https://hemerotecadigital.uanl.mx/files/original/312/20798/Ingenierias_2017_Ano_20_No_75_Abril-Junio.pdf b75bada19d707efdef50f495e489a0bc PDF Text Text ��75 Contenido Abril-Junio de 2017, Año XX, No. 75 2 Directorio 3 Editorial: Habilidades de expresión escrita de los estudiantes Juan Antonio Aguilar Garib 7 Real time simulation with software relay models Carlos A. López, Víctor H. Ortíz, Daniel Ibarra 19 Adaptive autoreclosaure to increase system stability and reduce stress to circuit breakers Jörg Blumschein, Yilmaz Yelgin, Andrea Ludwig 32 Umbral para discriminar entre corriente inrush y falla interna en un transformador de potencia Martha N. Acosta Montalvo, Héctor Esponda Hernández, Manuel A. Andrade Soto, Ernesto Vázquez Martínez 43 Performance analysis of line differential protection using MPLS networks Lifan Yang, Thomas Rudolph, Min Li, Motaz Elshafi 54 CT saturation and its influence on protective relays Roberto Cimadevilla, Ainhoa Fernández 76 Eventos y reconocimientos 77 Tesistas titulados de Maestría en la FIME-UANL 78 Información para colaboradores 79 Código de ética Ingenierías, Abril-Junio 2017, Año XX, No. 75 � �DIRECTORIO Ingenierías, Año XX N° 75, abril-Junio 2017. Es una publicación trimestral, editada por la Universidad Autónoma de Nuevo León, a través de la Facultad de Ingeniería Mecánica y Eléctrica. Domicilio de la Publicación: Facultad de Ingeniería Mecánica y Eléctrica, Pedro de Alba S/N, Edificio 7, San Nicolás de los Garza, Nuevo León, México, C.P. 66450. Teléfono: +52 (81) 83294020 Ext. 5854, Fax +52 81 83320904. Editor responsable: Dr. Juan Antonio Aguilar Garib. Reserva de derechos al uso exclusivo No. 04-2011-101411064600-102, ISSN: 1405-0676. Número de certificado de licitud de título y contenido: 15,525, otorgado por la Comisión Calificadora de Publicaciones y Revistas Ilustradas de la Secretaría de Gobernación. Registro de marca ante el Instituto Mexicano de la Propiedad Industrial: En trámite. Impresa por: Impresos Baez, Jesús M. Garza 3219 Oriente, Col Francisco I. Madero, Monterrey Nuevo León, México, C.P. 64560. Fecha de terminación de impresión: 15 de abril de 2017. Tiraje: 800 ejemplares. Distribuido por: Universidad Autónoma de Nuevo León, a través de la Facultad de Ingeniería Mecánica y Eléctrica, Pedro de Alba S/N, Edificio 7, San Nicolás de los Garza, Nuevo León, México, C.P. 66450. Las opiniones expresadas por los autores no necesariamente reflejan la postura del editor de la publicación. Prohibida su reproducción total o parcial de los contenidos e imágenes de la publicación sin previa autorización del Editor. Impreso en México Todos los derechos reservados © Copyright 2017 revistaingenierias@uanl.mx UNIVERSIDAD AUTÓNOMA DE NUEVO LEÓN Mtro. Rogelio G. Garza Rivera Rector M.A. Carmen del Rosario de la Fuente García Secretario General Dr. Juan Manuel Alcocer González Secretario Académico Dr. Celso José Garza Acuña Secretario de Extensión y Cultura Lic. Antonio Ramos Revillas Director de Editorial Universitaria FACULTAD DE INGENIERÍA MECÁNICA Y ELÉCTRICA Dr. Jaime A. Castillo Elizondo Director Dr. Juan Antonio Aguilar Garib Editor responsable M.C. Cyntia Ocañas Galván M.C. Jesús G. Puente Córdova Redacción Gregoria Torres Garay Tipografía y formación M.A. José Luis Martínez Mendoza Diseño M.C. Adán Ávila Cabrera Fotografía Ing. Cosme D. Cavazos Martínez Webmaster María de los Ángeles Baez Acuña Impresor Allyson Lynette Calderón Vega / Claudia Yanely Rivera Reyna Auxiliares CONSEJO EDITORIAL INTERNACIONAL Dr. Mauricio Cabrera Ríos, Puerto Rico. UPRM / Dr. Cezar Henrique Gonzalez, Brasil. UFPE, Recife-Pernambuco / Dra. Ruth Kiminami, Brasil. UFSC, San Pablo / Dr. Juan Jorge Martínez Vega, Francia. Universidad de Toulouse III / Dr. Juan Miguel Sánchez, USA. UT-Austin / Dr. Zarel Valdez Nava, Francia. UPS-INPT-LAPLACE-CNRS CONSEJO EDITORIAL MÉXICO Dr. Jesús González Hernández, CIDESI / Dr. Benjamín Limón Rodríguez, FIC-UANL / Dr. José Rubén Morones Ibarra, FCFM-UANL / Dr. Ubaldo Ortiz Méndez, FIME-UANL / Dr. Miguel Ángel Palomo González, FACPYA-UANL / Dr. Félix Sánchez De Jesús, ICBI-UAEH / Dr. Ernesto Vázquez Martínez, FIME-UANL. COMITÉ TÉCNICO Dr. Efraín Alcorta García, FIME-UANL / Dr. Rafael Colás Ortiz, FIME-UANL / Dr. Jesús De León Morales, FIME-UANL / Dr. Virgilio Ángel González González, FIME-UANL / Dr. Carlos Alberto Guerrero Salazar, FIME-UANL / Dra. Oxana Vasilievna Karissova, FCFM-UANL / Dr. Francisco Eugenio López Guerrero, FIME-UANL / Dr. Martín Edgar Reyes Melo, FIME-UANL / Dr. Roger Z. Ríos Mercado, FIME-UANL / Dr. Juan Ángel Rodríguez Liñán, FIME-UANL. � Ingenierías, Abril-Junio 2017, Año XX, No. 75 �Editorial: Habilidades de expresión escrita de los estudiantes Juan Antonio Aguilar Garib Universidad Autónoma de Nuevo León, Facultad de Ingeniería Mecánica y Eléctrica juan.aguilargb@uanl.edu.mx En toda profesión existe, a diferentes niveles, la necesidad de expresarse en forma escrita, de manera que la habilidad de escritura no es una requerimiento exclusivo de periodistas, abogados o encargados de relaciones públicas. En muchos trabajos es necesario escribir reportes, notas y hasta recados simples. Es notorio que el interés de los estudiantes por continuar desarrollando la competencia de escritura, especialmente en áreas que no están relacionadas directamente con las letras, disminuye conforme avanzan en los grados escolares hacia la carrera. No es raro que consideren inapropiado que se les critique su ortografía o redacción, aunque no se les sancione en las evaluaciones, en clases que según ellos no tienen algo que ver con literatura. En ocasiones utilizan el mismo argumento de los estudiantes de secundaria, señalando que si la clase no es de español no debería haber problema, sin darse cuenta que la falta de ortografía puede modificar el sentido de lo que pretenden expresar, y que aun tratándose de comunicación informal, un mensaje plagado de errores ortográficos y gramaticales burdos no es la mejor carta de recomendación para mostrar capacidad y diligencia. Hay reportes de que un buen número de estudiantes en los niveles de educación superior en diferentes áreas del conocimiento tienen deficiencias graves en su habilidad para comprender textos, así como en su escritura, 1-3 a pesar de que la alfabetización tiene un carácter prioritario en los programas educativos. Esta deficiencia ya ha sido reconocida y en las instituciones se siguen diferentes estrategias para atenderla. Aunque se entiende que este problema debe ser atacado desde la educación básica, en los niveles de educación superior en lo inmediato, afecta el rendimiento escolar y en el futuro, el desempeño del graduado. El propósito de este mensaje es dar a la expresión escrita su validación tácita, al mismo tiempo que se presentan algunas sugerencias o recomendaciones de lo que los estudiantes pueden hacer ahora, que ya están en el nivel superior, en lo que parece ser una etapa tardía para abordar problemas de formación básica. Ya que se ha mencionado a la alfabetización, conviene situar el contexto que provoca una aparente falta de correlación entre su alto valor, de 94.5% en México, 4 y la deficiencia que se observa en los estudiantes de educación superior. La alfabetización se da a través de un proceso por el que se enseña a las personas la lectura y escritura de una lengua, usualmente la materna. La manera en que se mide, de acuerdo a la UNESCO, es conforme al porcentaje de Ingenierías, Abril-Junio 2017, Año XX, No. 75 � �Habilidad de expresión escrita de los estudiantes / Juan Antonio Aguilar Garib la población mayor de 15 años de edad que puede leer, escribir y comprender un texto sencillo y corto sobre su vida cotidiana. 5 En algunos países se consideran también las operaciones simples de aritmética. Este texto puede ser tan sencillo como un recado, 6 por lo que es posible que en realidad se esté evaluando la habilidad para descifrar el código de la escritura, identificar las palabras, sin que haya un nivel de comprensión que corresponda al de los textos que deben leerse y escribirse durante la formación en la educación superior. La prueba PISA (Programa de Evaluación Internacional de Alumnos) considera que los estudiantes que han concluido la educación obligatoria (primaria y secundaria) ya tienen las habilidades fundamentales para una participación plena en las sociedades modernas. 7 La prueba ENLACE (Evaluación Nacional del Logro Académico en los Centros Escolares), 8 por otra parte, considera la competencia lectora en sectores específicos de la población escolar. En ambos casos se obtiene una valoración más apegada a las habilidades de comunicación en eventos de la vida diaria, que la que se deduce de la tasa de alfabetización. Suponiendo que el problema fueran los cursos que se imparten en los niveles básicos y que las revisiones fueran implementadas inmediatamente, eso no sería útil para los estudiantes que ya se encuentran en los niveles superiores, puesto que no es posible regresarlos a los niveles básicos para que aprendan a expresarse de manera escrita, entonces se proponen aquí una serie de acciones que pueden ser seguidas por ellos desde el nivel en el que se encuentren actualmente. Lo primero que los estudiantes deben hacer es dejar de culpar a su paso por los niveles básicos en el que afirman que no aprendieron a escribir porque no les enseñaron, ya que aun siendo esto posible, esta disculpa los hace olvidar que el aprendizaje es una acción que le corresponde a los estudiantes. Siendo la escritura una habilidad dentro de las competencias de comunicación solamente hay una manera de desarrollarla, y esa es mediante la práctica. Así, sin justificar aquí que no les hubieran enseñado como muchos de los que tienen estas deficiencias dicen; cada tarea, cada reporte, cada anotación que hicieron en sus cuadernos fueron ocasiones de práctica. Como ya se encuentran en el nivel superior, entre lo que les queda está aprovechar los talleres de lectura y redacción, y los cursos de comunicación y de apreciación de las artes, a los que pudieran tener acceso desde sus programas educativos. De los cursos que he mencionado, es probable que el de apreciación de las artes les suene alejado de las habilidades de lectura de comprensión y escritura, pero si consideramos que el arte se refiere a la actividad en la que el hombre recrea, con una finalidad estética, un aspecto de la realidad o un sentimiento valiéndose de la materia, la imagen o el sonido, con las bellas artes: arquitectura, danza, escultura, música, oratoria, pintura y poesía, entonces no se podrá negar su contribución al desarrollo de la creatividad, tan necesaria en el arte de acomodar las palabras, para que formen las oraciones y los párrafos que expresen claramente lo que deseamos. Otra recomendación es tener presente que la habilidad para escribir requiere un buen conocimiento del idioma, lo que implica un buen vocabulario junto con el manejo de la ortografía y gramática, para luego construir, oraciones y párrafos correctos gramaticalmente, además de la estructura adecuada que expresa la información, discusión, conclusión o postura que se desea transmitir. � Ingenierías, Abril-Junio 2017, Año XX, No. 75 �Habilidad de expresión escrita de los estudiantes / Juan Antonio Aguilar Garib Tomar cursos formales de escritura es una buena idea, pero ya sea que se tomen o no, para entender estos elementos se requiere leer, así se va aprendiendo cómo se describen las cosas, cómo se escriben las palabras, de manera que de tanto leerlas se note inmediatamente si están mal escritas. Existe la idea, y hasta la presunción, de que los estudiantes de otras épocas leían más. Eso es cierto porque las condiciones tecnológicas eran otras, es posible que los estudiantes actuales no alcancen a comprender un entorno en el que no había manera de “cortar y pegar” de la manera en que se hace ahora, y todavía más atrás en el tiempo, no había fotocopias a precios accesibles. Parece que fue hace mucho, pero son condiciones que muchos profesores aún activos conocieron. Entonces, los estudiantes debían comprar los libros o ir a la biblioteca, allí tomaban notas de los libros, resumían, subrayaban, en fin, estaban obligados a leer y sintetizar. Aún la copia de la tarea de un compañero exigía por lo menos leerla para escribirla, no se trataba simplemente de cortar y pegar, así que lo que los estudiantes hacían, pasaba por sus mentes y aprendían. Con la ayuda de la tecnología se ha dejado de leer y escribir para generar notas, así que otra sugerencia es leer y obligarse a tomar notas y escribir resúmenes, aunque no sean un encargo para calificación. En una recomendación más, ya se mencionó que la escritura es una habilidad dentro de las competencias de comunicación y que la única manera de desarrollarla es mediante la práctica. No se puede considerar como práctica la repetición del mismo ejercicio muchas veces sin variación alguna, ya que podría ser que se estuviera haciendo mal y que se perfeccione la manera de hacerlo así. La verdadera práctica consiste en evaluar y criticar el trabajo hecho, de manera que haya una retroalimentación que sirva para ser creativos y corregir, o para reforzar, según sea el caso, la escritura. Otro punto, que sería muy bueno que fuera obvio, es que además de conocer las reglas de la escritura se debe tener el material que se va a escribir, si se trata de un reporte de práctica, ya se realizó la tarea de revisar, razonar y ordenar lo que se va a escribir. Es común que los estudiantes den mucha más atención a las instrucciones de presentación: tamaño de portada, de letra y detalles que, aunque importantes, no tienen que ver con la escritura y siempre se pueden arreglar. Si se va a escribir de ciencia, arte, cultura o deporte, se debe saber de ciencia, arte, cultura o deporte. Las recomendaciones anteriores llevan implícita la condición de dedicar tiempo para practicar y desarrollar la habilidad para escribir. Así que, si los estudiantes dedican a sus clases únicamente el tiempo asignado a las sesiones, entonces difícilmente podrán cumplir con el objetivo de su curso y mucho menos con actividades para el desarrollo de la escritura. Es necesario que los estudiantes comprendan que esa calidad les exige dedicar la mayor parte de su tiempo a asuntos académicos, como leer y escribir, dentro y fuera de las aulas. Algunos estudiantes se desaniman cuando no obtienen buenas notas en sus trabajos escritos, pero no deben hacerlo, lo que es necesario es poner atención a lo que no estuvo bien hecho sin esperar a que el maestro aborde nuevamente el tema, es posible que nunca lo haga porque ya pasó, entonces es conveniente dedicar el tiempo necesario para revisar lo que se ha escrito. Una mala calificación en un trabajo no significa por sí misma una descalificación del individuo, quien Ingenierías, Abril-Junio 2017, Año XX, No. 75 � �Habilidad de expresión escrita de los estudiantes / Juan Antonio Aguilar Garib debe trabajar en ajustar para mejorar, si se hacen siempre las mismas cosas se obtendrán los mismos resultados. Aprovecho la posibilidad de que los lectores críticos encuentren estas recomendaciones parecidas a algo que pudieran haber leído o escuchado en alguna parte, para extender una última recomendación, asegurándoles que cualquier semejanza con material que hayan conocido antes es mera coincidencia, producida tal vez, por la generosa dosis de documentos disponibles sobre este tema. Esta última recomendación tiene que ver con el temor a escribir trabajos que no sean originales, ya que es frecuente que se les insista a los estudiantes que razonen lo que escriben para que sea propio, que no copien sus trabajos ni entre ellos ni de otras fuentes. Es cierto que el plagio debe ser denunciado y castigado, pero la originalidad absoluta es prácticamente inalcanzable si lo que sabemos proviene de la lectura de temas de los que se ha escrito ampliamente. Ante esta situación, lo que debemos hacer es escribir de forma honesta con apego a la verdad según nuestros mejores conocimientos, sin preocuparse de que el producto pudiera parecerse a otros trabajos ¿de cuántas maneras se podría describir una práctica de laboratorio que está estandarizada? Si se escribe con honestidad, entonces resaltarán las diferencias en la discusión y conclusión de los resultados o planteamientos, ya sea en una práctica, en un discurso o en un artículo, ya que la percepción de las cosas es siempre individual, la originalidad de los escritos se hará evidente. Las recomendaciones mencionadas corresponden a las acciones que inciden en hábitos de estudio que pueden ser tomados por los estudiantes, quienes son los verdaderos responsables de construir esta competencia, hoy mismo, sin que requieran un cambio extraordinario en el ámbito del ambiente escolar, pero sí en el de sus intereses. Los profesores de los niveles superiores no podemos ser los responsables de cerrar la brecha en sus habilidades de escritura, pero sí debemos ser ejemplo con nuestro comportamiento en términos de nuestra expresión oral y escrita, al mismo tiempo que promovemos el interés entre los estudiantes por superarse en sus competencias de comunicación a través de la mejora de su habilidad de escritura, lo cual repercutirá positivamente en su desempeño académico y profesional. 1. Fregoso Peralta, Gilberto. Los problemas del estudiante universitario con la redacción. Un estudio de caso en los niveles de licenciatura y de maestría, Revista de Educación y Desarrollo, octubre-diciembre, 2007. 2. Hernández Zamora, Gregorio. Escritura académica y formación de maestros ¿por qué no acaban la tesis? Tiempo de Educar, Vol. 10, Núm. 19, Universidad Autónoma del Estado de México, enero-junio, 2009, pp. 11-40. 3. Domínguez Y., J. Guillermo ¿Por qué no escriben textos los estudiantes? (Parte 1) Revista del Centro de Investigación. Universidad La Salle, vol. 5, núm. 19, juliodiciembre, 2002, pp. 85-98. 4. INEGI. Encuesta Intercensal 2015. 5. Indicadores de educación, especificaciones técnicas. UNESCO Institute of Statistics, 2009. 6. Metodología de la serie histórica censal, INEGI, 2016. 7. Resultados PISA México 2015. 8. SEP, Prueba ENLACE 2014. � Ingenierías, Abril-Junio 2017, Año XX, No. 75 �Real time simulation with software relay models Carlos A. López, Víctor H. Ortíz, Daniel Ibarra Universidad de Guadalajara, Guadalajara, Jal., México RESUMEN Este artículo describe una variedad de modelos de software en plataformas corriendo a tiempo real, enfocados para obtener una simulación con la visualización de acciones automáticas durante condiciones de falla. Los modelos desarrollados pueden adquirir información en forma dinámica (voltaje y corriente) de un sistema de potencia virtual implementado. El voltaje y la corriente de fase se capturan como señales con forma de onda, y son accesibles para esquemas de protección integrados como modelos de software en software en un laso. Eso es, el uso de un código para representar algoritmos de relé a distancia y responder mientras la simulación en tiempo real tiene lugar. PALABRAS CLAVE Simulación en tiempo real, sistema de protección, prueba HIL, liberado de falla. ABSTRACT The article describes a variety of software models running in a real-time simulation platform, the focus is get a long time simulation with the visualization of automatic actions during a fault condition. The models developed can acquire dynamic information (voltage and current) from the virtual Power System implemented. The voltage and current three phase signals are captured as waveforms, and they are accessible to protection schemes integrated as software models into software in the loop process. It means the use of code for representing distance relays algorithms and responding while ongoing realtime simulation. KEYWORDS Real-time simulation, protection system, HIL test, fault clearance. INTRODUCTION Real-time simulators provide very useful data in the field of testing hardware, measurement units and protection systems. It would not only provide information in terms of optimal design and functionality of protection relays, also provide global information to develop Wide Area Monitoring Systems (WAMS) [1]. The WAMS is then essential data to execute remotely a logic designated to provoke automatic reconfiguration of the system topology. Anyway, each particular problem must be pre-evaluated in depth considering several transient scenarios. Ingenierías, Abril-junio 2017, Vol. XX, No. 75 � �Real time simulation with software relay models/ Carlos A. López, et al. In order to make use of long term simulations in a real-time platform, as first step, the user have to organize a multi-rate representation. Several factors are important to consider before getting a global model. Experts suggest organize by zones in order to manage for some elements equations with maximum detail and other treated as equivalent.2 The implementation is based in Matlab-Simulink® and RT- Lab® blocks prepared to manage a test power system. Those elements at remote areas would be a standard model for stability transient analysis. It will be noticed subsystems for representing generators, transformers, loads, etc. The simulation emphasizes in the use of waveforms passing through primary protection zone, and the transmission lines are planned as distributed parameter models without frequency dependence. An overview of OPAL-RT Simulator is presented in figure 1 and more information about it in.3 The event of fault clearance is present for less than one second. For that reason, the time of interest for testing software relays is in the order of 50 to 2000 ms. Within this period of time are detected the operation states showed in figure 2 and listed below: 1) Pre-fault condition, 2) Fault insertion, 3) Trip for one line terminal at closer substation, 4) Fault clearance at both terminals, 5) Breakers reclosing if it is possible, and 6) Stability response in post-fault condition. The purpose of this work is to simulate local intelligence taking action to complete automatically the fault clearance. The output signal for controlling the trip on breakers is with software in the loop process since acquire the voltage and currents signals, processing they inside the relay and its algorithms and produce the output signal to local breaker. Fig. 1. Overview of real-time simulator OPAL-RT®.4 � Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Real time simulation with software relay models/ Carlos A. López, et al. Fig. 2. Time scale with actions to fault clearance process in transmission lines. Original software models for protection relays and phasor measurement units (PMU’s) are programmed and tested within the real-time simulation platform, the last one with the goal to make long time registers. As mentioned above, the contribution is that the analyst does not program the different sequences, but only the beginning of the failure and its type. Technical challenges are presented in the process of adding the response action of the relay automatically by adding settings, given via software interface commands and prepare this interface to make acquisition data from analog signals in physical devices. The Section 2 shows the models implemented in OPAL-RT® platform, continuous and discrete models. In Section 3, the process to make the fault clearance process is explained previous to propose the test case that is presented in Section 4. The last part is the conclusions, Section 5. IMPLEMENTATION IN REAL-TIME PLATFORM Real-time simulation is a modern way for the design and improvement of electrical apparatus to Power Systems. With the evolution of computing technologies have appeared simulation tools to produce simultaneously slow and fast dynamics transient phenomena.5 Researchers beforehand complete the activity for a model validation and authentication of parameters. The enhanced models produce a formulation in the form: (1) (2) where A is a square matrix; x becomes the states, B is a matrix with parameters for altering u sources. Real-time platforms choose the trapezoidal rule as solver since it allows to engineers deal with different time steps according to zones of observation. By this manner a multi-rate simulation improve the exploitation of distributed parallel computing. Nowadays real-time platforms can provide amplified signals to feed directly protection relays with analog outputs. In this work has been programmed Ingenierías, Abril-junio 2017, Vol. XX, No. 75 � �Real time simulation with software relay models/ Carlos A. López, et al. all software models for representing the protection scheme and PMU’s as well a software models. Thus software in the loop (SIL) simulation is performed into an OPALRT® platform. In general this tool assigns: 1) One master computation subsystem block, 2) N number of slaves subsystems, and 3) One console subsystem as interface with user. In this work the center of attention is evaluate algorithms representing SIL process and the sceneries to use real signals for voltage and current and interact with physical breaker in a HIL simulation. By one hand are acting those enhanced models of Power System. While simulation is running, the topology suffers a modification by order of virtual relays programmed with the intention to create the signal for controlling breaker. Figure 3 is presented to make understandable the strategy to manage oriented objects in Matlab-Simulink® environment. The test system is separated as hierarchical subsystems as mentioned before with one master and some slaves. These are declared with relationship to an interior time step (∆T 1, ∆T 2, ∆T 3, ∆T 4 and ∆T 5). Slave system to group generators Evidently synchronous machines have a dynamic response slower in comparison with those for transformers and transmission lines. Thus based on this feature, it is organized slaves to manage zones A and C grouping generator models. The model includes a local control for each generator, figure 4 shows the implementation, by this manner the simulation can be extended with duration for several seconds. All required parameters are given by user through Simulink libraries.6 Line models on remote zones In order to optimize cycle time, slaves are prepared to concentrate models for remote areas. Perceive that a normal grid is with several interconnections, however Fig. 3. Software separation for implementation in real-time platform. 10 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Real time simulation with software relay models/ Carlos A. López, et al. Fig. 4. Simulink model of synchronous machines. some lines are where protection scheme do not respond. Since protection scheme programmed is based on distance relay logic, it still responds to faults over the adjacent transmission line, thus beyond that position it is a common practice to use a Π model for representing the transmission lines. In other way, models as universal line model (ULM) in which it has distributed parameters and it can even analyze the frequency dependency between the same parameters. For this case a model is used of aerial line only of distributed parameters without including the detail of the frequency, taking as a base the formulation of aerial lines for electromagnetic transient simulations,7 where: (3) (4) where the convolution denoted by ⊗ can be expressed by: (5) Protection relay scheme Become aware the need to declare a protection zone D, where relay should take action demand distributed parameter models for transmission lines. This zone considers an internal fault, so main transmission line in study is divided in two segments in order to insert a fault.8 The fault model is with a value given by the user as a resistance. Internally this slave (zone D) will be propagating waveforms with bandwidth by the order of 20 kHz, thus the internal time step is 50 µs. In figure 3 is noticeable arrow connections among slaves. It is presented a relationship between the zone D and discrete devices. The simulated signals pass through current transducer models considering saturation. By this manner we add distinctive noise in the input of relay models. The protection relay is the decision device; this decision is made by a logic process in three steps: fault distance (FD), zone discrimination (ZD) and fault classification (FC). Ingenierías, Abril-junio 2017, Vol. XX, No. 75 11 �Real time simulation with software relay models/ Carlos A. López, et al. Figure 4 shows the scheme or distance relay implemented. If these three algorithms find a fault condition in the protected element then the trip signal is sent to a CB, this device is the actuator that can perform the opening and/or reclosing. The detail in distance relay algorithms can be reviewed in.8 Digital filter and phasor calculation inside the protection relay A sinusoidal waveform can be characterized by a complex number know as phasor. Consider a pure sinusoidal signal given by: (6) The phasor illustration of (6) is set as: (7) Note that the signal frequency (ω) is not explicitly stated in the phasor representation. The magnitude of the phasor is the rms value and its phase is ϕ. This representation implies that the signal remains stationary at all times, the magnitude, frequency and phase do not change. The most common technique for determining the phasor of an input signal is to use data samples taken from the waveform and apply the discrete Fourier transform (DFT) or the fast Fourier transform (FFT), since sampled data are used to represent the input signal, it is essential that antialiasing filters be applied.7 If xk, k = 0, 1, . . . , N −1 are the N samples of the input signal taken over one period of the waveform input signal, and the phasor is given by (8) the peak value of the fundamental frecuency thus obtained is then convert to rms value by dividing by . The phase angle of the phasor is the angle between the time when the first sample is taken and the peak of the input but nonharmonic components and any other noise leads an error in estimation of the phasor, this error of estimation due to these effects has been discussed in the literature. Figure 5 represents the interconnection between current and voltage transducer (CT and VT, respectively) with the transmission line, bus and protection relay, in this case the output of relay go to the breaker (52, ANSI nomenclature). Synchrophasor The term synchrophasor is used to describe a phasor which has been estimated at a specific instant time known as time tag. In order to obtain simultaneous measurements in a Wide Area of power system, it is necessary to synchronize these time tags.10 The synchronization is achieved by using a sampling clock signal provided by a GPS receiver. The connection diagram to have a PMU is similar to figure 5 to protection relay; the difference is that PMU have an extra input with the GPS data and the output go thought optical fiber channel until the concentrator in a Electrical 12 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Real time simulation with software relay models/ Carlos A. López, et al. Fig. 5. Structural diagram block connection diagram of PMU. Control Center or someplace like that. The phasor data concentrator (PDC) receives and time-synchronizes phasor data from multiple PMUs to produce a real-time, time-aligned output data stream. APDC can exchange phasor data with PDCs at other locations. Through use of multiple PDCs, multiple lavers of concentration can be implemented within an individual synchrophasor data system.11 Implementation of PMUs The implementation is performing in OPAL-RT® platform with the RTLabv11.05® and Matlab®. The implementation consist in a FAA filter, this is emulate as a Butterworth function of Fifth order with a 480 Hz as a cutting frequency. The A/D converter is used to transfer the analog signal to digital samples, it uses a sampler with the capacity of 64 samples by cycle of fundamental frequency of input signal, in this case 60 Hz. The digital filter is used to remove all the harmonic component over the fundamental frequency, after it, the phasor is calculated as was show in the section before in this article. The Peak determination block is used to identify the peak of the input signal and recorded the tag time in this moment. The tag time in this case is getting by the Microsoft Windows® clock, which is synchronized through the internet to global time. The output of data in the common format that is presented in the IEEE C37.118.1-2011 Standard.10 In this case the output is showed as a display but it will be send by a fiber optic communication channel to its assigned PDC. Fault clearance process by protection relays in real-time simulation platform The protection system has the responsibility of detecting, classifying, locating and isolating faults in any element of power system; in this case the transmission line is the device to protect. The protection systems consist in three devices: transductors for voltage and current measurements (TP and TC), protection relays and circuit breakers (CB). As Ingenierías, Abril-junio 2017, Vol. XX, No. 75 13 �Real time simulation with software relay models/ Carlos A. López, et al. primary protection 1 (PP1) one distance relay at each end of the line is installed, without the communication channel between these two relays. In figure 6, the arrangement that is applied for the tests where the voltage meter and the current that feed the distance relays are presented, the relay emits the error signal to the switch which has the capacity to realize monopolar openings, finally a single-phase fault, 50% of the line and the transient character. The protection relay is the device in charge of performing the conversion process from analog signal of the transducers to digital signals in phasor format, can also deliver synchrophasor if the relay is added to the input of a GPS. The distance relay implemented has a decision time of 1.5 cycles and the CB of 5 cycles to perform the opening, the time is considering conditions of the fundamental frequency in 60 Hz for the tested case. Fig. 6. Scheme of primary protection system in the transmission line. Figure 7 shows the diagram for the real-time simulation platform implementation of the test case. In the master system, the generators, some transmission lines, circuit breakers; in one slave subsystem exits loads, transmission lines, buses and measurements are included; the second slave subsystem processed of signals measured for protection purposes in the distance relay and for synchrophasor recording and analyzing within the PDC, the distance relay has the ability to send signal to open the switches in case of failure. The Console subsystem is used as a monitoring space for raw and/or processed signals from the various nodes. The detail in the subsystem separation for any implementation in real-time platform OPAL-RT® can be consulted in.4 Test case To show the raised tools and the solutions obtained is proposed to use as a test system the Kundur model which is presented in figure 8. It is proposed to perform a single-phase ground fault in phase A between bus 1 and bus 2 in the double circuit that joins the two generation areas. The fault is located at 49% of the line from B1 to B2, the fault resistance is 10 Ω. The protection system consists of two distance relays, each at one terminal of the transmission line, a circuit breaker is associated with each relay and it has the ability to perform single-pole openings. 14 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Real time simulation with software relay models/ Carlos A. López, et al. Fig. 7. Scheme of RT- implementation by subsystems. Fig. 8. Kundur System implemented as test case. The registers are of magnitudes voltages and currents coming from PMUs, also the registers of phase sinusoidal signals of both voltages and currents are included; all measurements are seen from the bus 1. The PMU implemented has the ability to transmit voltages and currents in phase components as magnitude and phase, in this case only the magnitudes are shown from the bus 1. The PMU implemented has the ability to transmit voltages and currents in phase components as magnitude and phase, in this case only the magnitudes are shown. Figure 9 represents the magnitude of voltages in phase components, the register begin before the insertion the fault and stop three seconds after the CB finish their own operations. In the same way figure 10 displays the magnitude current phasor in ABC components; all those are digital values processed inside the PMU and delivered to a cup of 30 samples per second and for a better appreciation we proceeded to present in a graphical format instead of tabular. Figure 11 and figure 12 are the voltages and currents as sinusoidal signals before the signal processing as syncrophasor, in these signals is not possible in a clear form recognize the changes occurring in the magnitudes during the captured record, specifically in the voltage register. Ingenierías, Abril-junio 2017, Vol. XX, No. 75 15 �Real time simulation with software relay models/ Carlos A. López, et al. Fig. 9. Three phasor voltage signals. Fig. 10. Three phasor currents signals. Fig. 11. Three phase voltage signal in time. 16 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Real time simulation with software relay models/ Carlos A. López, et al. Fig. 12. Three phase current signal in time. CONCLUSION In this article is presented the process performed for fault clearance in a realtime Simulation platform; the work presents long-term records for voltage and current processing signals. The fault identification and the output opening or reclosing of breakers are due to the implementation of distance relays with the ability to respond when contingencies scenarios occurs in its primary protection zone. In this work has been obtained records of voltage and current signals by processing data in synchrophasor format, following the protocol presented by the IEEE Standard. The register is made in abc phase components instead of being only positive sequence; this is for voltage and current signals. Taking advantage of the recording capacity of the OPAL-RT® simulation platform, the sinusoidal voltage and current signals were stored and these are presented. In addition the acquisition system to protection relays and PMU has the ability to receive not only virtual signals from simulation, also with enable the analog channels in the real-time simulator can receive analog signals and processing them in the same form that the virtual. ACKNOWLEDGMENT Authors want to thank the National Science Foundation of Mexico (CONACYT) for the support of this project. REFERENCES 1. V. Terzija, G. Valverde, D. Cai, P. Regulski, V. Madani, J. Fitch, S. Skok, M. M. Begovic and A. Phadke, “Wide-Area Monitoring, Protection, and Control of Future Electric Power Networks ”, Proceedings of the IEEE, vol. 99, no. 1, pp. 80–93, 2011. 2. IEEE PES Task Force on Real-Time Simulation of Power and Enerdy Systems, “Applications of Real-Time Simulation Technologies in Power and Energy Systems”, IEEE Power and Enerdy Technology System Journal, vol. 2, no. 3, pp. 103-115, Sep. 2015. Ingenierías, Abril-junio 2017, Vol. XX, No. 75 17 �Real time simulation with software relay models/ Carlos A. López, et al. 3. [Online] http://www.opal-rt.com/system-emegasim/ 4. [Online] http://www.opal-rt.com/simulation-systems-overview/ 5. H.T. Su, K. W. Chan, L. A. Snider, T. S. Chung, and D. Z. Fang, “Recent Advancements in Electromagnetic and Electromechanical Gybrid Simulation” Proceedings of the 2004, Singapore, Nov. 2004. 6. [Online] https//www.mathworks.com/help/physmod/sps/motors-andgenerators-html?s_tid=gn_loc_drop 7. J. G. Proakis, D. G. Manolakis, Digital Signal Processing, 4th ed., Pearson Prentice Hall. 18 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Adaptive autoreclosaure to increase system stability and reduce stress to circuit breakers Jörg BlumscheinA, Yilmaz YelginA, Andrea LudwigB A B Siemens AG, Energy Management Division, Munich, Germany 50Hertz Transmission GmbH, Berlin, Germany RESUMEN El recierre automático es un elemento clave en el concepto de rejillas autosanadas. Según las estadísticas, una gran cantidad de fallas de transmisión y distribución son fallas temporales, las cuales desaparecen cierto tiempo después de la desenergización de las secciones de la red falladas. El recierre automático se utiliza para recuperar el estado original de la red sin intervención humana, y puede ser hecha en tres polos o en uno solo. En este artículo se describe la aproximación utilizada en Alemania, Polonia y Austria para liberar distintas fallas de fase a fase sin tierra, mediante el autocierre de un solo polo. En este caso se pueden llevar a cabo mediciones de voltaje durante el tiempo muerto del polo para determinar si el recierre fue exitoso. PALABRAS CLAVES Autocierre automático, autocierre de un solo polo, estabilidad de sistemas. ABSTRACT Automatic reclosure is a key element in the concept of self-healing grids. According to statistics, a large amount of faults in transmission and distribution networks are temporary faults. These faults disappear a certain time after deenergization of the faulted sections of the network. Automatic reclosure is used to recover the original status of the network without any human interaction. Automatic reclosure can be done as a three pole autoreclosure or a single pole autoreclosure. This paper describes an approach used in Germany, Poland and Austria to clear such phase to phase faults without ground by the means of a single pole autoreclosure. Also in this case voltage measurements during the single pole dead time can predict whether or not a reclosure will be successful. KEYWORDS Automatic reclosure, single pole autoreclosure, system stability. INTRODUCTION According to statistics, 80 to 85 percent of faults at transmission and distribution lines are temporary faults. Lightning is the most usual case for temporary faults but there are other reasons too. Swinging conductors contacting Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 19 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. each other caused by strong wind or shedding of ice can cause temporary phase to phase faults. Other well known reasons for temporary faults are related to bird streamers or vegetation reaching too close to the conductors due to lack of maintenance. These faults disappear a certain time after de-energization of the faulted sections of the network. Automatic reclosure is used to recover the original status of the network very fast and without any human interaction. Figure 1 is used to explain the basic principle of autoreclosure. A typical transmission line is connecting two parts of a network which are connected to the busses A and B. In figure 1a fault on the line is detected by the relays at both ends of the line measuring the currents I A and IB and the voltages U A and UB. As soon as the relays A and B detect that the fault is on the protected line they will send a trip command to open the associated circuit breaker CB like shown in figure 1b. At this time the automatic reclosing functions integrated in the relays A and B start the dead time of the autoreclosure. During this dead time the fault has the chance to extinguish. After the dead time is expired, the automatic reclosure function sends a close command to the associated circuit breaker. For transient faults a successful reclosure is mostly obtained with the first reclose cycle like shown in figure 1c. If the fault still persist the protective relays A and B will detect this and send a trip command again. A tree branch falling on a line for instance may need a second reclose cycle to burn up by the arc when the line is re-energized. For permanent faults caused by a broken conductor, the collapse of a line tower, trees falling onto a line or faults in cables a reclosure is not able to clear the fault. Therefore it is important to detect this condition and send a final trip to the circuit breaker. Based on experiences the most utilities apply automatic reclosure with one reclose cycle only. This is because the increasing chance to get a successful autoreclosure does not justify the stress to circuit breakers and system due to additional close-open cycles under full fault current in case of permanent faults. Autoreclosure can be distinguished as three pole autoreclosure or single pole autoreclosure like shown in figure 2. For three phase faults or phase to phase faults with ground all three phases must be isolated to clear the fault. For single phase to ground faults which are the great majority of faults in the transmission and distribution systems a single pole autoreclosure is sufficient to clear the fault. In Section 2 a scheme is described to use single pole tripping and autoreclosure also in case of phase to phase faults without ground like shown in figure 2. Today many utilities use single pole autoreclosure. Single pole autoreclosure has the following advantages compared to three pole autoreclosure: • Transport of energy possible during the dead time via the two remaining wires. • No synchronization needed before reclosing. • Enhanced system stability and reliability. In general the goal of the autoreclosure is to restore the line to service as quickly as possible. This goal has to be balanced with the negative effects of closing onto a fault which produces a lot of stress to the circuit breaker and the 20 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. electrical sytem. In Section 2 a scheme is explained how to prevent the closing of the second circuit breaker if the closing of the first circuit breaker was not successful. This scheme can help to reduce the negative impact that failed autoreclosure attempts have on the system. A scheme called “autoreclosure with adaptive dead time” is possible for three pole autoreclosure as well as for single pole autoreclosure like shown in figure 2. Section 3 will explain methods to improve single pole autoreclosure for single phase to ground faults by detecting whether or not a reclosure will be successful by means of secondary arc detection. Section 4 will explain the single pole autoreclosure scheme for phase to phase faults without ground and methods to detect whether or not the reclosure will be successful in this case. AUTORECLOSURE WITH AN ADAPTIVE DEAD TIME Autoreclosure with an adaptive dead time is a typical autoreclosing scheme for transmission lines also known as “leader follower autoreclosing scheme” 1. In this scheme the leader is defined as the line terminal that autorecloses first after a fixed dead time. The follower is the line terminal that recloses second and only if the reclosing of the leader was successful. In this scheme the leader is used to verify whether or not the fault is extinguished during the dead time of the autoreclosure. If the fault still persists the leader will open the associated circuit breaker again. In this case the follower does not attempt the autoreclosure which has the great advantage of reducing unnecessary stress to the circuit breaker at the follower end of the line. Figure 3 explains the behavior of the scheme in case of an unsuccessful autoreclosure. In figure 3a a fault condition is shown on a line protected by the relays called L (leader) and F (follower). Both relays detect the fault on the protected line and open the line by means of the associated circuit breakers shown in figure 3b. After the fixed dead time is expired only the leader recloses the breaker to verify Fig. 1. Basic principle of autoreclosure. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 21 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. Fig. 2. Autoreclosure schemes for different types of fault. whether or not the fault still persists. If the fault still persists like shown in figure 3c, the leader opens the circuit breaker to start another autoreclose cycle or send a final trip like shown in figure 3d. Figure 4 explains the same scenario for a successful autoreclosure. Different to figure 3c is that the fault does not persist after reclosing of the leader side shown in figure 4c. If there is a communication channel between the relays at both ends of the line the leader can send a “remote close command” to close the circuit breaker associated to the follower at the remote end of the line as shown in figure 4d. 2 Another solution for an autoreclosure with adaptive dead time without communication channel is explained in figure 5. A fault occurs at a transmission line like shown in figure 5a. Both relays detect the fault and open the associated circuit breaker like shown in figure 5b. Autoreclosure with fixed dead time is started at the leaders end only. After the fixed dead time is expired the leader closes the associated circuit breaker as shown in figure 5c. If the fault does not persist anymore the follower will detect a healthy voltage which indicates that the line was successfully re-energized from the remote end. Consequentially the autoreclose function in the follower device can close the circuit breaker like shown in figure 5d. This kind of adaptive autoreclosure requires that: • The voltage transformers are located on the line side of the circuit breaker at the follower end like shown in figure 5, • The leaders end of the line is strong enough to maintain a healthy voltage after reclosure. At 50 Hertz transmission the follower releases the close command if a voltage greater than 70% of nominal voltage is measured for more than 300 ms. SECONDARY ARC DETECTION A successful autoreclosure requires a dead time which exceeds the de-ionizing time, the time needed for the fault to extinguish. This time required for the deionizing of the fault path depends on several factors including: • Arcing time (fault duration). 22 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. • • • • Fault current. Weather conditions like wind, air humidity and air pressure. Circuit voltage. Capacitive coupling to adjacent conductors. In general the circuit voltage is the predominating factor influencing the deionizing time. For single pole autoreclosure there is another effect which has a significant influence to the success of the autoreclosure. The primary arc current is interrupted by disconnecting the faulted phase from the sources by opening the circuit breakers at both ends of the line. After this a secondary arc can prevent the fault clearance. During the single pole dead time capacitive and inductive coupling from the other two phases induces a voltage into the open phase conductor which Fig. 3. Unsuccessful autoreclosure. Fig. 4. Successful autoreclosure using a remote close command. feeds the secondary arc like shown in figure 6. The success of a single pole autoreclosure depends on the extinction of this secondary arc. On transmission and distribution lines the coupling between the two remaining phases and the open phase can be sufficient to maintain the secondary arc in the ionized air of the primary arc path. Depending on the above mentioned influencing factors like fault duration, fault current, atmospheric conditions and constructive parameters of the line the secondary arc may take longer to extinguish. In worst case the secondary arc does not extinguish at all during a single pole autoreclosure and reclosing in the presence of the secondary arc will only re-energize the fault. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 23 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. There are several methods to detect the presence of the secondary arc. All these methods are based on the simplified equivalent circuit shown in figure 7. The secondary arc is an arc between the open phase and ground which is fed by the two healthy phases via capacitive coupling. The voltage UM , measured at the disconnected phase is characterized by the ohmic nonlinear behavior of the secondary arc. If the secondary arc is extinguished the equivalent circuit is changing to a different model like shown in figure 8. The voltage UM , measured at the disconnected phase after extinguishing of the secondary arc is characterized by the linear capacitive behavior of the phase to ground capacitance of the open conductor. Fig. 5. Successful autoreclosure using line side voltage measurement. Figure 9 shows the current and voltage for a successful autoreclosure after a single phase fault on a transmission line. After tripping the line we can see that the fault current disappears. At the same time the voltage starts the typical nonlinear behavior of arcing. At a certain time the secondary arc extinguishes and the voltage is changing to a linear capacitive behavior. Finally voltage and current goes back to normal conditions after successful reclosing. Figure 10 however shows a case where the secondary arc does not extinguish during the single pole dead time. After reclosure the fault still persists which leads to a final trip of the protection. In 3 a method is described to detect the presence of the secondary arc using the relation between the fundamental com- ponent and the harmonics of the phase to ground voltage of the open phase. Figure 11 shows the harmonic content of the open phase voltage during the presence of the secondary arc on a 400 kV transmission line. Due to the nonlinear characteristic of the secondary arc there is a huge portion of 3rd, 5th and 7th harmonic. After the secondary arc is extinguished the voltage is rising up to 42 kV but without any harmonics like shown in figure 12. In 4 a method is described detecting the secondary arc based on the angle of the open phase voltage in relation to the other phase to ground voltages. Figure 13 shows the phasor diagram of the phase to ground voltages during the presence of the secondary arc for the same fault record of a 400 kV transmission line. Due to the ohmic characteristic of the secondary arc the voltage of the open phase lags the voltage phasor of the pre-fault voltage by 90°. After the secondary arc 24 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. is extinguished the voltage phasor is rising in magnitude and is located between the two healthy voltage phasors like shown in figure 14. Another method is given in 5 which detect the extinguishing of the secondary arc by evaluating the amplitude of the third harmonic component of the zero sequence voltage. All three methods were applied to a set of 46 fault records captured by real events from the transmission system of 50Hertz. At this time 50Hertz Transmission was using adaptive autoreclosure with a fixed dead time of 1.2 s for the leader for single phase faults. According to figure 15 the secondary arc was already extinguished after 0.2 s in many cases. Only in 3 cases the secondary arc needed 0.8 s to extinguish. By using adaptive autoreclosure with secondary arc detection the fixed dead time of 1.2 s for the leader could be reduced significantly in most cases. Fig. 6. Capacitive and inductive coupling between the three phases of a transmission line. There are other cases where the secondary arc does not extinguish even after 1.2 s. But secondary arc detection is also important for these cases. As soon as it is clear that a secondary arc and not a permanent fault was the reason for the unsuccessful reclosure, a manual closing of the line is permitted without a time consuming line patrol in advance. SINGLE POLE TRIPPING FOR PHASE TO PHASE FAULTS WITHOUT GROUND Under extreme weather conditions line swinging can cause an increasing number of phase to phase faults. These faults are mostly flash-arcs between two wires of a transmission or distribution line. Figure 16 shows a simplified equivalent circuit for such kind of faults. In 6 a scheme was protected by patent to clear phase to phase faults without ground by means of a single pole autoreclosure. It is obvious like shown in figure 17 that a single pole trip will clear a temporary phase to phase fault in most cases. In such schemes there are two options for single pole tripping in case of phase to phase faults without ground: trip leading phase or trip lagging phase. It must be ensured that all protective relays in a given network use the same phase preference for single pole trip in case of phase to phase faults. This scheme is successfully applied in Germany, Poland and Austria for many years to take the advantages of single pole autoreclosure also for phase to phase faults without ground. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 25 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. Figure 18 shows a fault record for a successful single pole autoreclosure for a phase to phase fault on the 220 kV system in Germany. After tripping of phase C the fault current in phase A and phase C disappears at the local end. Approximately 300 ms later also the voltage UC goes down indicating the isolation of the fault. Finally a successful reclosure brought the system back to normal conditions. According to 6 a successful isolation of the arc between the two faulted phases is given if the phase to ground voltage of the tripped phase is measured to be below a certain value for a given time like shown in figure 18. Figure 19 shows an unsuccessful single pole autoreclosure of phase C for a fault between phase A and phase C. Different to figure 18 the phase to ground voltage of the tripped phase C does not fall below a certain value for a given time. Detecting this condition an unsuccessful autoreclosure could be prevented in the future. Here also voltage measurements during the single pole dead time can predict whether or not a reclosure will be successful. Fig. 7. Simplified equivalent circuit of secondary arc, fed by capacitive coupling from the two healthy phases. Fig. 8. Simplified equivalent circuit after secondary arc is extinguished. 26 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. Fig. 9. Secondary arc extinguishing during single pole dead time. Fig. 10. Secondary arc not extinguishing during single pole dead time. Fig. 11. Harmonic content of voltage during the presence of the secondary arc. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 27 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. Fig. 12. Harmonic content of voltage after secondary arc is extinguished. Fig. 13. Phasor diagram of voltages during the presence of the secondary arc. Fig. 14. Phasor diagram of voltages after the secondary arc was extinguished. 28 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. Fig. 15. Time needed for the secondary arc to extinguish during the single pole dead time. Fig. 16. Simplified equivalent circuit for a phase to phase arcing fault without ground. Fig. 17. Simplified equivalent circuit for a single pole trip for a phase to phase fault without ground Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 29 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. Fig. 18. Fault record of successful single pole autoreclosure for a phase to phase fault. Fig. 19. Fault record of unsuccessful single pole autoreclosure for a phase to phase fault. CONCLUSION It was shown that using adaptive autoreclosure the system stability can be increased by adaptively shorten the dead time of the autoreclosure and prevent unnecessary reclosing onto faults. Several different methods were explained how to use voltage measurement during the single pole dead time to reduce unnecessary stress to the circuit breaker by reclosing onto faults. 30 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Adaptive autoreclosure to increase system stability and reduce stress to circuit breakers/ Jörg Blumschein, et al. REFERENCES 1. “IEEE Guide for Automatic Reclosing of Circuit Breakers for AC Distribution and Transmission Lines,” IEEE Std C37.104-2012 (Revision of IEEE Std C37.104-2002), pp. 1–72, 2012. 2. “SIPROTEC 5, Distance and Line Differential Protection, Breaker Management for 1-Pole and 3-Pole Tripping,” Siemens AG, Tech. Rep., 2015. 3. R. Luxenburger, P. Schegner, and A. Ludwig, “Beurteilung der Lichtbogenlöschung bei der einpoligen AWE in Höchstspannungsnetzen,” in Internationaler ETG-Kongress 2005, Dresden, 2005, pp. 1–9. 4. A. Guzmán, J. Mooney, G. Benmouyal, N. Fischer, and B. Kasztenny, “Transmission line protection system for increasing power system requirements,” in Proceedings of the International Symposium Modern Electric Power Systems (MEPS), 2010, Wroclaw, Poland, 2010, pp. 1–11. 5. S. Jamali and A. Parham, “New approach to adaptive single pole autoreclosing of power transmission lines,” IET Generation, Transmission & Distribution, vol. 4, no. 1, p. 115, 2010. [Online]. Available: http://digital-library.theiet. org/content/journals/10.1049/ iet-gtd.2009.0058 6. “Schaltungsanordnung zur Beseitigung zweipoliger Lichtbogenkurz- schlüsse in Hochspannungsnetzen,” 1972. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 31 �Umbral para discriminar entre corriente inrush y falla interna en un transformador de potencia Martha N. Acosta Montalvo, Héctor Esponda Hernández, Manuel A. Andrade Soto, Ernesto Vázquez Martínez Universidad Autónoma de Nuevo León, FIME, Doctorado en Ingeniería Eléctrica martha.acostamnt@uanl.edu.mx RESUMEN En este artículo se presenta el cálculo de un umbral que discrimina entre la corriente de energización y una falla interna en un transformador de potencia. Se forma una matriz que contiene las corrientes diferenciales de una señal de referencia y se obtienen los eigenvalores máximos, estos son utilizados para el cálculo del umbral. A partir de la matriz de corrientes diferenciales que contiene la señal de falla, se obtiene el eigenvalor dominante y se compara con el umbral, si la magnitud del eigenvalor dominante es menor al umbral se tiene una energización, de lo contrario se tiene una falla interna. PALABRAS CLAVE Umbral, eigenvalores, corriente inrush, transformador, protección diferencial. ABSTRACT This paper describes a threshold which discriminates between inrush current and internal fault in a power transformer. A matrix, which contains differential currents of a reference signal, is formed. The maximum eigenvalues of the matrix are obtained; these are used to calculate the threshold. The dominant eigenvalue is obtained from a matrix which contains differential currents of a fault signal; dominant eigenvalue is used for comparison with the threshold. If the magnitude of the eigenvalue is less than the threshold then it is an inrush current, otherwise it is an internal fault. KEYWORDS Threshold, eigenvalue, inrush current, transformer, differential protection. INTRODUCCIÓN El transformador es un dispositivo fundamental para la operación del sistema eléctrico de potencia. Cuando ocurre la energización del transformador, los dispositivos de protección pueden operar de forma incorrecta debido a la corriente inrush,1 la cual solo es detectada por los TCs de lado primario y esta es interpretada como una falla. 32 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Umbral para discriminar entre corriente inrush y falla interna.../ Martha N. Acosta Montalvo, et al. La correcta discriminación entre la corriente inrush y una falla interna supone un reto para la operación correcta de los dispositivos de protección. A lo largo del tiempo, se han propuesto diferentes métodos para dar solución a dicho problema, los cuales se pueden clasificar en tres grupos. El primer grupo emplea el contenido armónico de la corriente diferencial para bloquear la protección en escenarios de energización del transformador, conexión y desconexión.2-3 El segundo grupo utiliza el reconocimiento de la forma de onda de la corriente diferencial.4-5 El tercer grupo usa las señales de corriente y tensión como entradas.6-7 Algunos métodos basados en el reconocimiento de la forma de onda han propuesto diversos algoritmos, en los cuales la discriminación de la corriente inrush y una falla interna se realiza mediante un umbral. Mediante el análisis modal, en8 se establece que para eventos de corriente inrush, el eigenvalor dominante no supera un umbral de 0.4, mientras que para una falla interna el eigenvalor dominante es mayor al umbral. En,4 se calcula el factor de singularidad de la forma de onda (WSF) el cual obtiene la diferencia entre la forma de onda de la corriente y una forma de onda sinusoidal, cuando ocurre una falla la forma de onda de la corriente es aproximadamente una sinusoidal y WSF tiene valores muy cercanos a cero, cuando se tiene una corriente inrush WSF tendrá un valor elevado. Si WSF > 1 se tiene una corriente inrush, de lo contrario se tiene una falla interna.9 Propone el algoritmo de correlación mejorado, en el cual la corriente durante un ciclo es reorganizada como una nueva medición. Las corrientes de falla se distinguen de la corriente inrush si el coeficiente de correlación mejorado entre el primer medio ciclo y el medio ciclo restante es mayor a un umbral. El umbral es determinado con base en el límite de corriente inrush. En,10 se estructuran dos tipos de formas de onda sinusoidales normales de acuerdo con el valor y la posición del punto máximo de corriente en la zona de no saturación. Se calculan coeficientes de correlación (CC) entre la forma de onda original y dos formas de onda sinusoidales estructuradas. En base al CC se crean dos criterios, los cuales son comparados con un umbral propuesto de 0.8; si CC > 0.8 se tiene una corriente inrush, de lo contrario se tiene una falla interna. En 11 se calcula la función de auto-correlación (ACF) para las corrientes diferenciales, y después se calcula la desviación estándar de la ACF. Se define un criterio; el cual consiste en la diferencia entre el máximo y el mínimo de la desviación estándar de cada fase. Este resultado se compara con un umbral de 0.057; si el criterio es mayor al umbral, se tiene una corriente inrush, de lo contrario se trata de una falla interna. En este artículo, se propone el cálculo de un umbral para la correcta discriminación de la corriente inrush y una falla interna en el transformador con base en el análisis de eigenvalores. A partir de señales de referencia, las cuales contienen información de la energización del transformador, se obtiene el máximo pico de corriente inrush. A la matriz que contiene las corrientes diferenciales se le aplica un filtro delta para remover la componente de estado estable y se normaliza. De esta matriz, se obtiene una matriz simétrica, de la cual se calcula el eigenvalor mayor, el cual es utilizado para el cálculo del umbral. Durante la operación del algoritmo, a partir de las señales que se están midiendo, se obtiene el eigenvalor dominante y la magnitud de éste es comparada con el umbral calculado anteriormente. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 33 �Umbral para discriminar entre corriente inrush y falla interna.../ Martha N. Acosta Montalvo, et al. PRE-PROCESAMIENTO DE LAS SEÑALES Se realiza un pre-procesamiento de las señales de corriente diferenciales, el cual consiste en la aplicación de un filtro delta con el objetivo de remover el estado estable de la señal. Debido a que la corriente inrush puede alcanzar magnitudes mayores a las de la corriente de falla, las corrientes diferenciales se normalizan con respecto al valor absoluto de la corriente diferencial máxima.8 Algoritmo QR A k puede ser descompuesta en el producto de una matriz unitaria Q k y una matriz triangular superior R k , la cual es obtenida de la multiplicación H H de Q k A k dónde Q k = Q k . Ak +1 es formada por la multiplicación de R k Q k Una matriz [12]. Entonces: A1 = A (1) Ak = Qk R k (2) A k +1 = Q kH A k Q k = R k Q k (3) para k=1,2, …. La ecuación (3) se puede escribir como una transformación de similaridad: A k +1 = Q Hk Q Hk −1  Q1H AQ1  Q k −1Q k (4) La matriz A k tiende a una matriz triangular superior cuando k → ∞ y los elementos contenidos en la diagonal son los eigenvalores de A . Umbral Los eigenvalores de una matriz A∈ℝn×n simétrica son reales y no negativos denotados por ,y los eigenvectores derechos correspondientes u1 , u2 , , un son ortogonales.13 (5) Si se tienen una matriz B∈ℝ n×m , (5) se puede reescribir como (6) donde ⋅ denota la norma 2 y B T B da como resultado una matriz simétrica de n×n . Aplicando la raíz cuadrada en ambos lados de (6) (7) A partir de (7) se establece que para cada eigenvalor no nulo de la matriz generada por B T B se tiene un valor σ: (8) 34 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Umbral para discriminar entre corriente inrush y falla interna.../ Martha N. Acosta Montalvo, et al. el cual cumple con lo siguiente (9) (10) A partir de (10), se observa que σmax y σmin proporcionan un rango de las magnitudes que pueden tomar los eigenvalores de la matriz y (10) puede ser utilizado para establecer límites en la magnitud en los eigenvalores del sistema para ciertas condiciones de operación. Establecer el límite máximo de magnitud que pueden tomar los eigenvalores de una señal sin falla sirve como referencia para identificar disturbios en la zona de protección del transformador de potencia. Dicho lo anterior, se plantea el cálculo de un umbral que permita discriminar entre la corriente inrush y una falla interna. Se tiene una matriz AR que contiene las corrientes diferenciales de una señal de referencia. Mediante el algoritmo QR se obtiene los eigenvalores de AR , denotados por y mediante (9) se obtiene σmax1, σmax2, ...σmaxk. Una vez obtenidos los límites máximos de magnitud se calcula el umbral (11) donde k representa el número de σmax obtenidos a lo largo de la señal de referencia. Señal de Referencia En14 se determina que para una energización en 0º, la magnitud de la corriente inrush es la máxima, mientras que al energizar en 90º se obtiene la mínima magnitud de la corriente inrush. En15 se evaluó el impacto que tiene la impedancia de la fuente en la magnitud de la corriente inrush y se determinó que, ésta es muy sensible a la impedancia de la fuente, mientras la impedancia de la fuente sea menor el pico de la corriente inrush será mayor. La relación X/R de la fuente no afecta la magnitud de la corriente inrush. La señal de referencia debe contener el transitorio generado por la corriente inrush. Se considera la máxima corriente inrush esperada para el transformador de potencia con el fin de obtener un umbral que cumpla con la condición de discriminar entre la corriente inrush y una falla interna. Para obtener el máximo pico de corriente inrush, el ángulo de energización del transformador fue de 0º y la impedancia de la fuente es muy pequeña. Criterio de discriminación De (5) se observa que el eigenvalor mide la razón de cambio de la magnitud de sus eigenvectores asociados. Debido a esto se determina que la magnitud del eigenvalor dominante λd proporciona la mayor información de las variables de estado de la forma en la respuesta transitoria de onda.8 El eigenvalor dominante es el utilizado para el análisis del sistema. Se tiene una matriz que contiene las corrientes diferenciales de una señal que se está monitoreando. De esta matriz, se obtiene una matriz simétrica, de la cual se Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 35 �Umbral para discriminar entre corriente inrush y falla interna.../ Martha N. Acosta Montalvo, et al. obtiene el eigenvalor dominante mediante el algoritmo QR, y este eigenvalor dominante es comparado con el umbral calculado en (11). Para una condición de falla dentro de la zona de protección del transformador (falla interna) la magnitud del eigenvalor dominante es mayor al umbral. En cambio, para condiciones del sistema diferentes a una falla la magnitud del eigenvalor dominante es menor al umbral: λd > umbral (12) λd < umbral (13) Sistema de prueba El sistema de prueba utilizado para la evaluación del umbral se muestra en la figura 1, el cual se compone de un transformador de potencia conexión deltaestrella aterrizada, equivalente Thévenin en ambos lados del transformador; una línea de transmisión modelo π y una carga en el lado secundario del transformador; el sistema opera a 60 Hz. Los parámetros del sistema son obtenidos de,8 en la tabla I se describen los parámetros de los generadores y la carga; en la tabla II se describen los parámetros de la línea de transmisión y en la tabla III se describen los parámetros del transformador de potencia. Fig.1. Sistema de prueba. Tabla. I. Parámetros de los generadores y carga. Parámetro Unidad G1 G2 Carga Tensión línea-línea kV 115 13.8 13.8 Frecuencia Hz 60 60 60 MVA 90 3 80.06 Resistencia serie Ω 15 10 -- Resistencia paralelo Ω 1 1 -- Inductancias paralelo H 0.1 0.046 -- Grados 0 20 25.925 Potencia Ángulo Tabla. II. Parámetros del transformador de potencia. Parámetros Unidad Transformador trifásico de dos devanados Tensión en devanado kV 115/13.8 Frecuencia Hz 60 Potencia MVA 100 Conexión -- ∆ − Y aterrizado 36 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Umbral para discriminar entre corriente inrush y falla interna.../ Martha N. Acosta Montalvo, et al. Las simulaciones se realizaron a una razón de 64 muestras por ciclo lo cual equivale a una frecuencia de muestreo f s = 7.68 kHz. Tabla. III. Parámetros de la línea de transmisión. Parámetros Unidad Resistencia secuencia positiva Línea de Transmisión Ω Resistencia secuencia cero Ω Inductancia secuencia positiva H Inductancia secuencia cero H Capacitancia secuencia positiva F Capacitancia secuencia cero F Longitud 0.01273 km 0.3864 km 09337e-3 km 4.1264e-3 km 1274e-9 km 7.751e-9 km km 90 En la tabla IV se describen los parámetros de los transformadores de corriente utilizados en el esquema de protección diferencial del transformador. Tabla. IV. Parámetros de los transformadores de corriente. Característica Unidad T C l a d o T C l a d o primario secundario Relación de Transformación -- 500/5 4000/5 Resistencia secundario Ω 0.5 0.5 Inductancia secundario mH 0.8 0.8 Área m2 6.5 6.5 Longitud de la trayectoria m magnética 0.5 0.5 Flujo remanente T 0 0 Resistencia burden Ω 2 0.5 Inductancia burden mH 0.8 0.8 En la tabla V se resume el criterio de discriminación que se estableció en (12) y (13). Para la condición de energización del transformador, el umbral calculado mediante (11) es de 0.5 para el sistema de prueba de la figura 1. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 37 �Umbral para discriminar entre corriente inrush y falla interna.../ Martha N. Acosta Montalvo, et al. Tabla. V. Criterio de discriminación. Evento Magnitud del Eigenvalor Dominante Corriente Inrush Sobreexcitación [0, umbral] Falla Interna [umbral, ∞] RESULTADOS Con el fin de evaluar los criterios establecidos en la tabla V, se realizan simulaciones para diferentes condiciones de operación del sistema, las cuales se presentan a continuación. Energización El instante de energización y la curva de saturación del transformador impactan directamente la forma de onda y la magnitud de la corriente inrush. Para probar el umbral calculado, en la figura 2a) y 2b) se muestra la energización con un ángulo de 0º, 90º y 180º, respectivamente, con respecto a la señal de tensión. Se observa que para estas condiciones la magnitud del eigenvalor dominante no supera el umbral calculado. En la figura 2c) se muestra la energización cuando se ha realizado un cambio en la curva de saturación, para esta condición la magnitud del eigenvalor dominante es menor a la magnitud del umbral calculado. En todos los casos mostrados en la figura 2 se cumple con el criterio establecido en (13), discriminando correctamente la condición de energización respecto a una falla interna. Fig. 2. Energización del transformador a) a 90º, b) a 180º, y c) empleando TCs con curva de saturación modificada. 38 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Umbral para discriminar entre corriente inrush y falla interna.../ Martha N. Acosta Montalvo, et al. Sobreexcitación El transformador de potencia es propenso a condiciones de sobreexcitación debido a la caída de frecuencia del sistema y al aumento de la tensión aplicada. El transformador tolera máximo 110% de sobreexcitación con respecto a los valores nominales, esta sobreexcitación puede producir un incremento de la corriente inrush. Debido a esto se consideró probar el umbral calculado cuando el transformador se encuentra bajo esta condición, se realizó la energización del transformador y después se simuló una sobreexcitación debido a un 10% de la caída de frecuencia y un aumento de tensión del 50%. La figura 3 muestra el comportamiento del eigenvalor dominante ante la condición de sobreexcitación y se observa claramente que la magnitud del eigenvalor no supera el umbral calculado, cumpliéndose con el criterio en (13), i. e. no existe falla interna. Fig.3. Sobreexcitación después de la energización. Falla interna Las fallas dentro de la zona de protección del transformador deben ser detectadas correctamente y ser liberadas en el menor tiempo posible, ya que el transformador puede sufrir severos daños que repercuten en su tiempo de vida. Para probar el umbral calculado, se introduce una falla monofásica dentro de la zona de protección del relevador después de la energización. En la figura 4 se muestra el comportamiento del eigenvalor dominante ante dicha falla. La condición de falla interna es detectada correctamente, ya que la magnitud del eigenvalor dominante es mayor umbral al calculado. Para esta condición del sistema se cumple el criterio establecido en (12) y se inicia la operación de la protección diferencial. Falla interna con saturación en los TCs La forma de onda en el secundario de los TCs es distorsionada debido a la saturación del transformador de instrumento. La saturación en los TCs es causada por la presencia de fallas, los dispositivos de protección pueden tener retardo en la operación o simplemente tener una mala operación debido a esta condición. El umbral propuesto debe identificar la condición de falla aún con la saturación de los TCs. En la figura 5 se muestra el comportamiento del eigenvalor dominante ante la presencia de una falla interna monofásica, la cual causa la saturación de Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 39 �Umbral para discriminar entre corriente inrush y falla interna.../ Martha N. Acosta Montalvo, et al. los TCs. La magnitud del eigenvalor dominante ante la condición de falla es mayor a la del umbral calculado y se cumple el criterio establecido en (12), lo que indica una correcta discriminación de la falla interna. Fig. 4. Falla interna después de la energización. Fig. 5. Falla interna con saturación de los TCs después de la energización. CONCLUSIÓN Se calculó un umbral para la discriminación de una falla interna y la corriente inrush, el cual depende de una señal de referencia, la cual es obtenida a partir de la energización del transformador en un ángulo de 0° y la impedancia es muy pequeña, puesto que con estas condiciones se obtiene el caso más crítico de corriente inrush. El eigenvalor dominante del sistema fue utilizado para establecer los criterios de discriminación, obteniéndose un umbral de 0.5 de acuerdo a la condición base de energización. Se realizaron simulaciones con diferentes escenarios de operación para analizar y verificar la confiabilidad del umbral calculado. Puede verse (figura 4 y figura 5) que el umbral propuesto discrimina correctamente las fallas internas y se elimina la operación incorrecta de los dispositivos de protección ante condiciones de energización del transformador, sobreexcitación (figura 2 y figura 3), aun cuando se presente saturación de los TCs. 40 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Umbral para discriminar entre corriente inrush y falla interna.../ Martha N. Acosta Montalvo, et al. REFERENCIAS 1. L.-C. Wu, C.-W. Liu, S.-E. Chien, and C.-S Chen, “The Effect of Inrush Current on Transformer Protection,” in 2006 38th North American Power Symposium. IEEE, sep 2006, pp. 449-456. 2. J. Sykes and I. Morrison, “A Proposed Method of Harmonic Restraint Differential Protecting of Transformers by Digital Computer,” IEEE Transactions on Power Apparatus and Systems , vol.PAS-91, no.3, pp. 12661272, may 1972. 3. A. Guzman, S. Zocholl, G. Benmouyal, and H. Altuve, “A current-based solution for transformer differential protection: II. Relay description and evaluation,” IEEE Transactions on Power Delivery , vol. 17, no. 4, pp. 886893, oct 2002. 4. J. Ma, Z. Wang, and J. Wu, “A novel method for discrimination of internal faults and inrush currents by using waveform singularity factor,” in 2010 Conference Proceedings IPEC. IEEE, oct 2010, pp. 1035-1039. 5. R. J. N. Alencar, U. H. Bezerra, and A. M. D. Ferreira, “A method to identify inrush currents in power transformers protection based on the differential current gradient,” Electric Power Systems Research vol. 111, pp. 78-84, 2014. 6. A. G. Phadke and J. S. Thorp, “A New Computer-Based Flux-Restrained Current-Differential Relay for Power Transformer Protection,”. IEEE Power Engineering Review, vol. PER-3, no. 11, pp. 3624–3629, nov 1983. 7. G. Baoming, A. DeAlmeida, Z. Qionglin, and W. Xiangheng, “An Equivalent Instantaneous Inductance-Based Technique for Discrimination Between Inrush Current and Internal Faults in Power Transformers,” IEEE Transactions on Power Delivery, vol. 20, no. 4, pp. 2473 – 2482, oct 2005. 8. H. Esponda Hernández, “Identificación de la corriente inrush en un transformador de potencia utilizando análisis modal,” Maestría, Universidad Autónoma de Nuevo León, 2015. 9. Xiang-ning Liu, Pei Liu, and O. Malik, “Studies for identification of the inrush based on improved correlation algorithm,” IEEE Transactions on Power Delivery, vol.17, no. 4, pp. 901 – 907, oct 2002. 10. D. Q. Bi, X. A. Zhang, H. H. Yang, G. W. Yu, X. H. Wang, and W. J. Wang, “Correlation Analysis of Waveforms in Nonsaturation Zone-Based Method to Identify the Magnetizing Inrush in Transformer,” IEEE Transactions on Power Delivery, vol. 22, no. 3, pp. 1380 – 1385, jul 2007. 11. H. Samet, T. Ghanbari, and M. Ahmadi, “An Autocorrelation Function Based Technique for Discrimination of Internal Fault and Magnetizing Inrush Current in Power Transformers,” Electric Power Components and Systems, vol. 43, no. 4, pp. 399-411, feb 2015. 12. J. H. Wilkinson, The algebraic eigenvalue problem. Clarendon Press,1988. 13. R. L. Burden, J. D. Faires, Brooks/Cole Publishing Company., and Cengage Learning (Firm), Numerical analysis. Brooks/Cole, Cengage Learning, 2011. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 41 �Umbral para discriminar entre corriente inrush y falla interna.../ Martha N. Acosta Montalvo, et al. 14. R. Moon, and R. K. Dhatrak, “A Study of Effect of Magnetizing Inrush Current on Different Ratings Of Transformers,” International Journal of Advanced Research in Electrical, vol. 3, no. 4, pp. 9021-927, 2014. 15. S. Jazebi, F. de León, and N. Wu, “Enhanced Analytical Method for the Calculation of the Maximum Inrush Currents of Single-Phase Power Transformers,” IEEE Transactions on Power Delivery, vol. 30, no. 6, pp. 2590 – 2599, dec 2015. 42 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Performance analysis of line differential protection using MPLS networks Lifan YangA, Thomas RudolphB, Min LiC, Motaz ElshafiC Schneider Electric (China) Co., Ltd., Shanghai Shi, China Schneider Electric GmbH, Dresden, Germany C Cisco, Research Triangle Park, NC, USA A B RESUMEN Este trabajo valida el uso de redes IP/MPLS para cumplir los requerimientos de los esquemas de protección diferencial de líneas, y mostrar el efecto de una red IP/MPLS sobre el comportamiento del relé de protección bajo tres condiciones adversas: retardo asimétrico de canal; alta fluctuación (variación en el retardo de paquetes) y falla sobre el tiempo para cambio de carga. Se observan y registran indicadores clave de comportamiento para cada condición de prueba, incluyendo: retardo extremo-extremo, tiempo de disparo, y medida de corriente diferencial. La alineación de los datos se basa en la medición del viaje redondo de los mensajes de comunicación de los relés. Para el caso de aplicación utilizando un GPS basado en tiempo de muestreo sincronizado, la asimetría del canal no es un reto porque las mediciones tienen referencia de tiempo. Para el fin de la validación, se demuestra que los relé de protección existentes trabajan establemente sobre redes IP/MPLS bajo condiciones adversas. PALABRAS CLAVE Redes MPLS, esquemas de protección diferencial ABSTRACT This work validates the use of IP/MPLS networks to complay the requirements of digital line differential protection schemes, and shows the effect of an IP/MPLS network over the protection relay behavior under three adverse conditions: asymmetric channel latency; high jitter (packet delay variation) and failure path switchover time. For each test condition, key behavior indicators are observed and recorded, including the end-to-end delay, tripping time, and measurement of differential current. Impairment tools are used to inject an additional and artificial delay in one-direction or both directions of the path to introduce jitter. The data alignment is based on the locally measured round trip for communication messages by the relays. For the application case using a GPS-based time synchronized sampling, the channel asymmetry is not a challenging because the measurements contain absolute timestamps. By the end of the validation, it is demonstrated that existing protection relays works stably over IP/MPLS networks under adverse conditions. KEYWORDS MPLS network, schemes of differential protection Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 43 �Performance analysis of line differential protection using MPLS networks / Lifan Yanga INTRODUCTION Utilities, as operators of critical infrastructure, are responsible for the maintenance and control of the electrical power delivery and control equipment in the electrical grid at all times, regardless of circumstances. To achieve this goal, most distribution and transmission system operators have traditionally relied on private TDM- based solutions such as synchronous optical network (SONET) or synchronous digital hierarchy (SDH). These technologies delivered carrier-class performance, supported the deterministic traffic critical for grid operations, and were relatively straightforward for initial deployment. However, because of system upgrades and equipment end of life, time division multiplexing (TDM) infrastructures no longer support the long-term needs of utilities. Many were built and operated for specific applications or solutions, creating soloed infrastructures that make it more challenging and complex to integrate new systems and operational processes. This inflexibility necessitates the deployment of ever more specialized overlay networks, creating a spiral of continual increasing complexity. Such overbuilt networks are highly inefficient, require a great deal of manual administration, are more challenging to troubleshoot and increase operating and maintenance costs. As a result, such environments are actually less secure and increase operational risk over time. With the rapid development of smart grid technologies, the traditional TDM/ SDH communications transmission networks operated by electrical utilities face increasing challenges and cannot accommodate the communication and long-term evolution requirements. MPLS is a proven technology for network operators who need to support diverse legacy systems as well as modernize for next-generation applications. Enabling transparent integration of traditional and smart grid capabilities, MPLS facilitates transport of most forms of traffic. MPLS technology implements packet switching based on open communication standards widely used by telecommunications carriers and enterprise users. The technology features greater flexibility, efficiency and security. In MPLS networks, the bandwidth is dynamically shared for different services (e. g. video, voice, and intranet). Without additional constraints, the data exchange can be flexibly routed, resulting in variable latency. One of the most valued features of MPLS is that it allows utilities to perpetuate the use of existing TDM circuits on the same wide- area network (WAN) backbone with next-generation packet-based systems. This is achieved by running these legacy systems over an MPLS network using techniques such as circuit emulation with pseudo wire emulation edge-to-edge (PWE3). Enhanced by MPLS traffic engineering (TE) or MPLS transport profile (TP), networks can integrate virtually all forms of traffic without having to disruptively replace still-functioning older systems. This helps to unify the network management environment, making it significantly more cost-effective to administer. By running new applications alongside older systems on the same network, utilities can protect their current investment while transitioning the business to the smart grid. Utilities have traditionally accepted SONET/SDH for its ability to deliver highperformance connectivity. By contrast, packet solutions have sometimes been characterized as “best-effort networks,” especially in situations where they are based on T1 or low-bandwidth connectivity. But this not true for well-designed 44 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Performance analysis of line differential protection using MPLS networks / Lifan Yanga packet networks, especially not for high-speed MPLS networks designed with quality of service, traffic engineering, fault detection, and fast reroute (FRR) features. Digital line differential protection affords one of the highest requirements for communication channels in the field of power system protection. The inherent propagation latency, jitter, and asymmetry of an IP network should have no substantive impacts on the behavior of line protection. Moreover, a modern numerical relay keeps working in extreme cases like high jitter or severe asymmetrical latency due to its self-adaptive algorithms. In this paper, the usability of MPLS networks for such applications is evaluated and the results of lab tests are presented. TELECOMMUNICATION REQUIREMENTS FOR LINE DIFFERENTIAL PROTECTION Latency For channel-aided protection schemes, channel delay for transmitting protection messages should meet strict requirements. In North America, maximum of 10 ms latency budget is considered in practice for the communications portion to transport protection relay signals, independently of the distance/path.1 In China, National Standard2 defines that digital information one-way channel delay for transmission line teleprotection should be less than 12 ms. IEC/TR 61850-90-123 recommends one-way channel transmission time to be ≤ 10 ms. Communication channel delay impacts the time the protection takes to detect a fault and its tripping time. Latency of communication network channels consists of three parts: the interface delay between relay and communication equipment (including ingress and egress buffering and processing), communications equipment network delay (network nodes forwarding) and network physical medium latency (propagation delay). The ingress and egress buffering and processing delay depend on the type and speed of communication interface of protection relay. At the ingress of the communications channel, the communication device need to packetize these low-speed (56 kbps or n×64 kbps), synchronous messages of protection relays and transmit them onto high-speed (> 1 Gbit/s) IP communications network. At the egress, the communications device buffers and serializes the high-speed IP packets into the low-speed synchronous serial data referencing a common clock frequency shared with the ingress side. The ingress and egress latencies are generally between 2 ms and 3 ms in total. The packet transmission latency on a communications network is the total forwarding delay caused by the communications devices that the packets pass through. MPLS packet forwarding is implemented in such a way that the latency is very small, i. e. tens of microseconds per hop. The network physical medium latency refers particularly to the signal delay in the transmission medium. Modern power systems mostly utilize fiber optic communication networks as the backbone. The transmission delay is determined by the transmission distance, and it is also affected by fiber types, such as single- Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 45 �Performance analysis of line differential protection using MPLS networks / Lifan Yanga mode or multimode fiber and wave length, etc. Based on the propagation speed of an optical signal in the optical fiber, 1 ms delay can be estimated for every 200 km (125 miles) of distance traversed. Jitter requirements The average value of the communication network latency is not a sufficient criterion for line differential protection. Delay variation, also called jitter or packet delay variation (PDV), expresses how much the delay can vary. This was not an issue when relays were directly interconnected and wired but it becomes important in a packet switched network (PSN) infrastructure. Often, jitter is generated by the packet forwarding node waiting randomly for other high-priority traffic. PSN networks utilize quality of service (QoS) mechanisms, and data forwarding is based on priorities. Typically, messages from protective relays are marked as expedited forwarding (EF) which classifies packets as the highest priority. If an EF priority packet arrives while the communication device is processing an earlier packet, then the processing of that EF packet cannot start until earlier packet processing is completed. Waiting time depends on the size of the packet being processed; the larger the packet, the longer the wait time. Jitter can impact protection behavior, and even cause unpredictable errors in protection ping-pong scheme,4 an application sensitive to jitter. Therefore, in order to ensure consistency in differential protection performance, the requirement is that jitter must be as small as possible. IEC/TR 61850-90-15 defines three message performance classes, and the class TT1 (0.2 ms) providing the highest level of requirement can be used as a reference for current line differential protection. Symmetry requirements Line differential protection is usually based on the principle of a “pingpong” data synchronization algorithm; a prerequisite for this algorithm is the symmetrical latency of forward and reverse paths between two ends. The data alignment for line differential application is based on the locally measured round trip for communication messages. This commonly used method results in high requirements for the communication delay symmetry. When the delay is not equal in both directions, the error introduced in case of high through currents (e. g. external fault currents) can be estimated in the following formula: (1) in which ∆t is the difference of propagation delays in ms and f0 is the nominal system frequency. In case of ∆t = 1 ms and f0 = 50 Hz, a fake differential current appears at a level of 15% of through current. While the pickup current threshold is normally between 15% and 20% based on rating current, an additional asymmetrical latency between forward and reverse paths can lead to unnecessary starting or even mal-operation in case external faults occur. Similarly, in case of internal fault, the error introduced can be estimated in the following formula: 46 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Performance analysis of line differential protection using MPLS networks / Lifan Yanga (2) Notice that here δ% is negative which means differential current measured is smaller than true value. As a result, sensitivity of protection is degraded in this case. In a meshed network, messages between two ends can take different paths in both directions. In packet-switched networks, MPLS traffic engineering and MPLS-TP guarantee the same path for sending and receiving relay messages In the event of a failure along the communication path, both tunnel endpoints switchover to a backup path simultaneously Proper traffic engineering will ensure that the communication paths are symmetric even in the event of primary communication path failure. Reconfigurability requirements Reconfigurability or re-routing is a salient feature of modern communication networks. Fiber failure is one of main reasons for a packet not being received. 5 When a fiber failure occurs, the network must detect the failure and reconfigure to a backup path rapidly, if it is available. For SDH/SONET networks, ITU-T recommends that the switchover time be less than 50 ms. For teleprotection communication systems, there is no defined specifi- cation on switchover time. IEC 60834-1 does specify, however, that the probability of a “command” not being received within 10 ms should be < 10−4. The faster the switchover time, the lower the risk of protection relay failure to trip during a coinciding power system fault. VALIDATION RESULTS To assess the impact of MPLS communications on protection relay performance and to validate the interoperation of various technologies, a dedicated test bench was set up. The tests were performed using Schneider Electric Easergy MiCOM P5456 and Cisco ASR 900.7 In figure 1, the router network is setup with three possible paths between the Easergy MiCOM P545 relays: 1-hop path (green), 5-hop path (blue) and an 8-hop path (orange). Two MPLS-TP tunnels are defined as follows: 1. Tunnel 1 leverages the 1-hop path (green) as the working path, and the 5-hop path (blue) as the protect path. 2. Tunnel 2 leverages the 5-hop path (blue) as the working path, and the 8-hop path (orange) as the protect path. Test scheme and system configuration The service model for line differential protection implemented with C37.94 relay interfaces is illustrated in figure 2. Line differential protection relays connect to the router via an optical/electrical (O/E) interface unit (P-2M-L). The relays send proprietary telegrams to exchange current vectors in terms of use for line current differential protection schemes Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 47 �Performance analysis of line differential protection using MPLS networks / Lifan Yanga Fig. 1. Validated MPLS network topology. Fig. 2. Connection and service models. over C37.94, and O/E converters packetize these telegrams into an E1 2 Mbps circuit. The routers provide Circuit Emulation (CEM) services using TDM-based pseudowires over an MPLS network for transporting the teleprotection data between the two substations. Depending on the relay requirements, the TDMbased pseudowire can be config- ured to perform clear-channel circuits with structure-agnostic TDM over packet (SAToP) or structured circuits using circuit emulation services over packet-switched network (CESoPSN). The Cisco ASR903 router was configured for SAToP transport for P-2M-L E1 communications. Traffic-engineered forward and reverse paths between substation routers fulfill the pathsymmetry requirement for line current differential protection schemes employing channel-based synchronization. The substation routers support eight QoS queues per service, including two Priority Queues, and deep buffer sizes capable of accommodating highly bursty traffic in oversubscribed conditions. The TDM pseudowire is mapped to the highest priority queue (PQ1) ensuring that Teleprotection traffic experiences minimal packet delay variation (PDV) when traversing the network over the static tunnel. 48 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Performance analysis of line differential protection using MPLS networks / Lifan Yanga High-availability with sub-50 ms recovery against failures in the transport network is supported by MPLS-TP linear path-protection with hardware-based bi-directional forwarding detection (BFD) timers when MPLS-TP tunnels are used. Test results and analysis Channel latency Line differential protection devices continually monitor channel delay, and the measured results are displayed in figure 3. Each section of the figure depicts a one-way measurement calculated from the average of 250 test samples: • Relay communication latency back-to-back over a short fiber: 2.86 ms one-way. • Relay communication latency over a 1-hop MPLS network: 4.46 ms one-way. • Relay communication latency over a 5-hop MPLS network: 4.47 ms one-way. • Relay communication latency over a 8-hop MPLS network: 4.59 ms one-way. The network delay excludes the interface delay between relays and communication equipment (2.857 ms). The network delay does include, however, the delay introduced by O/E converters P-2M-L, which is negligible (around 20 µs). The network delay increase caused by traversing more hops is not easily noticeable, and in fact, the increase is almost negligible in the 5-hop case. Even while taking into account the network physical hediuh latency (propagation delay), and assuming signal transmission speed on an optical fiber is about 200 km/ms, (generating additional 2.5 ms delay on a distance of 500 km/310 miles), the overall channel delay easily falls within the 10 ms target. Asymmetric channel latency Impairment tools are used to inject a slowly increasing delay in one direction of the path. In case of internal fault, the differential currents measured become smaller than theoretical value as asymmetry increases. In figure 4a the difference of time delays of sending and receiving (X axis) increase from 0.1 to 5 ms, while the differential current deviation negatively reaches more than 6%. In another word, protection sensitivity becomes worse in this case. In case of throughout flow or external fault, a slight asymmetry will cause a significant error of differential current. As an example figure 4b shows that a 2 ms of asymmetric delay results in the 26% error of flowing current. If the current is significant enough, the artificial differential current will cause a mal-operation. However, modern protection relay can be immune to asymmetric latency when GPS-time stamped sampling and data alignment is enabled at all line ends. The latency of TX and RX can be estimated respectively and current vectors can be aligned correctly. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 49 �Performance analysis of line differential protection using MPLS networks / Lifan Yanga Fig. 3. Channel latencies of different paths. Fig. 4. Differential current deviation: (a) internal fault case, and (b) external fault case. 50 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Performance analysis of line differential protection using MPLS networks / Lifan Yanga High jitter A configurable de-jitter buffer is implemented on routers to compensate for the network delay variation. A larger de-jitter buffer effectively mitigates the risk of network jitter. However, increasing the de-jitter buffer also increases overall channel delay and tripping time as a side-effect. Artificial jitter is introduced by test tool using a Gaussian model, and the jitter is applied along with a baseline delay of 3 ms. This means at an instant in time where no jitter is applied, the channel latency will increase by 3 ms when compared to channel latency results tabulated in Channel latency. The applied jitter in the communication channel is observed by monitoring the channel delay variation. For each test scenario, 50 test samples of latency are recorded and statistical values are presented in the table I. The values show no obvious effect of network jitter on channel latency within certain limits when buffer mechanism works. The average values are portrayed in figure 5. When the de-jitter buffer is set to 2 ms, protection relays start reporting error messages when the jitter increases to 350 µs; When the de-jitter buffer is set to 3 ms, protection relays start reporting error message when the jitter increases to 950 µs; When the de-jitter buffer is 5 ms, no error messages are reported by the protection relays, even when the jitter increases to 2 ms. Failure path switchover The primary path (green path in figure 2) for teleprotection relay (TPR) traffic is protected with MPLS-TP 1:1 protection. When the link on the primary path failed, the convergence time is measured for the emulated bi-directional TPR traffic flow. When the link on the primary path recovered, the convergence time is also measured for the emulated bi-directional TPR traffic flow. MPLS-TP is provisioned to use BFD as its detection mechanism to quickly determine link/ path failure. As shown in figure 6, the primary path is broken at T1 and recovers at T2; and repeated at T3 and T4. The channel delay change is perceived (blue curve), jump from 4.5 to 6.6 ms when primary path failure; restore when primary path recover. Two error messages (orange curve) are reported at each moment of change. During the switchover, current is injected and make it outside of tripping zone but very close to the boundary. Obviously, differential current is calculated correctly because relay can detect the change of channel delay and use the new delay time for data alignment. The Ixia traffic tool indicates the time changeover is 6.74 ms, which is shorter than 25 ms, the threshold of trigger degraded mode of the protection relay. However modern line differential relays are able to master this situation. When the change of propagation time on communication channel is detected and exceeds a configurable value, the relay will upraise the threshold of differential function temporally to ensure the stability during channel switching or other abnormal conditions. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 51 �Performance analysis of line differential protection using MPLS networks / Lifan Yanga Fig. 5. Result of jitter test. Table I. Channel latency during jitter test. Fig. 6. Protection relay performance during path change over. 52 Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 �Performance analysis of line differential protection using MPLS networks / Lifan Yanga CONCLUSION AND SUGGESTIONS FOR IMPROVEMENT Based on intense tests connecting line differential protection with MPLS networks, it could be demonstrated that critical performance requirements of line differential protection applications are met by using the excellent communication infrastructure provided by MPLS networks. Also from protection relay perspective, commercial solutions already available are suitable to be used in this kind of communication networks. As seen from the test results presented in this paper, these main concerns from protection engineer point of view, like channel delay, traffic load, jitters and asymmetry are considered during the tests by using commercially available products and the results demonstrated a full mastering of these concerns. As technology is continuously evolving, lifecycle management is vital. Modular approaches based on existing standards applied on protection equipment provide the benefit of supporting existing traditional teleprotection communication technologies as well as migration paths towards an MPLS based communication. From customer side, one big benefit is that the relays do not have to be changed once it is decided to migrate the communication network to the new technology. This paper will serve as a reference for using MPLS networks for line differential protection applications allowing all kind of migration strategies. IP technology is also pushing the development of protection relay, especially protection traditional communication interface would be gradually IP-based, more efficient, more flexible, more reliable, and further enhance the overall performance of protective relay. REFERENCES 1. S. V. Achanta, R. Bradetich, and K. Fodero, “Speed and security considerations for protection channels,” Proceedings of the 42th Annual Western Protective Relay Conference, Oct. 2015. 2. GB/T 14285-2006, Technical code for relaying protection and security automatic equipment, Chinese Standard. 3. IEC/TR 61850-90-12, Communication networks and systems for power utility automation— Part 90-12: Wide Area Network Engineering Guidelines. 4. Network Protection & Automation Guide, Technical Report, Schneid-Electric, 2016 5. IEC/TR 61850-90-1, Communication networks and systems for power utility automation — Part 90-1: Use of IEC 61850 for the communica- tion between substations. 6. Easergy MiCOM P54x – User Manual, Technical Report, Schneider Electric, 2016. 7. ASR 903 data sheet, Technical Report, Cisco, 2016. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 53 �CT saturation and its influence on protective relays Roberto Cimadevilla, Ainhoa Fernández ZIV Grid Automation RESUMEN En este artículo se revisa el fenómeno de saturación CT y los factores que la afectan. Describe su influencia sobre las diferentes funciones de protección, tales como la sobre corriente, direccional, diferencial, y distancia, y explica las soluciones aplicadas por relé de última generación para incrementar la seguridad y dependabilidad durante la saturación CT . También muestra como dimensionar un CT basándose en los estándares de hoy en día. PALABRAS CLAVE CT saturación, CT dimensionamiento, dc offset, remanente, sobre correinte, direccional, diferencial, distancia. ABSTRACT This paper reviews the phenomena of Courrent Transformer saturation and the factors that affect it. It describes the influence CT saturation has on the different protection functions such as overcurrent, directional, differential and distance and explains the solutions applied by last generation relays to increase the security and dependability during CT saturation. The paper also shows how to dimension a CT based on the different standards used nowadays. KEYWORDS CT saturation, CT dimensioning, dc offset, remanence, overcurrent, directional, differential, distance. INTRODUCTION CT saturation occurs when the CT flux reaches the knee point of the magnetizing curve. It is affected by several factors such as fault current magnitude, CT burden, fault current dc offset, remanent flux, etc. CT saturation generates “gaps” in the current waveform, affecting any current-based protection function, either using instantaneous values or phasors, decreasing the current magnitude and leading its phase. CT saturation will affect the following protection functions: Overcurrent: Phase overcurrent units will tend to increase the tripping times as the phase current is underestimated. On the other hand, the directional units that supervise the overcurrent units will be affected by the angle error. Differential: CT saturation will create a false differential current during external faults affecting the protection stability. In the transformer differential protection the harmonic blocking may operate for internal faults with CT saturation, delaying the trip of the percentage differential protection. 54 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández The current magnitude reduction and the phase leading, both combined, may cause underreach or overreach of the distance units, depending on the shape of the distance characteristics. This paper explains the phenomena of CT saturation and the different factors that affect it. It also describes the solutions applied by last generation relays to cope with CT saturation, like saturation detectors, external fault detectors, directional comparison units, measurement algorithms based on peak values or on shorter windows, etc. It also explains how these solutions reduce the CT sizing requirements. CT FUNDAMENTALS CT equivalent circuit Figure 1 shows the equivalent circuit of a CT. Current i’1 represents the primary current referred to the secondary winding: (1) where N1 and N2 are the number of primary and secondary turns, respectively. As current i’1 is defined by the primary power system, i’1 is represented by a current source, therefore the primary leakage impedance can be removed as it does not have any effect in the CT circuit. On the other hand, the reactive components in the circuit may be neglected, considering a pure resistive burden (Xs ≈ 0 and Zburden = Rb ). The burden impedance is equal to the sum of the wire impedance, mostly resistive, and the relay impedance, negligible for digital relays. The circuit of figure 1 is, therefore, simplified to the circuit of figure 2. The power supplied by the CT is: (2) The magnetizing voltage will be: (3) Fig. 1. Equivalent circuit of a CT. CT saturation By applying the Lenz law, the magnetizing voltage em is: (4) Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 55 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández Fig. 2. Simplified equivalent circuit of a CT therefore (5) The Xm reactance is non-linear and the CT flux and the magnetizing current, im, follows the typical characteristic represented in figure 3. Fig. 3. CT magnetizing curve. During normal conditions the flux value will be lower than the saturation flux, therefore the magnetizing current, im , will be very low (Xm will be very high) and so i2 will practically equal i’1. However, if the flux increases because of a high secondary current i2 and/or a high Rct + Rb impedance, and it reaches the saturation flux, the magnetizing current will increase very much (X m will be very low), making i2 to have very reduced values. As the flux is sinusoidal, it will have values above and below the saturation flux, making the CT entering and leaving saturation. During the saturation periods, current i2 will be practically zero and current im will equal i’1. Figure 4 shows the secondary current, the magnetizing current and the flux in a CT. The red curves correspond to a saturated CT and the blue curves to a non-saturated CT (ideal CT with linear magnetizing characteristic). Note that during the CT saturation periods the flux is practically constant, the secondary current practically zero and the magnetizing current is practically equal to the secondary current of the ideal CT. Figure 5 shows the magnetizing curve of the real CT. 56 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández Fig. 4. Secondary current, magnetizing current and flux in a saturated and non-saturated CT (no dc offset included). Influence of dc offset There is another factor, apart from the current magnitude and the burden value, that makes the CT flux increase: the dc offset in the current. In the circuit of figure 6, the following equation is fulfilled: (6) When the switch S is closed, simulating the ocurrence of a fault, a transient state occurs during which the current follows (7) Fig. 5. Flux vs. magnetizing current curve. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 57 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández where Fig. 6. RL circuit. (8) (9) y (10) Let’s analyze the flux waveform when i’1 (t)=i(t). In order to simplify the latter equations we will consider θ=γ−ϕ. Analysis not considering Xm. If we consider Xm ≈ ∞, i2 = i’1, therefore Assuming φ(0) = 0 (11) Figure 7 shows, at the top, a current wave with and without dc offset and the flux that corresponds to each current wave at the bottom. As it can be observed when the current includes dc offset the flux grows very much, which can make the CT saturate. As the flux represents the area under the integrated waveform, when the latter one includes a dc offset, the positive and negative areas will be different (in the case shown in figure 7 the positive area is higher than the negative one), therefore they will not cancel each other, making the flux increase continuously. 58 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández Figure 8 shows the secondary current, the magnetizing current and the flux in a CT. The red curves correspond to a saturated CT and the blue curves to a non-saturated CT (ideal CT). In the saturated CT, when the flux reaches the saturation flux (see the magnetizing curve in figure 9), its growth is stopped, i2 current tends to zero and im current tends to equal i’1 current. The flux increase is stopped when it reaches the saturation flux by reducing the positive area below current i2, making i2 practically zero and, therefore, “cutting” the i2 waveform. When the flux is below the saturation flux, the CT leaves the saturation state, i2 gets normal values and im is practically zero. Saturation and non-saturation periods alternate until the CT definitively gets out of saturation. Fig. 7. Current and flux (a) with dc offset and (b) without dc offset. Analysis considering Xm . If Xm is considered, the flux can be calculated as (12) Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 59 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández Lm is calculated as (13) Calling (14) Fig. 8. Secondary current, magnetizing current and flux in a saturated and non-saturated CT (dc offset included). Fig. 9. Flux vs. Magnetizing current curve. 60 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández we can get.1 (15) Influence of flux remanence When the primary current in a CT is interrupted, a remanent flux will remain in the magnetic circuit. Its magnitude will depend on the flux in the CT core at the time of current interruption. If this flux was very high due to a high dc offset and/or a high symmetrical current and/or a high burden, the remanent flux will be high. Once the remanent flux has been established, very little is dissipated under normal conditions, therefore it will remain in the core until it is demagnetized. In non-gapped CTs the remanent flux may be up to 80% of the saturation flux. The remanent flux will affect the transient behavior of the CT. It will add or subtract the flux created by the current magnitude, burden value and dc offset. Depending on the sign of the remanent flux compared with the sign of the flux created by the other mechanisms, it will worsen or improve the CT transient response. The remanent flux should be especially considered for reclose onto fault conditions. Figure 10 shows the effect of a permanent fault on the CT wired to a line relay (the first plot represents the current and the second one the flux). It can be observed that the CT saturation is heavier when reclosing onto fault than during the initial fault. With regard to the remanence, there are three types of CTs: High remanence CT: no limit for the remanence is specified. The CT has no air-gap. The remanent flux can reach 80% of the saturation flux. Examples of such CTs are: class P, PX, TPX (IEC60044 and IEC61869-2), TPS (IEC600446), P or X (BS 3938) and non-gapped class C (IEEE C57.13). Low remanence CT: a remanence limit is specified. In classes PRLow remanence CT: a remanence limit is specified. In classes PR IEC61869-2) and TPY (IEC60044-6 and IEC61869-2) the specified remanence factor (remanent flux/saturation flux) must be lower than 10%. This type of CTs includes small air-gaps. No remanence CT: the remanent flux is practically negligible. Class TPZ (IEC60044-6 and IEC61869-2) is a non-remanence CT. This type of CT includes bigger air-gaps. CT dimensioning Class P of IEC 61869-2 standard The CT is specified with the following data: • Rated transformation ratio: the ratio of the rated primary current to the rated Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 61 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández secondary current, e. g. 600/5. • Rated power: power provided by the CT at rated current and rated burden, e.g. 10 VA. • Accuracy class: 5P and 10P defines a maximum composite error of 5% or 10% at the accuracy limit current (accuracy limit factor (ALF) multiplied by the rated current). • Accuracy limit factor: times the rated current, without dc offset, at which the accuracy class is fulfilled. • Secondary internal resistance: Rct in figure 2. The CT will be adequate if Ktotal = Kssc Kb Ktf Krem < ALF, where Kssc is the symmetrical short-circuit current factor, Kb is the burden factor, Ktf is the overdimensioning factor for dc offset, and Krem is the remanence overdimensioning factor. Symmetrical short-circuit current factor Kssc. It is the ratio between the maximum short circuit current and the rated current. Burden factor Kb. It is the ratio where Rn is the rated burden. Rn can be calculated from the CT rated power: (16) The accuracy limit factor is defined for the rated burden. For a different burden the maximum symmetrical current that assures the fulfillment of the accuracy class will be different than the accuracy limit current (it will be higher than the accuracy limit current if the burden is lower than the rated one and it will be lower if the burden is higher than the rated one). This condition is taken into account by the burden factor. Fig. 10. Current and flux during a trip-reclose-trip cycle. 62 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández Transient overdimensioning factor Ktf . It was seen in Influence of dc offset that the flux created by a current with dc offset (asymmetrical current) was much higher than the flux generated by a current without any dc component (symmetrical current). As the ALF factor is defined for a symmetrical current, an overdimensioning factor for asymmetrical currents must be considered. This factor will be given by which represents the ratio between the maximum total flux (sum of dc and ac fluxes) and the maximum ac flux. Considering (11), the maximum flux when there is no dc offset in the current is (17) The maximum flux when there is dc offset in the current is obtained for (maximum dc offset) and when the exponential has reached its limit (t → ∞): (18) therefore (19) Considering (15),1 (20) In some systems the X/R ratio may be very high (especially in points close to generators: X/R values of 20–30) giving rise to excessive overdimensioning factors that will require too large CTs, not viable, due to cost and space constraints. In this case it is not possible to assure that the CT is not saturated during the whole fault duration but it is assumed that the CT saturates after a timedelay from the fault inception is the value of the flux after the maximum saturation free time. Ktf can be calculated with (11) or (15). Considering (11), (21) Considering (15),1 (22) For saturation free times higher than 15 ms, the maximum flux will be obtained with θ = 0, however, for saturation free times lower than 15 ms, the maximum Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 63 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández flux will be obtained for other fault inception angles. Figure 11 shows the Ktf curves for three different θ values (0° – Ktf0 (maximum dc offset), 90° – Ktf1 (no dc offset) and 45° – Ktf2 ), considering T2 = 3 s and T1 = 0.125 s. It can be checked that approximately from t = 14 ms on, Ktf0 is always higher than Ktf1 and Ktf2. Figure 12 shows a zoom of figure 11 for times lower than 4 ms. In this case the maximum Ktf is Ktf1 with no dc offset. For each saturation free time tolerated by the protective relay the worst inception angle should be determined. Remanence overdimensioning factor Krem. The remanent flux may worsen the CT transient response if it has the same sign of the flux generated by the current magnitude, burden value and dc offset. This is considered by the remanence overdimensioning factor where Kr is the remanent factor (maximum remanent flux/saturation flux). (23) Fig. 11. Ktf for different θ values (Ktf0, Ktf1 and respectively). Ktf2 correspond to 0°, 90° and 45°, Fig. 12. Zoom of Fig. 11 (t = 0–4 ms). Class C of IEEE C57.13 standard The most common accuracy class in the IEEEC57.13 standard is the C class. The letter C is followed by a number that indicates the secondary voltage rating, which is defined as the CT secondary voltage that the CT will deliver when it is connected to a standard secondary burden at 20 times the rated secondary current, without exceeding a 10% ratio error. The common standard burdens for protection CTs are 1, 2, 4 and 8 Ω, which correspond, at 5 A rated current, to 100, 200, 400 and 800 V secondary rating voltages (for a C100 CT the voltage 64 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández at the 1 Ω burden will be 20×5×1 = 100 V). With the secondary voltage rating we can obtain the internal magnetizing voltage by adding the voltage drop in the secondary resistance (Rct ): (24) The dimensioning of an IEEE CT can be done by calculating Em as: (25) where K’total = KsscKtfKrem. If Em calc < Em rated =Vb+Rctx20I2n the CT will be valid. An easier deduction can be made considering that the ALF factor of a C class CT is always 20 (the 10% ratio error cannot be exceeded for a secondary current 20 times the rated current with the rated burden. If Ktotal < ALF the CT will be valid. Class X of BS3938 standard or Class PX of IEC61869-2 Class X CT is defined with: • Primary and secondary rated currents • Transformation ratio • Rated knee-point voltage • Magnetizing current at rated knee-point voltage • Resistance of secondary winding The rated knee-point voltage is defined as the minimum voltage, at rated frequency, applied to the CT secondary terminals which increased by a 10% causes an increase in the magnetizing current of 50% (see figure 13). The relationship between the rated knee-point voltage (Vknee) and the magnetizing voltage at the accuracy limit current with rated burden (Em rated ) is done by approximation, because the definition of the two voltages has no direct relation (Vknee has to do with the slope of the magnetizing characteristic and Em rated with the current composite error). It is normally considered that Em rated = 1.25 to 1.3 times Vknee. Fig. 13. Knee point voltage definition. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 65 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández Once Em rated is calculated it can be compared with Em calc = K' total I2n (Rct + Rb). The CT will be valid if Em calc < Em rated. INFLUENCE OF CT SATURATION ON PROTECTIVE RELAYS Overcurrent CT saturation introduces errors in the phasor estimation, decreasing the phasor magnitude and leading its phase.2 This makes the overcurrent units underestimate the current magnitude giving rise to delayed trips or no trips. Figure 14 shows the magnitude of the fundamental component of the saturated and non-saturated currents (|I A | and |I A sat |, respectively) shown in figure 4. A full-cycle DFT was used for the calculation. It can be seen that the value of the saturated current is much lower than the value of the non-saturated one. Figure 14 also shows the true rms value (|IA sat rms |); although it is higher than the magnitude of the fundamental frequency of the saturated current, it is still much lower than the magnitude of the fundamental component of the non-saturated current. For instantaneous units, if the time to saturation is lower than the time it takes for the overcurrent unit to pick-up the trip will be delayed. If the CT saturation is caused by an asymmetrical current, the delay will be approximately equal to the primary time constant, which defines the damping time of the dc offset. If Fig. 14. Magnitude of the fundamental components and true rms value of the saturated and non-saturated currents of figure 4. the CT saturation was due to a symmetrical current there will not be any trip. In order to increase the dependability of the instantaneous overcurrent units an operation based on the peak value can be used when CT saturation is detected. The implemented algorithm uses the saturation detector described in.3 This detector is based on the derivative of the measured current. At the time that CT saturation initiates there is a significant increase in the derivative. Given that the maximum value of the derivative of the current is where A is the maximum current and N the number of samples per cycle, when 66 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández where k is a constant, the saturation is detected. A is calculated as the greater of two consecutive maxima. The saturation detector will only operate when A exceeds a threshold. The detector includes a reset time of a cycle. The algorithm based on peak values calculates the dc offset with the difference between two consecutive peak values divided by two. The dc offset is removed to reduce the transient overreach. The overcurrent units based on this algorithm are only enabled when the saturation detector activates and two consecutive peak values have been measured since the activation of a fault detector; figure 15 shows the peak value measured once the dc offset is removed (Ipeak no dc). Although it is lower than the magnitude of the fundamental component of the non-saturated waveform calculated with a DFT it is much higher than the magnitude of the fundamental component or the true rms value calculated for the saturated waveform. The algorithm based on peak values requires a saturation free time lower than the algorithm based on full-cycle DFT. For a symmetrical CT saturation it just requires around a quarter of cycle saturation free time for faults in the limit of the reach. For an asymmetrical CT saturation it may require around halfcycle saturation free time. For symmetrical and asymmetrical CT saturation the full-cycle DFT will require one-cycle saturation free-time or higher times if it includes a dc offset filter. Fig. 15. Magnitude of the fundamental components, true rms value and peak value of the saturated and non-saturaded current of figure 4. For close-in faults the saturation free time for both algorithms, the one based on DFT and the one based on peak values will be much lower than for faults in the limit of the reach. The magnitude obtained with the DFT for the saturated current of figure 14 is around 50% of the magnitude of the non-saturated current. The peak value obtained with the described algorithm is around 90% of the magnitude of the non-saturated current. These values are normally high enough to be above the overcurrent unit pick-up value in the case of a close-in fault. CT remanence should be considered if reclosing is used. The effect of CT saturation on time-delayed overcurrent units is of less importance as the tripping times are normally much higher than the time the CT is in the saturation condition. Anyway, for high primary time constants additional coordination time may be included in Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 67 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández the settings to assure the selectivity. The algorithm based on peak values could also be applied to time-delayed overcurrent units. Directional unit Phase directional units normally have a high margin to distinguish the forward faults from the reverse ones. This margin is in the order of 90°. CT saturation makes the phasor angle increase; however it does not normally create an error higher than 90°. To increase the reliability a saturation detector as the one described in Overcurrent could be used in order to modify the directional characteristic by slightly reducing the characteristic angle (α in figure 16). Fig. 16. Directional unit. Special care should be taken in double breaker bays. In the circuit of Figure 17, if |Isum | is much smaller than |I1 | and |I1 |, the errors of the CTs can produce an inversion of the summed current. In this case any directional unit based on this summed current will operate erroneously. In the former situation directional units that compare the currents |I1| and |I1| can be implemented. These units may be based on phase currents or sequence currents. Reference4 describes the mentioned units. Fig. 17. Reverse fault in a double breaker bay. Figure 18a and figure 18b show the currents measured by a distance relay in the two CTs of a breaker and a half bay for a reverse AG fault. Figure 18c shows the sum current. As it can be checked, the current measured by CT-2 is slightly distorted, which generates a high distortion in the sum current. The enabled 68 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández directional unit was comparing the angle between the phase A positive-sequence voltage and phase A current. The evolution of this angle is shown in Figure 19. As it can be checked, during a small time, the angle is lower than (α + 90°, α is the characteristic angle, set to 84°), which makes the relay trip erroneously. The solution to this false trip was the implementation of a directional unit based on the angle between the currents measured in both CTs of the bay as explained in.4 The unit was easily implemented by means of the programmable logic. Differential units External faults For external faults, CT saturation increases the differential current and decreases the restraint current, moving the point (IDIF, IREST ) into the operation zone. The through fault current restraint of the percentage differential characteristic is normally not enough to cope with severe CT saturation, therefore units that complement the differential characteristic are normally required.3 Some common units are described below. Differential unit with instantaneous values. This unit is based on the ratio between the differential and restraint currents. It operates when this ratio is below a threshold. When the fault is external, during the time the CT is not saturated, the ratio will be very small. Figure 20 shows, for an external fault, and for the phase A, the currents at both ends of the protected element, 1 and 2, the differential and restraint currents. This external fault makes CT in winding 1 saturate. As it can be observed, since the activation of a fault detector (signal FDET, based Fig. 18. Currents measured by (a) CT-1 and (b) CT-2, and (c) sum current for a reverse fault in a breaker and a half bay. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 69 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández Fig. 19. Angle between Va1 and Ia. on a current change), there will be a consecutive number of samples for which the ratio IDIF /IREST is very low. If this happens, the external fault condition will be activated. More information can be found on.5–7 This unit requires a low saturation free time. Typical times are around 3–4 ms. Directional comparison units. Ref3 describes a directional comparison unit that uses the angle between the currents measured at each end of the protected element in order to determine if the fault is internal or external. When this angle is lower than 90° the fault is considered internal; on the contrary, if the angle is higher than 90° the fault is considered external (see figure 21). The angular comparison requires that the currents are above a minimum threshold. Two directional comparison units are described, one that operates with phase currents and another one that operates with positive-sequence pure fault current. Fig. 20. Currents measured (a) at both ends of the protected element, (b) differential and restrained currents, and (c) ratio between differential and restrained currents for an external fault with CT-1 saturated. 70 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández Fig. 21. Directional comparison criteria: (a) internal fault and (b) external fault. Phase directional comparison unit: the currents must be above the maximum load current. This condition assures that internal weak infeed faults do not generate an external fault indication. To check that the currents are above the load current an algorithm based on the current derivative is used. The phase directional comparison units may be implemented with instantaneous values instead of phasors, by comparing the sign of the currents at all the ends of the protected element.7 If any of the currents has opposite sign as compared with the rest of the currents during several samples an external condition is activated. Positive-sequence directional comparison unit: it uses pure fault current by removing the prefault current. This makes the unit compensate the load flow effect. The prefault current is taken two cycles before the activation of a fault detector, based on current changes. More information about the mentioned directional units can be found in.3 Both the directional units based on phasors and the directional unit based on instantaneous values tolerate very low saturation free times (typical values of 2–3 ms). This is because of the high margin the directional comparison units have to distinguish external faults from internal ones. Internal faults In transformer differential protections CT saturation for internal faults can make harmonic restraint / blocking operate, because of the harmonic content of the differential current during such faults. This will result in delayed trips. An unrestrained differential unit set above the maximum inrush current is normally used to increase the dependability. However, internal faults with CT saturation could happen for current values lower than the ones for the inrush currents. Reference8 describes a dynamic harmonic restraint/blocking that allows fast tripping during internal faults with CT saturation. The algorithm is based on an external fault detector (for inhibiting the second harmonic restraint / blocking) and underexcitation units (for inhibiting the fifth harmonic restraint/blocking). The second harmonic restraint/blocking is inhibited after a settable time from the Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 71 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández power transformer energization. It is enabled again during a settable time once the external fault detector has activated. The fifth harmonic blocking is inhibited if the ratio V / f is below a threshold. Inrush currents Due to the high dc offset and long time constants CT saturation during inrush is very common.8 Figure 22 shows a CT saturation registered by a busbar differential relay that tripped during an inrush condition. The first plot represents the currents measured by the three bays that fed the inrush current. The second plot represents the differential and restraint currents calculated by the busbar differential protection. In this case, as the phase bay currents were very low, the phase directional comparison unit was not able to activate the external fault condition. An algorithm to slightly increase the minimum pick-up value of the differential characteristic during transformer energization was included to increase the security for inrush conditions. Distance protection As CT saturation decreases the current magnitude and leads the current phase, it will increase the impedance magnitude and reduce its angle. This effect will tend to produce underreach of the distance units. The use of sub-cycle units will reduce the required saturation free time, as these units need lower times to take the trip decision9 (saturation free times of half-cycle to cycle depending on the fault location). Anyway, for faults in the limit of zone 1, the fast units will not operate, therefore the saturation free times will be increased to 1 cycle 1 and cycle. That is why the worst case scenario for CT dimensioning will be for a fault at the limit of zone 1. In case of reclose onto fault, the remanence factor should be considered. Taking into account that the close onto fault detector will extend zone 1, the worst case scenario will be for a fault at 100% of the line. As zone 1 will be extended to zone 2 (covering normally 120%) the saturation free times will be slightly lower than 1 cycle–1 and cycle. CT saturation does not only produce underreach of the distance protection but it can also produce overreach, mainly in quadrilateral characteristics. As the impedance angle is reduced, although its magnitude is increased, the impedance can enter the quadrilateral characteristic through the reactance line, as it can be seen in figure 23. Figure 24 shows the voltages and currents for an ABG fault at 100% of a line. The CT in phases A and B gets saturated. Figure 25 shows the measured AB impedance trajectory during the time the CT is entering and leaving saturation. The impedance was measured by a sub-cycle filter, based on half-cycle DFT. It also shows the sub-cycle distance characteristic related to zone 1 (reaching 80% of zone 1). The X point represents the AB impedance when the CT is not saturated. The error in the measured impedance when the CT is saturated is clearly seen. In order to avoid any overreach the saturation detector described in3 is implemented. The activation of the saturation detector tilts down the reactance line and reduces the resistive reach.9 72 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández Fig. 22. Phase currents and differential and restraint currents measured by a busbar differential protection during a transformer energization. Fig. 23. Measured impedance with and without CT saturation. Fig. 24. Voltages and currents for an ABG fault at 100% of the line with CT saturation. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 73 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández Fig. 25. AB impedance measured with sub-cycle filter for an ABG fault at 100% of the line with CT saturation. CONCLUSION CT saturation is affected by the fault current magnitude, the burden value, the dc offset in the fault current and the remanent flux in the CT core. These variables are considered in the following factors: symmetrical short-circuit current factor, burden factor, dc offset overdimensioning factor and remanence overdimensioning factor. Many times CT saturation cannot be avoided as it will require too large CTs, not viable because of costs and space constraints. In this case protection manufacturers should specify a saturation free time. New generation relays offer solutions to improve the reliability during CT saturation, relaxing the CT requirements. Some of this solutions are: –Overcurrent units based on peak values enabled when a CT saturation detector activates. –Directional units based on dual CT currents to improve the security during reverse faults in breaker and a half bays. –Complementary units to the differential units such as units that compare the ratio of differential/restraint currents and directional comparison units. –Dynamic harmonic restraint/blocking: allows disabling the harmonic restraint/blocking during internal faults with CT saturation. –Sub-cycle distance units complemented with a saturatiom detector that tilts the reactance line and reduces the resistive reach. REFERENCES 1. S. Holst and B. S. Palki, “Coordination of fast numerical relays and current transformers — Overdimensioning factors and influencing parameters,” ABB. 2. S. E. Zocholl and D. W. Smaha, “Current transformer concepts,” Proceddings 74 Ingenierías, Abril-junio 2017, Vol. XX, No. 75 �CT saturation and its influence on protective relays / Roberto Cimadevilla & Ainhoa Fernández 3. 4. 5. 6. 7. 8. 9. of the 19th Annual Western Protective Relay Conference, Spokane, Washington, Oct. 1992. R. Cimadevilla, “New protection units included in differential relays,” Proceddings of te 2011 PAC Conference, Dublin, Jun. 2011. R. Cimadevilla, “Protection for Breaker and a Half or Ring Bays,” Proceedings of the 2009 SEAPAC, Melbourne, Mar. 2009. R. Cimadevilla and S. López, “New Requirements for High Voltage Transformer Protection,” Proceedings of the 2009 SEAPAC, Mar. 2009. Instruction Manual, ZIV ‘‘Transformer Protection Model IDV’’, Z mudio, Spain, Reference BIDV1112Av07. Instruction Manual, ZIV ‘‘Busbar Protection Model DBN’’, Zamudio, Spain, Reference BDBN0901Av07. R. Cimadevilla, “Inrush currents and their effect on protective relays,” Proceedings of the 2013 TAMU Conference for Protective Relay Engineers, College Station, Texas, Apr. 2013. R. Cimadevilla and I. García, “Sub-cycle distance units: Design, testing and real operation,” Proceedings of the 2015 PAC Conference, Glasgow, June 2015. Ingenierías, Abril-Junio 2017, Vol. XX, No. 75 75 �Eventos y reconocimientos ING. GUADALUPE E. CEDILLO GARZA: FORJADOR DE INGENIEROS La Facultad de Ingeniería Mecánica y Eléctrica (FIME) celebró el pasado 20 de enero, los 60 años de vida académica, formando ingenieros en la FIME, del Ing. Guadalupe E. Cedillo Garza. La celebración estuvo encabezada por el Mtro. Rogelio G.Garza Rivera, Rector de la UANL; Dr. Jaime Arturo Castillo Elizondo, Director de la FIME y el Dr. Celso José Garza Acuña, Secretario de Extensión y Cultura; también estuvieron presentes ex rectores, ex directores, miembros de la Junta de Gobierno; profesores y alumnos. El Ing. Cedillo Garza fue Director de la FIME, Presidente de la Junta de Gobierno, es Profesor Emérito y maestro Decano. Se reconoció su actividad y constancia académica, El Ing. Guadalupe E. Cedillo con el libro “Forjador de Ingenieros” preparado por la FIME en su honor. Lo acompañan: el Rector de la UANL, Mtro. Rogelio G. Garza Rivera, y el Director de la FIME, Dr. Jaime Arturo Castillo Elizondo. 76 ya que continúa actualizándose y ofreciendo cursos de matemáticas en el área de ciencias básicas. RINDE PROTESTA COMO DIRECTOR DE LA FIME EL DR. JAIME ARTURO CASTILLO ELIZONDO El pasado 6 de abril de 2017, asumió el cargo de Director de la FIME para el período 2017-2020, el Dr. Jaime Arturo Castillo Elizondo, quien rindió protesta ante la Junta de Gobierno, presidida por el Dr. José Santos García Alvarado. Después de que le fue impuesta la Medalla “Venera”, el director expresó su compromiso para continuar el trabajo orientado a convertir a la FIME en la mejor facultad de ingeniería. El Rector de la UANL, Mtro. Rogelio G. Garza Rivera lo felicitó por haber asumido este cargo, a la vez que exhortó a los presentes para continuar en el esfuerzo en este segundo período. El Director de la FIME, Dr. Jaime Arturo Castillo Elizondo, dirigiendo un mensaje durante la ceremonia en la que rindió protesta. Ingenierías, Abril-Junio 2017, Año. XX, No. 75 �Tesistas titulados de Maestría en la FIME-UANL * Enero - Marzo 2017 Rafael Vela López, Maestría en Ciencias de la Ingeniería con orientación en Energías Térmica y Renovable. Modelado computacional de fluidos de la transición entre la zona de calentamiento y la zona de enfriamiento de un horno túnel, 13 de enero. Rosa Dolores Ortega Reyes, Maestría en Ciencias de la Ingeniería Mecánica con especialidad en Materiales. Comportamiento microestructural y mecánico en la aleación resistente al calor HR32 sometida a tratamiento de envejecimiento, 19 de enero. Julio Grimaldo Lee, Maestría en Ciencias de la Ingeniería Mecánica con especialidad en Materiales. Estudio del comportamiento termomecánico de una aleación Ti-Al- V-Fe para aplicaciones en la industria aeronáutica. 23 de enero Adriana Rodríguez Liñán, Maestría en Ciencias de la Ingeniería Mecánica con especialidad en Materiales. Estudio experimental de la soldadura por haz de electrones de acero para aplicaciones automotrices, 24 de enero. Ana Karen Vega Rivera, Maestría en Ciencias de la Ingeniería Automotriz. Obtención de cargas en el Trailer Hitch mediante simulación de dinámica vehicular en Tuvksim, 24 de enero. Alfonso Banda Urbina, Maestría en Ciencias de la Ingeniería Eléctrica. Análisis e implementación de estrategias de control tolerante, 25 de enero. Jorge Alberto Aguilar Navarro, Maestría en Ciencias de la Ingeniería Automotriz. Medición de esfuerzos residuales mediante la técnica de replicación de identaciones de mondoloques V8 de aluminio, 30 de enero. Claudia Susana Cázares González, Maestría en Ciencias de la Ingeniería Mecánica con especialidad en Materiales. Estudio del efecto de la velocidad de temple sobre la precipitación en aleaciones aluminio-silicio-magnesio. 22 de febrero. * Información proporcionada por la Coordinación de Titulación de Posgrado. Ingenierías, Abril-Junio 2017, Año XX, No. 75 77 �Información para colaboradores Se invita a profesionistas, profesores e investigadores a colaborar en la revista Ingenierías con: artículos de divulgación científica y tecnológica, artículos sobre los aspectos humanísticos del quehacer ingenieril y reportes de investigación. El envío de artículos a la revista Ingenierías para su publicación implica el ceder los derechos de autor a la UANL. Es requisito que las colaboraciones sean producto del trabajo directo de los autores estableciendo claramente su contribución; y que estén escritas en un lenguaje claro, didáctico y accesible. Las contribuciones no deberán estar redactadas en primera persona. Todos los artículos recibidos estarán sujetos a arbitraje de tipo doble anónimo siendo el veredicto inapelable. Los criterios aplicables a la selección de textos serán: originalidad, rigor científico, exactitud de la información, el interés general del tema expuesto y la claridad del lenguaje. Los artículos aprobados serán sujetos a revisión de estilo. CRITERIOS EDITORIALES Los autores de artículos de revisión o divulgación deberán contar con una producción directa reconocida en la temática de interés de la revista. Estos trabajos deben ofrecer una panorámica del campo temático, separar las dimensiones del tema, mantener la línea de tiempo y presentar una conclusión que derive del material presentado. No se aceptan reportes que muestren solamente mediciones. Los artículos deben presentar los resultados de las mediciones acompañados de su análisis detallado, un desarrollo metodológico original, una manipulación nueva de la materia o ser de gran impacto y novedad social. Sólo se aceptan modelos matemáticos que sean validados científicamente dentro del propio trabajo. No se aceptarán trabajos basados en encuestas de opinión o entrevistas, a menos que aunadas a ellas se realicen mediciones y se efectúe un análisis de correlación para su validación. No se aceptan protocolos de 78 investigación, proyectos, propuestas o trabajos de carácter especulativo. Los artículos a publicarse en partes, deben enviarse al mismo tiempo, pues se arbitrarán juntos. LINEAMIENTOS EDITORIALES Es requisito enviar para su consideración editorial: artículo, material gráfico y las fichas biográficas de cada autor en CD a la dirección de contacto, o por e-mail a revistaingenierias@uanl.mx . El artículo debe estar escrito en tipografía Times New Roman de 11 puntos a espacio sencillo en Word o algún formato compatible con este software. La extensión máxima es de 15 páginas, el título del artículo no debe exceder de 80 caracteres, el número máximo de autores por artículo es cinco. Los artículos deben incluir un resumen tanto en español como en inglés, de no más de 100 palabras, así como un máximo de 5 palabras clave tanto en español como en inglés. Las referencias deben incluir en orden, los siguientes datos: Autores o editores, título del artículo, nombre del libro o de la revista, lugar, empresa editorial, año de publicación, volumen y número de páginas. Las referencias deberán ir numeradas en el orden citado en el texto. El material gráfico (imágenes y gráficas) utilizadas en el artículo deberán enviarse en archivos individuales en formato .tif, .jpg o .bmp en una resolución mínima de 1800 X 1800 puntos (eso una resolución de 300 dpi y 15 cm en el lado más pequeño de la imagen). Las fichas biográficas deben ser de máximo 100 palabras. CONTACTO Facultad de Ingeniería Mecánica y Eléctrica de la Universidad Autónoma de Nuevo León, Edificio 7, 1er. piso, ala norte. Tel.: 8329-4000 Ext. 5854 Fax: 8332-0904 E-mail: revistaingenierias@uanl.mx Ingenierías, Abril-Junio 2017, Año. XX, No. 75 �Código de ética Autores Los autores deben presentar una narración concisa y exacta del trabajo desarrollado, así como una discusión objetiva de su significado intelectual y científico. Los autores deben abstenerse de ofrecer los mismos manuscritos que se encuentren en consideración por otras publicaciones. Los autores deben incluir en su manuscrito detalles suficientes y referencias a fuentes de información públicas para hacer posible la reproducción del trabajo por terceros. Los autores deben abstenerse de presentar críticas personales en sus trabajos. Los autores deben citar aquellas publicaciones que son antecedentes esenciales para comprender el trabajo. Los autores deben abstenerse de incluir información que hayan obtenido mediante comunicación privada que no se localice en publicaciones. Los autores deben abstenerse de incluir información que hayan obtenido de manera confidencial sin el permiso explícito correspondiente. Los autores deben abstenerse de incluir información obtenida en el proceso de servicios confidenciales, tales como documentación para concursos o solicitudes de becas. Los autores deben abstenerse de citar publicaciones que no se relacionen o que sólo se relacionen remotamente con la materia. Los autores deben reconocer mediante una nota de agradecimiento el apoyo de las instituciones y organismos que hayan contribuido significativamente al desarrollo del trabajo, así como a colaboradores que hayan contribuido de manera importante, pero sin que hayan llegado a cumplir con el criterio de coautoría, si los hubiera. Los autores deben reconocer mediante una nota de agradecimiento el apoyo a colaboradores fallecidos que hayan contribuido de manera importante, pero sin que lleguen a cumplir con el criterio de coautoría, si los hubiera, señalando la fecha de su muerte. Los autores deben abstenerse de utilizar nombres ficticios o seudónimos. Los autores deben responsabilizarse del material que presentan en su manuscrito. Revisores Los revisores deben declinar cualquier invitación para evaluar un manuscrito si no se consideran calificados, carecen de tiempo para juzgar o se les presenta algún conflicto de intereses, tal como encontrarse vinculados estrechamente a los autores o al trabajo a evaluar. Los revisores deben manifestar al editor cualquier conflicto de intereses que detecten. Los revisores deben considerar un manuscrito enviado para revisión como un documento confidencial. Los revisores deben abstenerse de expresar críticas personales. Los autores deben abstenerse de incluir como autores a terceros que no cumplan con el criterio de coautoría, el cual consiste en la contribución significativa al desarrollo y preparación del trabajo. Los revisores deben explicar y apoyar sus juicios de manera suficiente para que el editor, los miembros de cuerpo editorial y los autores comprendan el fundamento de las observaciones. Los autores deben incluir a los coautores fallecidos que cumplan con el criterio de coautoría, asentando la fecha de su muerte. Los revisores deben abstenerse de utilizar o difundir información, argumentos o interpretaciones no publicadas contenidas en un manuscrito bajo consideración, Ingenierías, Abril-Junio 2017, Año XX, No. 75 79 �Código de ética excepto con el consentimiento expreso de los autores posteriormente al proceso de evaluación. Los revisores deben considerar en su revisión posibles errores o fallas de los autores al citar el trabajo relevante de otros. Los revisores deben informar al editor si encontraran alguna semejanza substancial entre el manuscrito y cualquier otro trabajo. Los revisores no deberán intentar contactar a los autores, si hubieran inferido su identidad, previamente a haber emitido su fallo. Editor El editor debe dar consideración justa e imparcial a todos los manuscritos ofrecidos para su publicación, juzgando cada uno de sus méritos científicos o tecnológicos, sin prejuicios de raza, género, religión, creencia, origen étnico, ciudadanía, filosofía o política del autor. El editor debe considerar un manuscrito enviado para revisión como un documento confidencial. El editor debe abstenerse de expresar crítica personal. El editor debe explicar y apoyar su juicio final para que los autores comprendan el fundamento de las observaciones. El editor debe abstenerse de utilizar la información no publicada, argumentos o interpretaciones desplegados en un manuscrito sometido, excepto cuando cuente con el permiso del autor. El editor deben abstenerse de desplegar información sobre un manuscrito en proceso de revisión o publicación a ninguna persona fuera de aquellos a los que se les solicite consejo profesional. El editor debe respetar la independencia intelectual de los autores. El editor debe procesar los manuscritos con diligencia. El editor debe ejercer su responsabilidad y la autoridad para aceptar o rechazar un artículo enviado para su publicación. 80 El editor debe delegar en los miembros del consejo editorial o comité técnico la autoridad para aceptar o rechazar un artículo enviado para su publicación en casos en que se presente conflicto de interés con el editor. El editor debe delegar la responsabilidad y autoridad editorial a alguno de los miembros de los consejos editoriales cuando él sea autor o coautor de un manuscrito que se somete a consideración de la revista. Cuerpo Editorial (Consejos Editoriales y Comité Técnico) Los miembros del cuerpo editorial deberán estar dispuestos a otorgar consejo al editor en las situaciones requeridas. Los miembros del cuerpo editorial deben declinar cualquier invitación para brindar consejo si se les presenta algún conflicto de intereses, tal como encontrarse vinculados estrechamente a los autores o al trabajo a evaluar. Los miembros del cuerpo editorial deben manifestar al editor cualquier conflicto de intereses que detecten. Los miembros del cuerpo editorial deben considerar un manuscrito enviado para revisión como un documento confidencial. Los miembros del cuerpo editorial deben abstenerse de expresar críticas personales. Los miembros del cuerpo editorial deben explicar y apoyar sus juicios de manera suficiente para que el editor, los miembros de cuerpo editorial y los autores comprendan el fundamento de las observaciones. Los miembros del cuerpo editorial deben abstenerse de utilizar la información no publicada, argumentos o interpretaciones desplegados en un manuscrito sometido, excepto cuando se cuente con el permiso del autor. Los miembros del cuerpo editorial deben abstenerse de desplegar información sobre un manuscrito en proceso de revisión o publicación a cualquier persona fuera de aquellos que se les solicite consejo profesional. Los miembros del cuerpo editorial deberán respetar la independencia intelectual de los autores. Ingenierías, Abril-Junio 2017, Año. XX, No. 75 ��� Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Ingenierías Description An account of the resource Revista de la Facultad de Ingeniería Mecánica y Eléctrica de la UANL. Publicada a principios de la década de los noventa, editada por Rafael Covarrubias Ortiz. Contiene información sobre las actividades académicas, estudiantiles y administrativas de la Facultad, así como investigación y difusión de la ingeniería. Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Título Uniforme Ingenierías Año de publicación El año cuando se publico 2017 Año Año de la revista (Año 1, Año 2) No es es año de publicación. 20 Número Número de la revista 75 Mes de publicación Mes cuando se publicó Abril-Junio Día Día del mes de la publicación 1 Periodicidad La periodicidad de la publicación (diaria, semanal, mensual, anual) Trimestral Relación OPAC https://www.codice.uanl.mx/RegistroBibliografico/InformacionBibliografica?from=BusquedaAvanzada&bibId=1751916&biblioteca=0&fb=20000&fm=6&isbn= Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Ingenierías, 2017, Año 20, No 75, Abril-Junio Creator An entity primarily responsible for making the resource González Treviño, José Antonio, Director Subject The topic of the resource Ciencia Tecnología Ingeniería Investigación Publicaciones periódicas Description An account of the resource Revista de la Facultad de Ingeniería Mecánica y Eléctrica de la UANL. Publicada a principios de la década de los noventa, editada por Rafael Covarrubias Ortiz. Contiene información sobre las actividades académicas, estudiantiles y administrativas de la Facultad, así como investigación y difusión de la ingeniería. Publisher An entity responsible for making the resource available Universidad Autónoma de Nuevo León, Facultad de Ingeniería Mecánica y Eléctrica Contributor An entity responsible for making contributions to the resource Aguilar Garib, Juan Antonio, Editor Ocañas Galván, Cyntia, Redacción Date A point or period of time associated with an event in the lifecycle of the resource 01/04/2017 Type The nature or genre of the resource Revista Format The file format, physical medium, or dimensions of the resource tex/pdf Identifier An unambiguous reference to the resource within a given context 2020838 Source A related resource from which the described resource is derived Fondo Universitario Language A language of the resource spa Relation A related resource http://ingenierias.uanl.mx/archivo.html Spatial Coverage Spatial characteristics of the resource. San Nicolás de los Garza, N.L., (México) Rights Information about rights held in and over the resource Universidad Autónoma de Nuevo León Rights Holder A person or organization owning or managing rights over the resource. El diseño y los contenidos de La hemeroteca Digital UANL están protegidos por la Ley de derechos de autor, Cap. III. De dominio público. Art. 152. Las obras del dominio público pueden ser libremente utilizadas por cualquier persona, con la sola restricción de respetar los derechos morales de los respectivos autores Corriente inrush Eigenvalores Prueba HIL Transformador Umbral