . . . . . . . . . . . . . 27 27 27 29 30 32 5 Simulation Results for different fault types 5-1 EMTP Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2 RTDS Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 37 45 6 Conclusions 6-1 Result Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2 Future Work and suggestions . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 49 51 51 A Simulation softwares used A-1 MATLAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2 EMTP/ATPDraw . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 53 53 A-3 RTDS/RSCAD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 B Symmetrical components and Sequence Networks B-1 Conversion Between the Phase and Symmetrical Component Domains . . . . . B-2 Sequence Networks for Faults . . . . . . . . . . . . . . . . . . . . . . . . . . 57 57 58 C Phasor Measurement Units C-1 Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C-2 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C-3 Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 61 62 62 Glossary Shreya Parmar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Master of Science Thesis List of Figures 1-1 Basic layout of power transmission lines. Courtesy: Institute for Energy Research 1 2-1 2-2 2-3 2-4 2-5 2-6 Classification of fault types in electrical power systems. . . . . . . . Single line to ground Fault . . . . . . . . . . . . . . . . . . . . . . Line to Line Fault . . . . . . . . . . . . . . . . . . . . . . . . . . Double line to ground Fault . . . . . . . . . . . . . . . . . . . . . An example of protections zones offered by a typical distance relay. Types of Fault Location Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 7 7 8 10 11 3-1 3-2 3-3 3-4 Three phase representation of a faulted line . . . . . . . . . . . Symmetrical Components Analysis for unbalanced faults . . . . Sequence circuits of the three phase faulted line respectively. . . Flowchart depicting approach towards fault location algorithms . . . . . . . . . . . . . . . . . . . . . . . . 20 22 23 25 4-1 4-2 4-3 4-4 4-5 4-6 4-7 4-8 EMTP-ATPDraw model used for simulation tests. . . . . . . . . . . . . . . RTDS-RSCAD model used for simulation tests. . . . . . . . . . . . . . . . EMTP-ATPDraw and RTDS-RSCAD source model used for simulation tests. EMTP-ATPDraw and RTDS-RSCAD line model used for simulation tests. . EMTP fault fault control switch used for simulation tests. . . . . . . . . . RTDS-RSCAD fault control block used for simulation tests. . . . . . . . . . EMTP algorithm MODEL used for FLA implementation. . . . . . . . . . . RSCAD CBuilder used for FLA implementation. . . . . . . . . . . . . . . . . . . . . . . 28 28 29 30 31 31 32 33 5-1 Single line diagram for FLA simulation tests. . . . . . . . . . . . . . . . . . . 5-2 Simulation waveforms for SLG fault. . . . . . . . . . . . . . . . . . . . . . . 35 37 Master of Science Thesis . . . . . . . . Shreya Parmar vi List of Figures . . . . . . . . 38 39 42 43 44 46 47 48 A-1 ATPDraw User Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2 RSCAD Power system user library . . . . . . . . . . . . . . . . . . . . . . . 54 55 B-1 Asymmetrical fault sequence network. . . . . . . . . . . . . . . . . . . . . . B-2 Three phase fault sequence network. . . . . . . . . . . . . . . . . . . . . . . 58 59 C-1 PMU placement in power systems 62 5-3 5-4 5-5 5-6 5-7 5-8 5-9 5-10 Simulation Simulation Simulation Simulation Simulation Simulation Simulation Simulation Shreya Parmar results for SLG fault in EMTP waveforms for LL fault. . . . results for LL fault in EMTP waveforms for LLG fault. . . results for LLG fault in EMTP results for SLG fault in RTDS results for LL fault in RTDS . results Parameter independent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . for LLG fault. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Master of Science Thesis List of Tables 2-1 Statistics for fault types on transmission lines as per the Nordic grid Report 2013. 6 2-2 Artificial Intelligence used in Power Systems . . . . . . . . . . . . . . . . . . 14 2-3 Comparison of Frequency and Time domain Analysis . . . . . . . . . . . . . 16 5-1 Network Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2 Line parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 36 5-3 Simulation result for SLG Fault - EMTP/ATP 60 kms . . . . . . . . . . . . . 37 5-4 Simulation result for SLG Fault - EMTP/ATP 100kms . . . . . . . . . . . . . 39 5-5 Simulation result for LL Fault - EMTP/ATP 60kms . . . . . . . . . . . . . . 40 5-6 Simulation result for LL Fault - EMTP/ATP 100kms . . . . . . . . . . . . . . 41 5-7 Simulation result for LLG Fault - EMTP/ATP 60kms . . . . . . . . . . . . . . 41 5-8 Simulation result for LLG Fault - EMTP/ATP 100kms . . . . . . . . . . . . . 43 5-9 Simulation result for SLG Fault - RTDS 100kms . . . . . . . . . . . . . . . . 5-10 Simulation result for LL Fault - RTDS 100kms . . . . . . . . . . . . . . . . . 5-11 Simulation result for LLG Fault - RTDS 100kms . . . . . . . . . . . . . . . . 45 46 47 Master of Science Thesis Shreya Parmar viii Shreya Parmar List of Tables Master of Science Thesis Preface The purpose of this thesis is to examine the application of impedance-based fault location methods for transmission lines using phasor measurements and test them in well known transient analysis softwares, so that their effectiveness can be assured. Different types of faults are applied to a transmission line model to analyze the accuracy of the two algorithms used for fault location detection. The first algorithm is based on the Telegrapher’s equation. This algorithm is dependent on the line parameters used and hence, the thesis project explores a second fault location algorithm which is line parameter independent and has the capability to locate faults on transmission lines, without requiring any line parameters and just the voltage and current measurements at the line end terminals. The report begins with a discussion on the importance of fault location algorithms in its introduction in Chapter 1. Then Chapter 2 reviews the basics of fault analysis and deliberates over the commonly used approaches to fault location detection. This chapter revises the popularly used relay protection schemes and how they can be deficient in accurate and quick fault location detection. A short description on the basics of the different types of fault location methodologies is provided to acquaint the readers with the developments and research in the vast topic of fault location algorithms. Chapter 3 develops two impedance based algorithms used in this project and gives a detailed derivation for both. The first algorithm uses line impedances and critical measurements at the line end terminals, where as the second algorithm uses only line terminal measurements. The simulation modelling and methodologies are elaborated in Chapter 4 citing the investigation established with practical use of the EMTP, RTDS and MATLAB simulation platforms. A transmission line model in a two source network is simulated on EMTP-ATP and RTDS and confirmed with MATLAB to validate the effectiveness of fault location estimations in Chapter 5. The results are included for a line length of 60 and 100 kilometers respectively for the most common asymmetrical fault types, accounting for effects of fault resistances. Lastly, Chapter 6 gives a conclusion to the report, as the analysis and summary are identified here. This chapter also discusses the future research to utilize the results from this thesis project. Master of Science Thesis Shreya Parmar x Shreya Parmar Preface Master of Science Thesis Acknowledgments I would like to extend my sincere thanks to my advisor, Dr. Dipl.-Ing. M. Popov, for his assistance, guidance and encouragement throughout my research. I take this opportunity to express gratitude to all of the members of my committee, Dr. ir A. Rodrigo Mor, Dr. ir Gert Rietveld from VSL and Dr. ir D. Jeltsema for their recommendations. This thesis is dedicated to my parents and all my friends whom I had the immense pleasure of knowing during my graduate study at TU Delft. Special thanks goes out to Anuj Shah, who was my rock throughout my Master study at Delft. Delft, University of Technology July 8th, 2015 Master of Science Thesis Shreya Parmar Shreya Parmar xii Shreya Parmar Acknowledgments Master of Science Thesis “What is a soul? It’s like electricity - we don’t really know what it is, but it’s a force that can light a room. ” — Ray Charles Chapter 1 Introduction Overhead transmission lines are the most convenient means for transportation of electrical energy from sources of generation to load centers so as to result in energy use and consumption. A pictorial representation of their use is depicted in Figure 1-1. Deregulation of the electricity market, economic and environmental requirements have pushed electrical utilities to operate transmission lines close to their maximum limits. Smooth operation of electric power transmission lines is essential to deliver minimally interrupted power supply to consumers, who have become more and more sensitive to power outages with the growth in worldwide technology. This necessitates reliable operation of power equipment and satisfaction of consumers. Engineers are hence pushed to design transmission networks that formulate power system protection schemes to reliably detect and isolate faults compromising the security of the system. Figure 1-1: Basic layout of power transmission lines. Courtesy: Institute for Energy Research Transmission and distribution lines experience faults that are caused by nature such as storms, lightning, snow, rain, insulation breakdown and short circuit faults caused by birds, tree branches and other external objects. In most cases, electrical faults manifest themselves as mechanical damage, which must be repaired before returning the line to Master of Science Thesis Shreya Parmar 2 Introduction service. Any fault, if not detected and isolated quickly will thus, grow into a system wide disturbance causing widespread outages and even subsequent blackouts. When a fault occurs on a power system, financial losses can be reduced and the line service can be maintained if the location of the fault can be accurately determined, especially when generation, transmission and distribution occurs over a longer distance or area, thereby improving the security and quality of the energy supply. Electric utility services can thus be maintained if the location of the fault on a line can be accurately determined. These deductions and suitability factors make fault locators and their algorithms, a critical tool in maintaining the competence of smart grids as is also mentioned by Kezunovic in [1]. 1-1 Importance of Fault Location Transmission and distribution lines experience both temporary and permanent faults. Temporary faults are mostly self cleared and permanent faults can be detected and mitigated with the help of traditionally available protective relay equipments. When a line is taken out of service because of faults, the connected loads are not supplied and sometimes, other lines are forced to supply the loads of the faulted line. It is also possible that a series of cascading trips of a protective relay occurs, faulting successively larger parts of the system, resulting in large power system blackouts. An example of such a situation is the recent blackout in The Netherlands’ northern grid, leading to no power in almost 1 million Dutch households [2]. Restoration of power supply after permanent faults can be done only after the responsible maintenance team finishes the repair of the damage caused by the fault. For this purpose, the fault position has to be known, otherwise the whole line has to be inspected to find the damage origin. This task becomes even more tedious, if high voltage transmission lines, running upto hundreds of kilometers are considered. Underground lines and cables have to be uncovered from under the ground, requiring more manpower and machines, and in populated areas, roads and passageways have to be blocked and dug. Thus, it is important that the location of a fault is either known or can be estimated with a reasonable accuracy. This allows saving of both money and time for the inspection and repair work, and aids towards better service by utilities due to the possibility of faster restoration of power supply. 1-2 Thesis Overview and Objectives In conjunction with the topics discussed in the preceding sections, fault location algorithms (FLAs) for transmission line faults form the backbone of this thesis project. The motivation of this project is to detect and determine the location of various types of faults on a Shreya Parmar Master of Science Thesis 1-2 Thesis Overview and Objectives 3 transmission line model, while considering both accuracy and speed. The main aim of the thesis is to implement the algorithms in EMTP and real time simulator RTDS and test for compelling results. There are various methods of fault detection, and this report focuses on impedance based fault location techniques. The thesis report aims to first familiarize the reader with the basics of fault analysis for electric power transmission lines, using available research from the past years. The successive chapters present a review on the most commonly and widely used fault location techniques for line protection. Thereafter, the report focuses on two fault location algorithms which use measurements made from line end terminals to locate fault distances . A detailed design, methodology and simulation tests for the algorithms are presented. Lastly, a comparison is made based on the simulation results of the two algorithms. The simulation studies are based on the EMTP/ATPDraw [3] and RTDS [4] platforms. Conclusions based on the results are discussed, along with the factors affecting the accuracy of the algorithms. The author also suggests future work for the research of fault location techniques and their simple integration into Phasor Measurement Units (PMU) and Synchronized Measurement Technology (SMT) technologies to make the electric utilities and grid system smarter and more reliable. Master of Science Thesis Shreya Parmar 4 Shreya Parmar Introduction Master of Science Thesis Chapter 2 Fault Location Techniques This chapter presents the basics of fault analysis and fault locators. Distance relays are analyzed with respect to their effectiveness in fault location detection and a comparison is made with conventional fault locators using numerical algorithms. A comprehensive review of all of the different methods of fault location on power systems is presented in this chapter. The advantages and disadvantages of each method of fault location detection and techniques for deriving them is discussed, as is the need for special measures to ensure the accuracy of the results obtained from the fault locators under certain network topologies or fault conditions. 2-1 Fault Analysis Figure 2-1: Classification of fault types in electrical power systems. Master of Science Thesis Shreya Parmar 6 Fault Location Techniques A fault is an interruption to the normal flow of current in a circuit. Large currents flow across the lines in a faulted condition, which results not only in financial losses to the suppliers and inconvenience to the customers, but also in severe cases, a complete shutdown of the grid supply. According to the Nordic Grid report of the year 2013 and the data presented in IEEE journal [5], the most common to occur type of fault in electrical transmission lines remains to be the line to ground fault, but the most severe faults are still of three phase nature. The Table 2-1 summarizes the same: Table 2-1: Statistics for fault types on transmission lines as per the Nordic grid Report 2013. Type of fault Nature Percentage occurrence Single Line-to-ground–SLG unbalanced 85% Line-to-Line–LL unbalanced 8% Double Line-to-ground–LLG unbalanced 5% Triple Line–LLL balanced 2% There are broadly two types of faults in transmission lines – transient and permanent faults, as is shown in Figure 2-1. A transient fault is no longer present if power is disconnected for a short time and then restored. Many faults in overhead power lines are transient in nature and power system protection devices operate to isolate the area of the fault, clear the fault and then the power-line can be returned to service. Typical examples of transient faults include: • momentary tree, bird or animal contact • lightning strike • conductor clashes A permanent fault can cause lasting damage to the transmission lines. To counter a permanent fault, the line first has to be isolated and then correction has to be made to the line. Some examples of the fault of permanent nature are: • direct lightning stroke on line • man-made damage • mechanical damage due to environment and age Shreya Parmar Master of Science Thesis 2-1 Fault Analysis 7 A symmetrical or balanced fault on a line affects each of its three phases equally. In transmission line faults, roughly 3-5% are symmetric in nature as seen in Table 2-1. This is in contrast to an asymmetrical or unbalanced fault, where the three phases are not affected equally. Common types of asymmetric faults, and their causes are: • line-to-ground fault (Figure 2-2) - a short circuit between one line and ground, often caused by physical contact, for example due to lightning or other storm damage. Figure 2-2: Single line to ground Fault • line-to-line fault (Figure 2-3) - a short circuit between lines, caused by ionization of air, or when lines come into physical contact, for example due to a broken insulator. Figure 2-3: Line to Line Fault • double line-to-ground fault (Figure 2-4) - two lines come into contact with the ground and each other, commonly due to storm damage. Master of Science Thesis Shreya Parmar 8 Fault Location Techniques Figure 2-4: Double line to ground Fault Shreya Parmar Master of Science Thesis 2-2 Fault locators and Relays 2-2 2-2-1 9 Fault locators and Relays Fault Locators A fault locator is a system designed for locating a fault with the highest possibly accuracy. It mainly augments the protection equipment, which applies the fault-location algorithms for estimating the distance to a fault and can be used for fault type identification and fault impedance calculations as well. Their implementation can be as stand-alone devices or as a part of a microprocessor relay, or in a offline fault analysis programs. The analysis and computation segments of the fault locators form the basis of its fault location and detection technique. This objective is achieved by analyzing the critical information provided to the fault locator with the help of mathematical manipulation and deriving an accurate result based on a plethora of fundamental techniques suitable to the system and resources being used. Thus, fault location algorithms are employed in the fault locators, which use various techniques to arrive at the single goal of locating a fault on the line from a reference, with consistent accuracy and ease. Fault locators when used with conventional protection schemes usually use the line data from typical recording equipments like digital relays and Intelligent Electronic Devices. Fault analysis is usually established with the bus phase and line voltage and current data, but it can also utilize information on circuit breaker status. 2-2-2 Fault Locators vs Relays Formulated to provide protection from faults in the transmission lines, there exists some important differences between fault locators using algorithms and the conventional fault location by relays. The points highlighted below are vital to the growth of fault location techniques and fault locators, as they highlight the insufficient nature of the conventional line protection devices and relays. • Accuracy – Protective relays usually define a general area known as a protective zone where the fault might have occurred. Fault locators pinpoint these areas with a certain percentage of error, if any. • Speed – Protective relays require very high speed operations to mitigate the spreading of faulted current to other parts of the power network, making use of circuit breakers and high speed communication devices, sometimes, sacrificing the relay system security and selectivity. Fault locators on the other hand use algorithms that calculate fault locations in several seconds, if not in minutes. • Data window – Relays use a fault interval between the inception of the fault to the clearing by a breaker, which takes several fundamental frequency cycles, resulting in a wide data window. Fault locators use the most compatible data windows to minimize Master of Science Thesis Shreya Parmar 10 Fault Location Techniques the scope of errors in calculations. However, this is of a considerable importance as it may affect the results of the algorithm. • Complexity of calculations – The high speed operation of protective relays render the calculations to be simpler. Fault locators do not posses any limitations of complexity. They can be relatively simple (based on impedance of line, or travelling wave methods) and may even increase in their complexity to incorporate a more versatile operation (use of neural networks and fuzzy logic, use of PMUs etc). 2-2-3 Distance Relays This section explains the basic working principles of a distance relay. Distance relays are most closely related to fault locators. They work to locate faults occurring in particular protective zones, as shown in Figure 2-5, making them the most common form of protection of transmission lines. Figure 2-5: An example of protections zones offered by a typical distance relay. The distance relay operation depends upon the predetermined value of voltage to current ratio which is impedance. The relay will only operate when this ratio becomes less than its predetermined value or impedance. Distance relays are placed in a particular zone and usually communicate with the main system and other relays using the pilot relaying scheme. As the impedance of a transmission line is directly proportional to its length, it can be concluded that a distance relay can only operate if fault has occurred within a predetermined distance. Distance relays are most useful for reasonable line lengths (upto 20 kms or so) because their operating characteristics are based on the line parameters which proves to be a major flaw, owing to the capricious nature of line parameters. Shreya Parmar Master of Science Thesis 2-3 Fault Location Algorithms 2-3 11 Fault Location Algorithms Traditional line fault detection used to heavily rely on visual inspections of the faulted line parts resulting in long and tedious foot or aerial patrols. These methods were expensive and prone to more errors. Thus, the shift to automatic fault locators was not only desired, but also natural. Fault location techniques can be generally classified into the following main categories: • based on fundamental-frequency currents and voltages, mainly on impedance measurement • based on traveling-wave phenomenon • knowledge-based approaches • based on high-frequency components of currents and voltages generated by faults The most widely used FLAs can be split into two main groups: impedance based and travelling-wave based. Impedance-based algorithms make use of the line parameters (such as resistance, inductance and conductance per unit length, and the line length) as well as voltage and current data from one or more line terminals to calculate the distance to the fault from a reference point or line terminal. Travelling-wave based algorithms utilize the theory that waves travel along a line from a fault at the speed of light to calculate the distance to a fault from a reference point and timing wave reaching the line terminal. More details on the types of fault location detection techniques are presented in the subsequent sections. Figure 2-6: Types of Fault Location Algorithms Master of Science Thesis Shreya Parmar 12 Fault Location Techniques 2-3-1 Impedance based Algorithms Impedance-based FLAs calculate fault distance using the per unit length impedance of the line, voltage and current data, and circuit analysis techniques, such as Kirchhoff’s voltage and current laws. Single-terminal and two-terminal algorithms are the two main groups of impedance based FLAs . Single-terminal FLA (STFLA), often cited as the first and earliest class of FLA use voltage and current data from one end of the transmission line only. They determine the fault distance by calculating the impedance of the transmission line as seen from one line terminal and then using the line parameters to convert that value into a distance measurement. To implement this, a microprocessor-based relay or other measuring device can be used to measure the voltages and currents at the local line terminal and then produce an estimate of the fault distance based on those measurements. The main advantage of using STFLA is that no communication of data is needed between line terminals. They are simpler to construct and the results require less computations making them fast. Two-terminal FLAs (TTFLA) tend to be more accurate since they are unaffected by the resistance of the fault or other factors that reduce the precision of STFLAs as shown in [6], [7] and [8]. The positive-sequence components can be used as opposed to zero-sequence components, which is advantageous as zero-sequence components can be harder to assess. TTFLA utilize similar methods from STFLA. Some disadvantages of using TTFLAs are as listed below: • require a means of gathering the data from line terminals at two locations before it can be analyzed. • takes longer to locate the fault due to their complexity. • implementation requires more measuring devices, communication channels and data storage devices. Impedance based fault location algorithms, unfortunately require very accurate line parameters and sequence impedance data. Line parameters also tend to vary with fault conditions and this tends to distort the accuracy of this particular implementation. 2-3-2 Travelling Wave based Fault Location Algorithms Travelling-waves occur after faults, switching, or lightning strikes. When a fault occurs along a transmission line, the voltage and current transients will travel towards the line terminals. By wave reflection theory, these transients will continue to bounce back and forth between the fault point and the two terminals for the faulted line until the postfault steady state is reached. Since the travelling-waves move along the transmission line at the speed of light, by accurately measuring the time taken for the travelling-wave to Shreya Parmar Master of Science Thesis 2-3 Fault Location Algorithms 13 propagate to the line terminals, the distance to the fault can be effectively found. An advantage of this method over impedance-based techniques is persistence to pre-fault line loading, fault and grounding resistance. Disadvantage of Travelling-wave Fault Location Algorithms (TWFLAs) is that they cannot be used on transmission corridors consisting of overhead lines and underground cables as the surge impedance changes drastically between them, resulting in large inaccuracies in the location of the fault. The accuracy of TWFLAs can be affected by errors in detection of the waves. Strong buses on the power system network influencing the voltage and current waveforms due to line impedances can reduce the amplitude of voltage waves making them harder to detect, and thus reducing the accuracy of the FLA. Single ended TWFLA use wave sensors at one line terminal as it does not require any form of communication between the line terminals, where as two-ended TWFLAs work by recording the exact time that the transient travelling wave from the fault reaches each line terminal where accurate timing is usually achieved by using a GPS system. Two-ended TWFLAss are known to be more accurate than their single-terminal counterparts and do not require as much signal processing at the sensors since they do not have to correlate the reflected waves to their original waves as they work by simple timing. However, two-ended TWFLAs are more expensive than single-ended TWFLAs because of their requirement for a communications link and synchronization between the two line terminals. This communications link and its inherent problems also make two-ended TWFLAs less reliable and less robust than single-ended TWFLAs. A high sampling time for computations and expensive implementations add to the overall demerits of using TWFLAs. More information on travelling wave based FLAs can be found in [9], [10] and a combined Impedance based and travelling wave based fault location is established in [11]. 2-3-3 Knowledge based Fault Location Algorithms Uncertainty of line parameter affecting variables, such as length of cables and unknown fault resistance, coupled with the complex structure of distribution management systems tends to make fault location through impedance and travelling wave techniques inaccurate. As a result of this, knowledge-based technique for locating faults have receiving attention from researchers in the last few years. In general, the technique requires information such as substation and distribution switch status, line measurements, atmospheric conditions, and information provided by fault detection devices installed along the distribution feeders. This information is analyzed using artificial intelligence methods to locate a fault. The three major knowledge based techniques used in power systems are based on the following: • Expert Sytem Techniques (XPS)– mainly used in power-system automation and control to decipher off-line tasks such as settings coordination, post–fault analysis and fault–section identification. Master of Science Thesis Shreya Parmar 14 Fault Location Techniques • Artificial Neural Networks (ANN) – well known for solving many engineering problems related to classification and optimization. Its ability to recognize complex patterns has made it possible to use in locating a fault where the training input can be from measurement data such as voltage, current, status of circuit breaker and feeder and the target output is the location of fault. • Fuzzy Logic systems (FL) – based on the theory of fuzzy sets that was developed for a domain in which definitions of activities and observations are fuzzy, or not well-described, without sharp boundaries. A table summarizing the key differences in the knowledge-based fault location methods used in power system is given in Table 2-2. Table 2-2: Artificial Intelligence used in Power Systems Approach Feature XPS Knowledge Used expert level ANN FL information information extracted extracted difficult convenient Troubleshooting and changes required, improving a relay but possible Self learning possible possible possible Robustness easy difficult easy Integration with a relay convenient difficult convenient Computations extensive dedicated hardware moderate required 2-3-4 High Frequency component based Fault Location Algorithms High-frequency transient signals generated in the range of Hz to kHz due to fault conditions can be applied to achieve high accuracy in fault location as shown in the work established in [12], [13]. This method of fault location detection, based on the high frequency voltage and current components, has been shown to be immune to power frequency phenomena such as power swings and current transformer saturations. This method mainly uses the fault-generated high-frequency signals, negating the problem of identifying multiple reflections of the travelling wave from bus-bars and the fault point, Shreya Parmar Master of Science Thesis 2-3 Fault Location Algorithms 15 as is seen in the travelling wave based FLAs in Section 2-3-2. Problems associated with fault-inception angle are addressed as the high-frequency signals associated with the fault arc do not vary with the point on the wave at which the fault occurs. In the scheme described in [13], a high speed sampling system is used to capture the fault-generated high-frequency transients. The high frequency components are used to identify the path of the fault (line or ground) and then compute the fault location along the identified path based on the power-frequency signals. The multi-phase transient signals are first decomposed into modal components. Then, the modal signals are decomposed into their wavelet components and the corresponding wavelet coefficients are obtained. These wavelet coefficients are used to extract the relevant signal features, which are subsequently used to identify the branch or path where the fault is located. Finally, the fault distance from the main substation is calculated using the information provided by the power-frequency signal. This technique for fault location detection is not widely used as the method is considered expensive and complex, since use of specially tuned filters for measuring high-frequency components is required. 2-3-5 Synchronous Sampling in Fault Location Algorithms With decrease in hardware costs and increase in the use of microprocessors in computations, newer and more accurate algorithms that utilize voltage and current from both the ends of the line are being developed. Digitization of power system components has made way for digitization of critical bus data such as bus voltages and line currents samples. Hence, synchronized sampling which refers to sampling of various parameters of the system synchronously, is gaining popularity as a means of transmission line data acquisition. Time synchronization allows synchronized real-time measurements of multiple remote measurement points on the grid and the algorithm uses the SMT which is an emerging concept in power system engineering [14]. Synchronized measurements makes the algorithm settings-free or less dependent on the line parameters. PMUs are considered to be the most suitable platform to implement SMT for protection of transmission lines as the paper [15] discusses the same. Advantages of using FLAs based on synchronized data sampling are: • they are not sensitive to changes in loading or environmental conditions. • they are not affected by conditions on other parts of the power system. • they are not affected by the fault resistance, mutual coupling in parallel lines. If synchronization devices are not available during measurements, the same can be established by detecting the phase shift in the time sampling of the data and then collectively applying the difference between the standard values to achieve a homogeneous result. Master of Science Thesis Shreya Parmar 16 Fault Location Techniques 2-3-6 Time and Frequency Domain Analysis The development of all the fault location types and techniques mentioned in the Section 2-3 is usually established in two domains– time or frequency. Both domains add their own merits and demerits into the overall analysis of fault location, as can be seen in Table 2-3 below. Table 2-3: Comparison of Frequency and Time domain Analysis Frequency Domain Use of Fourier Transforms or wavelet Time Domain Use of Differential equations. transforms. Uses only fundamental phasors. Uses TL parameters. Can be affected by DC decay. Unaffected by DC decay or changes ( although FFT corrects this ) in power system frequencies. Slower time convergence. Faster time convergence, but requires more processing power. However, this thesis project focuses on use of phasors for fault location detection. Hence, the frequency or spectral domain is used for analysis here. Majority of power system control devices are based on phasors. In an AC system, voltage, current and other fault loop parameters can be represented as sampled waveforms or sinusoids. These waveforms are represented in the frequency domain as phasors. Since a sinusoidal signal waveform can easily be represented as a phasor, Fourier Analysis can be implemented for phasor extraction. Colloquial factors that can affect the accuracy of fault location estimation which are common to all techniques of fault location detection can be summarized as the following: • Inaccurate line model • Affects of fault resistance and fault arcs • Uniform line impedance assumption • Neglecting pre-fault load flow and power balance • Neglecting fault inception angles • Measurement and sampling errors • Noise in line due to disturbances or auxiliary instruments Shreya Parmar Master of Science Thesis 2-3 Fault Location Algorithms 17 Taking account of the points mentioned above, we see that despite the advances in communication, processing and storage technology, many reasons still exist which contribute to failure of fault protection and detection devices. The need of smarter and more efficient fault locators, banking on lesser line specifications is noticeable. Master of Science Thesis Shreya Parmar 18 Shreya Parmar Fault Location Techniques Master of Science Thesis Chapter 3 Transmission Line Fault Location Algorithms Over the course of time, various fault location algorithms have been developed for estimating fault location with different techniques. The fault locater used to determine the fault on the transmission lines comprises of many of the approaches discussed in Chapter 2, Section 2-3. This chapter derives the impedance based fault location algorithms tested in this thesis report, one being dependent on line parameters and line end terminal measurements and the second, independent of the line parameters and dependent only on the line terminal measurements. Both algorithms use only the fundamental phasors of the line voltages and currents sampled at each end of the transmission line. They are designed with respect to locating all fault types, including balanced three-phase faults. 3-1 Line Parameter Dependent Algorithm Overhead transmission lines and underground cables are commonly found in densely populated cities. The line parameter dependent algorithm uses the Telegrapher’s equations to translate voltages and currents on an electrical transmission line and define them with respect to distance and time. The solution to Telegrapher’s equations is converted to a three phase solution using Clarck’s Modal transformation as the original set of phase variables are converted into a set of 0, α and β variables. This algorithm uses the work established in E. Sayed’s paper in [16] and by Z. Radojevic in [17]. Consider the three phase network in Figure 3-1, where a fault F occurs from one of the line’s phase to the ground. D is the total length of the transmission line whereas ` is the distance at which a fault F occurs from the sending end terminal S of the network. Master of Science Thesis Shreya Parmar 20 Transmission Line Fault Location Algorithms Figure 3-1: Three phase representation of a faulted line The same fault point can located at a distance (D − `) when seen from the receiving end terminal R of the line. For a single phase transmission line, Telegrapher’s equations are given as: ∂I ∂V +L = −RI ∂x ∂t ∂V ∂I C + = −GV ∂t ∂x (3-1) (3-2) where V , I, R, L, C and G are the voltage, current, resistance, inductance, capacitance and conductance of the line and the conductors respectively. The propagation constant γ and characteristic impedance of the line zc are: zc = q γ= q (R + jωL)/(G + jωC), ((R + jωL)(G + jωC)). From the solution of the telegraph equations in Equation 3-1, the sending end voltages and currents VS , IS can be described as : Vx Ix cosh(γ(D − x)) −zc sinh(γ(D − x)) Vs = − sinh(γ(D − x))/zc cosh(γ(D − x)) Is (3-3) where D is the total length of the line and x is any point on the line. Same approach is used to obtain the receiving end voltages and currents VR , IR . Shreya Parmar Master of Science Thesis 3-1 Line Parameter Dependent Algorithm 21 When a fault occurs ` km away from the sending end, by making use of above equations, the distance to the fault can be determined by: `= 1 tanh−1 (A/B) γ (3-4) Here, A and B are given as: A = VS cosh(γD) − zc IS sinh(γD) − VR B = IR zc − VS sinh(γD) − zc IS cosh(γD) The application of this algorithm for three phase systems, as shown in Figure 3-1, can be done by Modal decomposition of the solutions in Equation 3-4. By exercising the Clarke’s transformation, the single-phase solution can be extended to a stationary threephase reference frame as: V0 Vα Vβ 1 2 1 1 Va = −1 −1 Vb √ √ 0 3 − 3 Vc and I0 Iα = Iβ 1 2 1 1 Ia −1 −1 Ib √ √ 0 3 − 3 Ic Hence, the distance to fault becomes: `0,α,β = 1 tanh−1 (Ai /Bi ) γi (3-5) where i= 0, α and β. Consequently, `0 is the ground mode, `α and `β are the two areal q √ modes, γi = zi Yi , zC,i = zi /Yi and Ai = VS,i cosh(γi D) − zc,i ISi sinh(γi D) − VR,i Bi = IR,i zc,i + VS,i sinh(γi D) − zc,i IS,i cosh(γi D) (3-6) (3-7) Accurate fault location can be selected by the appropriate mode and the fault type in equation 3-5. `α is valid for all types of faults except line-line fault where the `β is selected. As can be noticed from the Equations 3-5 and 3-6, the fault location ` is dependent on the line parameters of the transmission line undergoing the faults. Master of Science Thesis Shreya Parmar 22 3-2 Transmission Line Fault Location Algorithms Line Parameter Independent or Parameterless Algorithm Understanding the ambiguous nature of line parameters, an algorithm which is capable of accurately determining the location of any fault on the transmission line without needing these parameters, becomes more flexible, reliable and user friendly. The parameterless algorithm in the paper [18] by Z. Radojevic and G. Preston uses only the fundamental phasors of the line voltages and currents measurements sampled at each end of the transmission line. Discreet Fourier Transform is used in the algorithm to transform the samples in frequency domain and then symmetrical components are applied to arrive at the correct fault location on a transmission line. The algorithm presented in this report is only applicable to common unbalanced faults mention in Section 2-1. An asymmetric or unbalanced faults can be easily analyzed with the use of symmetrical components, where decomposition of the unbalanced system into a sequence of balanced networks is done. Except for the balanced three-phase fault, all faults result in an unbalanced system. The key idea of symmetrical component analysis is to decompose the unbalanced system into three sequences of balanced networks. The networks are then coupled only at the point of the unbalance, that is, at the fault. Figure 3-2: Symmetrical Components Analysis for unbalanced faults Assuming three unbalanced voltage or current phasors, it is possible to represent each Shreya Parmar Master of Science Thesis 3-2 Line Parameter Independent or Parameterless Algorithm 23 phasor quantity. The symmetrical components shown in the Figure 3-2 can be described as: • Positive Sequence Components: Three phasors, equal in magnitude, displaced by 120◦ in phase, have the same sequence as the original phasors (abc). • Negative Sequence Components: Three phasors, equal in magnitude, displaced by 120◦ in phase, have the opposite sequence as the original phasors (acb). • Zero Sequence Components: Three phasors, equal in magnitude, have the same phase shift. Thus, by using the symmetrical components technique, the positive, negative and zero sequence symmetrical components of the voltages and currents from the sending and receiving end of the transmission line can be determined from the voltage and current phasors. In the three phase network represented in Figure 3-1, the equivalent positive and negative sequence circuits can be shown as in the figures below. (a) Positive sequence circuit (b) Negative sequence circuit Figure 3-3: Sequence circuits of the three phase faulted line respectively. Using Kirchoffs laws, from the Figure 3-3, we get the following equations: VpS − z`IpS = VpR − z(D − `)IpR VnS − z`InS = VnR − z(D − `)InR (3-8) (3-9) p,n p,n where Vp,n and Ip,n R , are positive and negative sequence voltages and currents S , VR , IS for sending and receiving ends respectively and z is the positive (and negative) sequence line impedance. Solving the Equations 3-8, we get the total line impedance as a representation of the voltage and current phasors measured at the line end terminals: (VpS − VpR )InR − (VnS − VnR )IpR IpR InR − InS IpR (VpS − VpR )InS − (VnS − VnR )IpS z(D − `) = IpR InR − InS IpR z` = Master of Science Thesis (3-10) (3-11) Shreya Parmar 24 Transmission Line Fault Location Algorithms To remove the influence of line impedance in the fault distance calculations, the distance to the fault, `, can now be expressed as a percentage of the line length D as: `% = ` ∗ 100 D We can also represent this as: `% = z` ∗ 100 z` + z(D − `) (3-12) Rearranging Equations 3-10 and 3-12, we get the final solution as: (VpS − VpR )InR − (VnS − VnR )IpR `% = ∗ 100 (VpS − VpR )(InS + InR ) − (VnS − VnR )(IpS + IpR ) (3-13) As can be seen from Equation 3-13, the fault location uses only the symmetrical components of the measured current and voltage phasors from the sending and receiving end of the transmission network and not the line impedance for its result. Though line parameters are used in initial decoupling of the faulted network, it can be negated and only critical data measurements are used to arrive at the final result. Thus, a parameter independent fault location algorithm is developed. This method to locate fault on a transmission line, highlights the use of SMT, which can be implemented most easily using PMUs in an actual grid system to receive voltage and current samples from measuring and recording devices installed at each end of the line. The presence of negative–sequence components in the calculations can be used to signify an asymmetrical fault. This information can also be used to effectively differentiate between the fault types. In summary, the fault location algorithms described in this chapter follow the basic flow as is represented in Figure 3-4. Both algorithms extract the voltage and current values from the line end terminals during simulations. Thereafter, the waveforms undergo sampling and phasor extraction (as shown in blocks in Figures 4-7 and 4-8 of Chapter 4), done with the use of DFT to convert the samples to a phasor domain. This is in coherence with a practical and real measuring system where the output of instrument transformers is always with respect to continuous time, where as the recording of signals is done in discrete time. Modal transformation and symmetrical transformation is applied to the sampled phasors in each respective algorithm and the final result is found as a representation of fault distance in terms of kilometers using the sending end terminal as reference. Shreya Parmar Master of Science Thesis 3-2 Line Parameter Independent or Parameterless Algorithm 25 Figure 3-4: Flowchart depicting approach towards fault location algorithms Master of Science Thesis Shreya Parmar 26 Shreya Parmar Transmission Line Fault Location Algorithms Master of Science Thesis Chapter 4 Modelling Methodologies The design of the power system models for this project was done with the intent of testing and implementation of the algorithms in different simulation environments. This chapter describes the models adopted for the simulation tests. The different line parameter modelling will be discussed, along with the line fault control design. The implementation of the algorithms mentioned in the Chapter 3 are actualized in both EMTP and RTDS environment softwares and the method to fulfill the same is discussed in this chapter. 4-1 Network Modelling The network consists of the high voltage buses fed by two AC sources of 416 KV rms line to line at the sending end and 400 KV rms line to line, at the receiving end. The basic layout for the simulation test model is as shown in Figures 4-1 and Figure 4-2. The two source network is driven by the AC voltages with a phase difference between them. The RTDS network model also has two purely resistive loads attached to it, to damp out transients due to fault currents in the terminal voltage and current waveforms. 4-2 Source Modelling A source model is often used to represent some portion of the power system in a simplified way. It usually generates a 3 phase power, power system frequency, sine wave behind an internal impedance. The RSCAD source model can be configured to use impedance format or discrete RL values whereas the ATPDraw model required an external connection for source impedance values. Both models are as shown in Figure 4-3. Master of Science Thesis Shreya Parmar 28 Modelling Methodologies Figure 4-1: EMTP-ATPDraw model used for simulation tests. Figure 4-2: RTDS-RSCAD model used for simulation tests. Shreya Parmar Master of Science Thesis 4-3 Transmission Line Modelling 29 Figure 4-3: EMTP-ATPDraw and RTDS-RSCAD source model used for simulation tests. 4-3 Transmission Line Modelling The design of a transmission line depends on four electrical parameters: • Series resistance • Series inductance • Shunt capacitance • Shunt conductance The series resistance relies basically on the physical composition of the conductor at a given temperature. The series inductance and shunt capacitance are produced by the presence of magnetic and electric fields around the conductors, and depend on their geometrical arrangement. The shunt conductance is due to leakage currents flowing across insulators and air. As leakage current is considerably small compared to nominal current, it is usually neglected, and therefore, shunt conductance is normally not considered for the transmission line modeling. The arrangement of the parameters representing the line depends upon the length of the line which defines a short-length line for lengths is less than 80 km and a medium length line for lengths between 80 km and 240 km. Lines longer than this range are considered to be long length lines. Both short and medium-length transmission lines use approximated lumped-parameter models and same idea was used for line modelling in the test simulations. The line parameters were modelled using Lumped parameters in EMTP and a faulted SHARC transmission line model in RTDS. The line model used in EMTP is the lumped parameter model. One block model was used to represent the sending end and another for the receiving end, making their total line length equal to the line length of the simulated transmission line. The R–L–C π-equivalent LINEPI3S and LINEZT–3 models were used for test simulations. Master of Science Thesis Shreya Parmar 30 Modelling Methodologies Figure 4-4: EMTP-ATPDraw and RTDS-RSCAD line model used for simulation tests. The TLINE program, a member of the RSCAD software tools is used to lay out and define the geometry and parameters of an N conductor travelling wave transmission line. The output of the TLINE program is used by DRAFT when a case involving the line is compiled. A default TLINE model was used for the modelling in RSCAD, which in theory, should not affect the algorithms’ accuracy, at least for the Parameter Independent one. Later, the line parameters were specified as in Table 5-2. Three conductor bergeron faulted transmission line model as shown in Figure 4-4 was used. Line lengths were modified using a Draft variable slider to simulate faults at different lengths of the transmission line model. 4-4 Fault Control Modelling The transmission line faults were simulated by the use of separate single phase switches or a single three phase switch in the EMTP model. ATPDraw supports most of the switch type elements and a time controlled design which can be used to monitor the fault current and voltages at the fault point in the line, was used. The specifications are TSWITCH for single or 3-phase time controlled switching and SWIT–3XT which is a three-phase time controlled switch operating independent of the phases. The fault branch connected to the bus in figure 4-2 were used to create faults at that bus in the RTDS simulations. The fault branch consists of a switch whose open resistance is very large and whose closed resistance was specified at different values. The type of fault, line–ground or line–line, is specified in the fault component CONFIGURATION menu. For a SLG fault type, any phase can be selected and the model will include a fault branch in the L–G PARAMETERS menu. Fault initiation and removal was established by using control components utilizing an arrangement of a pulse generator control function and a push–button. The pulse width specified for the pulse generator function determines the fault duration. The length which is being faulted and the fault resistance can be adjusted with the help of interactive sliders. These were directly used in the RunTime window. Shreya Parmar Master of Science Thesis 4-4 Fault Control Modelling 31 Figure 4-5: EMTP fault fault control switch used for simulation tests. Figure 4-6: RTDS-RSCAD fault control block used for simulation tests. Master of Science Thesis Shreya Parmar 32 4-5 Modelling Methodologies FLA Block The fault location algorithm was coded in the MODELS simulation language in ATP. ATPDraw supports only a simplified usage of MODELS. The algorithms were written in the model–file and ATPDraw applied INPUT/OUTPUT section of MODELS along the block as instructed in the code file. As shown in the Figure 4-7, the critical values of voltages and currents from the line end terminals were extracted via splitter connections and fed into the MODEL block. In this block, the algorithms were implemented, including the sampling and phase transformations. Figure 4-7: EMTP algorithm MODEL used for FLA implementation. In RTDS simulations, the algorithms discussed in Chapter 3 are applied to the network model using a User Defined Component (UDC) block, known as the CBuilder. RSCAD uses CBuilder software module for the FLAs making use of the control system type components, compiled via the GPC Processor Cards available with the RTDS. The node voltages and line currents were extracted with node point and sampled in UDC blocks. Then these samples were exported to the algorithm implementation block as shown in Figure 4-8 where the final calculations were done. Both parameter dependent and parameter independent algorithms were used the FLA blocks described above to extract the voltage and current waveforms from the sending and receiving line end terminals. The extracted values were then sampled, and processed via the hard coding of the CBuilder blocks and the output was set as the fault location, which was as estimated by each algorithm. Shreya Parmar Master of Science Thesis 4-5 FLA Block 33 (a) Calculation block for Parameter Independent(b) Calculation block for Parameter dependent algorithm. algorithm. Figure 4-8: RSCAD CBuilder used for FLA implementation. Master of Science Thesis Shreya Parmar 34 Modelling Methodologies Shreya Parmar Master of Science Thesis Chapter 5 Simulation Results for different fault types In this chapter, the simulation results for the two fault location algorithms presented in the previous sections is exhibited. The algorithm tests were carried out by simulation analysis using the ATPDraw and RSCAD v2.025 software on the RTDS for a total line length of 60 km and 100 km. All calculations were established by taking the sending end of the line as reference. The schematic of the transmission line shown in Figure 5-1 is used to represent the basic modelling for the power system network, as explained in Chapter 4 previously. US and UR represent the active AC power sources. z` and z(D − `) are the sending and receiving end impedances with respect to fault distance ` respectively. The algorithms calculate the fault distance ` from the sending side as reference using the fundamental phasors of the line voltage and currents extracted using DFT, as is already explained in Figure 3-4. Figure 5-1: Single line diagram for FLA simulation tests. Master of Science Thesis Shreya Parmar 36 Simulation Results for different fault types The network parameters used for the simulation tests were common and are given in Table 5-1 and the line constants used for the simulations, are as in Table 5-2. Table 5-1: Network Parameters Parameters Sending End US Receiving End UR ULL,rms (KV) 416 400 φ (◦ ) 0 -20 R (Ω) 1.0185892 0.6366183 L (H) 0.0509295 0.0318309 R0 (Ω) 2.0371785 1.2732366 L0 (H) 0.1018589 0.0636618 Table 5-2: Line parameters Parameters p-n sequence zero sequence R (Ω/km) 0.065 0.195 L (mH/km) 0.95493 2.86479 C (µ/km) 0.008688 0.004762 Z (Ω/km) 0.3317+j0.41634 0.1972+j0.3699 Transmission line faults were simulated at different locations along the line with the use of a 3 phase controlled switch (in ATPDraw modelling) and a fault control logic (in RTDS/RSCAD modelling) as shown in Section 4. Fault was set to occur at 40 ms in all simulation cases with respect to the total simulation time of 200 ms. Percentage error, which is represented for the parameter dependent algorithm as PDE and the the parameter independent or parameterless algorithm as PLE in the simulation results were calculated using: (actual value − calculated value) error percentage(%) = 100 ∗ actual value Shreya Parmar Master of Science Thesis 5-1 EMTP Simulation Results 5-1 37 EMTP Simulation Results In EMTP simulations, the faults were set to occur at 40 ms. It was assumed that the line was loaded before the fault inception. The sampling frequency was fs = 25 kHz. The data window size was 20 ms. This corresponds to N = 500 samples per data window. It is assumed the synchronization error is equal to 0 degrees. The simulation results are tabulated where the results obtained by both algorithms are as shown. The plots for fault location at various distances are given, where all calculations have been done for Rf = 0.01Ω. Sending end Voltage (kV) Receiving end Voltage (kV) 400 400 phase A phase B phase C 200 0 0 −200 −200 −400 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 phase A phase B phase C 200 −400 0 0.02 0.04 0.06 0.08 t (s) Sending end Current (kA) 0.12 0.14 0.16 0.18 Receiving end Current (kA) 40 20 phase A phase B phase C 20 0 −20 −10 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 phase A phase B phase C 10 0 −40 0.1 t (s) −20 0 0.02 0.04 0.06 0.08 t (s) 0.1 0.12 0.14 t (s) Figure 5-2: Simulation waveforms for SLG fault. The results for an SLG fault with different Rf is tabulated as below, for a transmission line of 60 kms. Table 5-3: Simulation result for SLG Fault - EMTP/ATP 60 kms Actual Rf = 0.01Ω Fault (km) P DE% P LE% P DE% 10 14.488 1.838 14.329 1.755 14.352 2.695 14.342 0.447 20 4.535 4.345 4.525 4.345 -16.78 4.65 4.495 4.99 30 1.223 4.007 1.226 4.04 1.22 4.42 1.22 5.63 40 -0.402 3.272 -0.402 3.312 -0.402 3.765 -0.402 5.4 50 -1.318 2.428 -1.322 2.466 -1.322 2.99 -1.322 4.842 Master of Science Thesis Rf = 1.0Ω Rf = 10.0Ω P LE% P DE% P LE% Rf = 50.0Ω P DE% P LE% Shreya Parmar 0.16 0.18 38 Simulation Results for different fault types 50 Fault Fault Fault Fault Fault Fault Distance calculated ( km ) 45 40 at at at at at 50 40 30 20 10 km km km km km 35 30 25 20 15 10 5 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 t (s) 90 Fault Fault Fault Fault Fault Fault Fault Fault Fault 80 Fault Distance calculated( km ) 70 at at at at at at at at at 90km 80km 70km 60km 50km 40km 30km 20km 10km 60 50 40 30 20 10 0 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 t (s) Figure 5-3: Simulation results for SLG faults in a 60 km and 100 km transmission line model with parameter independent algorithm in EMTP simulations. Shreya Parmar Master of Science Thesis 5-1 EMTP Simulation Results 39 Table 5-4: Simulation result for SLG Fault - EMTP/ATP 100kms Actual Fault Rf = 0.01Ω Rf = 1.0Ω Rf = 10.0Ω (km) P DE% P LE% P DE% 10 88.046 0.615 16.112 0.267 16.145 0.96 16.614 1.702 20 6.58 3.6 6.485 3.46 6.505 3.54 6.495 4.045 30 3.316 3.734 3.28 3.686 3.29 3.8 3.283 4.467 40 1.69 3.717 1.69 3.465 1.682 3.627 1.68 4.435 50 0.718 3.094 0.72 3.11 0.718 3.304 0.718 4.224 60 0.221 2.678 0.55 2.7 0.08 2.9167 0.08 3.918 70 -0.37 2.154 -0.372 2.262 -0.372 2.497 -0.372 3.558 80 -0.696 1.797 -0.701 1.812 -0.702 2.065 -0.702 3.166 90 9.831 1.341 -0.92 1.354 -0.926 1.638 -0.926 2.756 P LE% P DE% P LE% P DE% P LE% Sending end Voltage (kV) Receiving end Voltage (kV) 400 400 phase A phase B phase C 300 200 100 phase A phase B phase C 300 200 100 0 0 −100 −100 −200 −200 −300 −400 Rf = 50.0Ω −300 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 −400 0 0.02 0.04 0.06 t (s) 0.08 0.1 0.12 0.14 0.16 0.18 t (s) Sending end Current (kA) Receiving end Current (kA) 100 15 phase A phase B phase C 50 phase A phase B phase C 10 5 0 0 −5 −50 −10 −100 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 −15 0 0.02 t (s) 0.04 0.06 0.08 0.1 0.12 0.14 t (s) Figure 5-4: Simulation waveforms for LL fault. Master of Science Thesis Shreya Parmar 0.16 0.18 40 Simulation Results for different fault types Table 5-5: Simulation result for LL Fault - EMTP/ATP 60kms Actual Fault P DE% P LE% 10 16.518 -3.19 15 12.4 -0.133 20 9.78 0.495 25 7.8 1.364 30 6.18 1.563 35 4.791 1.622 40 3.582 1.67 45 2.522 1.655 50 1.606 1.674 55 1.529 1.630 (km) Shreya Parmar Master of Science Thesis 5-1 EMTP Simulation Results 41 Table 5-6: Simulation result for LL Fault - EMTP/ATP 100kms Actual Fault P DE% P LE% 10 19.184 -4.04 15 18.868 -0.906 20 13.05 -0.435 25 11.376 0.684 30 10.03 0.926 35 8.88 1.051 40 7.862 1.117 45 6.937 1.146 50 6.086 1.154 55 5.292 1.147 60 4.55 1.131 65 3.850 1.187 70 3.191 1.078 75 2.569 1.045 80 1.985 1.01 85 1.44 0.972 90 0.94 0.937 (km) Table 5-7: Simulation result for LLG Fault - EMTP/ATP 60kms Actual Fault Rf = 0.01Ω (km) P DE% P LE% Rf = 1.0Ω P DE% Rf = 10.0Ω P LE% P DE% Rf = 50.0Ω P LE% P DE% P LE% 10 13.90 -7.54 14.113 -9.25 14.332 -19.44 14.333 -19.94 20 4.315 -2.505 4.49 -3.22 4.49 -8.3 4.46 -8.3 30 1.217 -1.1 1.22 -1.47 1.22 -4.703 1.21 -4.703 40 -0.402 -0.12 -0.395 -0.372 -0.402 -2.75 -0.4 -2.75 50 -1.276 0.702 -1.32 0.436 -1.322 -1.406 -1.32 -1.406 Master of Science Thesis Shreya Parmar 42 Simulation Results for different fault types 55 Fault Fault Fault Fault Fault 50 Fault Distance calculated( km ) 45 at at at at at 50km 40km 30km 20km 10km 40 35 30 25 20 15 10 5 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 t (s) 150 Fault Distance calculated( km ) Fault Fault Fault Fault Fault Fault Fault Fault Fault at at at at at at at at at 90km 80km 70km 60km 50km 40km 30km 20km 10km 100 50 0 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 t (s) Figure 5-5: Simulation results for LL faults in a 60 km and 100 km transmission line model with parameter independent and dependent algorithms in EMTP simulations. Shreya Parmar Master of Science Thesis 5-1 EMTP Simulation Results 43 Sending end Voltage (kV) Receiving end Voltage (kV) 400 400 phase A phase B phase C 300 200 100 200 100 0 0 −100 −100 −200 −200 −300 −400 phase A phase B phase C 300 −300 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 −400 0 0.02 0.04 0.06 t (s) 0.08 0.1 0.12 0.14 0.16 0.18 t (s) Sending end Current (kA) Receiving end Current (kA) 20 30 phase A phase B phase C 15 10 phase A phase B phase C 20 10 5 0 0 −10 −5 −20 −10 −30 −15 −20 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 −40 0 0.02 0.04 0.06 t (s) 0.08 0.1 0.12 0.14 0.16 t (s) Figure 5-6: Simulation waveforms for LLG fault. Table 5-8: Simulation result for LLG Fault - EMTP/ATP 100kms Actual Fault Rf = 0.01Ω (km) P DE% P LE% P DE% 10 15.573 -4.95 16.112 -4.87 16.118 -4.96 16.116 -5.04 20 6.48 -0.215 6.485 -0.335 6.49 -0.37 6.49 4.535 30 3.283 0.583 3.28 0.613 3.28 0.606 3.25 0.5034 40 1.68 0.827 1.678 0.955 1.678 0.948 1.66 0.86 50 0.716 1.074 0.71 1.092 0.718 1.086 0.71 1.01 60 0.08 1.121 0.08 1.13 0.078 1.125 0.541 1.063 70 -0.374 1.113 -0.361 1.124 -0.373 1.107 -0.368 1.062 80 -0.702 1.065 -0.702 1.073 -0.693 1.053 -0.695 1.027 90 -0.926 0.989 -0.926 0.88 -0.926 0.977 -0.915 0.968 Master of Science Thesis Rf = 1.0Ω Rf = 10.0Ω P LE% P DE% Rf = 50.0Ω P LE% P DE% P LE% Shreya Parmar 0.18 44 Simulation Results for different fault types 100 Fault Fault Fault Fault Fault Fault Fault Fault Fault 90 Fault Distance calculated( km ) 80 70 at at at at at at at at at 90km 80km 70km 60km 50km 40km 30km 20km 10km 60 50 40 30 20 10 0 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 t (s) 100 Fault Fault Fault Fault Fault Fault Fault Fault Fault 90 Fault Distance calculated( km ) 80 70 at at at at at at at at at 90km 80km 70km 60km 50km 40km 30km 20km 10km 60 50 40 30 20 10 0 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 t (s) Figure 5-7: Simulation results for LLG fault with parameter independent and dependent algorithms in EMTP simulations.. Shreya Parmar Master of Science Thesis 5-2 RTDS Simulation Results 5-2 45 RTDS Simulation Results The sampling for real time simulations was achieved in a different manner than its EMTP counterpart. The time step was specified as 40 µs to achieve the desired sampling window of N = 500 samples for the 200 ms plot duration. Fault was set to occur at 40 ms and synchronization error was assumed to be 0. The load flow was run to find the initial voltage magnitude and angle for each bus in the circuit. Subsequent RunTime results were taken into account. The RTDS simulation results are presented in the succeeding tables, along with the plots for both algorithms at various distances for Rf = 0.01Ω. Table 5-9: Simulation result for SLG Fault - RTDS 100kms Actual Fault P DE% P LE% 10 65.245 25.06 15 -16.666 -0.366 20 7.3 -0.263 25 -0.188 -0.2 30 -5 -0.14167 35 -3.171 -0.1 40 -1.687 -0.068 45 0.971 -0.04 50 -1.4 -0.02 55 -1.272 -0.001 60 0.416 0.016 65 -0.230 0.023 70 1.271 0.021 75 0.44 0.04 80 1.062 0.043 85 1.202 0.043 90 -0.05 0.044 95 -1.766 0.037 (km) Master of Science Thesis Shreya Parmar 46 Simulation Results for different fault types 160 Fault Fault Fault Fault Fault Distance calculated( km ) 140 at at at at 70km 50km 30km 10km 120 100 80 60 40 20 0 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 t (s) Figure 5-8: Simulation results for SLG fault using parameter independent algorithm in RTDS simulations. Table 5-10: Simulation result for LL Fault - RTDS 100kms Actual Fault P DE% P LE% 5 -17 -1.08 10 1.04 -0.55 15 -0.653 -0.367 20 -0.39 -0.275 30 -2.78 -0.15 40 -0.19 -0.0675 50 -1.786 -0.02 60 -1.681 0.0138 70 0.18 0.032 80 -0.020 0.043 90 -1.22 0.041 (km) Shreya Parmar Master of Science Thesis 5-2 RTDS Simulation Results 47 Figure 5-9: Simulation results for LL fault with Parameter independent algorithms in RTDS simulations.. Table 5-11: Simulation result for LLG Fault - RTDS 100kms Actual Fault P DE% P LE% 5 -36.8 -1.1 10 -62 -0.56 15 -72.866 -0.373 20 -35.488 -0.262 25 -24.388 -0.194 30 -14.516 -0.15 35 -17.391 -0.098 40 -2.467 -0.052 45 -12.372 -0.040 50 -14.288 -0.019 60 -9.278 0.012 70 -11.388 0.034 80 -11.353 0.043 90 -11.957 0.041 (km) Master of Science Thesis Shreya Parmar 48 Simulation Results for different fault types Figure 5-10: Simulation results Parameter independent algorithms for LLG fault. Shreya Parmar Master of Science Thesis Chapter 6 Conclusions This chapter discusses the fault location algorithms’ results and draws conclusions based on the simulation outcomes presented in Chapter 5. 6-1 Result Analysis The algorithms derived in Chapter 3 use two approaches to correctly locate a fault’s location in a transmission line. The parameter dependent algorithm, uses line parameters and Clarke’s transformation to assess the correct location for a fault. This algorithm can be applied to transmission corridors constituting of overhead transmission lines and underground cables. The second algorithm, does not use the parameters of the line, and hence, proves to be more amicable in locating fault distances when compared with the parameter dependent algorithm. Both algorithms were tested extensively on EMTP/ATPDraw platforms and then in real time environment on the RTDS with the help of system modelling done in RSCAD v2.025. The results of the algorithm on the two environments were evaluated to find the usability of the algorithms. Faults were simulated on different line lengths in both short and medium transmission lines using three phase control switches and fault control logic. When faults were made closer to the sending end terminals which was the reference for all simulations, a large value of error was observed in almost all simulation cases. This is because of the fault currents in the line phases, collect in a concentrated manner at the end terminals. Since the fault is being made closer to one of the line terminals which is also a measurement point, a higher value of fault current gives rise to errors in measurements and hence, an incorrect result. The algorithms maintained accuracy of 1% for all possible fault resistances in ground faults, only exacting a higher error value of fault location in parameter dependent algorithm, Master of Science Thesis Shreya Parmar 50 Conclusions because of its dependence of the line parameters, which can change with fault resistance, even more so for a fault resistance as high as 50Ω. In the RTDS simulations, oscillations around the correct answer were observed in the RunTime window, when fault location calculations with respect to time were to be plotted. This could be due to synchronization errors, that is, using a time step as low as 40 µs to get 500 samples for the time window in the DFT analysis duration, which increases the computations. However, in RTDS simulation tests, as the initial algorithm results were showing dummy peaks in the final value graphical plot, an 8 pole Low Pass filter with a cut-off frequency of 100 Hz was used, which resulted in a more stable result. Also, as shown in the result tabulations in Chapter 5, fault simulations for ground faults in Parameter Independent algorithm showed more error for higher fault resistance values towards the receiving end of the line terminal. This can be explained as, due to the fault being closer to the receiving end terminal, more fault current flows to the resistive load at that end, making the voltage phasors at the terminal less in magnitude and thereby, affecting the fault location calculations, which in turn is directly dependent on the magnitude of the voltage and current phasors. Errors in fault location results are smaller towards the middle of the total line length than the line end terminals also because of a homogeneity in the fault currents at the middle of the line, than at the line end terminals. Errors were also introduced because of use of a basic DFT algorithm. Tests done using the Fast Fourier Transform (FFT) in MATLAB showed that the calculations were not just faster, but the graphical plots also showed less or no garbage values, which could be due to the high frequency components in the closely spaced samples. This is also because the fundamental phasors for the sampled values were more accurate than those in the DFT. The FFT is calculated more precisely because the reduction of redundant calculations results in less round-off error, which also eliminates noise and redundant values from the calculations. The reason for using DFT in the algorithms on EMTP and RTDS is however, because of the restrictions in the functions library that both softwares provide. The algorithms were derived using assumptions that the shunt capacitance of the transmission line can be neglected due to the algorithms being tested on short and medium line lengths. However, when the algorithms were tested, the lumped RLC and distributedparameter line models that include the shunt capacitance were used and the algorithms were found to remain accurate in EMTP results. The algorithms mostly maintained their accuracies for different fault types, proving their flexibility and robustness. Lastly, the simulation results and algorithm analysis proves that the parameter independent or parameterless algorithm is not only more accurate, but flexible and robust enough to be used under different system conditions and topologies, which is an advantage over FLAs that can only be used for one fault type or that are sensitive to line parameters and fault resistance. Shreya Parmar Master of Science Thesis 6-2 Future Work and suggestions 6-2 51 Future Work and suggestions The simplicity of the algorithms used in this thesis makes it their best feature, as they can be programmed or hard coded and implemented into any intelligent electronic device at the line terminals and will accurately locate any fault that occurs on the line, even when data sampling time-synchronization is lost. The FLAs are compatible with most devices using SMT and the algorithms discussed in Chapter 3 use Phasor extraction methods, hence their implementation into PMUs could result in more accurate and faster fault location detection. The major disadvantage of these algorithms is that their accuracy can be reduced when the data sampling synchronization is lost and a high fault resistance is introduced in the system. Therefore, a synchronization check can be established for the sampled data for correcting their outputs when the synchronization is lost. Further research should be carried out for a better understanding of the harmonics induced by fault arcs, which were assumed to be negligible in all simulation cases, and their affect on rest of the power system, and in turn the fault location detected by the numerical algorithms. The phasor-domain fault location algorithm presented in Section 3-2 should be further tested for symmetrical three phase faults types. A suitable sampling frequency and data window should be selected for the algorithms’ calculations since this can directly affect the accuracy of the results and the speed of the calculations. A suitable sampling time and a high data window will always be beneficial, as it will render the algorithm to be more accurate, since more data will negate the affect of the incorrect values, as a whole. However, a higher sampling time will compromise the computation time of the algorithms’. Hence, studies can be conducted and optimization tools can be used to arrive at the most suitable combination of the two parameters. The algorithms could be improved if more realistic line models including the shunt capacitance, were used in its derivation. The algorithms can be tested on untransposed and longer transmission line lengths to validate the accuracy of their results. Using a network model with more than two line end terminals and more generation and load sources could also determine the algorithms accuracy in a practical electric grid usage. Lastly, citing the real time test simulation results, the algorithms can be implemented and tested using a microprocessor relay or any IED to help in their error improvement. 6-3 Summary As a result of continuous exposure to unpredictable atmospheric and man-made conditions, the occurrence of faults on transmission lines is a common problem. Fault location algorithms are important tools for expediting repairs after the occurrence of a fault on a transmission line. This thesis explains the importance of protection of transmission lines and how existing relay line protection devices can be insufficient in doing the same. The Master of Science Thesis Shreya Parmar 52 Conclusions basics of fault analysis is discussed and the main categories of the fault location algorithm techniques are critiqued. This thesis report gives a detailed derivation of two fault location schemes for transmission line networks. The first algorithm uses line parameters to locate faults. The second algorithm is independent of the line parameters. Both algorithms can use synchronized voltage and current measurements from the line end terminals. Extensive simulations were carried out using ATPDraw, MATLAB and RTDS to evaluate the performance of both algorithms under SLG, LL and LLG faults which are the most common types of asymmetrical faults on transmission lines. From the errors observed in both FLAs, we see that the parameter-independent fault location algorithm gives a better accuracy over the parameter dependent algorithm. The simplicity of parameter independent fault location detection is complimented by its capability of being easily practiced with existing power system protection schemes. Shreya Parmar Master of Science Thesis Appendix A Simulation softwares used A-1 MATLAB MATLAB (matrix laboratory) is a multi-paradigm numerical computing environment and fourth-generation programming language. Developed by MathWorks, MATLAB allows matrix manipulations, plotting of functions and data, implementation of algorithms, creation of user interfaces, and interfacing with programs written in other languages, including C, C++, Java, Fortran and Python. Although MATLAB is intended primarily for numerical computing, an optional toolbox uses the MuPAD symbolic engine, allowing access to symbolic computing capabilities. An additional package, Simulink, adds graphical multi-domain simulation and Model-Based Design for dynamic and embedded systems. MATLAB was largely used for graphical plots made in the report and to check the results for corrective purposes. A-2 EMTP/ATPDraw ATPDraw is a graphical, mouse-driven preprocessor to the ATP version of the Electromagnetic Transients Program (EMTP) on the MS-Windows platform. In ATPDraw the user can construct an electrical circuit using the mouse and selecting components from menus, then ATPDraw generates the ATP input file in the appropriate format based on "what you see is what you get". The simulation program ATP and plotting programs can be integrated with ATPDraw. A license is required to use the solver ATP. Master of Science Thesis Shreya Parmar 54 Simulation softwares used Figure A-1: ATPDraw User Interface A-3 RTDS/RSCAD Real Time Digital Simulator or RTS as the abbreviation recommended by IEEE committee on real-time simulator applied for power systems provides power systems simulation technology for fast, reliable, accurate and cost-effective study of power systems with complex High Voltage Alternating Current (HVAC) and High Voltage Direct Current (HVDC) networks. The RTS is a fully digital electromagnetic transient power system simulator that operates in real time. The system’s graphical user interface, proprietary software and mathematical algorithms can simulate any modern electric power grid configuration. As new equipment or components are added or subtracted from the simulator’s configuration, the model instantly updates. For example, researchers can run simulated system-failure scenarios such as a control system cyber intrusion or a physical damage event such as a terrorist attack or natural disaster and instantly detect the order and reasoning for why dedicated relays, breakers or substations failed. Because the simulator functions in real time, the power system algorithms are processed to continuously produce output conditions that represent conditions in a real network. Realtime simulation is significant because the user can test physical devices using the ’Hardware in Loop’ method, which is important for testing of many equipments and algorithms, which may be difficult to achieve in an actual High Voltage line. Also, simulations can be run faster since simulation of the system’s response over 1 second is computed in exactly 1 second. This is note worthy as all calculations required to determine the power system’s state are completed in a time exactly equal to the simulation time step. Power systems Shreya Parmar Master of Science Thesis A-3 RTDS/RSCAD 55 networks can be built and compiled in the Draft case and the simulations are established on the RunTime window. Figure A-2: RSCAD Power system user library Master of Science Thesis Shreya Parmar 56 Shreya Parmar Simulation softwares used Master of Science Thesis Appendix B Symmetrical components and Sequence Networks B-1 Conversion Between the Phase and Symmetrical Component Domains Any set of phase quantities can be converted into symmetrical components, where α is defined as 1∠120◦ , as follows: 1 I0 I 1 I2 = 1 1 Ia 1 1 α α2 I b 3 2 1 α α Ic where I0 , I1 , and I2 are the zero, positive, and negative sequence components, respectively. The above transformation equation shows the symmetrical component transformation in terms of currents, but the same equations are valid for voltages as well. This results in the following equations: 1 I0 = (IA + IB + IC ) 3 1 I1 = (IA + αIB + α2 IC ) 3 1 I2 = (IA + α2 IB + αIC ) 3 Master of Science Thesis Shreya Parmar 58 Symmetrical components and Sequence Networks Similarly, the inverse phase vectors can be achieved as: 1 Ia I b Ic = 1 1 I0 1 1 α2 α I α 3 2 Iβ 1 α α which leads to: IA = (I0 + I1 + I2 ) IB = (I0 + α2 I1 + αI2 ) IC = (I0 + αI1 + α2 I2 ) Thus, a set of unbalanced system variables can be defined in terms of three balanced systems and the symmetrical transformation may be used to convert phase voltages (or currents) to symmetrical component voltages (or currents) and vice versa. These conversions are valid for an A-phase base, which can be used for A-phase-to-ground, B-phase-to-Cphase, B-phase-to-C-phase-to-ground, and three-phase faults. B-2 Sequence Networks for Faults Sequence networks for the faulted system are always dependent on the type of fault, where separate symmetrical or balanced networks after symmetrical transformation are linked at the fault point. Sequence network connections for common shunt fault types are shown in this section where Zf is defined as the fault impedance from each phase to the common point, and Zg is defined as the impedance from the common point to ground. The Zg base term is only significant when Zf differs per phase or if the line impedance to the fault point is different between phases. (c) Double line to ground fault (b) Double line fault network.network. (a) Single line fault network. Figure B-1: Asymmetrical fault sequence network. Shreya Parmar Master of Science Thesis B-2 Sequence Networks for Faults 59 Figure B-2: Three phase fault sequence network. Master of Science Thesis Shreya Parmar 60 Shreya Parmar Symmetrical components and Sequence Networks Master of Science Thesis Appendix C Phasor Measurement Units Critical data, like voltage and current in power system calculations can be represented as complex numbers. This in turn can be represented as ’phasors’ which contains a magnitude and direction, or angle. C-1 Basics Conventional measurement instruments measure only the magnitude of these phasor quantities. Measurements requiring synchronization and accurate time stamping, as in the case of phasor measurements for fault location detection, have phase angle measurement errors as a prevalent phenomenon. As a solution to this problem, PMUs measure the magnitude and angles of phasors with accurate time stampings with respect to the system frequency. The measurements are synchronized with a clocking signal obtained continuously from a global positioning system (GPS). Thus, the base operation or control station for the concerned grid of power system is able to receive the synchronous data from each PMU in real time. The location of malfunctioning circuits or transmission lines can be immediately identified if phase differences between different PMUs are detected. In synchronous measurement technology, PMUs are often referred to as ’Synchrophasors’ and they are termed to be the most important measuring devices in the future of Smart Grids. PMUs can be used as stand alone devices or can be easily integrated with existing power system protection and monitoring devices. Some examples of such PMUs are L60, D60 models of the General Electric. The demerits of using SCADA over PMU based measuring devices are as given below: • measurements are not synchronized due to inaccurate time stamping. Master of Science Thesis Shreya Parmar 62 Phasor Measurement Units • slower rate of measurements – upto 1 sample per sec. • limited dynamic performance monitoring for the power system C-2 Applications • state estimation • fault detection • line parameter calculations • congestion management • real time control and monitoring in Wide Area Monitoring Systems C-3 Placement Figure C-1: PMU placement in power systems PMU exacts the voltage phasor of the bus where it is installed and current phasors of all branches incident to that bus. The PMU measurements are given from different buses, which are synchronized by the common clock signal from global positioning system (GPS). A PMU and its affiliated communication equipments are generally costly. So it is neither economical nor necessary to install PMUs at all system buses. Consequently, the placement and optimal numbers of PMU units becomes of maximal importance to ensure no errors in the measurements. Much research has been established in this direction. The paper [19] covers most of the methods used for hierarchy based PMU placement methods. Shreya Parmar Master of Science Thesis Bibliography [1] M. Kezunovic, “Smart fault location for smart grids,” Smart Grid, IEEE Transactions on, vol. 2, no. 1, pp. 11–22, 2011. [2] http://www.reuters.com/article/2015/03/27/us-dutch-power-outages, 2015. [3] http://www.atpdraw.net/, 2015. [4] https://www.rtds.com, 2015. [5] S. Tamronglak, S. Horowitz, A. Phadke, and J. Thorp, “Anatomy of power system blackouts: preventive relaying strategies,” Power Delivery, IEEE Transactions on, vol. 11, no. 2, pp. 708–715, 1996. [6] M. M. Hashim, H. W. Ping, and V. Ramachandaramurthy, “Impedance-based fault location techniques for transmission lines,” in TENCON 2009-2009 IEEE Region 10 Conference, pp. 1–6, IEEE, 2009. [7] M. Idris, M. Mustafa, and Y. Yatim, “Effective two-terminal single line to ground fault location algorithm,” in Power Engineering and Optimization Conference (PEDCO) Melaka, Malaysia, 2012 Ieee International, pp. 246–251, IEEE, 2012. [8] J. Izykowski, E. Rosolowski, and M. Saha, “Method and device for fault location in a two-terminal transmission or distribution power line,” May 22 2012. US Patent 8,183,871. [9] C. Christopoulos, D. Thomas, and A. Wright, “Scheme, based on travelling-waves, for the protection of major transmission lines,” in IEE Proceedings C (Generation, Transmission and Distribution), vol. 135, 1988. [10] F. H. Magnago and A. Abur, “Fault location using wavelets,” Power Delivery, IEEE Transactions on, vol. 13, no. 4, pp. 1475–1480, 1998. Master of Science Thesis Shreya Parmar 64 Bibliography [11] E. Ngu and K. Ramar, “A combined impedance and traveling wave based fault location method for multi-terminal transmission lines,” vol. 33, pp. 1767–1775, Elsevier, 2011. [12] Z. Bo, A. Johns, and R. Aggarwal, “A new non-unit protection scheme based on fault generated high frequency current signals,” 1995. [13] Z. Bo, G. Weller, and M. Redfern, “Accurate fault location technique for distribution system using fault-generated high-frequency transient voltage signals,” in Generation, Transmission and Distribution, IEE Proceedings-, vol. 146, pp. 73–79, IET, 1999. [14] A. G. Phadke and J. S. Thorp, Synchronized phasor measurements and their applications. Springer, 2008. [15] D. Novosel and K. Vu, “Benefits of pmu technology for various applications,” Zbornik radova sedmog simpozija o sustavu vođenja EES-a HK CIGRE, Cavtat, vol. 5, no. 8.11, 2006. [16] E. Sayed Tag El Din, M. M. Abdel Aziz, M. Gilany, et al., “Fault location scheme for combined overhead line with underground power cable,” Electric power systems research, vol. 76, no. 11, pp. 928–935, 2006. [17] M. Popov, G. Rietveld, Z. Radojevic, and V. Terzija, “An efficient algorithm for fault location on mixed line-cable transmission corridors,” in International Conference on Power Systems Transients (IPST), 2013. [18] Z. Radojević, C. Kim, M. Popov, G. Preston, and V. Terzija, “New approach for fault location on transmission lines not requiring line parameters,” in Proceeding of IPST Conference, 2009. [19] J. Zhong and K. Wong, “A hierarchical analysis of phasor measurement unit placement optimization in transmission network,” in Universities Power Engineering Conference (AUPEC), 2011 21st Australasian, pp. 1–4, IEEE, 2011. Shreya Parmar Master of Science Thesis Glossary FLA Fault Location Algorithm(s) PMU Phasor Measurement Unit(s) SMT Synchronised Measurement Technology(s) DFT Discrete Fourier Transform GPS Global Positioning System EMTP Electro-magnetic Transients Programme ATP Alternate Transients Programme RTDS Real Time Digital Simulator SLG Single line-to-ground LLG Line-to-line-to-ground LL Line-to-line STFLA Single-terminal Fault Location Algorithm(s) TTFLA Two-terminal Fault Location Algorithm(s) TWFLA Travelling Wave Fault Location Algorithm SCADA Supervisory Control and Data Acquisition XPS Expert Sytem Techniques ANN Artificial Neural Networks FL Fuzzy Logic systems Master of Science Thesis Shreya Parmar 66 Glossary AC Alternate Current UDC User Defined Component IED Intelligent Electronic Device Shreya Parmar Master of Science Thesis 67 List of Symbols and Variables R Line resistance L Line inductance C Line capacitance G Line shunt conductance F Fault point on a line D Total length of transmission line ` Fault location calculated by the algorithm(s) S Sending end of transmission line R Receiving end of transmission line γ Propagation constant Zc Characteristic impedance of the transmission line Vx Voltage at any point x on the transmission line VS Voltage from Sending end of the transmission line Ix Current at any point x on the transmission line IS Current from Sending end of the transmission line VR Voltage from Receiving end of the transmission line IR Current from Receiving end of the transmission line Va,b,c Three phases of voltage Ia,b,c Three phases of current Vα,β,0 Three modes of the voltage after applying Clarcke’s transformation Iα,β,0 Three phases of current after applying Clarcke’s transformation z Impedance of the transmission line Vp Positive sequence component of voltage Vn Negative sequence component of voltage Master of Science Thesis Shreya Parmar 68 Glossary Ip Positive sequence component of current In Negative sequence component of current fs Sampling frequency N Number of samples used Rf Fault resistance Shreya Parmar Master of Science Thesis