criterios de optimización y diseño de un recticador trifásico para
Post on 12-Feb-2017
239 Views
Preview:
TRANSCRIPT
Universidad Politecnica de MadridCentro de Electronica Industrial
Proyecto Fin de Master
Criterios de Optimizacion y Diseno de un RectificadorTrifasico para Aplicaciones Aeronauticas
Marcelo Alexis Silva Faundez
Master en Electronica Industrial
Universidad Politecnica de Madrid
Centro de Electronica Industrial
Escuela Tecnica Superior de Ingenieros Industriales
Departamento de Automatica, Ingenierıa Electronica e Informatica Industrial
Madrid, Octubre 2011
Universidad Politecnica de Madrid
Centro de Electronica Industrial
Escuela Tecnica Superior de Ingenieros Industriales
Departamento de Automatica, Ingenierıa Electronica e Informatica Industrial
Master en Electronica Industrial
Criterios de Optimizacion y Diseno de un RectificadorTrifasico para Aplicaciones Aeronauticas
Autor: Marcelo Alexis Silva Faundez
Director: Jesus A. Oliver
Madrid, Octubre 2011
Proyecto Fin de Master
5
A mi querida Madre,Q. E. P. D.
6
CONTENIDO
7
8 Contenido
CONTENIDO
1. INTRODUCCION 9
1.1. Historia del Reabastecimiento en Vuelo en Aviones ........................... 9
1.2. Actual Arquitectura del Sistema de Realimentacion .......................... 12
1.3. Perfil de Carga del Sistema .......................................................... 12
1.4. Desarrollo de la Tesis .................................................................. 13
2. OPTIMIZACION A NIVEL DE ARQUITECTURA DEL SISTEMA 17
2.1. Arquitectura del Sistema de Alimentacion Propuesta......................... 17
2.1.1. Distribucion electrica en aplicaciones aeronautica .................... 17
2.2. Estudio de Arquitecturas para el Sistema de Alimentacion ................. 18
2.3. Unidad de Almacenamiento de Energıa. .......................................... 20
2.3.1. Dimensionamiento de la Baterıa .......................................... 21
2.3.2. Simulacion de la Arquitectura propuesta con Baterıa ............... 21
2.3.3. Dimensionamiento del Super Condensador ............................. 21
2.3.4. Simulacion de la Arquitectura Propuesta con Super Condensador 23
2.3.5. Dimensionamiento de Baterıa mas Super Condensador ............. 23
2.3.6. Simulacion de la Arquitectura Propuesta con Baterıa mas SuperCondensador ................................................................... 25
2.3.7. Resumen de las Unidades de Almacenamiento ....................... 26
3. SELECCION Y OPTIMIZACION DE LA TOPOLOGIA DEL SIS-TEMA 27
3.1. Rectificador Activo ..................................................................... 27
3.2. Familias de Rectificadores Activos ................................................. 27
3.3. Rectificadores bidireccionales ........................................................ 29
3.4. Rectificadores unidireccionales ...................................................... 29
3.5. Comparacion entre las topologıas unidireccionales con convertidoresDC-DC. ................................................................................... 29
3.5.1. Comparacion en eficiencia .................................................. 30
3.5.2. Comparacion en peso y volumen. ......................................... 31
3.5.3. Aspectos del sistema. ........................................................ 32
3.5.4. Resumen de la comparacion ................................................ 32
3.6. Rectificador Trifasico tipo Buck de Alta Eficiencia. ........................... 33
3.7. Aislamiento Electrico. ................................................................. 34
5
6 Contenido
3.8. Topologıa multi-celda.................................................................. 35
3.9. Optimizacion de la topologıa en cuanto numero de celdas ................... 39
3.10. Simulacion del rectificador de tres celdas......................................... 40
4. ANALISIS DEL RECTIFICADOR TIPO BUCK 45
4.1. Rectificador trifasico tipo buck ..................................................... 45
4.2. Principio de operacion................................................................. 45
4.2.1. Suposiciones ................................................................... 45
4.3. Estados de Conduccion................................................................ 45
4.4. Modulacion Vectorial .................................................................. 49
5. NUEVAS CONSIDERACIONES DE DISENO DEL FILTRO EMIDE RECTIFICADOR TRIFASICO PARA APLICACIONES AERONAUTI-CAS 53
5.1. Filtro EMI................................................................................ 54
5.2. Topologıa del Convertidor ............................................................ 55
5.3. Estandar EMI en avionica: MIL-STD 461E .................................... 56
5.4. Diseno del Filtro EMI ................................................................. 58
5.4.1. Espectro de la corriente de entrada del rectificador .................. 58
5.4.2. Frecuencia de corte del filtro ............................................... 59
5.4.3. Consideraciones para el condensador .................................... 59
5.4.4. Nuevas consideraciones para el del diseno del condensador ....... 60
5.5. Optimizacion del Filtro .............................................................. 62
5.5.1. Estimacion de del peso y las perdidas del filtro ....................... 62
5.5.2. Estimacion de las perdidas en los MOSFETs.......................... 64
5.5.3. Frecuencia de conmutacion optima ....................................... 64
5.5.4. Conclusiones.................................................................... 67
6. PLATAFORMA DE PRUEBA 69
6.1. Tarjeta de Potencia del Rectficador................................................ 71
6.1.1. Layout de la Tarjeta de Potencia, version 1. ........................... 71
6.1.2. Layout de la Tarjeta de Potencia, version 2. ........................... 71
6.2. Tarjeta de Control del Convertidor ................................................ 75
6.3. Filtro EMI................................................................................ 77
7. RESULTADOS EXPERIMENTALES 79
7.1. Formas de Onda de las conmutaciones ............................................ 79
7.2. Rendimiento, Factor de Potencia y Distorsion Armonica .................... 79
7.3. Formas de onda de Corriente y tensiones de lınea ............................. 79
8. CONCLUSIONES 85
9. APENDICE 87
9.1. Articulos Publicados en Congresos................................................. 87
9.1.1. Articulo presentado en el congreso SAAEIG 2011. ................... 87
9.1.2. Articulo presentado en el congreso ECCE 2011. ...................... 93
Contenido 7
9.2. Codigos Fuentes del DSP .............................................................100
10.REFERENCIAS 117
8 Contenido
Capıtulo 1
INTRODUCCION
LA presente tesis esta enmarcada en el proyecto Boom Power supervisado porAirbus, el cual consiste en la actualizacion y mejoramiento de los sistemas de
alimentacion de un sistema que permite el reabastecimiento de combustible de unaaeronave en vuelo.
En el trabajo se presenta un estudio del sistema actual de potencia para el re-abastecimiento en vuelo, posteriormente se realiza un estudio y optimizacion a nivelde arquitecturas para el mejoramiento del sistema. Una vez optimizado el sistema anivel de arquitectura se realiza una optimizacion a nivel de topologıa que incorporaentre sus variables el analisis del filtro EMI y la division del rectificador en pequenosrectificadores en paralelo para ası obtener un mınimo peso, volumen y perdidas delsistema en su globalidad. Finalmente se presenta implementacion de la nueva arqui-tectura y topologıa propuesta incluyendo resultados experimentales.
1.1. Historia del Reabastecimiento en Vuelo en Aviones
La capacidad de una aeronave para mantenerse en vuelo y alcanzar mayores dis-tancias siempre fue una variable que se deseo aumentar. Los problemas tecnicos eranmuchos, pero hacia mitad del siglo XX se logro lo que parecıa imposible: construiraviones que pudieran ser reabastecidos de combustible en vuelo, sin tener que aterrizary despegar. El reabastecimiento aereo de combustible cambio, ası, todas las reglas delcombate aereo y terrestre.
El reabastecimiento implica que un avion cisterna o tanquero le provea a otrosaviones de combustible, estando ambos en el aire y con los motores encendidos. Estole permite al avion receptor mantenerse en vuelo por mas tiempo, logrando alcanzardistancias que antes hubieran resultado imposibles. Estos sistemas hacen que un avionpueda estar volando, teoricamente, de manera indefinida, siendo limitados primero porla fatiga de la tripulacion (la cual puede ser rotada, en el caso de ciertos bombarderos)y luego por ciertas cuestiones de mantenimiento e ingenierıa del aparato en sı.
El reabastecimiento permite que los aviones obtengan varias ventajas tacticas:
que cualquier avion alcance distancias que antes no podrıa haber alcanzado,particularmente por no estar disenado para ello;
que cualquier avion, particularmente un caza, bombardero o de ataque a tierra,permanezca mas tiempo en el aire, a la espera de ayudar a fuerzas amigas;
9
10 Capıtulo 1 – Introduccion
que el avion pueda despegar con carga maxima de armamento, pero poco com-bustible, en mejores condiciones de seguridad y con una pista de extensionnormal. Generalmente el peso maximo de carga de un avion no le permite man-tenerse en el aire, y mucho menos despegar; de manera que el avion puedelevantarse con mucho peso de armas para llenar sus combustibles mas tarde, ala ida o a la vuelta de sus misiones.
Estas ventajas traen aparejadas otras no menores. Por ejemplo, permite que loscazas no gasten puntos fuertes en llevar tanques de combustible desechables, los cualesperjudican su aerodinamica y deben ser lanzados en caso de entrar en combate aereo.
La capacidad de permanecer mas tiempo en el aire es un verdadero multiplicadorde fuerza, ya que permite que un avion haga el trabajo de dos o tres, evitando idas yvenidas para aterrizajes y despegues.
La figura 1.1(a) muestra una fotografıa del sistema de reabastecimiento en vuelofabricado por Airbus. En la figura 1.1(b) se nuestra el primer contacto exitoso logradopor Airbus con su sistema de reabastecimiento en vuelo.
En una breve explicacion, el procedimiento de reabastecimiento en vuelo disenadopor Airbus consiste en lo siguiente. Los pilotos de ambos aviones aproximan y alin-ean los aviones, una vez los aviones esten una posicion adecuada, el avion tanquerocomienza el descenso controlado de la pertiga, la pertiga esta sujeta al avion medianteun pivote que permite el movimiento en 360 grados, de esta forma, para controlar elmovimiento vertical de la pertiga se utiliza un motor electrico y para controlar losmovimiento laterales se utilizan dos alerones que estan en el extremo de la pertiga,cada uno de estos alerones son accionados por motores electricos. Una vez la pertigaeste en el lugar indicado con una manguera telescopica se da alcance al avion receptor;esta manguera telescopica es tambien controlada por un motor electrico.
1.1 Historia del Reabastecimiento en Vuelo en Aviones 11
Figura 1.1. Fotografıa del sistema de reabastesimiento en vuelo
12 Capıtulo 1 – Introduccion
1.2. Actual Arquitectura del Sistema de Realimentacion
El sistema de abastecimiento de combustible en vuelo, es accionado mediantemotores AC alimentados y controlados a traves de drives desde una red trifasica de115V a 400Hz. la arquitectura electrica del sistema es mostrada en la figura 1.2, enesta figura tambien se incluye la potencia aparente y el peso de cada uno de los drives.
115V@400Hz
AC Motor
AC Motor
AC Motor
AC Motor
3
Drive2kVA 11kg
Drive2kVA 11kg
Drive4kVA 39kg
Drive3kVA 39kg
Figura 1.2. Arquitectura actual del sistema de alimentacion
Cada uno de los motores es accionado mediante un drive; los drives son alimenta-dos desde la red trifasica 115V a 400Hz. En la figura 1.3 muestra como es la estructurainterna de estos drives. Los drives internamente rectifican la tension alterna utilizandoun puente de diodos. A la salida del puente de diodo se generan 270V en continua,llamada de ahora en adelante enlace de continua (DC-Link). En una ultima etapaun inversor transforma la tencion interna continua en una tension alterna adecuada,variable en frecuencia y amplitud que alimenta y controla los motores AC.
Ademas, los drives presentan filtros de entrada que permiten que el convertidorcumpla con los estandares aeronauticos.
Asimismo los drives contienen resistencias en el enlace de continua; estas evitansobre tension en el enlace de continua cuando los motores actuan como generadores;lo que se debe a que la energıa regenerada no puede ser regresada a la red debido alpuente de diodos, ası esta energıa es disipada mediante resistencias.
1.3. Perfil de Carga del Sistema
La operacion de reabastecimiento en vuelo dura 10 minutos desde el momento enque la pertiga va descendiendo hasta que ambos aviones estan acoplados y existe eltraspaso de combustible de un avion al otro. La figura 1.4 muestra el perfil de cargadel sistema.
1.4 Desarrollo de la Tesis 13
Figura 1.3. Estructura interna de los Drives
Este proceso de divide en tres fases, la primera fase corresponde a la puesta enmarcha de la perdiga, es decir, el descenso controlado de esta.
La segunda fase corresponde a la operacion en vuelo libre de la perdiga, es decir,cuando la manguera telescopica que va dentro de la perdiga va descendiendo hastaque se acopla con el avion receptor de combustible. Esta es la etapa mas exigente enterminos de potencia, en esta etapa existe un pico de potencia de 34kW que dura 2segundos; este evento corresponde a una accion de emergencia, en la que la retraccionde la manguera se realiza a la velocidad maxima. Si bien esta accion va a ser requeridasolo casos esporadicos de emergencia, el sistema debe ser disenado para cumplir coneste requisito a cabalidad.
La tercera fase es cuando ambos aviones estan acoplados, en esta etapa existeregeneracion de energıa; lo que se expresa como potencia negativa en el perfil decarga. Esta regeneracion se debe a que al estar ambos aviones acoplados se produceuna transferencia de energıa mecanica en electrica producto de que el sistema derealimentacion frena el movimiento relativo entre ambos aviones.
El perfil de carga muestra la potencia total del sistema de reabastecimiento en vue-lo, por lo que no existe informacion acerca de los picos de potencia o las regeneracionde energıa de cada motor; sin embargo como se mostrara en el siguiente capıtulo,una de las actualizaciones del sistema consiste en la unificacion electrica del sistema,de esta forma con el perfil de carga mostrado en 1.4 se tiene toda la informacionnecesaria para el diseno del nuevo sistema.
1.4. Desarrollo de la Tesis
La tesis se desarrolla en 8 capıtulos, detallados
Capıtulo 1
Presenta la importancia de los sistemas de reabastecimientos en vuelo en aviones.Ademas se muestra la arquitectura y convertidores utilizados para la alimentaciondel sistema.
Capıtulo 2
En el capıtulo 2 se muestra la optimizacion a nivel arquitectura del sistema, paralo cual se han estimado los pesos y volumenes para diferentes arquitecturas de
14 Capıtulo 1 – Introduccion
Figura 1.4. Perfil de carga del sistema de reabastecimiento de combustible en vuelo
acuerdo a las especificadores electricas del sistema.
1.4 Desarrollo de la Tesis 15
Capıtulo 3
Se muestra el estudio empleado para la seleccion de la topologıa, el estudioincluye estimaciones de perdidas y peso del sistema a diferentes frecuencias deconmutacion y numero de celdas en paralelo
Capıtulo 4
En el capıtulo 4 se analiza el rectificador tipo buck explicado su modo de fun-cionamiento y modulacion.
Capıtulo 5
Se muestra la optimizacion realizada del filtro EMI y rectificador en conjunto,ademas incluye una nueva metodologıa de diseno del filtro para la obtencion defactor de potencia unitario.
Capıtulo 6
En este capıtulo se expone la plataforma utilizada para la pruebas experimen-tales.
Capıtulo 7
En este capıtulo se muestran los resultados experimentales.
Capıtulo 8
Durante el transcurso del proyecto parte del trabajo realizado ha sido publicadoen congresos internacionales.
16 Capıtulo 1 – Introduccion
Capıtulo 2
OPTIMIZACION A NIVEL DEARQUITECTURA DEL SISTEMA
En este capıtulo se muestra la evolucion de la optimizacion del sistema desde unaoptimizacion a nivel de arquitectura.
2.1. Arquitectura del Sistema de Alimentacion Propuesta
Como se explica anteriormente, el sistema de reabastecimiento en vuelo consta devarios motores accionados mediante drives, los cuales son alimentados directamentedesde la red. Al estar cada motor accionado por un drive diferente no se puede reuti-lizar la energıa que los motores regeneran, ası este se debe desperdiciar en resistenciasque aumentan el peso del sistema y disminuyen.
Con el objeto de reutilizar la energıa regenerada, se propone conectar los enlacesde continua (270V) de todos los drives, como se muestra en la figura 2.7. De estamanera que se obtienen dos importantes beneficios:
Cuando algun motor trabaje como generador, la energıa regenerada puede serabsorbida por otro motor sin provocar una sobretension en el enlace de continua.
El rectificador solo debe entregar la potencia media de la carga, dejando queun dispositivo almacenador de energıa, baterıa y/o super condensador absorbala los valles de potencia negativa y entregue los picos de potencia a la carga.
Para llevar a cabo esto la arquitectura cambia someramente, los motores son ac-cionados mediante inversores, y estos a su vez, son alimentados mediante tensioncontinua. Los motores y estan fısicamente en lugares distantes, lo que obliga a re-alizar una distribucion electrica a 270Vdc. ¿Es o no posible distribuir a 270Vdc enaplicaciones aeronauticas?
2.1.1. Distribucion electrica en aplicaciones aeronautica
La distribucion electrica en aeronautica se realiza principalmente en continua a28V o en trifasica a 115V a 400Hz, sin embargo hoy en dıa existe una fuerte ten-dencia en los nuevos disenos de aeronaves militares y aeronaves no tripuladas haciael concepto de avion mas electrico (MEA). Por un lado esto es una consecuenciade las sustituciones de los equipos convencionales, que dependen de la neumatica,mecanica e hidraulica, por equipos que dependen de la energıa electrica. Este factor
17
18 Capıtulo 2 – Optimizacion a nivel de arquitectura del sistema
Figura 2.1. Arquitectura actual del sistema de alimentacion
ha provocado aumento de los equipos, que requieren de energıa electrica. Ademas,estos cambios proporcionan un mejor rendimiento del sistema debido al aumento dela fiabilidad, menor mantenimiento, la eficiencia en la conversion de la energıa y laeficiencia total del sistema. Por otro lado, hay un aumento del numero de equiposque dependen de la energıa electrica y forman parte de los diferentes sistemas de laaeronave (avionica, comunicaciones, vigilancia), como por ejemplo: radares, camarasde infrarrojos y aparatos de radio electro-opticos , etc [1–6].
Al incrementar la demanda de energıa electrica en aviones, se ha visto en lanecesidad de elevar la tension de distribucion para disminuir el diametro y a suvez el peso de los cables de distribucion. Por esta razon durante la ultima decadaen aeronautica se ha empezado a utilizar corriente continua en alta tension a 270Vganando cada dıa un mayor protagonismo en la distribucion electrica de los aviones[4, 5].
2.2. Estudio de Arquitecturas para el Sistema de Alimentacion
Con el objeto de mejorar el sistema de alimentacion actual de reabastecimientoen vuelo, se han estudiado diversas posibles arquitecturas electricas con el fin deoptimizar el sistema a nivel de arquitectura en terminos de volumen y peso.
Las arquitecturas seleccionadas para su analisis se muestran en las figuras 2.2 y2.3 . Estas arquitecturas son denominadas de la siguiente manera: 1B, 1BC, 1C1,1C2, 2B, 2BC1, 2BC2 y 2C1, donde el prefijo 1 indica que la arquitectura tiene unaconfiguracion en cascada y el prefijo 2 indica que la arquitectura tiene una configu-racion en paralelo. La letra B y C indican que la arquitectura utiliza como unidad dealmacenamiento una baterıa y/o un super condensador respectivamente.
Para el dimensionamiento de los elementos en cada arquitectura, es decir, rectifi-cador, convertidor DC-DC, baterıa y super condensador, se ha considerado el siguienteescenario:
2.2 Estudio de Arquitecturas para el Sistema de Alimentacion 19
Figura 2.2. Arquitecturas de alimentacion en cascada
Figura 2.3. Arquitecturas de alimentacion en en paralelo
La potencia media exigida en el perfil de carga mostrado en el capıtulo anterior(figura 1.4) sera suministrada por el rectificador
Los picos y valles de potencia seran suministrados y absorbidos por la unidadde almacenamiento de energıa.
Para todas las arquitecturas se ha dimensionado cada elemento, y a su vez, se haestimado el peso y el volumen de cada uno de estos elementos de acuerdo al estadodel arte en tecnologıas de baterıas, super condensadores y convertidores comercialesen el mundo aeronautico. La figura 2.4 muestra un resumen de los resultados de estasestimaciones.
La diferencia entre potencia media y potencia pico en esta aplicacion es enorme, lapotencia media es 7kW y el pico de potencia es 34kW, ası al ser el pico de potencia tan
20 Capıtulo 2 – Optimizacion a nivel de arquitectura del sistema
Figura 2.4. Resumen de los resultados de las estimacions de peso y volumen de lasarquitecturas seleccionadas
grande todas las arquitecturas que contienen convertidores que conectan las unidadesde almacenamiento con la carga se ven altamente perjudicadas. De esta forma lasarquitecturas mas favorecidas son la 1B, 1BC y 1C2. La diferencia tanto en peso comoen volumen entre las arquitecturas convenientes y las no convenientes es inmensa, loque significa que sin realizar el estudio de arquitecturas se podrıa cometer un errory seleccionar una arquitectura varias veces mas pesada que la optima sin importar eldiseno de los convertidores, baterıas y super condensadores.
Las arquitecturas seleccionas (1B, 1BC y 1C2) tienen en comun que son con-figuraciones en cascada y ademas sus unidades de almacenamiento de energıa estandirectamente conectadas a las cargas, siendo el tipo de unidad de almacenamientola unica diferencia entre ellas. Entonces para saber si unidad de almacenamiento deenergıa que mejor cumple los requisitos del sistema e una baterıa, super condensadoro una combinacion de ambos, se han estudiado con mas detalles cada uno de estasarquitecturas.
2.3. Unidad de Almacenamiento de Energıa.
Para el almacenamiento de energıa se han estudiado tres alternativas, utilizarsuper-condensadores, baterıas o una combinacion de ambos. En esta seccion se pre-sentara el diseno para cada uno de estos elementos y posteriormente simulaciones quepermitan el analisis en detalle de cada arquitectura.
2.3 Unidad de Almacenamiento de Energıa. 21
2.3.1. Dimensionamiento de la Baterıa
Para el diseno de la bacteria se utilizo el siguiente criterio:
La densidad de potencia de una baterıa de ion litio es de 1kW/kg, y la densidadde energıa es 70Wh/kg.
El sistema es mas exigente en terminos de potencia que de energıa, esto significaque la baterıa debe ser disenada de acuerdo a la potencia maxima instantanea.
Suponiendo que la potencia del rectificador es 10kW, la maxima potencia de-mandada desde la baterıa es 24kW. Considerando la densidad de potencia1kW/kg, la baterıa requerida tiene un peso de 24kg
24kg de baterıa equivalen a 24*70Wh = 1680Wh de capacidad. La tension enla salida es 270V, por lo tanto la baterıa requerida es de 6Ah a 270V.
La corriente en el pico de potencia es 100A, es decir, la descarga maxima serealizara a 17C, lo que es aceptable para este tipo de baterıas.
2.3.2. Simulacion de la Arquitectura propuesta con Baterıa
Para estudiar el comportamiento de la arquitectura con baterıa se ha simulado anivel de sistema la arquitectura completa, con la baterıa disenada y el perfil de carga,la figura 2.5, muestra los resultados de la simulacion.
En la simulacion el rectificador es controlado en modo corriente y con otro lazoexterno el rectificador controla el estado de carga de la baterıa (SOC) de forma demantenerlo en el 70 % de su carga maxima. La potencia maxima esta limitada a10kW, cuando la potencia de salida es mayor a 10kW, la baterıa entrega la potenciarestante.
En el pico de potencia (34kW) La baterıa se descarga solo del 70 % a 69 %. Unavez el pico de corriente ha pasado el rectificador continua entregando 10kW hastallevar la baterıa a su 70 % de su carga nominal. La caıda de tension es de menosde 2V llegando la tension a 268V lo que en la practica es despreciable considerandola maxima caıda de tension permitida es hasta 250V. Durante el resto de tiempo labaterıa no tiene un rol importante ya que toda la potencia requerida por la cargaen entregada por el rectificador, salvo en los transitorios y durante la regeneracionde energıa (cerca del segundo 500). En esta etapa la energıa regenerada es absorbidapor la baterıa y su estado de carga comienza a aumentar sobre el 70 %. Durante estetiempo el rectificador deja de entregar energıa haciendo la corriente cero hasta quetoda la energıa absorbida por la baterıa sea devuelta hacia la carga y su SOC regreseal 70 %. Una vez ocurra esto el rectificador vuelve a operar normalmente entregandola potencia hacia la carga.
2.3.3. Dimensionamiento del Super Condensador
Para el diseno del super condensador se utilizo el siguiente criterio:
Los super condensadores, al contrario de las baterıas, tienen una alta densidadde potencia a cambio de una baja densidad de energıa. La densidad de potenciade un super condensador es de 5kW/kg y una densidad de energıa es 3Wh/kg.
22 Capıtulo 2 – Optimizacion a nivel de arquitectura del sistema
Figura 2.5. Simulacion de la arquitectura con baterıa como unidad almacenadora
El sistema es mas exigente en terminos de potencia que de energıa, esto significaque el super condensador debe ser disenado de acuerdo a la potencia maximainstantanea.
La corriente pico es de 100A, y la tension mınima de salida es de 250V, por lotanto el condensador requerido es:
C =2s · 100A
270V − 250V= 10F
Por lo tanto el super condensador requerido es un condensador de 10F a 270V.
2.3 Unidad de Almacenamiento de Energıa. 23
Un condensador de 10F a 270V almacena una energıa de 0,5 ·C ·V 2. De acuerdocon la densidad de energıa de estos componentes, el peso del super condensadores el siguiente:
0,5 · C · V 2
Densidad Energıa=
0,5 · 10F · 270V 2
3 · 3600Ws/kg= 34kg
2.3.4. Simulacion de la Arquitectura Propuesta con Super Condensador
Para estudiar el comportamiento de la arquitectura con super condensador se hasimulado la arquitectura completa, con el super condensador disenado y el perfil decarga, la figura 2.6 muestra los resultados de la simulacion.
En esta arquitectura el sistema de control varia un poco en comparacion a laarquitectura con baterıa. En esta arquitectura existen dos lazos de control anidados,en un lazo interno se controla la corriente de salida del rectificador y en un lazoexterno se controla la tension en el condensador donde su consigna es 270V.
La potencia del rectificador es limitada a 10kW al igual que en la simulacionanterior, cuando la potencia de carga es superior a 10kW, el condensador entrega lapotencia restante. Cuando esto sucede la tension en el condensador disminuye, y porende en el bus de salida tambien disminuye. En el pico de 34kW la caıda de tensionllega a 250V, como se esperaba de acuerdo al diseno del super condensador.
Cuando se acaba el pico de 34kW, el rectificador continua entregando la maximapotencia hasta llevar la tension a 270V en el super condensador; el sistema tarda 50sen volver a recuperar la tension a 270V.
Durante el resto del tiempo el condensador solo entrega o absorbe energıa en lostransitorios; salvo en el momento en que la potencia de carga es negativa, en esemomento la potencia regenerada es absorbida completamente por el condensador yel rectificador deja de entregar energıa. Una vez la potencia de carga vuelve a serpositiva toda la energıa absorbida por el super condensador es devuelta a la carga sinnecesidad de desperdiciar esta energıa en resistencias de sobre tension como ocurrelos drives actuales.
2.3.5. Dimensionamiento de Baterıa mas Super Condensador
La idea de esta arquitectura es repartir potencia entre la baterıa y super con-densador en el pico de potencia. Para el diseno de la baterıa y super condensador seutilizo el siguiente criterio:
Una baterıa de 2Ah puede entregar una corriente de hasta 40A. El peso de unabaterıa de 2Ah a 270V es de 11kg
Entonces los restantes 60A deben ser entregados por el super condensador du-rante 2s, donde la caıda de voltaje no debe superar los 20V(270-250V).
C =2s · 60A
270V− 250V= 6F
El peso de un super condensador de 6F a 270V es de 20kg
24 Capıtulo 2 – Optimizacion a nivel de arquitectura del sistema
Figura 2.6. Simulacion del la arquitectura utilizando un supercondensador comounidad almacenadora
2.3 Unidad de Almacenamiento de Energıa. 25
Figura 2.7. Simulacion del la arquitectura utilizando una baterıa y super condensadorcomo unidad almacenadora
En resumen, el peso en unidades de almacenamiento para esta topologıa e de31kg, ademas de la circuiterıa necesaria para limitar la corriente maxima de labaterıa.
2.3.6. Simulacion de la Arquitectura Propuesta con Baterıa mas SuperCondensador
Al igual que para las arquitecturas anteriores se ha simulado la arquitectura conbaterıa mas super condensador, en la figura 2.7 se muestran los resultados de lasimulacion
Como se ve en la figura 2.7 en el momento del pico de 34kW, la baterıa entrega
26 Capıtulo 2 – Optimizacion a nivel de arquitectura del sistema
una corriente de 40A y el resto de corriente requerida por la carga es entregada porel condensador.
Luego del pico de potencia, el sistema tarda 10s en llevar nuevamente la tensioncerca de los 270V (269V), ası con una combinacion de baterıa y super condensadorel sistema necesita 5 veces menos tiempo en volver al punto de operacion que en laarquitectura con solo super condensadores. Esto permitirıa un eventual segundo picode 34kW luego de 10s.
Durante el resto del tiempo, la baterıa y el super condensador se reparten lacorriente en los transitorios y la tension del bus de salida permanece en 270V conleves deviaciones en los transitorios.
En el momento de regeneracion, la mayor parte de la energıa regenerada es alma-cenada en la baterıa, ya que su integral de corriente negativa es mayor. Esto se debea tiene una mayor capacidad de almacenamiento de energıa a una misma desviacionde tension que el super condensador.
2.3.7. Resumen de las Unidades de Almacenamiento
La tabla 2.1 muestra un resumen con los datos mas importante de los resulta-dos obtenidos de la estimaciones de peso y de las simulaciones de las arquitecturasseleccionadas.
Tabla 2.1. Resumen de resultados de peso de las unidades de almacenamiento deenergıa para las tres arquitecturas estudiadas.
Storage Unit Capacity Weight Voltage Drop Over Voltage
Battery 6Ah 24kg 268V 271V
µCapacitor 3F 10 kg 205V 285V
µCapacitor 10F 34 kg 250V 271V
Batt+µCap 2Ah,6F 11+20kg 250V 275V
El resumen muestra un diseno con un super condensador de 3F, el cual si bienes por lejos el menos pesado, no es suficiente para cumplir con el lımite de caıdade tension. Con este super condensador el peso ganado no compensa los beneficiosde tener un bus de continua en principio interno que cumpla con los estandaresaeronauticos, que en un futuro pueda ser utilizado en otras aplicaciones dentro delavion.
Entre la arquitectura con baterıa y la con super condensador, claramente la mejoropcion es la baterıa, ya que ademas de ser 10kg mas liviana, tiene la ventaja que luegodel pico de potencia 34kW el estado de carga de la baterıa cambia solo marginalmentelo que permitirıa otra eventual descarga de 34kW sin necesidad de esperar tiempo aque el sistema vuelva al punto normal de trabajo, como ocurre con el super conden-sador.
Entre la solucion con baterıa o con baterıa mas super condensador, la mejor opcionsigue siendo con baterıa por simplicidad, ya que la otra arquitectura ademas de tenermas elementos, se le debe incluir la electronica para poder limitar la corriente de labaterıa, lo cual incrementa el peso, complejidad y perdidas del sistema.
Capıtulo 3
SELECCION Y OPTIMIZACIONDE LA TOPOLOGIA DEL
SISTEMA
3.1. Rectificador Activo
Hasta el momento el rectificador considerado es un puente de diodos, lo que sig-nifica que el rectificador no es controlado, es por esto que se necesita un convertidorcontinua a continua que regule la potencia del rectificador, para ası dejar que lasunidades de almacenamiento entreguen energıa en los picos de potencia. Adicional-mente en la entrada del rectificador puente de diodos se necesita un filtro LC quepermita tener un THD admisible en aplicaciones aeronauticas. En [7–9] muestrancomo disenar estos filtros para tener un THD menor al 5 % que es lo que se exigeen aplicaciones aeronauticas. Sin embargo el factor de potencia es mermado y puedellegar a ser menor a 0.8, lo que no es admisible.
Para resolver todos estos problemas en conjunto se propone cambiar el puente dediodos y convertidor continua a continua a un rectificador trifasico activo que, controlela potencia de entrada y la energıa en la unidad de almacenamiento, permita un THDmenor al 5 % y un factor de potencia superior al 95 %. Que tipo de convertidor activose discutira mas adelante en la presente tesis.
3.2. Familias de Rectificadores Activos
En la figura 3.2(b) las dos grandes familias de rectificadores activos, los reductores(buck) y los elevadores (boost) [10, 11]. Dentro del grupo de los reductores los masutilizados son el bidireccional de 6 switches y el unidireccional de 3 switches. Por otrolado dentro de la familia de los elevadores los mas importantes son el bidireccionalde 6 switches y el unidireccional de 3 switches o rectificador Viena. Las topologıas deestos rectificadores se muestran en la figura 3.1.
En esta aplicacion se necesita un rectificador que convierta una tension trifasicade 115V eficaz fase a neutro a una frecuencia de 400Hz en una tension continua de270V, cabe destacar que esta relacion 115V en la entrada y 270V en la salida es laque se obtiene directamente con un rectificador puente de diodos.
Al ser los rectificadores activos elevadores o bien reductores se debe necesariamente
27
28 Capıtulo 3 – Seleccion y optimizacion de la topologıa del sistema
Figura 3.1. Topologıas de rectificadores tipo boost y tipo buck
Figura 3.2. (a) Configuracion en dos etapas, (b) rectificadores trifasicos activos tipobuck y tipo boost
optar por una topologıa de dos etapas 3.2(a), ya que con una red trifasica de 115Vun rectificador activo tipo elevador generara en la salida una tension superior a 270V,asimismo un rectificador activo tipo reductor generara una tension de salida inferiora 270V.
Como se muestra en la figura 3.2(a) con un rectificador tipo buck se obtieneuna tension continua de 200V, en una segunda etapa se eleva esta tension con unconvertidor de continua a continua a 270V. Por otro lado, con un rectificador tipoboost se obtiene una tension continua 400V y en una segunda etapa se baja la tenciona 270V con un convertidor reductor.
3.3 Rectificadores bidireccionales 29
Para analizar y averiguar cual de estas configuraciones, un rectificador buck masun DCDC boost o un rectificador boost mas un DC-DC buck, es la que mejor cumplelos requisitos del sistema, se analizaran primero los rectificadores bidireccionales de6 switches y posteriormente los unidireccionales de 3 switches .
3.3. Rectificadores bidireccionales
Las topologıas bidireccionales son ampliamente las mas utilizadas aplicacionesindustriales, especialmente la topologıa elevadora, ya que al ser bidireccionales puedentrabajar como rectificador o inversor. En particular en accionamiento de maquinaselectricas donde las cargas actuan como motores y generadores presentan la granventaja de poder regenerar energıa hacia la red. Aplicaciones en donde se puedeencontrar este tipo topologıas son por ejemplo montacargas, gruas, elevadores y UPS[11–14]
Si bien en nuestra aplicacion existe regeneracion de energıa, cuando ambos avionesestan acoplados, esta energıa no se puede devolver hacia la red ya que los estandaresaeronauticos no los permiten. Por lo tanto esta ventaja de los rectificadores bididi-recciones sobre los unidirecciones no aplica en nuestra aplicacion en particular.
El rectificador elevador es un convertidor fuente de tension de dos niveles ( +V dc/2y −V dc/2). El rectificador reductor es un convertidor fuente de corriente tambien dedos niveles (+Idc y −Idc).
3.4. Rectificadores unidireccionales
Este tipo de convertidores son utilizados en la actualidad como un reemplazode un rectificador de diodos convencionales para proporcionar un factor de potenciaunitario, reduccion del THD en corriente alterna y una tension continua regulada desalida [11].
El rectificador elevador es un convertidor fuente de tension de tres niveles (+V dc/2, 0 y −V dc/2). El rectificador reductor es un convertidor fuente de corri-ente tambien de tres niveles (+Idc, 0 y −Idc). Al tener estos convertidores un nivelmas de tension y corriente que sus versiones bidireccionales, sus filtros de entrada sonnotoriamente mas pequenos en peso y volumen.
Por lo tanto como las topologıas unidireccionales presentan una mayor densidad depotencia que sus versiones bidirecciones, son las mas apropiadas en nuestra aplicacion.
3.5. Comparacion entre las topologıas unidireccionales conconvertidores DC-DC.
Un estudio acabado de las dos topologıas que nos interesa incluido el convertidorDC-DC es mostrado en las referencias [15, 16]. Para analizar y posteriormente elegirla topologıa nos basaremos es el estudio realizado en estos artıculos.
La figura 3.3 muestra las topologıas a analizar, en (a) se muestra la topologıarectificador trifasico unidireccional tipo buck mas un boost continua a continua, cabedestacar que ambos convertidores comparten los inductores L0+ y L0−, estos corre-sponden al filtro de salida del rectificador buck y al inductor de entrada del boost.
30 Capıtulo 3 – Seleccion y optimizacion de la topologıa del sistema
Figura 3.3. (a)Topologıa rectificador trifasico tipo buck mas boost DC-DC,(b)Topologıa rectificador trifasico tipo boost mas buck DC-DC
En la figura 3.3(b) se muestra la topologıa completa de rectificador unidireccionaltrifasico tipo boost mas un buck de dos fases continua a continua.
Para asegurar un balance entre densidad de potencia y eficiencia total del sistemasimilar para ambas topologıas se han tomado en cuenta las siguientes consideracionesen el diseno de los convertidores:
La comparacion entre ambas topologıas se realiza en tres parametros, eficienciatotal del sistema, peso y volumen y aspectos del sistema.
3.5.1. Comparacion en eficiencia
En la figura 3.4 se muestra la comparacion de eficiencia de ambas topologıasa diferentes tensiones de salida. A baja tension de salida la topologıa boost+buckpresenta una mejor eficiencia, sin embargo a alta tension de salida la configuracionbuck+boost presenta mejor eficiencia. En nuestra aplicacion la tension de salida es
3.5 Comparacion entre las topologıas unidireccionales con convertidores DC-DC. 31
Figura 3.4. Comparacion de la eficiencia total de ambas topologıas [15,16]
Figura 3.5. (a) Configuracion en dos etapas, (b) rectificadores trifasicos activos tipobuck y tipo boost [15,16]
fija a 270V por lo que en terminos de eficiencia ambas topologıas presenta desempenossimilares.
3.5.2. Comparacion en peso y volumen.
El peso y volumen de los rectificadores corresponden principalmente a los compo-nentes pasivos y los disipadores. Como las eficiencias de ambos sistemas son similares,sus perdidas tambien son practicamente las mismas por consiguiente el pesos y volu-men de sus disipadores son iguales.
Para resaltar las diferencias de ambas topologıas, los disipadores no estan incluidosen comparacion mostrada en la figura 3.5. La topologıa buck+boost es mas ventajosasobre la boost+buck debido al peso y volumen de los componentes pasivos de potencia.Esto debido al gran numero de inductores que requiere la topologıa boost+buck donde
32 Capıtulo 3 – Seleccion y optimizacion de la topologıa del sistema
los inductores del filtro de entrada son comparables a los inductores almacenadoresde energıa en la topologıa buck+boost (L0+ y L0). Ademas los condensadores en elDC-link de la topologıa boost+buck son relativamente grande en comparacion conlos condensadores del filtro de entrada de la topologıa buck+boost.
Es importante destacar que para esta comparacion el peso y volumen de los filtroEMI no han sido incluidos, y en principio el filtro de la topologıa boost+buck esmas pequeno que la buck+boost, sin embargo esta diferencia no es en ningun caso lasuficiente como para que se inviertan los resultados mostrados en 3.5.
3.5.3. Aspectos del sistema.
Ademas de la eficiencia, peso y volumen tambien los sistemas han sido contrasta-dos en terminos de complejidad en cuanto al numero de componentes y al control.
La tabla 3.1 resume el numero de elemento que compone cada topologıa. Latopologıa buck+boost presenta menor numero de transistores de potencia y de diodos.En terminos de complejidad en el control el rectificador tipo buck es mas sencillogracias a que es un convertidor fuente de corriente y por ende no requiere un lazointerno de corriente.
Otro beneficio importante de la topologıa buck+boost es que tiene una encendidodirecto sin necesidad de circuitos de precarga de los condensadores del DC-link.
En ambas topologıas se puede limitar la corriente en caso de que exista un cor-tocircuito en la salida, ademas ambas topologıas trabajan con factor de potenciaunitario incluso cuando la red esta altamente desbalanceada.
Tabla 3.1. Numero de componentes necesarios para la realizacion de las topologıasbuck+boost y boost+buck respectivamente.
Elemento Buck+Boost Boost+Buck
Transistores de Potencia 4 5
Diodos 13 20
Condensadores de almacenamiento 1 3
Inductores de almacenamiento 2 2
Sensores de Voltaje 3 4
Sensores de Corriente 1 3
3.5.4. Resumen de la comparacion
En terminos de eficiencia ambas topologıas tienen desempenos equivalentes. Ladensidad de potencia de la topologıa buck+boost es el doble de la topologıa boost+buckcuando no se incluye el filtro EMI, ademas la topologıa buck+boost es menos com-pleja en terminos del numero de semiconductores y elementos pasivos, y no necesitaningun circuito de precarga. Por todas estas razones la topologıa mas conveniente ennuestra aplicacion es el rectificador trifasico tipo buck.
3.6 Rectificador Trifasico tipo Buck de Alta Eficiencia. 33
3.6. Rectificador Trifasico tipo Buck de Alta Eficiencia.
Hasta el momento solo se ha considerado el rectificador unidireccional tipo buckcon tres switches controlados, sin embargo existe otra version con seis switches con-trolados de mayor eficiencia, la figura 3.6 muestra ambas topologıas.
La mayor ventaja del la topologıa con tres switches en contraste con la de seisswitches es que con un solo switch se puede obtener corriente de entrada positiva,negativa y cero dependiendo de la tension de entrada del rectificador. En la versionde seis switches, segun la tension de entrada se debe conmutar el switch de la ramapositiva o negativa lo que en principio complica un poco el control, sin embargo enesta ultima topologıa la corriente circula en la ida por el MOSFET y un diodo, y enla vuelta por otro MOSFET y diodo; por otro lado, en la version con tres MOSFETsla corriente circula en la ida por el MOSFET y dos diodos, al igual que en el retornode la corriente. Por lo tanto la topologıa de seis switches tiene menores perdidas porconduccion lo que la hace mas eficiente.
Para la optimizacion a nivel topologıa se han considerado ambas topologıas, sin
Figura 3.6. (a) Rectificador trifasico unidireccional de tres switches. (b) Rectificadortrifasico unidireccional tipo buck de alta eficiencia.
34 Capıtulo 3 – Seleccion y optimizacion de la topologıa del sistema
embargo, en cualquier caso la topologıa entre ambas es la mas eficiente.
3.7. Aislamiento Electrico.
Dentro de las especificaciones existe la necesidad de aislamiento galvanico entre laentrada (AC) y la salida del rectificador (DC), para cumplir este requisito es necesariointroducir un transformador; este transformador puede ser situado en la entrada delrectificador, o bien en el convertidor de continua a continua. La figura 3.7 ilustracomo serian las topologıas con los transformadores.
Al posicionar el transformador en la entrada del rectificador 3.7(a), el transfor-mador no solo permite la aislacion sino que tambien eleva tension de entrada alrectificador lo suficiente para prescindir el convertidor DC-DC elevador. A pesar dela simplicidad de esta topologıa tiene un gran inconveniente el cual es el peso y volu-men del transformador ya que este es de baja frecuencia, por lo que sin mayor analisisse desestima esta opcion.
Una segunda opcion es incluir el transformador en convertidor DC-DC, en la lit-eratura existen varios tipos de convertidores con transformador como por ejemploel Full brigde, Half-brigde, Push-Pull, Flyback, Forward etc. De acuerdo al nivel depotencia de la aplicacion (10kW) de entre todos estos convertidores se ha decididoutilizar el Full brigde. La topologıa del sistema completo con el rectificador y conver-tidor DC-DC se muestra en la figura 3.7(b); en esta configuracion el transformador esde alta frecuencia, del orden de los cientos de kHz por lo que este es considerablemente
EMIFilter
400Hz
Isolation
60-180kHz
EMIFilter
Isolation
Figura 3.7. (a) Rectificador trifasico tipo buck de alta eficiencia con transformador a400Hz. (b) Rectificador trifasico tipo buck de alta eficiencia con transformador en elconvertidor DCDC.
3.8 Topologıa multi-celda 35
menor en peso y volumen que su equivalente de baja frecuencia.
3.8. Topologıa multi-celda
Incluyendo el aislamiento se abre la posibilidad dividir el convertidor de 10kW endos o mas convertidores en paralelo, de esta forma se obtienen varios beneficios comopor ejemplo, se aumenta la fiabilidad del sistema debido a que si algun MOSFET odiodo falla, se desactiva la correspondiente celda, el sistema entra en modo degradado,no puede entregar la potencia nominal pero sı una fraccion importante de ella.
Al poner rectificadores en paralelo el filtro de entrada se ve altamente beneficiado,ya que los niveles de corriente de entrada se aumentan en dos por cada rectificadoradicional, es decir con dos rectificadores en paralelo la corriente de entrada obtenidatiene 5 niveles y con tres en paralelo se obtienen 7 niveles de corriente. Los multi-ples niveles se obtienen la entrelazar las corrientes de entradas de los convertidoresdesfasando las senales portadoras en la generacion de las PWM.
En la figura 3.8(a) se muestra la corriente punzante de tres niveles obtenida conun rectificador, en 3.8(b) se muestra el espectro de esta corriente donde se puededestacar que el primer armonico corresponde a la frecuencia de conmutacion, en estecaso 60kHz, la magnitud de este armonica es 178dbµV.
En la figura 3.9(a) se muestra la corriente punzante de cinco niveles obtenidacon dos rectificador, en 3.9(b) se muestra el espectro de esta corriente donde se puededestacar que el primer armonico corresponde a dos veces la frecuencia de conmutacion120kHz, la magnitud de este armonica continua siendo 178dbµV. Sin embargo este esel mejor escenario en el cual ambos rectificadores se reparten exactamente la mismacantidad de carga, lo cual puede ser muy difıcil de lograr, para analizar un esce-nario mas realista se ha simulado el sistema incorporado un desbalance entre ambosrectificadores.
La figura 3.10 muestra la corriente y su espectro cuando los rectificadores estandesbalanceados, en el espectro aparece una componente en 60kHz la cual puede afectaro no en el diseno del filtro de entrada dependiendo de la magnitud del desbalance decarga.
36 Capıtulo 3 – Seleccion y optimizacion de la topologıa del sistema
0 0.5 1 1.5 2 2.5
x 10−3
−60
−40
−20
0
20
40
60
104
105
106
107
40
60
80
100
120
140
160
180
frequency Hz
dB µ
V
Current spectrumMIL−STD−461ECISPR 22 Class ACISPR 22 Class B
Figura 3.8. (a)Forma de onda de la corriente conmutada de tres niveles correspondientea un rectificador trifasico de tres niveles.(b) Espectro de la corriente de tres nivelesconmutada.
3.8 Topologıa multi-celda 37
0 0.5 1 1.5 2 2.5
x 10−3
−60
−40
−20
0
20
40
60
104
105
106
107
40
60
80
100
120
140
160
180
frequency Hz
dB µ
V
Current spectrumMIL−STD−461ECISPR 22 Class ACISPR 22 Class B
Figura 3.9. (a)Forma de onda de la corriente conmutada de cinco niveles correspon-diente a dos rectificadores trifasicos de tres niveles en paralelo.(b) Espectro de lacorriente de cinco niveles conmutada.
38 Capıtulo 3 – Seleccion y optimizacion de la topologıa del sistema
0 0.5 1 1.5 2 2.5
x 10−3
−60
−40
−20
0
20
40
60
104
105
106
107
40
60
80
100
120
140
160
180
frequency Hz
dB µ
V
Current spectrumMIL−STD−461ECISPR 22 Class ACISPR 22 Class B
Figura 3.10. (a)Forma de onda de la corriente conmutada de cinco niveles correspon-diente a dos rectificadores trifasicos de tres niveles en paralelo desbalanceados.(b)Espectro de la corriente de cinco niveles conmutada cuando los rectificadores estandesbalanceados.
3.9 Optimizacion de la topologıa en cuanto numero de celdas 39
3.9. Optimizacion de la topologıa en cuanto numero de cel-das
Como se menciono en la seccion anterior, al dividir el convertidor en varios conver-tidores mas pequenos trabajando en paralelo se consiguen varias ventajas, las cualesdeben ser cuantificadas para encontrar la configuracion optima.
Para llevar a cabo la cuantificacion se ha estimado el filtro de entrada (EMI Filter),rectificador (Rectifier) en dos versiones la con tres y con seis transistores (3 SW y 6SW respectivamente), Full-Bridge en dos versiones, la alimentada en corriente (CF)y la alimentada en tension (VF). La tabla 3.2 un resumen de las estimaciones parauna celda, dos celda (dos convertidores en paralelo) y tres celdas (tres convertidoresen paralelo).
El rectificador de seis transistores presenta siempre una mejor eficiencia que suversion de tres transistores.
Con una configuracion con una celda los convertidores DC-DC se ven muy afec-tados en peso, por lo que se desestima una celda. Al pasar a dos celdas el pose de loscomponentes disminuye considerablemente, de hecho el DC-DC VF disminuye en un70 %, ademas las perdidas en el rectificador de 6 SW disminuyen en un 15 %, en elresto de los elementos las perdidas se mantienen sin mayor variacion.
Al pasar de dos a tres celdas no hay mayor variacion en perdidas y peso en todoslos componentes, salvo en el filtro ya que disminuye sus perdidas en un 40 %, encuanto al numero de MOSFETs y diodos en el rectificador estos disminuyen ya quecon dos celdas es necesario utilizar al menos dos MOSFTEs en paralelo para disminuirla perdidas y evitar que la temperatura de este se eleve por sobre la establecida enestandares aeronauticos (lımite esta en 112C con una temperatura ambiente de 70C).Por lo tato la fiabilidad de la topologıa con 3 celdas es mayor que la con dos celdas.Ademas la topologıa con 3 celdas tiene una ventaja adicional en cuanto a la fiabilidad,si una celda falla el sistema puede seguir entregando el 66 % de la potencia nominalcon lo que se puede cubrir gran parte del perfil de carga 1.4, a diferencia de la topologıade dos celdas en la solo quedarıa disponible el 50 % del la potencia nominal.
Por lo tanto la topologıa escogida es una de tres celdas, con un filtro de entradade dos etapas, los rectificadores son tipo buck de 6 transistores y los convertidoresde continua a continua full-bridge alimentado en corriente, la figura 3.11 muestra latopologıa completa (sin el filtro de entrada).
40 Capıtulo 3 – Seleccion y optimizacion de la topologıa del sistema
Tabla 3.2. Resumen de caracterısticas de la topologıa a implementar.Cell Element Type Weight Losses MOSFETs Diodes
EMI Filter 3 Stages 4.3kg 79W - -Rectifer 3 SW 1.5kg 545W 18 26
1 Cell Rectifer 6 SW 1.54kg 470W 18 14DC-DC VF 5.3kg 241W 24 12DC-DC CF 5.3kg 190W 21 12
EMI Filter 3 Stages 2.5kg 65W - -Rectifer 3 SW 1.0kg 557W 18 26
2 Cell Rectifer 6 SW 1.0kg 404W 24 14DC-DC VF 1.6kg 252W 32 12DC-DC CF 4.1kg 185W 26 12
EMI Filter 2 Stages 2.9kg 38W - -Rectifer 3 SW 1.0kg 521W 18 39
3 Cell Rectifer 6 SW 1.04kg 400W 18 21DC-DC VF 1.4kg 256W 36 12DC-DC CF 2.5kg 203W 27 12
3.10. Simulacion del rectificador de tres celdas
Para la simulacion del convertidor se ha utilizado el software PSIM, este es unsoftware de circuitos electricos especializado en electronica de potencia. Las princi-pales ventajas son su simplicidad y versatilidad. El rectificador tipo buck tiene unamodulacion particular, con PSIM esta modulacion se puede realizar mediante leguajeen C, como se realiza en un procesador digital de senales, esto facilita enormementeel traslado del codigo desde el simulador hacia la plataforma de pruebas.
Para la simulacion se consideraron tres rectificadores, cada uno de ellos trifasicos,en paralelo como muestra la 3.11. La figura 3.12(a) muestra la corriente en una delos inductores de continua, en ella se ve la forma diente de sierra de la corriente yuna componente de baja frecuencia en la envolvente de la corriente. Esta componentede baja frecuencia es generada por la tension en el diodo de libre circulacion, comose aprecia en la figura 3.12(b) cuando el diodo no conduce el diodo debe soportarla maxima tension instantanea la que desde luego contiene una componente de bajafrecuencia.
La figura 3.12(c) muestra las senales de gobierno de los MOSFETs, en ella solose muestran tres senales para los seis MOSFETs, esto se debe a que solo se necesitanestas tres senales desde el DSP, con simple hardware cada una de estas senales detransforman en dos dependiendo del signo de la tension de cada rama. Como se puedever en esta figura en cada periodo de conmutacion solo dos MOSFETs conmutan, elrestante permanece encendido durante todo el periodo.
La figura 3.13(a) muestra la forma de onda la corriente de red y su correspondientetension de red, como se aprecia la corriente es sinusoidal, y el factor de potencia espracticamente unitario. La figura 3.13(b) se muestra la corriente conmutada antes delfiltro, como se ve esta tiene 7 niveles de corriente, estos niveles son generados graciasal entrelazado de corriente en los tres rectificadores en paralelo, cada uno de estos
3.10 Simulacion del rectificador de tres celdas 41
Figura 3.11. Topologıa complete propuesta, la cual incluye tres rectificadores trifasicostipo buck de seis MOSFTES mas tres Full-bridge alimentados en tension
rectificadores tiene una corriente de tres niveles como se ve en la figura 3.13(c), enesta figura se puede ver tambien el entrelazado en las corrientes al desfasar 2π/3 unasde otras.
42 Capıtulo 3 – Seleccion y optimizacion de la topologıa del sistema
Figura 3.12. Resultados de simulacion. (a) Corriente en uno de los inductores. (b)Tension en el diodo de libre circulacion. (c) Senales de gobierno de los MOSFETs
3.10 Simulacion del rectificador de tres celdas 43
Figura 3.13. Resultados de simulacion. (a) Corriente y tension de linea. (b) Corri-ente conmutada de 7 niveles demandada por el rectificador sin filtrar. (c) Corrienteconmutada de 3 niveles en las tres celdas con entrelazado entre ellas
44 Capıtulo 3 – Seleccion y optimizacion de la topologıa del sistema
Capıtulo 4
ANALISIS DEL RECTIFICADORTIPO BUCK
4.1. Rectificador trifasico tipo buck
Este rectificador fue inventado en el 2000 [17], y desde su comienzo ha sido muyutilizando en publicaciones tanto en congresos como revistas. Entendiendo que dentrola topologıa seleccionada la parte mas interesante es el rectificador se realizara unanalisis de esta topologıa
4.2. Principio de operacion
4.2.1. Suposiciones
Para el siguiente analisis se tomaran en cuenta las siguientes suposiciones
Las tensiones en los condensadores de entrada son sinusoides puras, es decir:
vR,N = VNcos(ωN t),
vS,N = VNcos(ωN t− 2π/3),
vT,N = VNcos(ωN t− 4π/3),
Estas tres ecuaciones pueden ser descritas en una sola ecuacion vectorial
~vN = Vn exp(jφN)
donde φN = ωN t y ωN es la frecuencia angular de la red.
La corriente en el inductor y la tension en el condensador de salida son con-stantes
4.3. Estados de Conduccion.
Para el estudio de los estados de conduccion, en un periodo de la frecuenciafundamental se divide en 12 sectores los cuales son definidos de la siguiente forma:
45
46 Capıtulo 4 – Analisis del Rectificador tipo Buck
Figura 4.1. Topologıas de rectificadores tipo boost y tipo buck
Sector 1 : vR > 0 > vS > vT Sector 7 : vT > vS > 0 > vR
Sector 2 : vR > vS > 0 > vT Sector 8 : v,T > 0 > vS > vR
Sector 3 : vS > vR > 0 > vT Sector 9 : vT > 0 > vR > vS
Sector 4 : vS > 0 > vR > vT Sector 10 : vT > vR > 0 > vS
Sector 5 : vS > 0 > vT > vR Sector 11 : vR > vT > 0 > vS
Sector 6 : vS > uT > 0 > uR Sector 12 : vR > 0 > vT > vS
Para lograr un mejor entendimiento de los sectores en la figura 4.1 se muestravisualmente cada uno de estos sectores en fucion de las tensiones de entrada.
Dada la simetrıa del sistema se procedera al analisis exhaustivo de todos los esta-dos de conmutacion para el sector 1. Para los demas sectores el analisis es equivalente.
El siguiente analisis se analizan todos los posibles estados de conmutacion delrectificador, para la formacion del espacio vectorial deseado de la corriente de entrada,de la que se derivan finalmente los ciclos de trabajo r. Para ilustrar las relacionesde una tension de entrada se considera la siguiente situacion vR > 0 > vS > vT(sector 1, ver figura 4.2). Debido a la simetrıa de las tensiones de entrada y la mismaestructura de las tres fases de las condiciones de entrada del convertidor analogicoprevalecera para los demas sectores. Para mayor claridad, se discutiran aquı solo enel sector 1 ya que para los demas sectores las conclusiones son equivalentes
En principio, existen 23 = 8 posibles estados de conmutacion, cada uno de estosestados tiene asociado un circuito electrico diferente en el cual la distribucion decorrientes dependera de la tension de la red.
Con el estado sj = (000), es decir, todos los transistores apagados, todas lascorrientes de entrada son iguales a cero, la corriente continua circula por el diodo de
4.3 Estados de Conduccion. 47
Figura 4.2. Recopilacion de todos los estados de conduccion posibles en el rectificadortipo buck.
48 Capıtulo 4 – Analisis del Rectificador tipo Buck
libre circulacion (ver la figura 4.2(a)). El vector de corriente de entrada es:
irec,(000) = 0
Ademas del estado irec,(000) = 0, tambien para los estados sj = (100) = (010) =(001) la corriente de entrada tambien es cero, debido a que al estar solo un transistorencendido, la corriente de entrada tiene camino de ida o bien de regreso pero noambos, por lo tanto con existe transferencia de energıa desde la red alterna hacialas cargas continuas. Como en el estado anterior el diodo de circulacion permite laconduccion de corriente continua (ver Figura 4.2 (b) - (d)). El vector de corriente deentrada es:
irec,(100) = irec,(010) = irec,(001) = 0
Para el resto de los estados al estar dos o los tres transistores estan cerrados, siexiste un circuito electrico cerrado donde puede circular corriente. Cuando el inter-ruptor de la fase R y S estan encendidos sj = (110), la corriente circula por la fase Ry regresa por la fase S, debido a que VR > VS .
Con la definicion del espacio vectorial
irec =2
3(irec,R + ej2π/3 · irec,S + ej4π/3 · irec,T )
considerando que para el estado sj = (110), la corriente de entrada es irec,R = I yirec,S = −I se obtiene el siguiente vector
irec,(110) = I · 2√3e−jπ/6
Analogamente, para sj = (011), es decir, con los transistores de las fases S yT encendidos (vease la figura 4.2(f)) el espacio vectorial de corriente obtenido es elsiguiente
irec,(011) = I · 2√3j,
debido a la vS > vT crea un flujo de corriente de fase en fase S y T , una tension delcircuito intermedio v = vS − cT .
Los dos ultimos estados de conmutacion sj = (101) = (111) generan el mismovector de corriente de entrada (vease la Figura 4.2(g) - (h)), esto debido a que vS > vTpot lo tonto el diodo Ds,− queda polarizado inversamente, por lo tanto aunque inclusoel transistor de la fase S este cerrado, no circula corriente atreves de el. El espaciovectorial obtenido con ambos estados de conduccion es el siguiente
irec(101) = irec(111) = I · 2√3ejπ/6
Por lo tanto, con los ocho posibles estados de conmutacion se analizados se con-cluye que:
4.4 Modulacion Vectorial 49
Figura 4.3. Representacion de los vector espaciales disponibles en el sector 1 y laformacion del espacio vectorial de corriente deseado i∗rec.
Con cuatro estados de conmutacion (sj = (000)(100)(010), (001)) la corrientede entrada es cero y la corriente en el inductor de continua circula atreves deldiodo de libre circulacion.
Para los otros cuatro estados de conmutacion (sj = (110), (011), (101), (111))hay un flujo de corriente entre el lado de corriente alterna y el lado corrientecontinua, se suministra energıa hacia el lado de corriente continua. Por lo tanto,estos cuatro estados en adelante, los estados activos.
4.4. Modulacion Vectorial
El espacio vectorial de las tensiones de los condensadores de filtro vC , como semenciono antes, en una primera aproximacion, el voltaje de entrada del espacio devector con el angulo de fase φn = ωn ·t y una amplitud igual a la amplitud de la UN , lafase de tensiones de red equivalente a vC ≈ vN (2,3). Ahora, la corriente de red deseadaen el espacio vectorial es i∗N . Si, como se menciono anteriormente, el cambio de fasedel filtro de entrada se desprecia, entonces el componente de frecuencia fundamentaldel vector de entrada actual espacio i∗rec, (1) es exactamente igual a la actual red deespacio vectorial i∗N . El angulo de fase actual y el φC1 cantidad, es decir, la amplitudde la fase de entrada del rectificador de corriente de I∗rec, (1) o red de la actual fasefundamental I∗N , (1) se establecen de forma directa utilizando el vector de espaciodisponible.
50 Capıtulo 4 – Analisis del Rectificador tipo Buck
En la figura 4.3, se muestra una representacion grafica del espacio de los estadosdisponibles en el sector 1. El vector de corriente deseado se construye mediante losvectores de corriente adyacentes, Irec(110) y el Irec,(101) = Irec,(111). Con un tercervector no activo (que genere una corriente cero en la entrada) se regula la magnitudde la corriente deseada ajustando del ındice de modulacion (M = 0 . . . 1). Ademasde ajustar la magnitud de la corriente de entrada, el ındice de modulacion controlatension continua de salida del rectificador, en nuestra aplicacion en particular al estarconectada la salida del rectificador con un convertidor continua a continua, el ındicede modulacion se dejara fijo y el control de tension de salida se realizara con elconvertidor DC-DC.
El estado activo de conduccion irec,(011) no es utilizado en el sector 1, ya que alno ser adyacente a la corriente de referencia, no es capaz de generar el vector dereferencia.
Resumiendo los vectores a utilizar en la modulacion en el sector 1 son iref,(101),iref,(110), iref,(111). Existen en la literatura diversos metodos de modulacion dependi-endo el orden que se modulan estos vectores. Dependiendo de la modulacion se puedeobtener menor perdidas de conmutacion o menor rizado en la corriente o bien entension en el inductor o condensadores respectivamente [18–20].
En este trabajo se decidio utilizar la modulacion que minimiza las perdidas porconmutacion. En la figura 4.4 se muestra los estados de conmutacion, la corriente en elinductor y tension en el diodo de libre circulacion para esta modulacion en el sector 1.Esta modulacion es simetrica respecto a TP /2 donde TP es el periodo de conmutacion.Para esta modulacion se inicia con los tres transistores encendidos (111), la corrienteen el inductor aumenta y la tension en el diodo de libre circulacion corresponde a lamaxima tension instantanea entre lınea, en el sector 1 es vRT . Luego el transistor en lafase T se abre, la corriente continua subiendo pero con una pendiente menor debido aque la tension en el diodo disminuye a la segunda tension instantanea mas alta entrelıneas (vRS). Luego se abre el transistor de la fase R, de esta manera el diodo delibre circulacion comienza a conducir, su tension se hace cero (V F ) y la corriente elinductor disminuye. Cabe destacar que el transistor en la fase S permanece encendidodurante todo el periodo, es por esta razon que con esta modulacion se minimizan lasperdidas por conmutacion
4.4 Modulacion Vectorial 51
Figura 4.4. Representacion de los vector espaciales disponibles en el sector 1 y laformacion del espacio vectorial de corriente deseado i∗rec.
52 Capıtulo 4 – Analisis del Rectificador tipo Buck
Capıtulo 5
NUEVAS CONSIDERACIONESDE DISENO DEL FILTRO EMI DE
RECTIFICADOR TRIFASICOPARA APLICACIONES
AERONAUTICAS
Los filtros EMI corresponden a una gran parte del sistema en terminos de volumeny peso, la figura . Es evidente que en aplicaciones aeronauticas ambas variables tienenuna gran importancia sobre todo el peso, es por esto que es necesario realizar unproceso de optimizacion del filtro para minimizar el peso y volumen.
Figura 5.1. Ejemplos de rectificadores trifasicos tipo buck con sus respectivos filtrosEMI de modo diferencial [21–24]
53
54Capıtulo 5 – Nuevas consideraciones de Diseno del Filtro EMI de Rectificador Trifasico para Aplicaciones
Aeronauticas
5.1. Filtro EMI
El filtro de entrada en un convertidor conmutado tiene tres funciones [23]:
Asegurar una forma de onda sinusoidal en las corrientes de entrada filtrando losarmonicos generados por las conmutaciones.
Disminuir la interferencia electromagnetica generada hacia otros dispositivoscercanos al convertidor.
Evitar susceptibilidad electromagnetica provocada por equipos alrededor y lade sı mismo.
Sobre este ultimo punto, organizaciones internacionales han definido estandares,los cuales han sido considerados para el diseno de los filtros de compatibilidad electro-magnetica (EMC) en sistemas de electronica de potencia. Sin embargo, al insertar unfiltro de entrada, este influye en la funcionalidad, estabilidad y tamano del sistemacompleto. Por lo tanto, para el diseno del filtro se debe tomar en consideracion losiguiente:
Cumplir con los estandares internacionales de compatibilidad electromagnetica.
Limitaciones fısicas en tamano versus energıa de almacenamiento de los com-ponentes del filtro.
Suficiente amortiguamiento pasivo para evitar oscilaciones cuando el convertidoresta operando en vacio.
Mınimas perdidas en los componentes resistivos del amortiguamiento del filtro.
Obtencion del mayor factor de potencia.
Evitar que la resonancia del filtro este a un multiplo de la frecuencia de con-mutacion.
Minimizacion de la impedancia de salida del filtro para asegurar estabilidad enel sistema y minimizar las restricciones en el diseno de control.
Minimizacion del costo del filtro.
Desde luego, los requisitos son parcialmente contradictorios y, por lo tanto, nose puede cumplir todos al mismo tiempo. Ademas, se debe considerar los siguientesaspectos que dificultan el diseno del filtro
El desconocimiento de la impedancia de lınea desplaza la frecuencia de resonan-cia o introduce nuevos circuitos resonantes con bajo amortiguamiento.
Diferentes topologıas del filtro pueden ser realizados para cumplir con los req-uisitos en atenuacion.
Un modelo en frecuencia y en tiempo del receptor de prueba de EMC es nece-sario para predecir el cuasi-picos de la corriente de entrada.
Disponibilidad de valores discretos de capacitancia y de baja tolerancia complicael procedimiento de optimizacion del filtro.
El filtro tiene influencia sobre la estabilidad del sistema completo.
5.2 Topologıa del Convertidor 55
Todas estos requisitos y dificultares deben ser abordados en el diseno final de unfiltro EMI.
En el presente capıtulo se expondra un nuevo metodo para el calculo de las ca-pacitancias e inductancia del filtro EMI, con el objetivo de cumplir los estandaresEMI en avionica y alcanzar un factor de potencia unitario. Ademas se muestra unaoptimizacion del rectificador y el filtro EMI en conjunto para lograr un buen balanceentre perdidas y peso del convertidor.
5.2. Topologıa del Convertidor
El filtro EMI es disenado para un rectificador buck trifasico, fuente de corriente,(figura 5.2) . Sus principales caracterısticas son las siguientes:
Corrientes de entrada sinusoidales
Encendido directo sin necesidad de un circuito de precarga.
Proteccion de sobre corriente cuando la salida es cortocircuitada.
Alta densidad de potencia.
Gracias a estas cualidades este convertidor es un gran candidato para ser imple-mentado en aplicaciones aeronauticas. Por lo tanto es requisito disenar un filtro paraque el rectificador cumpla los requisitos establecidos para este tipo de aplicaciones.
Three-phase buck-type PWM Rectifier
EMC input Filter
Figura 5.2. Topologıa del Rectificador Buck trifasico
56Capıtulo 5 – Nuevas consideraciones de Diseno del Filtro EMI de Rectificador Trifasico para Aplicaciones
Aeronauticas
Figura 5.3. lımites de armonicos segun MIL-STD 461E, CISPR A, CISPR B
5.3. Estandar EMI en avionica: MIL-STD 461E
El objetivo final construir un rectificador para ser utilizado que permita cumplircon los estandares para aplicaciones aeronauticas, para lograr este objetivo el disenodel filtro de entrada toma gran relevancia, dado que los equipos electricos para avionestienen altos requisitos en el campo de la compatibilidad electromagnetica.
Principalmente existen dos estandares que regulan el ruido diferencial de unequipo, la CISPR y la MIL-STD. La CISPR es dirigida a equipos industriales deuso comercial, y la MIL-STD esta dirigida a equipos militares. En la figura 5.3 semuestran los lımites para los armonicos de ambos estandares.
En el mundo industrial el estandar a cumplir es el CISPR, este estandar ponelımites a partir de los 150Khz, por esta razon, comunmente se utiliza una frecuenciade conmutacion bajo este rango por ejemplo 28kHz, de este modo el primer, segundo,tercero, cuarto y quinto armonicos (140kHz) quedan fuera del rango del estandar, deesta forma el primer armonico a considerar en el diseno del filtro es el sexto armonico(168kHz) como se puede ver en 5.4.
Sin embargo, si el equipo esta regulado bajo la MIL-STD 461E, que es el casopara equipos en avionica, todos los armonicos (desde los 10kHz hasta 10Mhz) quedandentro de rango comprendido de la MIL-STD 461E, ası el filtro EMI debe ser disenadopara atenuar el primer armonico de la frecuencia de conmutacion.
El estandar en aplicaciones aeronauticas comienza en 10Khz, ¿Es posible utilizaruna frecuencia de conmutacion inferior a 10kHz para que el filtro EMI se vea beneficia-do?. La respuesta es no, ya que como muestra la figura 5.5 en aplicaciones industrialesla frecuencia de red es 50 o 60Hz y el estandar comienza en 150kHz (3000 veces lafrecuencia de red), considerando que la frecuencia de conmutacion debe ser muchomayor que la de red para no generar exceso distorsion e inferior al estandar pararelajar el diseno del filtro, por lo que hay espacio suficiente para colocar la frecuenciaconmutacion, por ejemplo en 28Khz (560 veces la frecuencia de red). Por otro ladoen aplicaciones aeronauticas la frecuencia de la red es 400Hz y el estandar comienzaen 10kHz, solo 25 veces la frecuencia de red, por lo tanto no se puede seleccionar unafrecuencia mucho mayor que la de la red e inferior a la del estandar.
5.3 Estandar EMI en avionica: MIL-STD 461E 57
104 105 106 10720
40
60
80
100
120
140
frequency Hz
dBµ
V
Current spectrumCISPR 22 Class ACISPR 22 Class B
Critical Harmonic
Figura 5.4. Espectro de la corriente de entrada de un rectificador trifasco tipo buckcontratado con los lımites de los estandares industriales
Grid FrequencySwitching frequency
Standard frequency starts here
Indsutrial Applications
Standard frequency starts here
Grid Frequency
Aircraft Applications
Figura 5.5. Comparacion entre frecuencia de red y frecuencia del estandar a cumpliren aplicaciones industriales y aeronauticas
58Capıtulo 5 – Nuevas consideraciones de Diseno del Filtro EMI de Rectificador Trifasico para Aplicaciones
Aeronauticas
Figura 5.6. Espectro de la corriente en la red con el convertidor conmutando a 25kHz
La gran diferencia entre el rango de frecuencias de la MIL-STD 461E y CISPRconlleva a disenar el filtro con un nuevo criterio en el que se estudie cuanto es el efectoen perdidas y peso de la frecuencia de conmutacion del filtro, para ası seleccionar laque tenga un optimo balance entre perdidas y peso
Para esto se han estimado las perdidas en el convertidor y peso del filtro EMI auna serie de frecuencias de conmutacion 25kHz, 40 kHz, 60kHz, 80kHz y 100 kHz
5.4. Diseno del Filtro EMI
5.4.1. Espectro de la corriente de entrada del rectificador
El rectificador de 10kW el cual esta dividido en 3 convertidores 3.33KW, estosconvertidores son fuente de corriente de tres niveles, Para disenar el filtro se ha simu-lado el rectificador utilizando el software de simulacion PSIMr. Los datos obtenidoshan sido procesados en MATLABrpara obtener el espectro de la corriente de entra-da en unidades de dBµV. En la figura 5.6 se muestra la respuesta en frecuencia delvoltaje en dBµV conmutando el convertidor a 60kHz.
A 60Khz el estandar indica que el armonico debe ser inferior a 86dBµV cuando latension de linea es 115V. Sı ademas consideramos un margen de 6 dB, las atenuacionesrequeridas para las frecuencias de conmutacion seleccionadas, se muestran en la tabla5.1
5.4 Diseno del Filtro EMI 59
Tabla 5.1. Atenuacion requerida para cumplir con el estandar aeronautico.fsw Lım MIL-STD Atenuacion requerida
25kHz 86µVdB 168,4− 86 + 6 = 76,4µVdB
40kHz 82µVdB 168,4− 82 + 6 = 80,4µVdB
60kHz 78µVdB 168,4− 78 + 6 = 84,4µVdB
80kHz 76µVdB 168,4− 76 + 6 = 86,4µVdB
100kHz 74µVdB 168,4− 74 + 6 = 88,4µVdB
Figura 5.7. Respuesta en frecuencia de un filtro LC pasa bajos. Donde n es el numerode etapas del filtro
5.4.2. Frecuencia de corte del filtro
La figura 5.7 muestra la relacion entre la frecuencia de corte del filtro, la frecuenciade conmutacion y la atenuacion. La frecuencia de corte del filtro en funcion de laatenuacion requerida a la frecuencia de conmutacion es la siguiente:
ωcutoff =1√L · C
=2π · fs√
10Att[dB]/(20n)(5.4.1)
L · C =1096.4[dB]/(20n)
(2π · fskHz)2, (5.4.2)
Donde n es el numero de etapas del filtro. Con esta ecuacion se obtiene unarelacion para el producto LC, la otra ecuacion que se necesita para disenar el filtrose obtiene de acuerdo con el factor de potencia.
5.4.3. Consideraciones para el condensador
En [17] y [25] el condensador de entrada es disenado para limitar la potenciareactiva. La ecuacion (5.4.3) entrega un valor maximo para el condensador C comofuncion de la potencia reactiva (en porcentaje de la potencia nominal PN ). Usualmente
60Capıtulo 5 – Nuevas consideraciones de Diseno del Filtro EMI de Rectificador Trifasico para Aplicaciones
Aeronauticas
esta potencia es limitada del 5 % al 10 % de PN de forma de asegurar un alto factorde potencia
Figura 5.8. Filtro LC pasa bajos. Donde n es el numero de etapas del filtro
C ≤ (0,05..,0,1) · PNω · U2
N,l−l,rms=
(0,05..,0,1) · 10kW
2π · 400Hz(√
3 · 115V)2= 4,96..,9,92µF. (5.4.3)
Ası, un buen valor para C es 4.7µF ya que es menor que 4,96µF, ecuacion. (5.4.3).Al fijar el condensador se obtiene el valor de L segun la ecuacion de la frecuencia decorte (5.4.2); de esta manera L es 3.7mH utilizando un filtro de dos etapas.
La figura 5.8 muestra el circuito equivalente del sistema visto desde la red. Laexpresion analıtica de la impedancia es presentada en la ecuaciones (5.4.4) y (5.4.5).
Zeq1stg = ωL+1
ωC + 1/R(5.4.4)
Zeq2stg = ωL+1
ωC + 1/Zeq1stg, (5.4.5)
Donde Zeq1stg and Zeq2stg son las impedancias equivalente para un filtro de unay dos etapas respectivamente. La resistencia correspondiente para una potencia desalida de 3,3kW a 115V, es R = 11,9Ω.
Con este filtro, el factor de potencia a potencia nominal es solo 0.66( cos(∠Zeq2stg)), pero de acuerdo con ecuacion 5.4.3 deberıa ser mayor que 0.99. El factor de potenciano corresponde con las consideraciones de diseno porque la ecuacion 5.4.3 no incluyeel efecto del inductor el cual es despreciado a 50 o 60Hz. En aplicaciones aeronauticasla frecuencia de red es 400Hz , a esta frecuencia el efecto inductivo no puede serdespreciado.
5.4.4. Nuevas consideraciones para el del diseno del condensador
Para saber la influencia del filtro L-C en el factor de potencia, la parte real eimaginaria del filtro son mostrados separadamente en las ecuacion. (5.4.6) y (5.4.7),
5.4 Diseno del Filtro EMI 61
<Zeq1stg =R
1 + (ω · C ·R)2(5.4.6)
=Zeq1stg = ωL− ωC
R−2 + ω2C2. (5.4.7)
El factor de potencia es unitario cuando =Zeq1stg = 0. Con esta condicion y laecuacion de la frecuencia de corte del filro (5.4.1), el filtro queda determinado con lassiguientes ecuacion (5.4.8) y (5.4.9)
C =1
R√ω2cutoff − ω2
(5.4.8)
L =1
ω2cutoffC
. (5.4.9)
Ası el factor de potencia del rectificador es unitario utilizando un filtro de unaetapa. Ademas si (ω · C · R)2 << 1 (esto aplica cuando el condensador el del or-den de los µF), la impedancia equivalente un filtro de dos etapas es aproximada-mente R (Zeq1stg ≈ R, eq. (5.4.6) entonces Zeq2stg = Zeq1stg eq. (5.4.6), por ende=Zeq2stg = =Zeq1stg = 0. Por lo tanto, independiente del numero de etapas delfiltro, al disenarlo con las ecuaciones 5.4.8) y (5.4.9), el factor de potencia es unitarioa potencia nominal. Utilizando estas ecuaciones el condensador y el inductor del filtroqueda are C = 10µF y L = 1,4mH. En comparacion con el metodo clasico de diseno,el metodo propuesto es menos voluminoso y mas liviano debido a tener menor induc-tancia. La figura ?? muestra el espectro de corriente utilizando el filtro disenado. Elespectro de corriente comple con la MIL-STD-461E en todo el rango de frecuencias.
Como se menciono en el parrafo anterior, con el metodo propuesto se obtienefactor de potencia unitario solo a potencia nominal, sin embargo, en la vecindad dela potencia nominal el factor de potencia va a ser alto, esta vecindad esta definidapor la inecuacion (ω · C · R)2 << 1, es decir cuanto menor sea el valor de C mayorsera la vecindad o rango de potencia de carga donde el factor de potencia sea alto. Enla figura 5.10 se muestra como varia el factor de potencia para diferentes potenciasdemandadas, a baja potencia el factor de potencia decae asintoticamente a cero, estose debe a que a baja carga la red ve practicamente solo el filtro L − C. la figura5.10(a) tres disenos diferentes utilizando una configuracion en dos etapas, uno en elcual se obtiene factor de potencia unitario a plena carga, otro a media carga y otroa baja carga, de estos disenos el que tiene un mayor rango de alto factor de potenciaes el disenado a media carga, sin embargo, este diseno tiene mayor inductancia queel a plena carga, lo que se traduce en mayor peso, y como el objetivo es optimizar enpeso, para la optimizacion del filtro y rectificador en conjunto se utilizaran disenos aplena carga.
62Capıtulo 5 – Nuevas consideraciones de Diseno del Filtro EMI de Rectificador Trifasico para Aplicaciones
Aeronauticas
Figura 5.9. Espectro de la corriente de entrada del rectificador trifasico tipo buck conel filtro disenado
Tabla 5.2. Frecuencia de corte para el filtro EMI en funcion del numero de etapas yfrecuencias de conmutacion
Frecuencia de corte 25kHz 40kHz 60kHz 80kHz 100kHz
Una etapa 792Hz 1.1kHz 1.2Hz 1.4kHz 1.6kHz
Dos etapas 4.5kHz 6.3kHz 8.6Hz 10.7Hz 12.6Hz
Tres etapas 7.9Hz 11.7Hz 16.4Hz 20.9Hz 25.1Hz
Cuatro etapas 10.6kHz 15.9Hz 22.7Hz 29.2Hz 35.5Hz
Cinco etapas 12.5Hz 19.1Hz 27.6Hz 35.7Hz 43.7Hz
5.5. Optimizacion del Filtro
En [23] la frecuencia de conmutacion para esta topologıa seleccionada es 28kHz porque hasta el quinto armonico (140kHz) esta fuera del rango del estandar a cumplir(150kHz 30MHz). Sin embargo en aplicaciones aeronauticas este metodo de “ocul-tamiento” de armonicos no puede ser aplicado porque la MIL-STD-461E comienzaen 10kHz. La frecuencia de conmutacion sera determinada mediante un analisis deltrade-off del volumen/peso and perdidas por conmutacion, para esto el peso y el vol-umen y las perdidas seran estimadas para las frecuencias de interes. Ası obtener undiseno con un buen balance entre tamano y perdidas.
5.5.1. Estimacion de del peso y las perdidas del filtro
En [26] se demuestra que el mınimo volumen para un filtro de multiples etapasse logra colocando todas las etapas a la misma frecuencia de corte, mas aun, losinductores y capacitores deben ser los mismos tambien.
La tabla 5.2 muestra las frecuencias de corte del filtro requeridas para cumplircon MIL-STD-461E.
5.5 Optimizacion del Filtro 63
0 500 1000 1500 2000 2500 3000 35000.8
0.85
0.9
0.95
1
Power Factor vs Output Power for 2 stages filter @ 60kHz
Po
wer
Fac
tor
Power W
0 500 1000 1500 2000 2500 3000 35000.8
0.85
0.9
0.95
1
Power Factor vs Output Power for 3 stages filter @ 60kHz
Po
wer
Fac
tor
Power W
0 500 1000 1500 2000 2500 3000 35000.8
0.85
0.9
0.95
1
Power Factor vs Output Power for 4 stages filter @ 60kHz
Po
wer
Fac
tor
Power W
P = 0.33kWP = 1.65kWP = 3.3kW
Figura 5.10. factor de potencia en funcion de la carga demandada con un: (a) filtrode dos etapas, (b)filtro de tres etapas y (c) filtro de cuatro etapas
64Capıtulo 5 – Nuevas consideraciones de Diseno del Filtro EMI de Rectificador Trifasico para Aplicaciones
Aeronauticas
Los filtro de una a cinco etapas han sido disenados utilizando las ecuaciones(5.4.1), (5.4.8) y(5.4.9). Luego con los inductores y capacitores obtenidos, los pesos yvolumenes para todos los filtro son estimados.
La figura 5.11 (a) y (b). muestra los resultados de los pesos y las perdidas es-timadas. Los filtros de una etapa son excesivamente pesados en comparacion a losdemas filtros por lo que de desestiman. Los filtros de tres, cuatro y cinco etapas en-tregan mejoras marginales en terminos de peso con respecto al filtro de dos etapas.En cuanto a las perdidas un filtro de dos etapas entrega mayor eficiencia que unocon mayor numero de etapas. Por lo tanto para esta aplicacion el numero de etapasoptimo es dos. Para saber la frecuencia optima es necesario observar las perdidas enlos semiconductores.
5.5.2. Estimacion de las perdidas en los MOSFETs
Las perdidas por conduccion son estimadas utilizando los estreses de corrientesen los semiconductores obtenidos en [25]; para las perdidas por conmutacion se hanutilizado estimaciones de los tiempos de coexistencia entre de la tesion drenador-fuente y la corriente en el MOSFET [27]. Estos calculos se ha realizado para variosMOSFETs utilizando una base de datos de infineon. Ademas para estos calculos seha considerado incluir MOSFETs en paralelo la estimacion incluye desde uno hastacuatro MOSFET en paralelo para cada fase. Para todas las combinaciones posiblesentre numero de MOSFETs y frecuencia de conmutacion se ha seleccionado la opcioncon menor perdidas. El resumen de los resultados de las estimaciones se muestran enla figura fig. 5.12, la figura (a) muestra las perdidas totales en todos los MOSFETs, la(b) muestra las perdidas por MOSFETs y la (c) muestra la temperatura en la junturade cada MOSFETs considerando una temperatura ambiente de 70C.
Para un mismo numero de MOSFETs en paralelo las total de perdidas incrementaa mayor frecuencia de conmutacion. A baja fecuencia (25kHz o 40kHz) al pasar de unMOSFET a dos MOSFETs en paralelo las perdidas disminuyen considerablemente, dela misma forma a alta frecuencias (80kHz y 100kHz) al pasar de dos a tres MOSFETsen paralelo disminuyen las perdidas considerablemente.
Al incrementar el numero de MOSFETs en paralelo no solo perdidas totales dis-minuyen sino que tambien disminuyen drasticamente las perdidas en cada MOSFETs.Las temperatura maxima permitida por los fabricantes de estos MOSFETs es 150Cpero por temas de fiabilidad la temperatura maxima en aplicaciones aeronauticas esun 75 % de lo estipulado por el fabricante, es decir 112C. La figura 5.12(c) muestraambos lımites de temperatura, ası es necesario utilizar al menos dos MOSFETs enparalelo, o de otra forma la temperatura serıa mayor o muy cercana a los lımites.Ademas por temas de fiabilidad es mejor utilizar la menor cantidad de MOSFETs enparalelo, por lo que dos MOSFETs en paralelo la mejor opcion.
5.5.3. Frecuencia de conmutacion optima
Desde el punto de vista del peso del filtro, la mejor configuracion es con dos etapas,por otro lado en cuanto a las perdidas, temperaturas y fiabilidad en los MOSFETs eloptimo es utilizar dos MOSFETs en paralelo. Considerando fijos estos dos paramet-ros y analizando el compromiso de diseno entre las perdidas y el peso a distintas
5.5 Optimizacion del Filtro 65
0
5
10
15
20
25
30
35Filter Losses Estimations
Loss
es W
0
500
1000
1500
2000
2500
3000
3500
Wei
ght g
r
Filter Weight Estimations
Single StageTwo StageThree StageFour StageFive Stage
Figura 5.11. Estimacion de peso y perdidas de filtro de uno, dos, tres, cuatro y cincoetapas a diferentes frecuencias de conmutacion
66Capıtulo 5 – Nuevas consideraciones de Diseno del Filtro EMI de Rectificador Trifasico para Aplicaciones
Aeronauticas
0
20
40
60
80
100
W
Total Losses in MOSFETs
0
5
10
15
20
25
30Losses in each MOSFET
W
80
100
120
140
160Temperature in MOSFETs
C
1 MOSFET2 MOSFETs in Parallel3 MOSFETs in Parallel4 MOSFET in Parallel
Temperature limit according to manufacturer
Reability temperature limit
25kHz 40kHz 60kHz 80kHz 100kHz
25kHz 40kHz 60kHz 80kHz 100kHz
25kHz 40kHz 60kHz 80kHz 100kHz
Figura 5.12. estimacion de peerdidas en los MOSFTES considerando una, dos, tres yhasta cuatro MOSFETs en paralelo a diferentes frecuencias de conmutacion
5.5 Optimizacion del Filtro 67
frecuencias de comutacion se busta un buen balance peso-perdidas.A 25kHz el peso del filtro es exesivo, por lo cual se descarta esta frecuencia de con-
mutacion. El peso a 40kHz y 60kHz es practicamente el mismo al igual que a 80kHzy 100kHz, considerando que las perdidas aumentan con la frecuencia se descartan60kHz y 100kHz. De acuerdo con las especificaciones de nuestra aplicacion en par-ticular lo que se gana en peso compensa lo que se pierde en energıa cuando se pasade 40kHz a 80kHz, por lo tanto la frecuencia de conmutacion optima para nuestraaplicacione es 80kHz.
5.5.4. Conclusiones
Este capıtulo introduce nuevas consideraciones en el filtro EMI de entrada para una three-phase buck-type pulse-width modulation rectifier para aplicaciones aeronauti-cas. En este tipo de aplicaciones los estandares EMI a cumplir son mas estrictos quelos estandares industriales, sobre todo en el rango de frecuencias. En la MIL-STD-461E el rango de frecuencias empieza en 10kHz, esto impide que la frecuencia deconmutacion sea puesta debajo del rango, debido a esto el filtro se debe disenar demanera de atenuar la frecuencia de conmutacion. En aplicaciones industriales la fre-cuencia a atenuar es por lo general el sexto armonico de la frecuencia de conmutacion.
Debido a esta restricciones es necesario disenar un optima frecuencia de con-mutacion de manera de balancear las perdidas en los semiconductores y el volumen ypeso del sistema completo. Para lograr esto se han estimado las perdidas para variasfrecuencias de conmutacion utilizando una base de dados de MOSFETs, ademas seha considerado la opcion de utilizar hasta cuatro MOSFET en paralelo. Luego se hanestimado los pesos de los filtro utilizando componentes industrial para el diseno defiltros EMI.
Con los resultados obtenidos la solucion seleccionada ha sido, utilizar una frecuen-cia de conmutacion de 80kHz utilizando dos MOSFET en paralelo.
Ademas se ha propuesto una nueva metodologıa de diseno para el filtro EMI demanera tal de obtener un factor de potencia unitario a potencia nominal.
68Capıtulo 5 – Nuevas consideraciones de Diseno del Filtro EMI de Rectificador Trifasico para Aplicaciones
Aeronauticas
Capıtulo 6
PLATAFORMA DE PRUEBA
Figura 6.1. Fotogracia de la plataforma de pruebas utilizada en las pruebas experi-mentales
Para la construccion y posterior prueba de convertidor se ha preparado unaplataforma de prueba, ver figura 6.1. Esta plataforma esta compuesta de:
una fuente de alimentacion trifasica de tension y frecuencia variable, la cualpueden entregar hasta 5A RMS por fase a 400Hz,
un vatımetro digital trifasico para la medicion de potencia, factor de potenciay harmonicos,
un par de bancos resistivos que son utilizados como cargas,
69
70 Capıtulo 6 – Plataforma de Prueba
un PC para la programacion y posterior depuracion del codigo implementadoen el DSP,
fuentes de laboratorio DC para la alimentacion de la tarjetas
y polımetros digitales.
La primera parte del proyecto correspondio al diseno, implementacion y pruebasde rectificador trifasico con su correspondiente filtro EMI, en principio sin incluir elconvertidor DC-DC.
En el presente capıtulo se mostraran las tres tarjetas disenadas para las pruebasexperimentales, dos tarjetas para el rectificdor, una de potencia y otra de senal, yotra tarjeta para el Filtro EMI.
6.1 Tarjeta de Potencia del Rectficador 71
6.1. Tarjeta de Potencia del Rectficador
La figura 6.2 muestra el esquematico del rectificador, con este esquematico se handisenado dos PCBs con diferentes layout para estudiar la influencia de las inductanciasparasitas en las perdidas por conmutacion.
6.1.1. Layout de la Tarjeta de Potencia, version 1.
El primer layout se diseno con dos MOSFETs en paralelo tratando de dejarlos lomas cerca posible uno del otro, la inductancia parasita en el lazo MOSFET, Diodo,Condensador es de 342nH estimado con una herramienta de analisis de elementosfinitos Q3D de ansoft.
6.1.2. Layout de la Tarjeta de Potencia, version 2.
Un segundo layout fue disenado para disminuir la inductancia parasita, en estelayout se velo por dejar lo mas cerca posible el grupo MOSFET, Diodo y Condensador,asi de disminuyo a mas de cuatro veces. La inductancia parasita obtenida es de 83nH
72 Capıtulo 6 – Plataforma de Prueba
11
22
33
44
DD
CC
BB
AA
Title
Num
ber
Revi
sion
Size A4 Date
:27
/07/
2011
Shee
t o
fFi
le:
C:\U
sers\
..\Es
quem
atic
o El
ectri
co R
ectif
icad
or C
elda
1.S
chDo
cDr
awn
By:
Q13
IPW
60R1
65CP
Q10
IPW
60R1
65CP
Q14
IPW
60R1
65CP
Q11
IPW
60R1
65CP
Q15
IPW
60R1
65CP
Q12
IPW
60R1
65CP
6Ga
te_R
2
Vcc_
leg_
R2
GND_
leg_
R2
Vcc_
leg_
S1
GND_
leg_
S1
GND_
PO
Vcc_
leg_
R1
GND_
leg_
R1
Vcc_
leg_
T1
GND_
leg_
T1
Gate
_R1
Sour
ce_R
1
Drai
n_R1
Gate
_S1
Sour
ce_S
1
Drai
n_S1
Gate
_T1
Sour
ce_T
1
Drai
n_T1
Vcc_
leg_
R1
GND_
leg_
R1
Vcc_
leg_
S1
GND_
leg_
S1
Vcc_
leg_
T1
GND_
leg_
T1
R_af
ter_
Filte
rS_
afte
r_Fi
lter
T_af
ter_
Filte
r
Buck
type
Rec
tifie
r
100u
H
L10
Indu
ctor
ETD
49
100u
H
L11
Indu
ctor
ETD
49
3.3u
F
C10
Epco
s Ser
ie B
329
6R1
0Ga
te_R
1
Vcc_
leg_
R1
GND_
leg_
R1
6Ga
te_S
2
Vcc_
leg_
S2
GND_
leg_
S2
6Ga
te_S
1
Vcc_
leg_
S1
GND_
leg_
S1
6Ga
te_T
2
Vcc_
leg_
T2
GND_
leg_
T2
6Ga
te_T
1
Vcc_
leg_
T1
GND_
leg_
T1
D11
STTH
30R0
4
D16
STTH
30R0
4
D10
STTH
30R0
4
D15
STTH
30R0
4
D12
STTH
30R0
4
D17
STTH
30R0
4
D13
STTH
30R0
4
15v_
inpu
t
15v_
inpu
t
15v_
inpu
t
0v_i
nput
0v_i
nput
0v_i
nput
vout
_270 vo
ut_0
Sour
ce_R
1So
urce
_R2
Sour
ce_S
1So
urce
_S2
Sour
ce_T
1So
urce
_T2
R1+
R1-
S2-
R1+
R1-
S1+
S2-
T1+
T2-
1uF
C19
1uF
C20
1uF
C15
1uF
C16
1uF
C11
1uF
C12
Gate
_R2
Gate
_S2
Gate
_T2
Drai
n_R2
Drai
n_S2
Drai
n_T2
Vcc_
leg_
R2
GND_
leg_
R2
15v_
inpu
t
0v_i
nput
Vcc_
leg_
S2
GND_
leg_
S2
15v_
inpu
t
0v_i
nput
Vcc_
leg_
T2
GND_
leg_
T2
15v_
inpu
t
0v_i
nput
R2+
R2-
S1-
S2+
T1-
T2+
1 2 3 4 5 6 7 8 9 10 11 12
P10
Head
er 1
2
R2-
R2+
S1-
S1+
T1-
T1+
S2+
T2-
T2+
Vcc_
leg_
T2
GND_
leg_
T21u
FC2
11u
FC2
2
Vcc_
leg_
S2
GND_
leg_
S21u
FC1
71u
FC1
8
Vcc_
leg_
R2
GND_
leg_
R21u
FC1
31u
FC1
4
D14
STTH
30R0
4
Cann
ot
open
fil
e C:
\Use
rs\M
arc
elo\
Des
kto
p\dc
dcs
ymbo
l.jpg
Vout
+
Vout
-
Vin+
Vin-
Isola
ted
DC-D
C 11
MEV
1S12
15SC
Cann
ot
open
fil
e C:
\Use
rs\M
arc
elo\
Des
kto
p\dc
dcs
ymbo
l.jpg
Vout
+
Vout
-
Vin+
Vin-
Isola
ted
DC-D
C 15
MEV
1S12
15SC
Cann
ot
open
fil
e C:
\Use
rs\M
arc
elo\
Des
kto
p\dc
dcs
ymbo
l.jpg
Vout
+
Vout
-
Vin+
Vin-
Isola
ted
DC-D
C 13
MEV
1S12
15SC
Cann
ot
open
fil
e C:
\Use
rs\M
arc
elo\
Des
kto
p\dc
dcs
ymbo
l.jpg
Vout
+
Vout
-
Vin+
Vin-
Isola
ted
DC-D
C 10
MEV
1S12
15SC
Cann
ot
open
fil
e C:
\Use
rs\M
arc
elo\
Des
kto
p\dc
dcs
ymbo
l.jpg
Vout
+
Vout
-
Vin+
Vin-
Isola
ted
DC-D
C 14
MEV
1S12
15SC
Cann
ot
open
fil
e C:
\Use
rs\M
arc
elo\
Des
kto
p\dc
dcs
ymbo
l.jpg
Vout
+
Vout
-
Vin+
Vin-
Isola
ted
DC-D
C 12
MEV
1S12
15SC
NC A K NCGN
D5
out
6
out
7
Vcc
8
Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\De
skto
p\Si
8220
.jpg
Isodr
iver
Si1
0
Si82
20 Is
olat
ed D
river
NC A K NCGN
D5
out
6
out
7
Vcc
8
Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\De
skto
p\Si
8220
.jpg
Isodr
iver
Si1
1
Si82
20 Is
olat
ed D
river
NC A K NCGN
D5
out
6
out
7
Vcc
8
Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\De
skto
p\Si
8220
.jpg
Isodr
iver
Si1
4
Si82
20 Is
olat
ed D
river
NC A K NCGN
D5
out
6
out
7
Vcc
8
Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\De
skto
p\Si
8220
.jpg
Isodr
iver
Si1
2
Si82
20 Is
olat
ed D
river
NC A K NCGN
D5
out
6
out
7
Vcc
8
Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\De
skto
p\Si
8220
.jpg
Isodr
iver
Si1
3
Si82
20 Is
olat
ed D
river
NC A K NCGN
D5
out
6
out
7
Vcc
8
Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\De
skto
p\Si
8220
.jpg
Isodr
iver
Si1
5
Si82
20 Is
olat
ed D
river
Sour
ce_R
2So
urce
_S2
Sour
ce_T
1
PIC1001PIC1002COC10
PIC1101PIC1102COC11
PIC1201PIC1202COC12
PIC1301PIC1302CO
C13
PIC1401PIC1402COC14
PIC1501PIC1502COC15
PIC1601PIC1602COC16
PIC1701PIC1702CO
C17
PIC1801PIC1802COC18
PIC1901PIC1902COC19
PIC2001PIC2002COC20
PIC2101PIC2102CO
C21
PIC2201PIC2202COC22
PID1001 PID1002COD10
PID1101 PID1102COD11
PID1201 PID1202COD12
PID1301 PID1302COD13
PID1401 PID1402COD1
4
PID1501 PID1502COD15
PID1601 PID1602COD16
PID1701 PID1702COD17
PIIsod
river
Si1001
PIIsod
river
Si1002
PIIsod
river
Si1003
PIIsod
river
Si1004
PIIsod
river
Si1005
PIIsod
river
Si1006
PIIsod
river
Si1007
PIIsod
river
Si1008
COIsodriver Si10
PIIsod
river
Si1101
PIIsod
river
Si1102
PIIsod
river
Si1103
PIIsod
river
Si1104
PIIsod
river
Si1105
PIIsod
river
Si1106
PIIsod
river
Si1107
PIIsod
river
Si1108
COIsodriver Si11
PIIsod
river
Si1201
PIIsod
river
Si1202
PIIsod
river
Si1203
PIIsod
river
Si1204
PIIsod
river
Si1205
PIIsod
river
Si1206
PIIsod
river
Si1207
PIIsod
river
Si1208
COIsodriver Si12
PIIsod
river
Si1301
PIIsod
river
Si1302
PIIsod
river
Si1303
PIIsod
river
Si1304
PIIsod
river
Si1305
PIIsod
river
Si1306
PIIsod
river
Si1307
PIIsod
river
Si1308
COIs
odri
ver
Si13
PIIsod
river
Si1401
PIIsod
river
Si1402
PIIsod
river
Si1403
PIIsod
river
Si1404
PIIsod
river
Si1405
PIIsod
river
Si1406
PIIsod
river
Si1407
PIIsod
river
Si1408
COIsodriver Si14
PIIsod
river
Si1501
PIIsod
river
Si1502
PIIsod
river
Si1503
PIIsod
river
Si1504
PIIsod
river
Si1505
PIIsod
river
Si1506
PIIsod
river
Si1507
PIIsod
river
Si1508
COIsodriver Si15
PIIsolat
ed DC0DC
1001
PIIsol
ated D
C0DC 1
002PII
solate
d DC0D
C 1003
PIIsolat
ed DC0DC
1004
COIsolated DC0DC 10
PIIsol
ated D
C0DC 1
101
PIIsolat
ed DC0DC
1102
PIIsolat
ed DC0DC
1103
PIIsol
ated D
C0DC 1
104
COIsolated DC0DC 11
PIIsolat
ed DC0DC
1201
PIIsol
ated D
C0DC 1
202PII
solate
d DC0D
C 1203
PIIsolat
ed DC0DC
1204
COIsolated DC0DC 12
PIIsol
ated D
C0DC 1
301
PIIsolat
ed DC0DC
1302
PIIsolat
ed DC0DC
1303
PIIsol
ated D
C0DC 1
304
COIsolated DC0DC 13
PIIsolat
ed DC0DC
1401
PIIsol
ated D
C0DC 1
402PII
solate
d DC0D
C 1403
PIIsolat
ed DC0DC
1404
COIsolated DC0DC 14
PIIsol
ated D
C0DC 1
501
PIIsolat
ed DC0DC
1502
PIIsolat
ed DC0DC
1503
PIIsol
ated D
C0DC 1
504
COIsolated DC0DC 15
PIL1001
PIL1002
COL1
0
PIL1
101
PIL1
102
COL1
1
PIP1001
PIP1002
PIP1
003
PIP1
004
PIP1005
PIP1006
PIP1007
PIP1
008
PIP1
009
PIP10010
PIP10011
PIP10012CO
P10
PIQ1001
PIQ1002PIQ1003COQ10
PIQ1101
PIQ1102PIQ1103COQ11
PIQ1201
PIQ1202PIQ1203COQ12
PIQ1301
PIQ1302PIQ1303COQ13
PIQ1401
PIQ1402PIQ1403 COQ14
PIQ1501
PIQ1502PIQ1503 COQ15
PIR1001
PIR1002
COR1
0PIR1101
PIR1102
COR1
1
PIR1201
PIR1202
COR1
2
PIR1301
PIR1302
COR13
PIR1401
PIR1402
COR1
4PIR1501
PIR1502
COR15
PIIsol
ated D
C0DC 1
002
PIIsolat
ed DC0DC
1102
PIIsol
ated D
C0DC 1
202
PIIsolat
ed DC0DC
1302
PIIsol
ated D
C0DC 1
402
PIIsolat
ed DC0DC
1502
NL0v
0inp
ut
PIIsolat
ed DC0DC
1001
PIIsol
ated D
C0DC 1
101
PIIsolat
ed DC0DC
1201
PIIsol
ated D
C0DC 1
301
PIIsolat
ed DC0DC
1401
PIIsol
ated D
C0DC 1
501
NL15v0input
PID1501PIQ1302 NLDra
in0R
2
PID1601PIQ1402 NLDra
in0S
2NL
Drai
n0T2
PIQ1001
PIR1002NL
Gate
0R1
PIQ1301
PIR1102NL
Gate
0R2
PIQ1101
PIR1202
NLGate0S1
PIQ1401
PIR1302NL
Gate0S2
PIQ1201
PIR1402
NLGate0T1
PIQ1501
PIR1502NL
Gate0T2
PIC1101PIC1201
PID1002
PIIsod
river
Si1005
PIIsol
ated D
C0DC 1
003
PIQ1003
NLGND0le
g0R1
NLSo
urce
0R1
PIC1301PIC1401
PIIsod
river
Si1105
PIIsolat
ed DC0DC
1103
PIQ1002 PIQ1303NLDr
ain0
R1
NLGND0leg0R2
NLR0
after0
Filter
NLSo
urce
0R2
PIC1701PIC1801
PIIsod
river
Si1305
PIIsolat
ed DC0DC
1503
PIQ1102 PIQ1403NLDrai
n0S1
NLGND0leg0S2
NLS0
after0Fi
lter
NLSo
urce
0S2
PIC2101PIC2201
PIIsod
river
Si1505
PIIsolat
ed DC0DC
1303
NLGND0leg0T2
NLSo
urce
0T2
PID1302PID1402
PID1502PID1602
PID1702PI
L110
1
PID1001PID1101
PID1201
PID1301PID1401
PIL1001
PID1701PIQ1502
PIIsod
river
Si1001
PIIsod
river
Si1004
PIIsod
river
Si1006
PIIsod
river
Si1007PIR1001
PIIsod
river
Si1101
PIIsod
river
Si1104
PIIsod
river
Si1106
PIIsod
river
Si1107PIR1101
PIIsod
river
Si1201
PIIsod
river
Si1204
PIIsod
river
Si1206PIR1201
PIIsod
river
Si1207
PIIsod
river
Si1301
PIIsod
river
Si1304
PIIsod
river
Si1306PIR1301
PIIsod
river
Si1307
PIIsod
river
Si1401
PIIsod
river
Si1404
PIIsod
river
Si1406PIR1401
PIIsod
river
Si1407
PIIsod
river
Si1501
PIIsod
river
Si1504
PIIsod
river
Si1506PIR1501
PIIsod
river
Si1507
PIIsod
river
Si1002
PIP1
004
NLR10
PIIsod
river
Si1003
PIP1
003
NLR1
0
PIIsod
river
Si1102
PIP1002
NLR20
PIIsod
river
Si1103
PIP1001
NLR2
0
PIIsod
river
Si1202
PIP1
008
NLS1
0
PIIsod
river
Si1203
PIP1007
NLS10
PIIsod
river
Si1302
PIP1006
NLS2
0
PIIsod
river
Si1303
PIP1005
NLS20
PIC1501PIC1601
PID1102
PIIsod
river
Si1205
PIIsol
ated D
C0DC 1
403
PIQ1103
NLGND0leg0S1
NLSo
urce
0S1
PIC1901PIC2001
PID1202
PIIsod
river
Si1405
PIIsol
ated D
C0DC 1
203
PIQ1202PIQ1203 PIQ1503NLDrai
n0T1
NLGN
D0leg0
T1
NLSo
urce
0T1
NLT0after0Filter
PIIsod
river
Si1402
PIP10012
NLT10
PIIsod
river
Si1403
PIP10011
NLT10
PIIsod
river
Si1502
PIP10010
NLT20
PIIsod
river
Si1503
PIP1
009
NLT20
PIC1102PIC1202
PIIsod
river
Si1008
PIIsolat
ed DC0DC
1004
NLVcc0leg0R1
PIC1302PIC1402
PIIsod
river
Si1108
PIIsol
ated D
C0DC 1
104
NLVc
c0leg0
R2
PIC1502PIC1602
PIIsod
river
Si1208
PIIsolat
ed DC0DC
1404
NLVcc0leg0S1
PIC1702PIC1802
PIIsod
river
Si1308
PIIsol
ated D
C0DC 1
504
NLVcc0leg0S2
PIC1902PIC2002
PIIsod
river
Si1408
PIIsolat
ed DC0DC
1204
NLVcc0leg0T1
PIC2102PIC2202
PIIsod
river
Si1508
PIIsol
ated D
C0DC 1
304
NLVcc0leg0T2
PIC1001
PIL1
102
NLvout00
PIC1002
PIL1002
NLvout
0270
Figura 6.2. Esquematico del Rectificador
6.1 Tarjeta de Potencia del Rectficador 73
PA102
PA101
CO1
PA202
PA201
CO2
PAC102
PAC101
COC1
PAC201
PAC202
COC2
PAC301
PAC302
COC3
PAC401
PAC402
COC4
PAC601 PAC602 COC6
PAC702
PAC701
COC7
PAC901
PAC902
COC9
PAC100
1
PAC1002COC10
PAC1202
PAC1201COC12
PAC1502
PAC150
1COC15
PAC1702
PAC1701
COC17
PACR102
PACR101
COCR1
PACS102 PACS101
COCS1
PACT102 PACT101
COCT1
PAD100
PAD102
PAD101
COD1
PAD200
PAD202
PAD201
COD2
PAD300
PAD302
PAD301
COD3
PAD401
PAD402
PAD400
COD4
PAD500
PAD502
PAD501
COD5
PAD600
PAD602
PAD601
COD6
PAD700
PAD702
PAD701
COD7
PADCDC101
PADCDC102
PADCDC103PADCDC104
CODCDC1
PADCDC201
PADCDC202
PADCDC20
3
PADCDC204
CODCDC2
PADCDC301
PADCDC302
PADCDC303
PADCDC304
CODCDC3
PADCDC401
PADCDC402
PADCDC40
3
PADCDC404
CODCDC4PADCDC
501
PADCDC502
PADCDC503
PADCDC504
CODCDC5PADCD
C601
PADCDC602
PADCDC603
PADCDC604
CODCDC6
PADriver 105
PADriver 106
PADriver 107
PADriver 108
PADriver 104
PADriver 103
PADriver 102
PADriver 101
CODriver 1
PADriver 205
PADriver 206
PADriver 207
PADriver 208
PADriver 204
PADriver 203
PADriver 202
PADriver 201
CODriver 2
PADriver 305
PADriver 306
PADriver 307
PADriver 308
PADriver 304
PADriver 303
PADriver 302
PADriver 301
CODriver 3
PADriver 405
PADriver 406
PADriver 407
PADriver 408
PADriver 404
PADriver 403
PADriver 402
PADriver 401
CODriver 4
PADriver 505
PADriver 506
PADriver 507
PADriver 508
PADriver 504
PADriver 503
PADriver 502
PADriver 501
CODriver 5
PADriver 605
PADriver 606
PADriver 607
PADriver 608
PADriver 604
PADriver 603
PADriver 602
PADriver 601
CODriver 6
PAL101
PAL102
COL1
PAL201
PAL202
COL2
PALEM102
PALEM101
COLEM1
PAP101
PAP102
PAP103
PAP104
PAP105
PAP106
PAP107
PAP108
PAP109
PAP1010
PAP1011
PAP1012
COP1
PAQ101
PAQ102
PAQ103
PAQ100
COQ1
PAQ201
PAQ202
PAQ203
PAQ200
COQ2
PAQ301
PAQ302
PAQ303
PAQ300
COQ3
PAQ401
PAQ402
PAQ403
PAQ400
COQ4
PAQ501
PAQ502
PAQ503
PAQ500
COQ5
PAQ601
PAQ602
PAQ603
PAQ600
COQ6
PAQ701
PAQ702
PAQ703
PAQ700
COQ7
PAQ801
PAQ802
PAQ803
PAQ800
COQ8
PAQ901
PAQ902
PAQ903
PAQ900
COQ9
PAQ1001
PAQ1002
PAQ1003
PAQ1000
COQ1
0PAQ1
101PAQ1
102PAQ1
103
PAQ1100 COQ11
PAQ1201
PAQ1202
PAQ1203
PAQ1200 COQ12
PAR102
PAR101
COR1
PAR201
PAR202
COR2
PAR301PAR302
COR3
PAR401PAR402
COR4
PAR501
PAR502
COR5
PAR601PAR602
COR6
PAVout102
PAVout101
COVo
ut1
PAVout202
PAVout201
COVout2
PA201
PA202
PADCDC102
PADCDC202
PADCDC302
PADCDC402
PADCDC502
PADCDC602
PA101
PA102
PADCDC101
PADCDC201
PADCDC301
PADCDC401
PADCDC501
PADCDC601
PAC402
PAC902
PAC1701
PACR101
PACR102
PADCDC103
PADriver 405
PAQ102
PAQ402
PAQ703
PAQ1003
PAD501PAQ7
02
PAQ1002
PAC302
PAC401
PAC1201
PACS101
PACS102
PADCDC20
3
PADriver 505
PAQ202
PAQ502
PAQ803
PAQ1103
PAD601PAQ8
02
PAQ1102
PAC301
PAC701
PAC901PACT1
01
PACT102
PADCDC303
PADriver 605
PAQ302 PAQ602
PAQ903
PAQ1203
PAD701PAQ9
02
PAQ1202
PAQ101
PAQ401
PAR102
PAQ701
PAQ1001
PAR402
PAQ201
PAQ501 PAR2
02
PAQ801
PAQ1101
PAR502
PAQ301
PAQ601
PAR302
PAQ901
PAQ1201
PAR602
PAC150
1
PAD102
PADCDC40
3
PADriver 105
PAQ103
PAQ403
PAC100
1
PAD202
PADCDC503
PADriver 205
PAQ203
PAQ503
PAC601
PAD302
PADCDC603
PADriver 305
PAQ303
PAQ603
PAD402
PAD502
PAD602
PAD702
PALEM101
PAD101
PAD201
PAD301
PAD401
PAL101
PADriver 106
PAR101
PADriver 206PAR201
PADriver 306
PAR301
PADriver 406
PAR401
PADriver 506
PAR501
PADriver 606
PAR601
PAL201
PALEM102
PADriver 102
PAP104
PADriver 103
PAP103
PADriver 402
PAP102
PADriver 403
PAP101
PADriver 202
PAP108
PADriver 203
PAP107
PADriver 502
PAP106
PADriver 503
PAP105
PADriver 302
PAP1012
PADriver 303
PAP1011
PADriver 602
PAP1010
PADriver 603
PAP109
PAC1502
PADCDC404
PADriver 108
PAC1702
PADCDC104
PADriver 408
PAC1002
PADCDC504
PADriver 208
PAC1202
PADCDC204
PADriver 508
PAC602
PADCDC604
PADriver 308
PAC702 PADCDC304
PADriver 608
PAC101
PAC201
PAL202
PAVout201
PAVout202
PAC102
PAC202
PAL102
PAVout101
PAVout102
Figura 6.3. Layout de la tarjeta de potencia version 1. La inductancia parasita dellazo condensador, MOSFET y Diodo es de 342nH
74 Capıtulo 6 – Plataforma de Prueba
PAC101
PAC102
COC1
PAC202
PAC201
COC2
PAC302
PAC301
COC3
PAC402
PAC401
COC4
PAC902
PAC901
COC9
PACR102
PACR101
COCR1
PACS102
PACS101
COCS1
PACT102
PACT101
COCT1
PAD101
PAD102
PAD100
COD1
PAD201
PAD202
PAD200
COD2
PAD301
PAD302
PAD300
COD3
PAD401
PAD402
PAD400
COD4
PAD501
PAD502
PAD500
COD5
PAD601
PAD602
PAD600
COD6
PAD701
PAD702
PAD700
COD7
PAD1200
PAD1202
PAD1201
COD12
PAL102
PAL101
COL1
PAL202
PAL201
COL2
PALEM101
PALEM102
COLE
M1
PAQ101
PAQ102
PAQ103
PAQ100
COQ1
PAQ201
PAQ202
PAQ203
PAQ200
COQ2
PAQ301
PAQ302
PAQ303
PAQ300
COQ3
PAQ701
PAQ702
PAQ703
PAQ700
COQ7
PAQ801
PAQ802
PAQ803
PAQ800
COQ8
PAQ901
PAQ902
PAQ903
PAQ900
COQ9
PAVout101
PAVout102
COVout1
PAVout201
PAVout202
COVo
ut2
PAC402
PAC902
PACR101
PACR102
PAQ102 PAQ7
03
PAD501 PAQ702
PAC302
PAC401
PACS101
PACS102
PAQ202 PAQ803
PAD601 PAQ802
PAC301
PAC901
PACT101
PACT102
PAQ302
PAQ903
PAD701
PAQ902
PAQ101
PAQ701
PAQ201
PAQ801
PAQ301
PAQ901
PAD102
PAQ103
PAD202
PAQ203
PAD302PAQ3
03
PAD402
PAD502
PAD602
PAD702
PAD1202
PALEM101
PAD101
PAD201
PAD301
PAD401
PAD1201
PAL101
PAL201
PALEM102
PAC101
PAC201
PAL202
PAVout201
PAVout202
PAC102
PAC202
PAL102
PAVout101
PAVout102
Figura 6.4. Layout de la tarjeta de potencia version 2. La inductancia parasita dellazo condensador, MOSFET y Diodo es de 83nH
6.2 Tarjeta de Control del Convertidor 75
6.2. Tarjeta de Control del Convertidor
La tarjeta de control tiene dos objetivos principales, primero acondicionar lassenales de los sensores de tension del convertidor y transformar los pulsos del DSPde tension 0 y 3.3V a pulsos de corriente de 0 y 10mA, la figura 6.5 muestra elesquematico utilizado en esta tarjeta.
Los sensores de de tension son de la marca LEM modelo LV25-P, entre sus car-acterısticas se puede destacar su alta precision, bajo desfase a 400Hz y el aislamientoentre la tension medida y la senal de medicion. El circuito de acondicionamientode senal atenua y suma offset a la senal del sensor de manera de cumplir el rangodinamico de tension permitida por los conversores analogos a digital del DSP (0 a3.3V).
Para transformar los pulsos de tension de las salidas PWM del DSP en pulsos decorrientes se ha utilizado simples circuitos en base a espejos de corriente. Ademas seincluyen diodos schottky y condensadores en la base del transistor para disminuir eltiempo de descarga y carga de este.
76 Capıtulo 6 – Plataforma de Prueba
11
22
33
44
55
66
77
88
DD
CC
BB
AA
Title
Num
ber
Revi
sion
Size A3 Date
:27
/07/
2011
Shee
t o
fFi
le:
C:\U
sers\
..\M
edic
ion_
Tens
ion.
SchD
ocDr
awn
By:
1
23
4
Q1 PN29
07A
GND
Vcc+
5V
GND
GND
signa
l_R1
GND_
R1
Q3 BC54
7
10uF
C58
Cap
sup
1K7
R36
Res1
500
R27
Res1
PWM
_R1
2.2n
FC6
2Di
ode
Scho
ttky1
1N57
11W
1
23
4
Q2 PN29
07A
GND
Vcc+
5V
GND
GND
signa
l_R2
GND_
R2
Q4 BC54
7
10uF
C59
Cap
sup
1K7
R37
Res1
500
R28
Res1
PWM
_R2
2.2n
FC6
3Di
ode
Scho
ttky2
1N57
11W
1
23
4
Q5 PN29
07A
GND
Vcc+
5V
GND
GND
signa
l_S1
GND_
S1
Q7 BC54
7
10uF
C67
Cap
sup
1K7
R49
Res1
500
R47
Res1
PWM
_S1
2.2n
FC7
0Di
ode
Scho
ttky3
1N57
11W
1
23
4
Q6 PN29
07A
GND
Vcc+
5V
GND
GND
signa
l_S2
GND_
S2
Q8 BC54
7
10uF
C68
Cap
sup
1K7
R50
Res1
500
R48
Res1
PWM
_S2
2.2n
FC7
1Di
ode
Scho
ttky4
1N57
11W
1
23
4
Q9 PN29
07A
GND
Vcc+
5V
GND
GND
signa
l_T1
GND_
T1
Q17
BC54
7
10uF
C72
Cap
sup
1K7
R53
Res1
500
R51
Res1
PWM
_T1
2.2n
FC7
4Di
ode
Scho
ttky5
1N57
11W
1
23
4
Q16
PN29
07A
GND
Vcc+
5V
GND
GND
signa
l_T2
GND_
T2
Q18
BC54
7
10uF
C73
Cap
sup
1K7
R54
Res1
500
R52
Res1
PWM
_T2
2.2n
FC7
5Di
ode
Scho
ttky6
1N57
11W
Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\Des
kto
p\LV
25-
P.jp
g
+HT
HT-
-Vcc
Med
+Vcc
LEM
Vol
tage
1
LV25
-P Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\Des
kto
p\LV
25-
P.jp
g
+HT
HT-
-Vcc
Med
+Vcc
LEM
Vol
tage
2
LV25
-P
150K
R16
150K
R29
150K
R41
150K
R46
150
R26 15
0R4
5
GND
GND
Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\De
skto
p\TL
084.
jpg
-Vcc
Vcc+
TL1
Op A
mp
150k
R19
75k
R915
0kR1
7
GND
1kR7VC
C+15 GN
D
22k
R8
GND
10k
R18
Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\De
skto
p\TL
084.
jpg
-Vcc
Vcc+
TL2
Op A
mp
150k
R44
75k
R40
150k
R42
GND
22k
R39
GND
10k
R43
GND
1kR38
VCC+
15
D1
D Ze
ner 7
.5V
D2 D Ze
ner 7
.5V
Med
_Vrs
Med
_Vst
10uF
C79
10uF
C64
GND
GND
10uF
C65
10uF
C69
GND
GND
33nF
C66
GND33
nFC6
1
GND
VCC+
15V
VCC-
15V
VCC+
15V
VCC-
15V
R_af
ter_
Filte
r
S_af
ter_
Filte
r
T_af
ter_
Filte
r
Cann
ot o
pen
file
C:\U
sers\
Mar
celo
\Des
ktop
\dsp
.jpg
ADC1
ADC2
PWM
_R1
PWM
_R2
PWM
_S1
PWM
_S2
PWM
_T1
PWM
_T2
DSP
F1
DSP
F283
35
Med
_Vrs
Med
_Vst
PWM
_R1
PWM
_R2
PWM
_S1
PWM
_S2
PWM
_T1
PWM
_T2
500n
FC7
8
500n
FC8
1
500n
FC8
3
500n
FC8
7
200n
FC6
0
200n
FC8
4
200n
FC7
6
200n
FC8
5
200n
FC8
820
0nF
C89
500n
FC8
0
500n
FC7
7
GND
GND
500n
FC8
6
500n
FC8
2
GND
GND
PIC5801PIC5802CO
C58
PIC5901PIC5902COC59
PIC6001 PIC6002COC60
PIC6101 PIC6102COC61
PIC6201
PIC6202
COC62
PIC630
1PIC630
2
COC63
PIC6401 PIC6402COC64 PIC6501 PIC6502
COC65
PIC6601 PIC6602COC66
PIC6701PIC6702COC67
PIC6801PIC6802CO
C68
PIC6901 PIC6902COC69
PIC700
1PIC700
2
COC70
PIC7101
PIC7102
COC71
PIC7201PIC7202COC72
PIC7301PIC7302CO
C73
PIC740
1PIC740
2
COC74
PIC7501
PIC7502
COC75
PIC7601 PIC7602COC76
PIC7701 PIC7702COC77
PIC7801 PIC7802COC78
PIC7901 PIC7902COC79
PIC8001 PIC8002COC80
PIC8101 PIC8102COC81
PIC8201PIC8202COC82
PIC8301 PIC8302COC83
PIC8401 PIC8402COC84
PIC8501 PIC8502COC85
PIC8601 PIC8602COC86
PIC8701 PIC8702COC87
PIC8801 PIC8802COC88
PIC8901 PIC8902COC89
PID101
PID102
COD1
PID201
PID202
COD2
PIDiod
e Scho
ttky10
1PID
iode S
chottk
y102
CODiode Schottky1
PIDiod
e Scho
ttky20
1PID
iode S
chottk
y202
CODiode Schottky2
PIDiod
e Scho
ttky30
1PID
iode S
chottk
y302
CODiode Schottky3
PIDiod
e Scho
ttky40
1PID
iode S
chottk
y402
CODiode Schottky4
PIDiod
e Scho
ttky50
1PID
iode S
chottk
y502
CODiode Schottky5
PIDiod
e Scho
ttky60
1PID
iode S
chottk
y602
CODiode Schottky6
PIDSP F100
PIDSP F101
PIDSP F102
PIDSP F103
PIDSP F104
PIDSP F105
PIDSP F106
PIDSP F107
CODSP F1
PILE
M Vo
ltag
e100
PILEM
Voltag
e101
PILEM
Voltag
e102
PILE
M Vo
ltag
e104
COLEM Voltage1
PILE
M Vo
ltag
e200
PILE
M Vo
ltag
e201
PILE
M Vo
ltag
e202
PILE
M Vo
ltag
e204
COLEM Voltage2
PIQ101PIQ102PIQ103
PIQ104CO
Q1
PIQ201PIQ202PIQ203
PIQ204CO
Q2
PIQ30B
PIQ30C PIQ30ECOQ3
PIQ40B
PIQ40C PIQ40ECOQ4PIQ501
PIQ502PIQ503PIQ504
COQ5
PIQ601PIQ602PIQ603
PIQ604CO
Q6
PIQ70B
PIQ70C PIQ70ECOQ7
PIQ80B
PIQ80C PIQ80ECOQ8
PIQ901PIQ902PIQ903
PIQ904CO
Q9
PIQ1601PIQ1602PIQ1603
PIQ1604COQ16
PIQ170B
PIQ170C PIQ170ECOQ1
7PIQ180B
PIQ180C PIQ180ECOQ18
PIR7
01PI
R702
COR7
PIR801 PIR802COR8
PIR901
PIR9
02COR9
PIR1601
PIR1602
COR1
6PIR1701
PIR1
702
COR17
PIR1801PIR1802 COR18
PIR1901
PIR1902
COR19
PIR2601 PIR2602COR26
PIR2701PIR2702 COR27
PIR2801PIR2802 COR2
8
PIR2901
PIR2902
COR2
9
PIR360
1PI
R3602
COR3
6PI
R3701
PIR370
2CO
R37
PIR3801
PIR3802
COR3
8
PIR3901 PIR3902COR39
PIR4
001
PIR4002COR40
PIR4101
PIR4102
COR4
1PIR4201
PIR4
202
COR42
PIR4301PIR4302 COR43
PIR4401
PIR4402
COR44
PIR4501PIR4502 COR45
PIR4601
PIR4602
COR4
6
PIR4701PIR4702 COR4
7
PIR4801PIR4802 COR48
PIR4901
PIR490
2CO
R49
PIR500
1PI
R5002
COR5
0
PIR5101PIR5102 COR5
1
PIR5201PIR5202 COR52
PIR5301
PIR530
2CO
R53
PIR540
1PI
R5402
COR5
4
PITL100
PITL101
PITL102
PITL103
PITL104
PITL105
PITL106
PITL107
COTL1
PITL200
PITL201
PITL202
PITL203
PITL204
PITL205
PITL206
PITL207
COTL2
PIC5801PIC5901
PIC6002
PIC6102
PIC6402 PIC6501
PIC6602
PIC6701PIC6801
PIC6902
PIC7201PIC7301
PIC7602
PIC7701PIC7801
PIC7901
PIC8002
PIC8102
PIC8202PIC8301
PIC8402PIC8502
PIC8602
PIC8702
PIC8802PIC8902
PID101
PID201
PIQ30EPIQ40EPIQ70E
PIQ80E
PIQ170EPIQ180E
PIR2602
PIR4501
PITL102
PITL104
PITL202
PITL204
NLGN
D0R1
NLGN
D0R2
NLGND0S1
NLGND0S2
NLGN
D0T1
NLGN
D0T2
PIDSP F100
PIR1802PITL106
NLMe
d0Vr
s
PIDSP F101
PIR4302PITL206
NLMe
d0Vs
t
PIC6101
PILEM
Voltag
e102
PIR1901
PIR2601
PIC6201PIDiod
e Scho
ttky10
2 PIQ30B
PIR3602
PIC630
1PIDiod
e Scho
ttky20
2 PIQ40B
PIR370
2
PIC6401PIC8101
PILEM
Voltag
e102 PIC6502
PIC8302PI
LEM
Volt
age2
04
PIC6601PI
LEM
Volt
age2
02
PIR4401
PIR4502
PIC6901PIC8701
PILE
M Vo
ltag
e202
PIC700
1PIDiod
e Scho
ttky30
2 PIQ70B
PIR490
2
PIC7101PIDiod
e Scho
ttky40
2 PIQ80B
PIR5002
PIC740
1PIDiod
e Scho
ttky50
2 PIQ170B
PIR530
2
PIC7501PIDiod
e Scho
ttky60
2 PIQ180B
PIR5402
PIC7702PIC7802
PIC7902PI
LEM
Volt
age1
04
PIC8001PITL103
PID102PI
R701
PIR1701
PID202PIR3801
PIR4201
PIDiod
e Scho
ttky10
1 PIQ30CPIR2701PID
iode S
chottk
y201 PIQ40CPIR2801
PIDiod
e Scho
ttky30
1 PIQ70CPIR4701PID
iode S
chottk
y401 PIQ80CPIR4801
PIDiod
e Scho
ttky50
1 PIQ170CPIR5101PID
iode S
chottk
y601 PIQ180CPIR5201
PILE
M Vo
ltag
e100
PIR1602 PIL
EM Vol
tage10
1PIR2902 PILE
M Vo
ltag
e200
PIR4102 PILE
M Vo
ltag
e201
PIR4602
PIQ101 PIR2702PIQ201 PIR2802PIQ501 PIR4702
PIQ601 PIR4802
PIQ901 PIR5102PIQ1601 PIR5202
PIR801
PIR1801PITL105
PIR802PIR901PITL100
PIR9
02PI
R170
2
PIR1902
PITL101
PIR3901
PIR4301PITL205
PIR3902PI
R400
1PITL200
PIR4002
PIR4
202
PIR4402
PITL201
PIC6202
PIDSP F102
PIR360
1NLPWM0R1
PIC630
2
PIDSP F103
PIR3701
NLPWM0R2
PIC700
2
PIDSP F104
PIR4901
NLPWM0S1
PIC7102
PIDSP F105
PIR500
1NLPWM0S2
PIC740
2
PIDSP F106
PIR5301
NLPWM0T1
PIC7502
PIDSP F107
PIR540
1NLPWM0T2
PIR1601
NLR0after0Filter
PIR2901
PIR4101
NLS0after0Filter
PIQ102 NLsi
gnal
0R1
PIQ202 NLsignal0R2
PIQ502 NLsignal0S1
PIQ602 NLsignal0S2
PIQ902 NLsignal0T1
PIQ1602 NLsignal0T2
PIR4601
NLT0after0Filter
PIC5802PIC5902
PIC6001
PIC6702PIC6802
PIC7202PIC7302
PIC7601PIC8401PIC8501
PIC8801PIC8901
PIQ103PIQ104
PIQ203PIQ204
PIQ503PIQ504
PIQ603PIQ604
PIQ903PIQ904
PIQ1603PIQ1604
NLVc
c05V
PIR7
02
PIR3802
PIC8201
PITL107
PITL207
PIC8601PITL203
Figura 6.5. Esquematico de la tarjeta de medicion de tension de lınea y conversorestension corriente
6.3 Filtro EMI 77
6.3. Filtro EMI
La figura 6.6 muestra el esquematico del filtro EMI de dos etapas implementado.
11
22
33
44
DD
CC
BB
AA
Title
Num
ber
Revi
sion
Size A4 Date
:27
/07/
2011
Shee
t o
fFi
le:
C:\U
sers\
..\EM
I Filt
er lo
sses.S
chDo
cDr
awn
By:
EMI T
hree
Sta
ge F
ilter
2.2n
FC55
DE2E
3KH2
22C5
7DE
2E3K
H222
1.6m
H
L2T6
0405
-S61
23-X
363
1.6m
HL1
4T6
0405
-S61
23-X
363
2.2n
FC52
DE2E
3KH2
22
C53
C54
1.6m
HL1
T604
05-S
6123
-X36
3
1.6m
HL9
T604
05-S
6123
-X36
3
38uH
L3M
agne
tics 5
8110
- Hih
g Fl
ux 6
0
38uH
L6
38uH
L12
1.5u
F
C1B3
2923
C310
5+ 1.5u
F
C9B3
2923
C310
5+
1.5u
FC5
B329
23C3
105+
R1
1.5u
FC5
0B3
2923
C310
5+
R5
1.5u
FC3
B329
23C3
105+
1.5u
F
C6 B329
23C3
105+
R2
GND
GND
Line
R
Line
S
Line
T
38uH
L4M
agne
tics 5
8110
- Hih
g Fl
ux 6
0
38uH
L7
38uH
L13
1.5u
F
C2B3
2923
C310
5+ 1.5u
F
C49
B329
23C3
105+
1.5u
FC7
B329
23C3
105+
R3
1.5u
FC5
1B3
2923
C310
5+
R6
1.5u
FC4
B329
23C3
105+
1.5u
F
C8 B329
23C3
105+
R4
Line
R a
fter E
MI f
ilter
Line
S a
fter E
MI f
ilter
Line
T a
fter E
MI f
ilter
5W
5W
5W
5W 5W
5W
6.8W
6.8W 6.
8W
6.8W
6.8W 6.
8W
2W 2W
2W2W
2W
2W
2W
2W
2W
2W
2W
2W
L5
L8
PIC101PIC102CO
C1PIC201PIC202
COC2
PIC301PIC302CO
C3PIC401PIC402
COC4
PIC501PIC502CO
C5PIC601PIC602
COC6
PIC701PIC702COC7
PIC801PIC802CO
C8
PIC901PIC902CO
C9PIC4901PIC4902
COC49
PIC5001PIC5002COC50
PIC5101PIC5102COC51
PIC5201PIC5202COC52
PIC5301PIC5302CO
C53
PIC5401PIC5402COC54
PIC5501PIC5502CO
C55
PIC5601PIC5602CO
C56
PIC5701PIC5702CO
C57
PIL101
PIL102
COL1
PIL201
PIL202
COL2
PIL3
01PI
L302
COL3
PIL401
PIL402
COL4
PIL501
PIL502
COL5
PIL6
01PI
L602
COL6
PIL701
PIL702
COL7
PIL801
PIL802
COL8
PIL901
PIL902
COL9
PIL1
201
PIL1
202
COL1
2PIL1301
PIL1302
COL13
PIL1401
PIL1402
COL1
4
PIR101PIR102 COR1
PIR201PIR202 COR2
PIR301PIR302 COR3
PIR401PIR402 COR4
PIR501PIR502 COR5
PIR601PIR602 COR6
PIC5201PIC5301
PIC5401PIC5501
PIC5601PIC5701
PIL101
NLLi
ne R
PIC202PIC402
PIC5502
PIL402
PIR302PIR402
NLLi
ne R
after
EMI
fil
ter
PIL501
NLLi
ne S
PIC201
PIC701
PIC4902
PIC5602
PIL702
PIR602
NLLi
ne S
after
EMI
fil
ter
PIL901
NLLi
ne T
PIC401
PIC801
PIC4901
PIC5101
PIC5702PIL1302
NLLi
ne T
after
EMI
fil
ter
PIC101
PIC501
PIC902
PIC5302
PIL6
02PIL801
PIR502
PIC102PIC302
PIC5202
PIL201
PIL3
02PIR102
PIR202
PIC301
PIC601
PIC901
PIC5001
PIC5402PI
L120
2PIL1401
PIC502PIR101PIC602PIR201
PIC702PIR301PIC802PIR401
PIC5002PIR501PIC5102PIR601
PIL102
PIL3
01PIL202
PIL401
PIL502
PIL6
01PIL701
PIL802
PIL902
PIL1
201
PIL1301
PIL1402
Figura 6.6. Esquematico del Filtro EMI de modo Diferencial
78 Capıtulo 6 – Plataforma de Prueba
Capıtulo 7
RESULTADOSEXPERIMENTALES
En este capıtulo se presentan formas de ondas experimentales obtenidas con elplataforma y convertidor construido.
7.1. Formas de Onda de las conmutaciones
La figura 7.1(a) muestra la corriente en los inductores de continua, la figura 7.1(b)muestra la forma de onda de la corriente en un MOSFETs y la figura 7.1(c) muestrala tension en el diodo de libre circulacion. Los picos de corriente y tension en lasconmutaciones hasta el momento no han sido gran problema en cuanto a las perdidasy a la fiabilidad del sistema pero son un problema que debe ser resulto en el futuro.
7.2. Rendimiento, Factor de Potencia y Distorsion Armonica
La figura 7.2(a) muestra la eficiencia del rectificador mas filtro EMI en funcionde la potencia de carga, la figura 7.2(b) muestra el factor de potencia en funcion dela potencia de carga y la figura 7.2(c) muestra el THD de la corriente y tension dered en funcion de la potencia de carga. Todas estas mediciones se han realizado hasta1.5kW que es el 45 % de la carga nominal, no se han realizado mediciones a mayorpotencia debido a que la fuente de alimentacion trifasica que disponemos no entregamas de 1.5kW a 400Hz.
7.3. Formas de onda de Corriente y tensiones de lınea
La figura 7.3(a) muestra la corriente en los inductores, la figura 7.3(b) muestra lacorriente de lınea del rectificador y la figura 7.3(c) muestra la tension de lınea de lafuente.
La envolvente de baja frecuencia en al corriente de los inductores corresponde conlas simulaciones mostradas en el capıtulo 4. La corriente y la tension de lınea no estanen fase, esto se debe la baja potencia de carga, solo el 40 % de la tension nominal. Apotencia nominal el desfase debe ser cero.
La corriente de lınea no es puramente sinusoidal, esto se debe a que la corriente deentrada es modulada mediante la medicion de la tension de entrada, y esta al no seruna sinodal pura la mosma modulacion inyecta los armonicos. Este efecto en principio
79
80 Capıtulo 7 – Resultados Experimentales
iba a ser atenuado con un filtro digital, sin embargo este no fue suficiente. Como partedel trabajo futuro esta la implementacion de nuevas tecnicas para la atenuacion delTHD como por ejemplo la implementacion de una PLL [28–30] o una modulacionSHE [31,32].
7.3 Formas de onda de Corriente y tensiones de lınea 81
0 10 20 30 40 50 604
6
8
10
Time µ s
Indu
ctor
Cur
rent
A
0 10 20 30 40 50 60−5
0
5
10
15
20
Time µ s
MO
SF
ET
Cur
rent
A
0 10 20 30 40 50 60
0
200
400
600
Time µ s
Fre
ewhe
elin
g D
iode
Vol
tage
V
Figura 7.1. Formas de onda de las conmutaciones
82 Capıtulo 7 – Resultados Experimentales
400 600 800 1000 1200 1400 16000.966
0.968
0.97
0.972
0.974
0.976
0.978
Eff
icie
ncy
Power Demand W
400 600 800 1000 1200 1400 1600
0.65
0.7
0.75
0.8
0.85
0.9
0.95
Power Demand W
Po
wer
Fac
tor
400 600 800 1000 1200 1400 16001
2
3
4
5
6
7
Power Demand W
TH
D %
Current THDVoltage THD
Figura 7.2. Resultado de las mediciones de eficiencia, factor de potencia y THD delrectificador hasta media carga
7.3 Formas de onda de Corriente y tensiones de lınea 83
0 0.5 1 1.5 2 2.5 32
4
6
8
10
Time ms
DC
Cur
rent
A
0 0.5 1 1.5 2 2.5 3−10
−5
0
5
10
Time ms
Line
Cur
rent
A
0 0.5 1 1.5 2 2.5 3−200
−100
0
100
200
Time ms
Line
Vol
tage
V
Figura 7.3. Formas de onda de la corriente en los inductores, corriente de lınea ytension de linea
84 Capıtulo 7 – Resultados Experimentales
Capıtulo 8
CONCLUSIONES
En la presente tesis se ha mostrado la arquitectura electrica actual utilizada enel sistema de reabastecimiento de combustible en vuelo disenado por Airbus. Luegode un estudio de este sistema se ha desarrollado sistematicamente un procedimientode optimizacion del sistema electrico para el mejoramiento de este, en terminos dedesempeno, peso y volumen.
Esta optimizacion se ha desarrollado a varios nieves, primero a un nivel de ar-quitecturas, como se muestra en el capıtulo 1 esta optimizacion permite seleccionaruna arquitectura que reduce enormemente tanto el peso como el volumen total delsistema, en comparacion con arquitecturas no optimas.
La arquitectura propuesta propone la unificacion de todos los motores para sertrabajados como su fuesen solo una carga, de esta forma el sistema se ve beneficiado encuanto a rendimiento ya que la energıa regenerada puede ser reutilizada y la demandade potencia del sistema desde el generador, ya que esta demanda se reduce a lapotencia media del sistema, en terminos numericos la potencia demanda es reducidade 34kW a solo 7kW.
Para el estudio de la arquitectura propuesta se han analizados los principalestipos de unidad de almacenamiento de energıa, como lo son baterıas y super con-densadores, para esto se ha simulado el sistema incluyendo modelos de baterıas ysuper condensadores. Como resultado de este analisis, las baterıas presentan mejoresprestaciones en peso y ademas permiten descargas maximas consecutivas algo quecon super condensadores no es posible.
Para la optimizacion a nivel de topologıas se han considerados varios paramet-ros. El rectificador unidireccional tipo buck presenta mejores caracterısticas que sushomologos elevadores ya que tienen mayor densidad de potencia, menor complejidad,en numeros de componentes y control, ademas prescinden de circuito adicional deprecarga de condensador. Dentro de los rectificadores tipo buck existe la version deseis switches que presenta una mayor eficiencia, por lo cual este ultimo es el que hasido seleccionado para nuestra aplicacion.
Ademas, en la optimizacion a nivel de topologıa se ha incluido la variable multiplesceldas, es decir, repartir la potencia total del sistema en varios rectificadores traba-jando en paralelo, ası se ve beneficiado el filtro de entrada y la fiabilidad del sistema.La optimizacion el sistema se ha realizado mediante estimaciones de perdidas y pesopara diferentes frecuencias de conmutacion con el objeto de obtener un buen balanceentre perdidas y peso. La topologıa optima obtenida es de tres celdas, cada una de3300W, el rectificador trifasico tipo buck de seis switches trabajando a una frecuencia
85
86 Capıtulo 8 – Conclusiones
de 60kHz acoplado a un convertidor Full-brigde alimentado en tension trabajando auna frecuencia de 180kHz y un filtro EMI de dos etapas comun para las tres celdas.
Ademas en el presente trabajo se han desarrollado prototipos de 3.3kW para lavalidacion del sistema. Resultados experimentales muestran que el sistema trabajacorrectamente hasta 1.5kW. Actualmente se esta trabajando para probar el rectifi-cador a potencia nominal, sin embargo los resultados obtenidos hasta ahora muestranun buen augurio para las pruebas posteriores debido a la alta eficiencia y baja dis-torsion del convertidor.
Capıtulo 9
APENDICE
9.1. Articulos Publicados en Congresos
9.1.1. Articulo presentado en el congreso SAAEIG 2011.
87
88 Capıtulo 9 – Apendice
EMI Filter Design of a Three-Phase Buck-TypePWM Rectifier for Aircraft Applications.
Marcelo Silva, Nico Hensgens, Jesus Oliver, Pedro Alou, Oscar Garcıa, and Jose A CobosCentro de Electronica Inductrial
Universidad Politecnica de MadridMadrid, Spain
Email: marcelo@silvas.cl
Abstract— An EMI filter for a three-phase buck-type mediumpower pulse-width modulation rectifier is designed. This filterconsiders differential mode noise and complies with MIL-STD-461E for the frequency range of 10kHz to 10MHz. In industrialapplications, the frequency range of the standard starts at150kHz and the designer typically uses a switching frequencyof 28kHz because the fifth harmonic is out of the range. Thisapproach is not valid for aircraft applications. In order to designthe switching frequency in aircraft applications, the power lossesin the semiconductors and the weight of the reactive componentsshould be considered. The proposed design is based on aharmonic analysis of the rectifier input current and an analyticalstudy of the input filter. The classical industrial design doesnot consider the inductive effect in the filter design because thegrid frequency is 50/60Hz. However, in the aircraft applications,the grid frequency is 400Hz and the inductance cannot beneglected. The proposed design considers the inductance and thecapacitance effect of the filter in order to obtain unitary powerfactor at full power. In the optimization process, several filtersare designed for different switching frequencies of the converter.In addition, designs from single to five stages are considered. Thepower losses of the converter plus the EMI filter are estimatedat these switching frequencies. Considering overall losses andminimal filter volume, the optimal switching frequency is selected.
Keywords: Three Phase Rectifier, EMI Filter, High powerfactor.
I. INTRODUCTION
The input filter in a PWM rectifier system has three pur-poses: 1) to ensure sinusoidally shaped input currents byfiltering the switching-frequency harmonics; 2) to attenuatethe electromagnetic interference with other electronic systems;3) to avoid susceptibility to electromagnetic emissions fromsurrounding systems and itself [1], [2], [3]. While designingan EMI filter for a power electronic system, the applicableEMI standards need to be considered.
Typically in industrial applications, the standard to complywith is CISPR 22 class B [4]. The frequency range consideredby this standard reaches from 150kHz to 30MHz. In [1], [5],systems with a switching frequencies (fs) of 28kHz and 18kHzrespectively have been designed. These fs have been chosenbecause they are sufficiently higher in comparison with thegrid (50 or 60Hz). In addition, the first, second, third, fourth,and fifth harmonic of the fs are out of the range of CISPR 22class B; thus, the first harmonic to consider in the input filter
104
105
106
107
-20
0
20
40
60
80
100
120
Current spectrum at f s = 28kHz
CISPR 22 Class A
CISPR 22 Class B
Critical Harmonic
168kHz)(
Frequency Hz
Fig. 1. Spectrum current using 28kHz switching frequency
design is the sixth harmonic at 168kHz (when fs = 28kHz).This can be seen in the fig 1.
This work introduces new considerations in the input filterdesign for a three-phase buck-type pulse-width modulationrectifier (fig. 2(a)) for aircraft applications. In this applicationthe standard to comply with is MIL-STD-461E [6]. Thisstandard is more restrictive than the CISPR 22, regulatinga wider range of frequencies from 10kHz to 10Mhz. Fig.2(b) shows the limits for MIL-STD-461E, CISPR 22 classA, and CISPR 22 class B. Due to the frequency range of theMIL-STD-461E and the fact that switching frequencies below10kHz would not be an optimal design, the rectifier switchingfrequency must be inside of the range. Therefore, the inputfilter must be designed in order to attenuate the switchingfrequency.
II. DESIGN OF THE INPUT FILTER
A. Converter Topology
The EMC input filter is designed for a three-phase, three-switch, current source (buck-type) PWM rectifier system,fig 2 (a), with sinusoidal input current, direct start-up, andovercurrent protection in case of an output short circuit. In [7]and [8] this topology has been considered for the realizationof the input stage of high-power telecommunications rectifier
Figura 9.1. Artıculo publicado en SAAEIG 2011. Pagina 1
9.1 Articulos Publicados en Congresos 89
Three-phase buck-type PWM Rectifier
EMC input Filter
104
105
106
107
0
20
40
60
80
100
120
140
160
180 X: 1.204e+005Y: 168.4
Frequency Hz
dB
µV
Current spectrum at fs
=
MIL−STD−461E
CISPR 22 Class A
CISPR 22 Class B
120kHz
Fig. 2. (a) Three-phase buck-type PWM rectifier Topology and EMI input Filter, (b) Current spectrum of the rectifier at 10kW without input filter
modules. All of these benefits have prompted the authors tointroduce this rectifier in aircraft applications.
In this work, a 10kW system for an aircraft application willbe designed.
B. Cutoff frequency of the input FilterIn order to know the desired attenuation, the topology needs
to be simulated without the input filter. Fig 2 (b) shows themeasured input current spectrum. This current is measuredutilizing a line impedance stabilizing network (LISN).
With a switching frequency (fs) of the converter at 120kHz,the first harmonic has an amplitude of 168.4 dBµV and theMIL-STD-461E limit is 84 dBµV. Considering a margin of6 dB, the required attenuation is 168.4 − 84 + 6(Margin) =96.4dBµV. The cut-off frequency as a function of attenuationand the switching frequency is given by:
ωcutoff =1√
L · C=
2π · fs√10Att[dB]/(20n)
(1)
L · C =1096.4[dB]/(20n)
(2π · 120kHz)2, (2)
where n is the number of the filter stages. Eq. 2 indicatesthe value of the product L · C as a function of the requiredattenuation at a certain frequency. To design the inductive andcapacitive values, it is necessary to take into consideration thepower factor of the rectifier.
C. Input Capacitor Consideration in industrial applicationIn [7] and [9] the input capacitor is designed in order to limit
the reactive power of the rectifier. Eq. (3) gives the maximumvalue for the input capacitor C as a function of the reactivepower (in percentage of the nominal power PN ). Usually thispower is limited to (5..10%) of the rated power in order toensure high power factor.
C ≤ (0.05...0.1) · PN
ω · U2N,l−l,rms
= 1.67...3.04µF, (3)
where ω is the grid frequency and UN,l−l,rms is the line toline input voltage (RMS). In aircraft applications, ω = 2π ·400rad/s and UN,l−l,rms = 115
√(3)V. Thus, a good value for the
capacitor is 1µF because the capacitance is lower than 1.67µF,eq. (3). Once the capacitance is fixed, the filter inductancecan be calculated with eq.(2); therefore, the inductor value is452µH using a two stage filter.
Fig. 3 (a) shows the equivalent circuit for the rectifierincluding the two stage filter impedance seen from the grid.The analytical expression of the impedances are as presentedin eqs. (4) and (5)
Zeq1stg = ȷωL +1
ȷωC + 1/R(4)
Zeq2stg = ȷωL +1
ȷωC + 1/Zeq1stg, (5)
where Zeq1stg and Zeq2stg are the equivalent impedances forsingle stage and two stage filters respectively. The resistancecorresponding the output power for 10kW at 115V is R = 4Ω.
With this filter, the power factor of the system is only 0.88 (cos(∠Zeq2stg) ), but according to eq. (3) should be higher than0.99. Therefore, the power factor does not correspond withthe design considerations because eq. (3) does not include theeffect of inductor, which can be neglected for a grid frequencyof 50 or 60 Hz. However, in aircraft applications the grid is400Hz [10]; at this frequency the effect of the inductor cannot be neglected anymore.
D. Proposed Consideration for the input capacitor
In order to know the influence of the single stage L-C filteron the power factor, the real and the imaginary part of theimpedance needs to be considered separately according to eqs.
Figura 9.2. Artıculo publicado en SAAEIG 2011. Pagina 2
90 Capıtulo 9 – Apendice
104
105
106
107
−20
0
20
40
60
80
100
frequency Hz
dB
µV
Current spectrum at fs
= 120kHz
MIL−STD−461E
CISPR 22 Class A
CISPR 22 Class B
Fig. 3. (a) Equivalent circuit for the rectifier including the input filter seen from the grid. (b) Current spectrum using the proposed filter.
(6) and (7),
ℜZeq1stg =R
1 + (ω · C · R)2(6)
ℑZeq1stg = ωL − ωC
R−2 + ω2C2. (7)
The unity power factor is obtained when ℑZeq1stg = 0.With this condition, and the cut-off frequency being as it isin eq. (1), the filter component is as presented in eqs. (8) and(9)
C =1
R√
ω2cutoff − ω2
(8)
L =1
ω2cutoffC
. (9)
Thus, the power factor of the rectifier is unity using a singlestage input filter. In addition, if (ω ·C ·R)2 << 1 (this applieswhen the capacitance is in the order of µF), the equivalentimpedance for the single stage filter is approximately R(Zeq1stg ≈ R). If Zeq1stg ≈ R, then Zeq2stg ≈ Zeq1stg
according to eqs. 4 and 5, and ℑZeq2stg = ℑZeq1stg = 0.Therefore, independent of the amount of the filter stages, whendesigning the filter according to eqs. (8) and (9), the powerfactor at full power is unity.
Using this proposed design method, the filter capacitanceand inductance are C = 5µF and L = 84µH. In comparisonwith the classical design, the proposed design is smaller due tothe inductor value. Fig. 3 (b) shows the current spectrum of therectifier including the designed input filter. The current spec-trum complies with the MIL-STD-461E in all the frequenciesof the range.
E. Power factor depending on the power demand of the load
The Three-Phase Buck-Type PWM Rectifier is a two quad-rant converter; when the input voltage is positive, the inputcurrent is positive and when the input voltage is negative
the input current is negative. In addition, the rectifier iscontrolled in order to obtain sinusoidal wave form currentsproportional to the input voltage. For this reason, the rectifierhas a resistive behavior at low frequencies (grid frequency),and cannot deliver or absorb reactive power. Because of this,the power factor depends on the input filter and the powerdemand of the load.
In principle, the filter was designed in order to obtain unitarypower factor at full power (10kW). The black line in fig 5 (a)shows the behavior of the power factor in full range of thepower demand; however, from 5kW to 10kW the power factoris relatively high (higher than 95%). On the other hand, if thefilter is designed at 5kW, the range of the high power factorincreases from 2.5kW to 10kW. This can be seen in the red linein fig 5 (a). If the filter is designed at 1kW, the range is reducedfrom 0.5kW to 2kW, showed with the blue line in fig 5(a). Thefigs 5 (b) and (c) show the same curves for three and four stagefilters respectively. The high power factor range increases withthe number of stages in the filter. In low power demand, thepower factor is inevitably low because the equivalent resistanceof the rectifier is neglected in comparison with the impedanceof the input filter; thus, the system is practically reactive.
In order to have high power factor in the wide range ofthe power demand, the input filter has to be designed in halfof the nominal power (in this case 5kW). However, generallythe power factor is measured at full power; therefore, in thispaper, the filter is designed in order to obtain unitary powerfactor at full power.
III. APPLICATION TO THE OPTIMIZATION OF THE INPUTFILTER
In [1] the switching frequency for this topology is selectedat 28kHz because the fifth harmonic (140kHz) is out ofthe standard range (150kHz - 30MHz). However, in aircraftapplications this method of hiding the switching frequencyharmonies below the standards frequency range can not ap-plied because the MIL-STD-461E starting at 10kHz.
Figura 9.3. Artıculo publicado en SAAEIG 2011. Pagina 3
9.1 Articulos Publicados en Congresos 91
2 Stage Filter
0
1000
2000
3000
4000
5000
6000
gr
Filter Total Weight
3 Stage Filter
4 Stage Filter
5 Stage Filter
kHz50 80kHz kHz120 160 200kHz kHz0
0.5
1
1.5
2
2.5Filter Total Volume
L
2 Stage Filter
3 Stage Filter
4 Stage Filter
5 Stage Filter
0
100
200
300
400
W
Total Power Losses in MOSFETs
1 Mosfet
2 Mosfet in Parallel
3 Mosfet in Parallel
4 Mosfet in Parallel
Switching frequencySwitching frequency Switching frequency
kHz50 80kHz kHz120 160 200kHz kHz kHz50 80kHz kHz120 160 200kHz kHz
Fig. 4. (a) and (b) Volume and Weigh estimation for single, two, three, four and five stage. (c) Power losses in semiconductor using single, two, three, fourMOSFETs in parallel
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 100000.8
0.85
0.9
0.95
1
Power Factor vs Output Power for 2 stages filter @ 120kHz
Po
wer
Facto
r
Power W
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 100000.8
0.85
0.9
0.95
1
Power Factor vs Output Power for 3 stages filter @ 120kHz
Po
wer
Facto
r
Power W
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 100000.8
0.85
0.9
0.95
1
Power Factor vs Output Power for 4 stages filter @ 120kHz
Po
wer
Facto
r
Power W
Unirtary PF at 1kW
Unirtary PF at 5kW
Unirtary PF at 10kW
(a)
(c)
(b)
Fig. 5. Power factor depending to the power demand. (a) for two phase filter,(b) for three phase filter, (a) for four phase filter.
The switching frequency will be determined by the trade-off between volume/weight and power losses, for that thesize/weight of the filter and the losses are going to estimatedfor different switching frequencies; thus, to obtain a designwith a good balance between size and losses.
A. Weight and Volume estimation of the filter
For multiple-stage LC filters the minimum volume isachieved by using the same cut-off frequency for all stages[11]. Table I shows the cut-off frequencies from single tofive stage filters for different switching frequencies in orderto comply with MIL-STD-461E. The cut-off frequencies forsingle stage filters are close to or lower than the grid frequency,which makes it impossible to employ a single stage filter
TABLE ICUT-OFF FREQUENCY OF THE FILTER FOR SINGLE TO FIVE STAGES AND
DIFFERENT SWITCHING FREQUENCIES IN ORDER TO COMPLY
MIL-STD-461E.
cutoff freq 50kHz 80kHz 120kHz 160kHz 200kHzSingle stage 308Hz 392Hz 467Hz 555Hz 618HzTwo stages 3.9kHz 5.6kHz 7.5kHz 9.4kHz 11.1kHz
Three stages 9.2kHz 13.6kHz 18.9kHz 24.2kHz 29.1kHzFour stages 14kHz 21.2kHz 30kHz 38.8kHz 47.1kHzFive stages 18kHz 27.6kHz 39.6kHz 51.5kHz 63kHz
solution.Filters with two to five stages are designed using equations
(1), (8), and (9). Then, with the inductance and capacitancevalues from resulting from these cut-off frequencies, the sizeand weight of the different filter solutions can be estimated.
For the weight estimation, the weight of the magnetic coresand of the wound wire are considered. The weight of thecapacitors is neglected, since it is much lower than that ofthe magnetic components.
For the volume estimation, first, the total surface area iscalculated by taking the sum of the all the elements; thisnumber is multiplied by 1.5. Then, the volume is obtainedby multiplying the height of the highest component by thetotal surface area to get the boxed volume.
The figs. 4 (a) and (b) show the estimation results. A twostage filter is not practical because the volume and weight areconsiderably bigger than for three, four, and five stage filters.Four and five stage filters provide only minimal improvements(if any) compared to three stage filters, and the number ofcomponents (and parasitic couplings between components) ismuch higher. Consequently, the three stage filter appears to bethe best solution.
B. Power losses estimation
The conducting losses are estimated using the currentstresses in the semiconductor [9]. The switching losses areestimated for considered switching frequencies [12] and [13].These calculations have been carried out using one, two,three, and four MOSFETs in parallel in order to decrease the
Figura 9.4. Artıculo publicado en SAAEIG 2011. Pagina 4
92 Capıtulo 9 – Apendice
conducing losses. For every combination of fs and the numberof MOSFET in parallel, the optimal device, with respect to thepower losses, has been selected from of a database availablecomponents. The power losses of the best MOSFET for everycombination switching frequency vs MOSFETs in parallel areshown in the fig. 4 (c).
For the same number of MOSFETs in parallel, the totallosses increase with the switching frequency, due to switchinglosses. The total losses decrease with the number of MOSFETin parallel; from one MOSFET to two MOSFETs in parallel(and from two to three MOSFETs in parallel), the powerlosses decrease considerably. However, to change from threeto four MOSFETs in parallel, the power losses are onlymarginally decreased. In addition, the reliability decreaseswith the number of MOSFETs; therefore, it is apparent thatusing three MOSFETs in parallel is good option. The totallosses increase consistently with the increase of the frequency;however, the weight does not decrease consistently. In the fig.4 [a], the filter at 80kHz and 120kHz, as well as the filter at160kHz and 200kHz, have practically the same weight usinga three stage filter. Therefore, 80kHz and 160kHz are betterswitching frequencies because they have better ratios of weightlosses. In order to be conservative in efficiency, the switchingfrequency selected is 80kHz.
IV. CONCLUSION
This work introduces new considerations in the input filterdesign for a three-phase buck-type pulse-width modulationrectifier for aircraft applications. For this type of applications,the EMI standard to comply is MIL-STD-461E. This standardis more restrictive than the CISPR 22 because of its frequencyrange, and it is not recommendable to consider a switchingfrequency below the standard range. The switching frequencyhas an impact in the trade-off between size of the input filterand the power losses; in order to obtain an optimum switchingfrequency the volume/weight and power losses have beenestimated for a 10kW system. According to these estimations,the best trade-off between volume/weight and power lossesis at 80kHz using three filter stages and three MOSFETs inparallel because the reduction of the filter size from 80kHz to120kHz is marginal in comparison with the increment of thelosses.
In addition, the classical design method for the input filterconsiders only grid frequencies of 50Hz or 60Hz; however,when the grid frequency is 400Hz, the power factor for thesystem is not close to unity. This work proposes a newconsideration in the filter design in order to obtain a unitarypower factor at full power independently of the number of thefilter stages.
Currently a 10kW three-phase buck-type pulse-width mod-ulation rectifier prototype is being built.
ACKNOWLEDGMENT
Thank you to Airbus and INDRA for the collaboration inthis work. Thank you to Lauren Hayes for revising the Englishin this paper.
REFERENCES
[1] T. Nussbaumer, M. L. Heldwein, and J. W. Kolar, “Differential modeinput filter design for a three-phase buck-type pwm rectifier based onmodeling of the emc test receiver,” vol. 53, no. 5, pp. 1649–1661, 2006.
[2] H. W. OTT, Electromagnetic Compatibility Engineering, Wiley, Ed.Wiley, 2009.
[3] R. P. CLAYTON, Introduction to Electromagnetic compatibility, secondedition. Wiley, 2006.
[4] C.I.S.P.R., Comite international special des perturbationsradioelectriques Std.
[5] A. Stupar, T. Friedli, J. Miniboandck, M. Schweizer, and J. Kolar,“Towards a 99% efficient three-phase buck-type pfc rectifier for 400 vdc distribution systems,” in Applied Power Electronics Conference andExposition (APEC), 2011 Twenty-Sixth Annual IEEE, march 2011, pp.505 –512.
[6] MIL-STD-461E, Requirements for the control of electromagnetic inter-ference characteristics of subsystems and equipment Std., August 1999.
[7] M. Baumann, U. Drofenik, and J. W. Kolar, “New wide input voltagerange three-phase unity power factor rectifier formed by integration of athree-switch buck-derived front-end and a dc/dc boost converter outputstage,” in Proc. INTELEC Telecommunications Energy Conf. Twenty-second Int, 2000, pp. 461–470.
[8] L. Malesani and P. Tenti, “Three-phase ac/dc pwm converter withsinusoidal ac currents and minimum filter requirements,” no. 1, pp. 71–77, 1987.
[9] T. Nussbaumer, M. Baumann, and J. W. Kolar, “Comprehensive designof a three-phase three-switch buck-type pwm rectifier,” vol. 22, no. 2,pp. 551–562, 2007.
[10] G. Gong, M. L. Heldwein, U. Drofenik, J. Minibock, K. Mino, andJ. W. Kolar, “Comparative evaluation of three-phase high-power-factorac-dc converter concepts for application in future more electric aircraft,”vol. 52, no. 3, pp. 727–737, 2005.
[11] M. L. Heldwein and J. W. Kolar, “Design of minimum volume emc inputfilters for an ultra compact three-phase pwm rectifier,” in COBEP07. The7th Brazilian Power Electronics Conference, 2007.
[12] Z. John Shen, Y. Xiong, X. Cheng, Y. Fu, and P. Kumar, “Powermosfet switching loss analysis: A new insight,” in Proc. 41st IAS AnnualMeeting Industry Applications Conf. Conf. Record of the 2006 IEEE,vol. 3, 2006, pp. 1438–1442.
[13] A. K. Dr. Dusan Graovac, Marco Purschel, MOSFET Power LossesCalculation Using the DataSheet Parameters, INFINEON, July 2006.
Figura 9.5. Artıculo publicado en SAAEIG 2011. Pagina 5
9.1 Articulos Publicados en Congresos 93
9.1.2. Articulo presentado en el congreso ECCE 2011.
94 Capıtulo 9 – Apendice
New Considerations in the Input Filter Design of aThree-Phase Buck-Type PWM Rectifier for Aircraft
Applications.Marcelo Silva, Nico Hensgens, Jesus Oliver, Pedro Alou, Oscar Garcıa, and Jose A Cobos
Centro de Electronica InductrialUniversidad Politecnica de Madrid
Madrid, SpainEmail: marcelo@silvas.cl
Abstract— An EMI filter for a three-phase buck-type mediumpower pulse-width modulation rectifier is designed. This filterconsiders differential mode noise and complies with MIL-STD-461E for the frequency range of 10kHz to 10MHz. In industrialapplications, the frequency range of the standard starts at150kHz and the designer typically uses a switching frequencyof 28kHz because the fifth harmonic is out of the range. Thisapproach is not valid for aircraft applications. In order to designthe switching frequency in aircraft applications, the power lossesin the semiconductors and the weight of the reactive componentsshould be considered. The proposed design is based on aharmonic analysis of the rectifier input current and an analyticalstudy of the input filter. The classical industrial design doesnot consider the inductive effect in the filter design because thegrid frequency is 50/60Hz. However, in the aircraft applications,the grid frequency is 400Hz and the inductance cannot beneglected. The proposed design considers the inductance and thecapacitance effect of the filter in order to obtain unitary powerfactor at full power. In the optimization process, several filtersare designed for different switching frequencies of the converter.In addition, designs from single to five stages are considered. Thepower losses of the converter plus the EMI filter are estimatedat these switching frequencies. Considering overall losses andminimal filter volume, the optimal switching frequency is selected.
Keywords: Three Phase Rectifier, EMI Filter, High powerfactor.
I. INTRODUCTION
The input filter in a PWM rectifier system has three pur-poses: 1) to ensure sinusoidally shaped input currents byfiltering the switching-frequency harmonics; 2) to attenuatethe electromagnetic interference with other electronic systems;3) to avoid susceptibility to electromagnetic emissions fromsurrounding systems and itself [1], [2], [3]. While designingan EMI filter for a power electronic system, the applicableEMI standards need to be considered.
Typically in industrial applications, the standard to complywith is CISPR 22 class B [4]. The frequency range consideredby this standard reaches from 150kHz to 30MHz. In [1], [5],systems with a switching frequencies (fs) of 28kHz and 18kHzrespectively have been designed. These fs have been chosenbecause they are sufficiently higher in comparison with thegrid (50 or 60Hz). In addition, the first, second, third, fourth,
104
105
106
107
20
40
60
80
100
120
140
frequency Hz
dB
µV
Current spectrum
CISPR 22 Class A
CISPR 22 Class B
Critical Harmonic
Fig. 1. Spectrum current using 28kHz switching frequency
and fifth harmonic of the fs are out of the range of CISPR 22class B; thus, the first harmonic to consider in the input filterdesign is the sixth harmonic at 168kHz (when fs = 28kHz).This can be seen in the fig 1.
This work introduces new considerations in the input filterdesign for a three-phase buck-type pulse-width modulationrectifier (fig. 2(a)) for aircraft applications. In this applicationthe standard to comply with is MIL-STD-461E [6]. Thisstandard is more restrictive than the CISPR 22, regulatinga wider range of frequencies from 10kHz to 10Mhz. Fig.2(b) shows the limits for MIL-STD-461E, CISPR 22 classA, and CISPR 22 class B. Due to the frequency range of theMIL-STD-461E and the fact that switching frequencies below10kHz would not be an optimal design, the rectifier switchingfrequency must be inside of the range. Therefore, the inputfilter must be designed in order to attenuate the switchingfrequency.
Figura 9.6. Artıculo publicado en ECCE 2011. Pagina 1
9.1 Articulos Publicados en Congresos 95
Three-phase buck-type PWM Rectifier
EMC input Filter
104
105
106
107
20
40
60
80
100
120
140
160
frequency Hz
dB
µV
Current spectrum
MIL−STD−461E
CISPR 22 Class A
CISPR 22 Class B
Fig. 2. (a) Three-phase buck-type PWM rectifier Topology and EMI input Filter, (b) Current spectrum of the rectifier at 3.3kW without input filter
II. DESIGN OF THE INPUT FILTER
A. Converter Topology
The EMC input filter is designed for a three-phase, three-switch, current source (buck-type) PWM rectifier system,fig 2 (a), with sinusoidal input current, direct start-up, andovercurrent protection in case of an output short circuit. In [7]and [8] this topology has been considered for the realizationof the input stage of high-power telecommunications rectifiermodules. All of these benefits have prompted the authors tointroduce this rectifier in aircraft applications.
In this work, a 3.3kW system for an aircraft application willbe designed.
B. Cutoff frequency of the input Filter
In order to know the desired attenuation, the topology needsto be simulated without the input filter. Fig 2 (b) shows themeasured input current spectrum. This current is measuredutilizing a line impedance stabilizing network (LISN).
With a switching frequency (fs) of the converter at 60kHz,the first harmonic has an amplitude of 146 dBµV and the MIL-STD-461E limit is 85 dBµV. Considering a margin of 6 dB,the required attenuation is 146 − 85 + 5(Margin) = 66dBµV.The cut-off frequency as a function of attenuation and theswitching frequency is given by:
ωcutoff =1√
L · C=
2π · fs√10Att[dB]/(20n)
(1)
L · C =1068[dB]/(20n)
(2π · 60kHz)2, (2)
where n is the number of the filter stages. Eq. 2 indicatesthe value of the product L · C as a function of the requiredattenuation at a certain frequency. To design the inductive andcapacitive values, it is necessary to take into consideration thepower factor of the rectifier.
C. Input Capacitor Consideration in industrial application
In [7] and [9] the input capacitor is designed in order to limitthe reactive power of the rectifier. Eq. (3) gives the maximumvalue for the input capacitor C as a function of the reactivepower (in percentage of the nominal power PN ). Usually thispower is limited to (5..10%) of the rated power in order toensure high power factor.
C ≤ (0.05...0.1) · PN
ω · U2N,l−l,rms
= 4.96...9.92µF, (3)
where ω is the grid frequency and UN,l−l,rms is the lineto line input voltage (RMS). In aircraft applications, ω =2π · 400 rad/s and UN,l−l,rms = 115
√(3)V. Thus, a good
value for the capacitor is 4.7µF because the capacitance islower than 4.96µF, eq. (3). Once the capacitance is fixed, thefilter inductance can be calculated with eq.(2); therefore, theinductor value is 3.7mH using a single stage filter.
Fig. 3 (a) shows the equivalent circuit for the rectifierincluding the two stage filter impedance seen from the grid.The analytical expression of the impedances are as presentedin eqs. (4) and (5)
Zeq1stg = ȷωL +1
ȷωC + 1/R(4)
Zeq2stg = ȷωL +1
ȷωC + 1/Zeq1stg, (5)
where Zeq1stg and Zeq2stg are the equivalent impedances forsingle stage and two stage filters respectively. The resistancecorresponding the output power for 3.3kW at 115V is R =11.9Ω.
With this filter, the power factor of the system is only 0.66 (cos(∠Zeq2stg) ), but according to eq. (3) should be higher than0.99. Therefore, the power factor does not correspond withthe design considerations because eq. (3) does not include the
Figura 9.7. Artıculo publicado en ECCE 2011. Pagina 2
96 Capıtulo 9 – Apendice
MIL
CISPR 22 Cl
CISPR 22 Cl
104
105
106
107
−40
−20
0
20
40
60
80
100
120
frequency Hz
dB
µV
Current spectrum
MIL−STD−461E
CISPR 22 Class A
CISPR 22 Class B
Fig. 3. (a) Equivalent circuit per phase for the rectifier including the input filter seen from the grid. (b) Current spectrum using the proposed filter.
effect of inductor, which can be neglected for a grid frequencyof 50 or 60 Hz. However, in aircraft applications the grid is400Hz [10]; at this frequency the effect of the inductor cannot be neglected anymore.
D. Proposed Consideration for the input capacitor
In order to know the influence of the single stage L-C filteron the power factor, the real and the imaginary part of theimpedance needs to be considered separately according to eqs.(6) and (7),
ℜZeq1stg =R
1 + (ω · C · R)2(6)
ℑZeq1stg = ωL − ωC
R−2 + ω2C2. (7)
The unity power factor is obtained when ℑZeq1stg = 0.With this condition, and the cut-off frequency being as it isin eq. (1), the filter component is as presented in eqs. (8) and(9)
C =1
R√
ω2cutoff − ω2
(8)
L =1
ω2cutoffC
. (9)
Thus, the power factor of the rectifier is unity using a singlestage input filter. In addition, if (ω ·C ·R)2 << 1 (this applieswhen the capacitance is in the order of µF), the equivalentimpedance for the single stage filter is approximately R(Zeq1stg ≈ R). If Zeq1stg ≈ R, then Zeq2stg ≈ Zeq1stg
according to eqs. 4 and 5, and ℑZeq2stg = ℑZeq1stg = 0.Therefore, independent of the amount of the filter stages, whendesigning the filter according to eqs. (8) and (9), the powerfactor at full power is unity.
Using this proposed design method, the filter capacitanceand inductance are C = 10µF and L = 1.4mH . In comparisonwith the classical design, the proposed design is smaller due to
the inductor value. Fig. 3 (b) shows the current spectrum of therectifier including the designed input filter. The current spec-trum complies with the MIL-STD-461E in all the frequenciesof the range.
E. Power factor depending on the power demand of the load
The Three-Phase Buck-Type PWM Rectifier is a two quad-rant converter; when the input voltage is positive, the inputcurrent is positive and when the input voltage is negativethe input current is negative. In addition, the rectifier iscontrolled in order to obtain sinusoidal wave form currentsproportional to the input voltage. For this reason, the rectifierhas a resistive behavior at low frequencies (grid frequency),and cannot deliver or absorb reactive power. Because of this,the power factor depends on the input filter and the powerdemand of the load.
In principle, the filter was designed in order to obtain unitarypower factor at full power (3.3kW). The black line in fig 4(a) shows the behavior of the power factor in full range of thepower demand; however, from 1.65kW to 3.3kW the powerfactor is relatively high (higher than 95%). On the other hand,if the filter is designed at 1.65kW, the range of the high powerfactor increases from 2.5kW to 3.3kW. This can be seen inthe red line in fig 4 (a). If the filter is designed at 330W,the range is reduced from 165W to 670W, showed with theblue line in fig 4(a). The figs 4 (b) and (c) show the samecurves for three and four stage filters respectively. The highpower factor range increases with the number of stages in thefilter. In low power demand, the power factor is inevitably lowbecause the equivalent resistance of the rectifier is neglectedin comparison with the impedance of the input filter; thus, thesystem is practically reactive.
In order to have high power factor in the wide range of thepower demand, the input filter has to be designed in half ofthe nominal power (in this case 1.65kW). However, generallythe power factor is measured at full power; therefore, in this
Figura 9.8. Artıculo publicado en ECCE 2011. Pagina 3
9.1 Articulos Publicados en Congresos 97
0 500 1000 1500 2000 2500 3000 35000.8
0.85
0.9
0.95
1
Power Factor vs Output Power for 2 stages filter @ 60kHz
Po
wer
Facto
r
Power W
0 500 1000 1500 2000 2500 3000 35000.8
0.85
0.9
0.95
1
Power Factor vs Output Power for 3 stages filter @ 60kHz
Po
wer
Facto
r
Power W
0 500 1000 1500 2000 2500 3000 35000.8
0.85
0.9
0.95
1
Power Factor vs Output Power for 4 stages filter @ 60kHz
Po
wer
Facto
r
Power W
P = 0.33kW
P = 1.65kW
P = 3.3kW
Fig. 4. Power factor depending to the power demand. (a) for two phase filter,(b) for three phase filter, (a) for four phase filter.
paper, the filter is designed in order to obtain unitary powerfactor at full power (3.3kW).
III. APPLICATION TO THE OPTIMIZATION OF THE INPUTFILTER
In [1] the switching frequency for this topology is selectedat 28kHz because the fifth harmonic (140kHz) is out ofthe standard range (150kHz - 30MHz). However, in aircraftapplications this method of hiding the switching frequencyharmonies below the standards frequency range can not ap-plied because the MIL-STD-461E starting at 10kHz.
The switching frequency will be determined by the trade-off between weight and power losses, for that the weight ofthe filter and the losses are going to estimated for differentswitching frequencies; thus, to obtain a design with a goodbalance between size and losses.
A. Weight and Losses estimation of the filter
For multiple-stage LC filters, the minimum volume isachieved by using the same cut-off frequency for all stages[11]. Table I shows the cut-off frequencies from single to fivestages filters for different switching frequencies in order tocomply with MIL-STD-461E.
Filters with single to five stage are designed using equations(1), (8), and (9). Then, with the inductance and capacitancevalues from resulting from these cut-off frequencies, the sizeand weight of the different filter solutions can be estimated.
0
5
10
15
20
25
30
35
Filter Losses Estimations
Lo
sses W
0
500
1000
1500
2000
2500
3000
3500
Weig
ht
gr
Filter Weight Estimations
Single Stage
Two Stage
Three Stage
Four Stage
Five Stage
Fig. 5. Filter weight and losses estimation for single, two, three, four andfive stages.
TABLE ICUT-OFF FREQUENCY OF THE FILTER FOR SINGLE TO FIVE STAGES AND
DIFFERENT SWITCHING FREQUENCIES IN ORDER TO COMPLY
MIL-STD-461E.
cutoff freq 25kHz 40kHz 60kHz 80kHz 100kHzSingle stage 792Hz 1.0kHz 1.2kHz 1.4kHz 1.6kHzTwo stage 4.5kHz 6.3kHz 8.6kHz 10.7kHz 12.6kHz
Three stage 7.9kHz 11.7kHz 16.4kHz 20.9kHz 25.1kHzFour stage 10.6kHz 15.9kHz 22.7kHz 29.2kHz 35.5kHzFive stage 12.5kHz 19.1kHz 27.6kHz 35.7kHz 43.7kHz
For the weight estimation, the weight of the magnetic coresand of the wound wire are considered. The weight of thecapacitors is neglected, since it is much lower than that ofthe magnetic components.
For the losses estimation, only the winding losses areconsidered. The core losses are neglected because they arein the range of mW.
The fig. 5(a) shows the weight estimation results. A singlestage filter is not practical because the weight are considerablybigger than two, three, four, and five stage filters. Three,four and five stage filters provide only minimal improvements(if any) compared to two stage filters, and the number ofcomponents (and parasitic couplings between components) ismuch higher and because of that the reliability for a two stagefilter is higher. In fig 5(b) the filter losses estimations areshown. The two stage filters have lower losses than the restof filters for all frequencies. Consequently, a two stage filterappears to be the best solution.
Figura 9.9. Artıculo publicado en ECCE 2011. Pagina 4
98 Capıtulo 9 – Apendice
0
20
40
60
80
100
W
Total Losses in MOSFETs
0
5
10
15
20
25
30
Losses in each MOSFET
W
80
100
120
140
160
Temperature in MOSFETs
C
1 MOSFET
2 MOSFETs in Parallel
3 MOSFETs in Parallel
4 MOSFET in Parallel
Temperature limit according to manufacturer
Reability temperature limit
25kHz 40kHz 60kHz 80kHz 100kHz
25kHz 40kHz 60kHz 80kHz 100kHz
25kHz 40kHz 60kHz 80kHz 100kHz
Fig. 6. Power losses in MOSFETs considering since one to 4 MOSFETsin parallel. (a) shows total losses in MOSFETs, (b) shows losses in eachMOSFETs and (c) shows the MOSFETs temperature
B. Power losses estimation
The conducting losses are estimated using the currentstresses in the semiconductor [9]. The switching losses areestimated for considered switching frequencies [12] and [13].These calculations have been carried out using one, two,three, and four MOSFETs in parallel in order to decrease theconducing losses. For every combination of fs and the numberof MOSFET in parallel, the optimal device, with respect tothe power losses, has been selected from from a database ofavailable components. The power losses of the best MOSFETfor every combination switching frequency vs MOSFETs inparallel are shown in the fig. 6(a). The figs 6 (b) and (c) showthe power losses per MOSFET and the MOSFET temperaturerespectively.
For the same number of MOSFETs in parallel, the totallosses increase with the switching frequency, due to switchinglosses. In this case, at low switching frequency (25kHz and40kHz) from one MOSFET to two MOSFETs in parallel, thepower losses decrease considerably. In the same way, at 80kHzand 100kHz from two to three MOSFETs, the total lossesdecrease.
When increasing the number of MOSFETs in parallel, it notonly decreases the losses, but also the losses per MOSFETsand MOSFET temperature decrease drastically. The tempera-ture limit according to manufacture is 150C degrees, but forreliability considerations the temperature limit used is 75%of the manufacture limit (red line in the fig. 6(c)). Thereforeaccording to the MOSFET temperature at least it is necessary
to use two MOSFET in parallel. In addition, the reliabilitydecreases with the number of MOSFETs; therefore, it isapparent that using two MOSFETs in parallel is the optimum.
C. Optimum switching frequency
From the point of view of the filter weight, the highestpossible switching frequency is better; however, from the pointof view of the power losses in the MOSFETs, lower switchingfrequency is better. This is a typical tradeoff between efficiencyand power density.
In summary, the optimum number of stages of the inputfilter number is two, and the optimum number of MOSFET inparallel is two. For a two stages filter from 40kHz to 60kHz,and from 80kHz and 100kHz, there is no important differencein terms of the weight, fig 5(a); however, the losses for twoMOSFETs in parallel increase constantly with the switchingfrequency, fig 6(a). Therefore, with a switching frequency of80kHz, the best balance between losses and weight is obtained.
IV. CONCLUSION
In aircraft applications, the EMI standard to comply is MIL-STD-461E. This standard is more restrictive than the CISPR22 because of its frequency range, and it is not recommendableto consider a switching frequency below the standard range asused in industrial applications.
This work introduces new considerations in the input fil-ter design for a three-phase buck-type pulse-width modu-lation rectifier for aircraft applications. In this application,the switching frequency has a major impact in the trade-off between size of the input filter and the power lossesmore than that of industrial applications because in industrialapplications, the switching frequency, the second, third, fourthand fifth harmonics are outside of the standard frequencyrange.
In addition, in aircraft applications, the classical designmethod for the input filter considers only grid frequencies of50Hz or 60Hz; however, when the grid frequency is 400Hz, thepower factor for the system is not close to unity because theindustrial approach does not considerer how the inductancesaffect the system. This work proposes a new consideration inthe filter design considering effects of inductances in in orderto obtain a unitary power factor at full power, independentlyof the number of the filter stages.
In order to obtain an optimum switching frequency, theweight and power losses have been estimated for a 3.3kWsystem. According to these estimations, the best trade-offbetween weight and power losses is at 80kHz using a twostage filter and two MOSFETs in parallel. As it is presentedin this paper the reduction of the filter size from 80kHz to100kHz is marginal in comparison with the increment of thelosses. The filter weight at the optimum switching frequency(80kHz) is a quarter of the filter weight at 25kHz, which is atypical switching frequency in industrial applications.
Figura 9.10. Artıculo publicado en ECCE 2011. Pagina 5
9.1 Articulos Publicados en Congresos 99
REFERENCES
[1] T. Nussbaumer, M. L. Heldwein, and J. W. Kolar, “Differential modeinput filter design for a three-phase buck-type pwm rectifier based onmodeling of the emc test receiver,” vol. 53, no. 5, pp. 1649–1661, 2006.
[2] H. W. OTT, Electromagnetic Compatibility Engineering, Wiley, Ed.Wiley, 2009.
[3] R. P. CLAYTON, Introduction to Electromagnetic compatibility, secondedition. Wiley, 2006.
[4] C.I.S.P.R., Comite international special des perturbationsradioelectriques Std.
[5] A. Stupar, T. Friedli, J. Miniboandck, M. Schweizer, and J. Kolar,“Towards a 99% efficient three-phase buck-type pfc rectifier for 400 vdc distribution systems,” in Applied Power Electronics Conference andExposition (APEC), 2011 Twenty-Sixth Annual IEEE, march 2011, pp.505 –512.
[6] MIL-STD-461E, Requirements for the control of electromagnetic inter-ference characteristics of subsystems and equipment Std., August 1999.
[7] M. Baumann, U. Drofenik, and J. W. Kolar, “New wide input voltagerange three-phase unity power factor rectifier formed by integration of athree-switch buck-derived front-end and a dc/dc boost converter outputstage,” in Proc. INTELEC Telecommunications Energy Conf. Twenty-second Int, 2000, pp. 461–470.
[8] L. Malesani and P. Tenti, “Three-phase ac/dc pwm converter withsinusoidal ac currents and minimum filter requirements,” no. 1, pp. 71–77, 1987.
[9] T. Nussbaumer, M. Baumann, and J. W. Kolar, “Comprehensive designof a three-phase three-switch buck-type pwm rectifier,” vol. 22, no. 2,pp. 551–562, 2007.
[10] G. Gong, M. L. Heldwein, U. Drofenik, J. Minibock, K. Mino, andJ. W. Kolar, “Comparative evaluation of three-phase high-power-factorac-dc converter concepts for application in future more electric aircraft,”vol. 52, no. 3, pp. 727–737, 2005.
[11] M. L. Heldwein and J. W. Kolar, “Design of minimum volume emc inputfilters for an ultra compact three-phase pwm rectifier,” in COBEP07. The7th Brazilian Power Electronics Conference, 2007.
[12] Z. John Shen, Y. Xiong, X. Cheng, Y. Fu, and P. Kumar, “Powermosfet switching loss analysis: A new insight,” in Proc. 41st IAS AnnualMeeting Industry Applications Conf. Conf. Record of the 2006 IEEE,vol. 3, 2006, pp. 1438–1442.
[13] A. K. Dr. Dusan Graovac, Marco Purschel, MOSFET Power LossesCalculation Using the DataSheet Parameters, INFINEON, July 2006.
Figura 9.11. Artıculo publicado en ECCE 2011. Pagina 6
100 Capıtulo 9 – Apendice
9.2. Codigos Fuentes del DSP
En este apartado se muestran los codigos fuentes mas importantes en la progra-macion tanto de los perifericos del convertidor como el algoritmo propiamente talutilizado para la generacion de los pulsos de disparo de los MOSFETs a partir de latension de entrada del rectificador trifasico.
/∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗ File: Main 3.c −− File for Lab 3∗ Devices: TMS320F2802x∗ Author: Marcelo Silva Faundez∗ History:∗ 26/07/2011 − original∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗/
#include ”DSP2802x Device.h” // Peripheral address definitions#include ”Lab.h” // Main include file
//−−− Global VariablesUint16 AdcBuf[ADC BUF LEN]; // ADC buffer allocationUint16 DEBUG TOGGLE = 1; // Used for realtime mode investigation test
/∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗ Function: main()∗∗ Description: Main function for C28x workshop labs∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗/void main(void)//−−− CPU Initialization
InitSysCtrl(); // Initialize the CPU (FILE: SysCtrl.c)InitGpio(); // Initialize the shared GPIO pins (FILE: Gpio.c)InitPieCtrl(); // Initialize and enable the PIE (FILE: PieCtrl.c)InitWatchdog(); // Initialize the Watchdog Timer (FILE: WatchDog.c)
//−−− Peripheral InitializationInitAdc(); // Initialize the ADC (FILE: Adc.c)InitEPwm(); // Initialize the EPwm (FILE: EPwm.c)
//−−− Enable global interruptsasm(” CLRC INTM, DBGM”); // Enable global interrupts and realtime debug
//−−− Main Loopwhile(1) // endless loop − wait for an interrupt
asm(” NOP”);
//end of main()
/∗∗∗ end of file ∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗/
/∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗ File: Adc.c −− File for Lab 3 and 4 (not used in Lab 2)
9.2 Codigos Fuentes del DSP 101
∗ Devices: TMS320F2802x∗ Author: Marcelo Silva Faundez∗ History:∗ 26/07/2011 − original∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗/
#include ”DSP2802x Device.h” // Peripheral address definitions#include ”Lab.h” // Main include file
/∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗ Function: InitAdc()∗∗ Description: Initializes the ADC on the F2802x∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗/void InitAdc(void)
asm(” EALLOW”); // Enable EALLOW protected register access
//−−− Reset the ADC module// Note: The ADC is already reset after a DSP reset, but this example is just showing// good coding practice to reset the peripheral before configuring it as you never// know why the DSP has started the code over again from the beginning).
AdcRegs.ADCCTL1.bit.RESET = 1; // Reset the ADC
// Must wait 2 ADCCLK periods for the reset to take effect.// Note that ADCCLK = SYSCLKOUT for F2802x/F2803x devices.
asm(” NOP”);asm(” NOP”);
//−−− Power−up and configure the ADCAdcRegs.ADCCTL1.all = 0x00E4; // Power−up reference and main ADC
// bit 15 0: RESET, ADC software reset, 0=no effect, 1=resets the ADC// bit 14 0: ADCENABLE, ADC enable, 0=disabled, 1=enabled// bit 13 0: ADCBSY, ADC busy, read−only// bit 12−8 0’s: ADCBSYCHN, ADC busy channel, read−only// bit 7 1: ADCPWDN, ADC power down, 0=powered down, 1=powered up// bit 6 1: ADCBGPWD, ADC bandgap power down, 0=powered down, 1=powered up// bit 5 1: ADCREFPWD, ADC reference power down, 0=powered down, 1=powered up// bit 4 0: reserved// bit 3 0: ADCREFSEL, ADC reference select, 0=internal, 1=external// bit 2 1: INTPULSEPOS, INT pulse generation, 0=start of conversion, 1=end of conversion// bit 1 0: VREFLOCONV, VREFLO convert, 0=VREFLO not connected, 1=VREFLO connected to B5// bit 0 0: Must write as 0.
DelayUs(1000); // Wait 1 ms after power−up before using the ADC
//−−− SOC0 configurationAdcRegs.ADCSAMPLEMODE.bit.SIMULEN0 = 0; // SOC0 in single sample mode (vs. simultaneous mode)AdcRegs.ADCSAMPLEMODE.bit.SIMULEN0 = 0; // SOC0 in single sample mode (vs. simultaneous mode)
AdcRegs.ADCSOC0CTL.bit.TRIGSEL = 6; // Trigger using ePWM1−ADCSOCAAdcRegs.ADCSOC0CTL.bit.CHSEL = 1; // Convert channel ADCINA0 (ch0)AdcRegs.ADCSOC0CTL.bit.ACQPS = 10; // Acquisition window set to (6+1)=7 cycles
AdcRegs.ADCSOC1CTL.bit.TRIGSEL = 6; // Trigger using ePWM1−ADCSOCA
102 Capıtulo 9 – Apendice
AdcRegs.ADCSOC1CTL.bit.CHSEL = 9; // Convert channel ADCINA0 (ch0)AdcRegs.ADCSOC1CTL.bit.ACQPS = 10; // Acquisition window set to (6+1)=7 cycles
AdcRegs.ADCSOC2CTL.bit.TRIGSEL = 6; // Trigger using ePWM1−ADCSOCAAdcRegs.ADCSOC2CTL.bit.CHSEL = 1; // Convert channel ADCINA0 (ch0)AdcRegs.ADCSOC2CTL.bit.ACQPS = 10; // Acquisition window set to (6+1)=7 cycles
AdcRegs.ADCSOC3CTL.bit.TRIGSEL = 6; // Trigger using ePWM1−ADCSOCAAdcRegs.ADCSOC3CTL.bit.CHSEL = 9; // Convert channel ADCINA0 (ch0)AdcRegs.ADCSOC3CTL.bit.ACQPS = 10; // Acquisition window set to (6+1)=7 cycles
AdcRegs.SOCPRICTL.bit.SOCPRIORITY = 4; // All SOCs handled in round−robin mode
//−−− ADCINT0 configurationAdcRegs.INTSEL1N2.bit.INT1CONT = 0; // ADCINT0 pulses generated only when ADCINT0 flag is clearAdcRegs.INTSEL1N2.bit.INT1E = 1; // Enable ADCINT0AdcRegs.INTSEL1N2.bit.INT1SEL = 3; // EOC0 triggers ADCINT0
PieCtrlRegs.PIEIER1.bit.INTx1 = 1; // Enable ADCINT0 in PIE group 1IER |= 0x0001; // Enable INT1 in IER to enable PIE group
//−−− Finish upAdcRegs.ADCCTL1.bit.ADCENABLE = 1; // Enable the ADCasm(” EDIS”); // Disable EALLOW protected register access
// end InitAdc()
//−−− end of file −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
/∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗ File: EPwm.c −− File for Lab 3 and 4∗ Devices: TMS320F2802x∗ Author: Marcelo Silva Faundez∗ History:∗ 26/07/2011 − original∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗/
#include ”DSP2802x Device.h” // Peripheral address definitions#include ”Lab.h” // Main include file
int periodo = 1000;
/∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗ Function: InitEPwm()∗∗ Description: Initializes the Enhanced PWM modules on the F2802x∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗/void InitEPwm(void)//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
9.2 Codigos Fuentes del DSP 103
//−−− Must disable the clock to the ePWM modules if you//−−− want all ePMW modules synchronized.//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
asm(” EALLOW”); // Enable EALLOW protected register accessSysCtrlRegs.PCLKCR0.bit.TBCLKSYNC = 0;asm(” EDIS”); // Disable EALLOW protected register access
//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−//−−− Configure ePWM1 to trigger the ADC at a 50 kHz rate//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
EPwm1Regs.TBCTL.bit.CTRMODE = 0x3; // Disable the timer
EPwm1Regs.TBCTL.all = 0xC033; // Configure timer control register// bit 15−14 11: FREE/SOFT, 11 = ignore emulation suspend// bit 13 0: PHSDIR, 0 = count down after sync event// bit 12−10 000: CLKDIV, 000 => TBCLK = HSPCLK/1// bit 9−7 000: HSPCLKDIV, 000 => HSPCLK = SYSCLKOUT/1// bit 6 0: SWFSYNC, 0 = no software sync produced// bit 5−4 11: SYNCOSEL, 11 = sync−out disabled// bit 3 0: PRDLD, 0 = reload PRD on counter=0// bit 2 0: PHSEN, 0 = phase control disabled// bit 1−0 11: CTRMODE, 11 = timer stopped (disabled)
EPwm1Regs.TBCTR = 0x0000; // Clear timer counterEPwm1Regs.TBPRD = periodo; // Set timer periodEPwm1Regs.TBPHS.half.TBPHS = 0x0000; // Set timer phase
//EPwm1Regs.CMPA.half.CMPA = (periodo>>2); // MuestreoEPwm1Regs.CMPB = 35;
EPwm1Regs.ETPS.all = 0x1100; // Configure SOCA// bit 15−14 00: EPWMxSOCB, read−only// bit 13−12 01: SOCBPRD, 01 = generate SOCB on first event// bit 11−10 00: EPWMxSOCA, read−only// bit 9−8 01: SOCAPRD, 01 = generate SOCA on first event// bit 7−4 0000: reserved// bit 3−2 00: INTCNT, don’t care// bit 1−0 00: INTPRD, don’t care
EPwm1Regs.ETSEL.all = 0xF000; // Enable SOCA to ADC// bit 15 1: SOCBEN, 1 = enable SOCB// bit 14−12 111: SOCBSEL, 100 compB incrementando// bit 11 0: SOCAEN, 1 = enable SOCA// bit 10−8 010: SOCASEL, 010 TBCTR=TBPRD// bit 7−4 0000: reserved// bit 3 0: INTEN, 0 = disable interrupt// bit 2−0 000: INTSEL, don’t care
EPwm1Regs.AQCTLB.all = 0x0801;// bit 15−12 0000: reserved// bit 11−10 10: CBD, 10 = set// bit 9−8 00: CBU, 00 = do nothing// bit 7−6 00: CAD, 01 = do nothing// bit 5−4 00: CAU, 10 = do nothing
104 Capıtulo 9 – Apendice
// bit 3−2 00: PRD, 00 = do nothing// bit 1−0 01: ZRO, 01 = clear
EPwm1Regs.DBCTL.bit.OUT MODE = 0; // Deadband disabledEPwm1Regs.PCCTL.bit.CHPEN = 0; // PWM chopper unit disabledEPwm1Regs.TZCTL.bit.TZA = 0x3; // Trip action disabled for output A
EPwm1Regs.TBCTL.bit.CTRMODE = 0x2; // Enable the timer in count up mode 0x0
//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−//−−− Configure ePWM1 for 2 kHz symmetric PWM on EPWM1A pin//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
EPwm2Regs.TBCTL.bit.CTRMODE = 0x3; // Disable the timer
EPwm2Regs.TBCTL.all = 0xC033; // Configure timer control register// bit 15−14 11: FREE/SOFT, 11 = ignore emulation suspend// bit 13 0: PHSDIR, 0 = count down after sync event// bit 12−10 000: CLKDIV, 000 => TBCLK = HSPCLK/1// bit 9−7 000: HSPCLKDIV, 000 => HSPCLK = SYSCLKOUT/1// bit 6 0: SWFSYNC, 0 = no software sync produced// bit 5−4 11: SYNCOSEL, 11 = sync−out disabled// bit 3 0: PRDLD, 0 = reload PRD on counter=0// bit 2 0: PHSEN, 0 = phase control disabled// bit 1−0 11: CTRMODE, 11 = timer stopped (disabled)
EPwm2Regs.TBCTR = 0x0000; // Clear timer counterEPwm2Regs.TBPRD = periodo; // Set timer periodEPwm2Regs.TBPHS.half.TBPHS = 0x0000; // Set timer phase
EPwm2Regs.CMPA.half.CMPA = (periodo−50); // Set PWM duty cycleEPwm2Regs.CMPB = (periodo−50);
EPwm2Regs.CMPCTL.all = 0x0002; // Compare control register// bit 15−10 0’s: reserved// bit 9 0: SHDWBFULL, read−only// bit 8 0: SHDWAFULL, read−only// bit 7 0: reserved// bit 6 0: SHDWBMODE, don’t care// bit 5 0: reserved// bit 4 0: SHDWAMODE, 0 = shadow mode// bit 3−2 00: LOADBMODE, don’t care// bit 1−0 10: LOADAMODE, 10 = load on zero or PRD match
EPwm2Regs.AQCTLA.all = 0x0090; // Action−qualifier control register A// bit 15−12 0000: reserved// bit 11−10 00: CBD, 00 = do nothing// bit 9−8 00: CBU, 00 = do nothing// bit 7−6 01: CAD, 01 = clear// bit 5−4 10: CAU, 10 = set// bit 3−2 00: PRD, 00 = do nothing// bit 1−0 00: ZRO, 00 = do nothing
9.2 Codigos Fuentes del DSP 105
EPwm2Regs.AQCTLB.all = 0x0900; // Action−qualifier control register A// bit 15−12 0000: reserved// bit 11−10 01: CBD, 00 = clear// bit 9−8 10: CBU, 00 = set// bit 7−6 00: CAD, 01 = do nothing// bit 5−4 00: CAU, 10 = do nothing// bit 3−2 00: PRD, 00 = do nothing// bit 1−0 00: ZRO, 00 = do nothing
EPwm2Regs.AQSFRC.all = 0x0000; // Action−qualifier s/w force register// bit 15−8 0’s: reserved// bit 7−6 00: RLDCSF, 00 = reload AQCSFRC on zero// bit 5 0: OTSFB, 0 = do not initiate a s/w forced event on output B// bit 4−3 00: ACTSFB, don’t care// bit 2 0: OTSFA, 0 = do not initiate a s/w forced event on output A// bit 1−0 00: ACTSFA, don’t care
EPwm2Regs.AQCSFRC.all = 0x0000; // Action−qualifier continuous s/w force register// bit 15−4 0’s: reserved// bit 3−2 00: CSFB, 00 = forcing disabled// bit 1−0 00: CSFA, 00 = forcing disabled
EPwm2Regs.DBCTL.bit.OUT MODE = 0; // Deadband disabledEPwm2Regs.PCCTL.bit.CHPEN = 0; // PWM chopper unit disabledEPwm2Regs.TZCTL.bit.TZA = 0x3; // Trip action disabled for output A
EPwm2Regs.TBCTL.bit.CTRMODE = 0x2; // Enable the timer in count up/down mode
//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−//−−− Configure ePWM1 for 2 kHz symmetric PWM on EPWM1A pin//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
EPwm3Regs.TBCTL.bit.CTRMODE = 0x3; // Disable the timer
EPwm3Regs.TBCTL.all = 0xC033; // Configure timer control register// bit 15−14 11: FREE/SOFT, 11 = ignore emulation suspend// bit 13 0: PHSDIR, 0 = count down after sync event// bit 12−10 000: CLKDIV, 000 => TBCLK = HSPCLK/1// bit 9−7 000: HSPCLKDIV, 000 => HSPCLK = SYSCLKOUT/1// bit 6 0: SWFSYNC, 0 = no software sync produced// bit 5−4 11: SYNCOSEL, 11 = sync−out disabled// bit 3 0: PRDLD, 0 = reload PRD on counter=0// bit 2 0: PHSEN, 0 = phase control disabled// bit 1−0 11: CTRMODE, 11 = timer stopped (disabled)
EPwm3Regs.TBCTR = 0x0000; // Clear timer counterEPwm3Regs.TBPRD = periodo; // Set timer periodEPwm3Regs.TBPHS.half.TBPHS = 0x0000; // Set timer phase
EPwm3Regs.CMPA.half.CMPA = (periodo>>1); // Set PWM duty cycleEPwm3Regs.CMPB = (periodo>>1);
EPwm3Regs.CMPCTL.all = 0x0002; // Compare control register// bit 15−10 0’s: reserved// bit 9 0: SHDWBFULL, read−only
106 Capıtulo 9 – Apendice
// bit 8 0: SHDWAFULL, read−only// bit 7 0: reserved// bit 6 0: SHDWBMODE, don’t care// bit 5 0: reserved// bit 4 0: SHDWAMODE, 0 = shadow mode// bit 3−2 00: LOADBMODE, don’t care// bit 1−0 10: LOADAMODE, 10 = load on zero or PRD match
EPwm3Regs.AQCTLA.all = 0x0090; // Action−qualifier control register A// bit 15−12 0000: reserved// bit 11−10 00: CBD, 00 = do nothing// bit 9−8 00: CBU, 00 = do nothing// bit 7−6 01: CAD, 01 = clear// bit 5−4 10: CAU, 10 = set// bit 3−2 00: PRD, 00 = do nothing// bit 1−0 00: ZRO, 00 = do nothing
EPwm3Regs.AQCTLB.all = 0x0900; // Action−qualifier control register A// bit 15−12 0000: reserved// bit 11−10 01: CBD, 00 = clear// bit 9−8 10: CBU, 00 = set// bit 7−6 00: CAD, 01 = do nothing// bit 5−4 00: CAU, 10 = do nothing// bit 3−2 00: PRD, 00 = do nothing// bit 1−0 00: ZRO, 00 = do nothing
EPwm3Regs.AQSFRC.all = 0x0000; // Action−qualifier s/w force register// bit 15−8 0’s: reserved// bit 7−6 00: RLDCSF, 00 = reload AQCSFRC on zero// bit 5 0: OTSFB, 0 = do not initiate a s/w forced event on output B// bit 4−3 00: ACTSFB, don’t care// bit 2 0: OTSFA, 0 = do not initiate a s/w forced event on output A// bit 1−0 00: ACTSFA, don’t care
EPwm3Regs.AQCSFRC.all = 0x0000; // Action−qualifier continuous s/w force register// bit 15−4 0’s: reserved// bit 3−2 00: CSFB, 00 = forcing disabled// bit 1−0 00: CSFA, 00 = forcing disabled
EPwm3Regs.DBCTL.bit.OUT MODE = 0; // Deadband disabledEPwm3Regs.PCCTL.bit.CHPEN = 0; // PWM chopper unit disabledEPwm3Regs.TZCTL.bit.TZA = 0x3; // Trip action disabled for output A
EPwm3Regs.TBCTL.bit.CTRMODE = 0x2; // Enable the timer in count up/down mode
//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−//−−− Configure ePWM1 for 2 kHz symmetric PWM on EPWM1A pin//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
EPwm4Regs.TBCTL.bit.CTRMODE = 0x3; // Disable the timer
EPwm4Regs.TBCTL.all = 0xC033; // Configure timer control register// bit 15−14 11: FREE/SOFT, 11 = ignore emulation suspend// bit 13 0: PHSDIR, 0 = count down after sync event// bit 12−10 000: CLKDIV, 000 => TBCLK = HSPCLK/1
9.2 Codigos Fuentes del DSP 107
// bit 9−7 000: HSPCLKDIV, 000 => HSPCLK = SYSCLKOUT/1// bit 6 0: SWFSYNC, 0 = no software sync produced// bit 5−4 11: SYNCOSEL, 11 = sync−out disabled// bit 3 0: PRDLD, 0 = reload PRD on counter=0// bit 2 0: PHSEN, 0 = phase control disabled// bit 1−0 11: CTRMODE, 11 = timer stopped (disabled)
EPwm4Regs.TBCTR = 0x0000; // Clear timer counterEPwm4Regs.TBPRD = periodo; // Set timer periodEPwm4Regs.TBPHS.half.TBPHS = 0x0000; // Set timer phase
EPwm4Regs.CMPA.half.CMPA = (periodo>>1); // Set PWM duty cycleEPwm4Regs.CMPB = (periodo>>1);
EPwm4Regs.CMPCTL.all = 0x0002; // Compare control register// bit 15−10 0’s: reserved// bit 9 0: SHDWBFULL, read−only// bit 8 0: SHDWAFULL, read−only// bit 7 0: reserved// bit 6 0: SHDWBMODE, don’t care// bit 5 0: reserved// bit 4 0: SHDWAMODE, 0 = shadow mode// bit 3−2 00: LOADBMODE, don’t care// bit 1−0 10: LOADAMODE, 10 = load on zero or PRD match
EPwm4Regs.AQCTLA.all = 0x0090; // Action−qualifier control register A// bit 15−12 0000: reserved// bit 11−10 00: CBD, 00 = do nothing// bit 9−8 00: CBU, 00 = do nothing// bit 7−6 01: CAD, 01 = clear// bit 5−4 10: CAU, 10 = set// bit 3−2 00: PRD, 00 = do nothing// bit 1−0 00: ZRO, 00 = do nothing
EPwm4Regs.AQCTLB.all = 0x0900; // Action−qualifier control register A// bit 15−12 0000: reserved// bit 11−10 01: CBD, 00 = clear// bit 9−8 10: CBU, 00 = set// bit 7−6 00: CAD, 01 = do nothing// bit 5−4 00: CAU, 10 = do nothing// bit 3−2 00: PRD, 00 = do nothing// bit 1−0 00: ZRO, 00 = do nothing
EPwm4Regs.AQSFRC.all = 0x0000; // Action−qualifier s/w force register// bit 15−8 0’s: reserved// bit 7−6 00: RLDCSF, 00 = reload AQCSFRC on zero// bit 5 0: OTSFB, 0 = do not initiate a s/w forced event on output B// bit 4−3 00: ACTSFB, don’t care// bit 2 0: OTSFA, 0 = do not initiate a s/w forced event on output A// bit 1−0 00: ACTSFA, don’t care
EPwm4Regs.AQCSFRC.all = 0x0000; // Action−qualifier continuous s/w force register// bit 15−4 0’s: reserved// bit 3−2 00: CSFB, 00 = forcing disabled// bit 1−0 00: CSFA, 00 = forcing disabled
108 Capıtulo 9 – Apendice
EPwm4Regs.DBCTL.bit.OUT MODE = 0; // Deadband disabledEPwm4Regs.PCCTL.bit.CHPEN = 0; // PWM chopper unit disabledEPwm4Regs.TZCTL.bit.TZA = 0x3; // Trip action disabled for output A
EPwm4Regs.TBCTL.bit.CTRMODE = 0x2; // Enable the timer in count up/down mode
//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−//−−− Enable the clocks to the ePWM module.//−−− Note: this should be done after all ePWM modules are configured//−−− to ensure synchronization between the ePWM modules.//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
asm(” EALLOW”); // Enable EALLOW protected register accessSysCtrlRegs.PCLKCR0.bit.TBCLKSYNC = 1; // TBCLK to ePWM modules enabledasm(” EDIS”); // Disable EALLOW protected register access
// end InitEPwm()
//−−− end of file −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
/∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗ File: SysCtrl.c −− File for Lab 2, 3 and 4∗ Devices: TMS320F2802x∗ Author: Marcelo Silva Faundez∗ History:∗ 26/07/2011 − original∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗/#include ”DSP2802x Device.h” // Peripheral address definitions#include ”Lab.h” // Main include file
/∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗ Function: InitSysCtrl()∗ Description: Initializes the F2802x CPU∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗/void InitSysCtrl(void)volatile Uint16 i; // General purpose Uint16volatile int16 dummy; // General purpose volatile int16
asm(” EALLOW”); // Enable EALLOW protected register access
//−−− Call the Device cal() function located in reserved OTP.// Device cal func ptr is a macro defined in the file Lab.h. This// macro simply defines Device cal func ptr to be a function pointer to// the correct address in the reserved OTP. Note that the device cal function// is automatically called by the bootloader. A call to this function is// included here just in case the bootloader is bypassed during development.
SysCtrlRegs.PCLKCR0.bit.ADCENCLK = 1; // Enable ADC peripheral clock(∗Device cal func ptr)(); // Call the device calibration routineSysCtrlRegs.PCLKCR0.bit.ADCENCLK = 0; // Disable ADC peripheral clock
9.2 Codigos Fuentes del DSP 109
//−−− Memory Protection ConfigurationDevEmuRegs.DEVICECNF.bit.ENPROT = 1; // Enable write/read protection
// Configure the clock sources
SysCtrlRegs.CLKCTL.all = 0x6400; // Configure the clock control register// bit 15 0: NMIRESETSEL, NMI reset select for missing clock detect, 0=direct reset, 1=NMI WD causes reset// bit 14 1: XTALOSCOFF, Crystal oscillator off, 0=on, 1=off// bit 13 1: XCLKINOFF, XCLKIN input off, 0=on, 1=off// bit 12 0: WDHALTI, WD halt ignore, 0=WD turned off by HALT, 1=WD ignores HALT// bit 11 0: INTOSC2HALTI, OSC2 halt ignore, 0=OSC2 turned off by HALT, 1=OSC2 ignores HALT// bit 10 1: INTOSC2OFF, OSC2 off, 0=on, 1=off// bit 9 0: INTOSC1HALTI, OSC1 halt ignore, 0=OSC1 turned off by HALT, 1=OSC1 ignores HALT// bit 8 0: INTOSC1OFF, OSC1 off, 0=on, 1=off// bit 7−5 000: TMR2CLKPRESCALE, Timer2 prescale, 000=/1// bit 4−3 00: TMR2CLKSCRSEL, Timer2 source, 00=SYSCLKOUT, 01=ext. OSC, 10=OSC1, 11=OSC2// bit 2 0: WDCLKSRCSEL, WD source, 0=OSC1, 1=ext. OSC or OSC2// bit 1 0: OSCCLKSRC2SEL, clock src2 select, 0=ext OSC, 1=OSC2// bit 0 0: OSCCLKSRCSEL, clock src1 select, 0=OSC1, 1=ext. OSC or OSC2
SysCtrlRegs.XCLK.all = 0x0040; // Configure the clocking register// bit 15−7 0’s: reserved// bit 6 1: XCLKINSEL, XCLKIN source, 0=GPIO38, 1=GPIO19 (default) (Note: GPIO38 is JTAG TCK!)// bit 5−2 0’s: reserved// bit 1−0 00: XCLKOUTDIV, XCLKOUT divide ratio, 00=SYSCLKOUT/4, 01=SYSCLKOUT/.2, 10=SYSCLKOUT, 11=off
//−−− Configure the PLL
// Note: The DSP/BIOS configuration tool can also be used to initialize the PLL// instead of doing the initialization here.
// Make sure the PLL is not running in limp modeif (SysCtrlRegs.PLLSTS.bit.MCLKSTS != 1) // PLL is not running in limp mode
SysCtrlRegs.PLLSTS.bit.MCLKOFF = 1; // Turn off missing clock detect before changing PLLCRSysCtrlRegs.PLLSTS.bit.DIVSEL = 0; // DIVSEL must be 0 or 1 (/4 CLKIN mode) before changing PLLCRSysCtrlRegs.PLLCR.bit.DIV = 0x000C; // PLLx12/4 (because DIVSEL is /4)
// Wait for PLL to lock.// During this time the CPU will run at OSCCLK/4 until the PLL is stable.// Once the PLL is stable the CPU will automatically switch to the new PLL value.// Code is not required to sit and wait for the PLL to lock. However,// if the code does anything that is timing critical (e.g. something that// relies on the CPU clock frequency to be at speed), then it is best to wait// until PLL lock is complete. The watchdog should be disabled before this loop// (e.g., as was done above), or fed within the loop.while(SysCtrlRegs.PLLSTS.bit.PLLLOCKS != 1) // Wait for PLLLOCKS bit to set
SysCtrlRegs.WDKEY = 0x0055; // Service the watchdog while waitingSysCtrlRegs.WDKEY = 0x00AA; // in case the user enabled it.
// After the PLL has locked, we are running in PLLx12/4 mode (since DIVSEL is /4).// We can now enable the missing clock detect circuitry, and also change DIVSEL// to /2. In this example, I will wait a bit of time to let inrush currents settle,
110 Capıtulo 9 – Apendice
// and then change DIVSEL from /4 to /2. This is only an example. The amount of// time you need to wait depends on the power supply feeding the DSP (i.e., how much// voltage droop occurs due to the inrush currents, and how long it takes the// voltage regulators to recover).SysCtrlRegs.PLLSTS.bit.MCLKOFF = 0; // Enable missing clock detect circuitryDelayUs(20/2); // Wait 20 us (just an example). Remember we’re running
// at half−speed here, so divide function argument by 2.SysCtrlRegs.PLLSTS.bit.DIVSEL = 3; // Change to /2 mode
else // PLL is running in limp mode// User should replace the below with a call to an appropriate function,// for example shutdown the system (since something is very wrong!).
asm(” ESTOP0”);
//−−− Configure the clocksSysCtrlRegs.LOSPCP.all = 0x0002; // Lo−speed periph clock prescaler, LOSPCLK=SYSCLKOUT/4
SysCtrlRegs.PCLKCR3.bit.GPIOINENCLK = 1; // GPIO input module is clockedSysCtrlRegs.PCLKCR3.bit.CPUTIMER2ENCLK = 1; // SYSCLKOUT to CPU Timer2 enabledSysCtrlRegs.PCLKCR3.bit.CPUTIMER1ENCLK = 1; // SYSCLKOUT to CPU Timer1 enabledSysCtrlRegs.PCLKCR3.bit.CPUTIMER0ENCLK = 1; // SYSCLKOUT to CPU Timer0 enabledSysCtrlRegs.PCLKCR3.bit.COMP2ENCLK = 1; // SYSCLKOUT to COMP2 enabledSysCtrlRegs.PCLKCR3.bit.COMP1ENCLK = 1; // SYSCLKOUT to COMP1 enabled
SysCtrlRegs.PCLKCR1.bit.ECAP1ENCLK = 1; // SYSCLKOUT to eCAP1 enabledSysCtrlRegs.PCLKCR1.bit.EPWM4ENCLK = 1; // SYSCLKOUT to ePWM4 enabledSysCtrlRegs.PCLKCR1.bit.EPWM3ENCLK = 1; // SYSCLKOUT to ePWM3 enabledSysCtrlRegs.PCLKCR1.bit.EPWM2ENCLK = 1; // SYSCLKOUT to ePWM2 enabledSysCtrlRegs.PCLKCR1.bit.EPWM1ENCLK = 1; // SYSCLKOUT to ePWM1 enabled
SysCtrlRegs.PCLKCR0.bit.SCIAENCLK = 1; // LSPCLK to SCI−A enabledSysCtrlRegs.PCLKCR0.bit.SPIAENCLK = 1; // LSPCLK to SPI−A enabledSysCtrlRegs.PCLKCR0.bit.I2CAENCLK = 1; // SYSCLKOUT to I2C−A enabledSysCtrlRegs.PCLKCR0.bit.ADCENCLK = 1; // SYSCLKOUT to ADC enabled
// TBCLKSYNC bit is handled separately in InitEPwm() since it affects ePWM synchronization.SysCtrlRegs.PCLKCR0.bit.HRPWMENCLK = 1; // SYSCLKOUT to HRPWM enabled
//−−− Configure the low−power modesSysCtrlRegs.LPMCR0.all = 0x00FC; // LPMCR0 set to default value
//−−− Finish upasm(” EDIS”); // Disable EALLOW protected register access
// end InitSysCtrl()
//−−− end of file −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
/∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗ File: DefaultIsr 3 4.c −− File for Lab 3 and 4∗ Devices: TMS320F2802x∗ Author: Marcelo Silva Faundez∗ History:∗ 26/07/2011 − original
9.2 Codigos Fuentes del DSP 111
∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗/
#include ”DSP2802x Device.h” // Peripheral address definitions#include ”Lab.h” // Main include file
int mode = 0;long coefa[] = 16384,31972,15616;
long coefb[] = 384,384;int offset0 = 2380;//(0.62V)int offset1 = 2380;//(0.62V)int fil 1[] = 0,0,0,0,0,0,0,0;int fil 2[] = 0,0,0,0,0,0,0,0;long buff1[192] = ;//vector de 192 datoslong buff2[192] = ;//vector de 192 datoslong vrs f k1 = 0;long vrs f k2 = 0;long vrs f k3 = 0;int vrs f = 0;long vrs f l = 0;long usa k0 = 0;long usa k1 = 0;long usa k2 = 0;long vst f k1 = 0;long vst f k2 = 0;long vst f k3 = 0;int vst f = 0;long vst f l = 0;long usb k0 = 0;long usb k1 = 0;long usb k2 = 0;int di = 12;int i = 0;int ip = 0;int h = 10;int aveg0 = 0;int aveg1 = 0;int Vrs = 0;int Vst = 0;int y1 = 0;int y2 = 0;int y3 = 0;int y4 = 0;int y5 = 0;int y6 = 0;int ay1 = 0;int ay2 = 0;int ay3 = 0;int abs x1 = 0;int abs x2 = 0;int abs x3 = 0;int Vrt = 0;int Vr = 0;int Vs = 0;int Vt = 0;long auxrs1 = 0;
112 Capıtulo 9 – Apendice
long auxrs2 = 0;long auxrs3 = 0;long auxrs4 = 0;long auxrs5 = 0;long auxst1 = 0;long auxst2 = 0;long auxst3 = 0;long auxst4 = 0;long auxst5 = 0;int outadc0 = 0;int outadc1 = 0;int outadc2 = 0;int outadc3 = 0;int outadc4 = 0;int outadc5 = 0;int outadc6 = 0;int outadc7 = 0;int outadc8 = 0;int outadc9 = 0;int outadc10 = 0;int outadc11 = 0;int Vrs 1 = 0;int Vrs 2 = 0;int Vrt 1 = 0;int Vrt 2 = 0;int avg1 = 0;int avg2 = 0;int i2 = 0;
//−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−interrupt void ADCINT1 ISR(void) // PIE1.1 @ 0x000D40 ADCINT1 (ADC INT1)
GpioDataRegs.GPBDAT.bit.GPIO32 = 1;
PieCtrlRegs.PIEACK.all = PIEACK GROUP1; // Must acknowledge the PIE group
//−−− Manage the ADC registersAdcRegs.ADCINTFLGCLR.bit.ADCINT1 = 1; // Clear ADCINT1 flag
avg1 = (AdcResult.ADCRESULT0 + AdcResult.ADCRESULT2)>>1;avg2 = (AdcResult.ADCRESULT1 + AdcResult.ADCRESULT3)>>1;
Vrs = avg1 − offset0;Vrt = avg2 − offset1;
//Filtro Digitalvrs f k3 = vrs f k2;vrs f k2 = vrs f k1;vrs f k1 = vrs f l;usa k2 = usa k1;
9.2 Codigos Fuentes del DSP 113
usa k1 = usa k0;usa k0 = (long)Vrs;auxrs1 = (7715∗vrs f k1)>>12;auxrs2 = (−3627∗vrs f k2)>>12;auxrs3 = (7505∗usa k0)>>17;auxrs4 = (−7505∗usa k2)>>17;vrs f l = auxrs1 + auxrs2 + auxrs3 + auxrs4;auxrs5 = (vrs f l+1020);//1500 es la amplitudvrs f = (int)auxrs5;//vrs f = Vrs;
vst f k3 = vst f k2;vst f k2 = vst f k1;vst f k1 = vst f l;usb k2 = usb k1;usb k1 = usb k0;usb k0 = (long)Vrt;auxst1 = (7715∗vst f k1)>>12;auxst2 = (−3627∗vst f k2)>>12;auxst3 = (7505∗usb k0)>>17;auxst4 = (−7505∗usb k2)>>17;vst f l = auxst1 + auxst2 + auxst3 + auxst4;auxst5 = (vst f l+1020);//1500 es la amplitud
vst f = (int)auxst5;//vst f = Vrt;
ip = 1 + ip;if (ip>=h)buff1[i] = auxst5;buff2[i] = (long)Vrt;i = i+1;ip = 0;if ( i>= 192)i = 0;
//Tension filtrada fase neutro, falta la division por 3Vr = (((2∗vrs f) + vst f))>>2;Vs = (((−1)∗vrs f) + vst f)>>2;Vt = (((−1)∗vrs f) − (2∗vst f))>>2;//Maximo valor paa Vr = 1200∗3/(sqrt(3)∗2) = 1039 equivale a 162V//1000 es el periodo, entonces existe un ciclo de trabajo de 1.1
// 1039∗7/8 = 909, es decir, ciclo de trabajo 0.9Vr = ((Vr∗27)>>4);Vs = ((Vs∗27)>>4);Vt = ((Vt∗27)>>4);
//Valor Absoluto
114 Capıtulo 9 – Apendice
if (Vr < 0)
abs x1 = (−1)∗Vr;else
abs x1 = Vr;
if (Vs < 0)
abs x2 = (−1)∗Vs;else
abs x2 = Vs;
if (Vt < 0)
abs x3 = (−1)∗Vt;else
abs x3 = Vt;
// Modulacionif (abs x1 >= abs x2 && abs x2 >= abs x3)ay1 = abs x3 + abs x2;ay2 = abs x2;ay3 = 3000;
else if (abs x1 >= abs x3 && abs x3 >= abs x2)
ay1 = abs x2 + abs x3;ay2 = 3000;ay3 = abs x3;
else if (abs x2 >= abs x1 && abs x1 >= abs x3)ay1 = abs x1;ay2 = abs x3 + abs x1;ay3 = 3000;
else if (abs x2 >= abs x3 && abs x3 >= abs x1)
ay1 = 3000;ay2 = abs x1 + abs x3;ay3 = abs x3;
9.2 Codigos Fuentes del DSP 115
else if (abs x3 >= abs x2 && abs x2 >= abs x1)
ay1 = 3000;ay2 = abs x2;
ay3 = abs x1 + abs x2;
elseay1 = abs x1;ay2 = 3000;ay3 = abs x2 + abs x1;
if (Vr > 0)
y1 = ay1;y4 = 0;
else
y4 = ay1;y1 = 0;
if (Vs > 0)
y2 = ay2;y5 = 0;
else
y5 = ay2;y2 = 0;
if (Vt > 0)
y3 = ay3;y6 = 0;
else
y6 = ay3;y3 = 0;
//SalidasEPwm2Regs.CMPA.half.CMPA = y4;//Negativo REPwm2Regs.CMPB = y1;//Positivo R
EPwm3Regs.CMPA.half.CMPA = y5;//Negativo S
116 Capıtulo 9 – Apendice
EPwm3Regs.CMPB = y2;//Positivo S
EPwm4Regs.CMPA.half.CMPA = y6;//Negativo TEPwm4Regs.CMPB = y3;//Positivo T
if (mode == 0)
EPwm1Regs.CMPA.half.CMPA = ((Vr>>1)+500);else if (mode == 1)
EPwm1Regs.CMPA.half.CMPA = ((Vs>>1)+500);else
EPwm1Regs.CMPA.half.CMPA = ((Vt>>1)+500);
GpioDataRegs.GPBDAT.bit.GPIO32 = 0;
//−−− end of file −−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
Capıtulo 10
REFERENCIAS
117
118 Capıtulo 10 – Referencias
REFERENCIAS
[1] J. A. Rosero, J. A. Ortega, E. Aldabas, and L. Romeral, “Moving towards a moreelectric aircraft,” IEEE Aerospace and Electronic Systems Magazine, vol. 22, no. 3, pp.3–9, 2007.
[2] J. S. Cloyd, “Status of the united states air force’s more electric aircraft initiative,”IEEE Aerospace and Electronic Systems Magazine, vol. 13, no. 4, pp. 17–22, 1998.
[3] W. G. Homeyer, E. E. Bowles, S. P. Lupan, P. S. Walia, and M. A. Maldonado, “Ad-vanced power converters for more electric aircraft applications,” in Proc. 32nd Interso-ciety Energy Conversion Engineering Conf. IECEC-97, vol. 1, 1997, pp. 591–596.
[4] D. Izquierdo, R. Azcona, F. del Cerro, C. Fernandez, and B. Delicado, “Electrical powerdistribution system (hv270dc), for application in more electric aircraft,” in Proc. Twenty-Fifth Annual IEEE Applied Power Electronics Conf. and Exposition (APEC), 2010, pp.1300–1305.
[5] R. I. Jones, “The more electric aircraft: the past and the future?” in Proc. IEE Colloqui-um Electrical Machines and Systems for the More Electric Aircraft (Ref. No. 1999/180),1999.
[6] K. Emadi and M. Ehsani, “Aircraft power systems: technology, state of the art, andfuture trends,” IEEE Aerospace and Electronic Systems Magazine, vol. 15, no. 1, pp.28–32, 2000.
[7] D. Alexa, A. Sirbu, D. M. Dobrea, and T. Goras, “Topologies of three-phase rectifierswith near sinusoidal input currents,” IEE Proceedings -Electric Power Applications, vol.151, no. 6, pp. 673–678, 2004.
[8] D. Alexa, A. Sirbu, and A. Lazar, “Three-phase rectifier with near sinusoidal inputcurrents and capacitors connected on the ac side,” Industrial Electronics, IEEE Trans-actions on, vol. 53, no. 5, pp. 1612–1620, 2006.
[9] D. Alexa, T. C. Goras, I. V. Pletea, R. Buzatu, M. Moisa, and R. Chiper, “Analysisof the two-quadrant converter having rectifier with near sinusoidal input currents andcapacitors connected on the ac side,” in Proc. Int. Symp. Signals, Circuits and SystemsISSCS 2009, 2009, pp. 1–4.
[10] T. F. J. W. Kolar, “The essence of three-phase pfc rectifier systems,” in Proceedings ofthe 33rd IEEE International Telecommunications Energy Conference (INTELEC 2011),Amsterdam, Netherlands, October 9-13 2011.
[11] B. Singh, B. N. Singh, A. Chandra, K. Al-Haddad, A. Pandey, and D. P. Kothari, “Areview of three-phase improved power quality ac-dc converters,” Industrial Electronics,IEEE Transactions on, vol. 51, no. 3, pp. 641–660, 2004.
[12] H. Kohlmeier, O. Niermeyer, and D. F. Schroder, “Highly dynamic four-quadrant ac mo-tor drive with improved power factor and on-line optimized pulse pattern with promc,”Industry Applications, IEEE Transactions on, no. 6, pp. 1001–1009, 1987.
119
120 Referencias
[13] J. W. Dixon, A. B. Kulkarni, M. Nishimoto, and B.-T. Ooi, “Characteristics of acontrolled-current pwm rectifier-inverter link,” Industry Applications, IEEE Transac-tions on, no. 6, pp. 1022–1028, 1987.
[14] B. T. Ooi, J. W. Dixon, A. B. Kulkarni, and M. Nishimoto, “An integrated ac drivesystem using a controlled-current pwm rectifier/inverter link,” Power Electronics, IEEETransactions on, vol. 3, no. 1, pp. 64–71, 1988.
[15] T. Nussbaumer and J. W. Kolar, “Comparison of 3-phase wide output voltage rangepwm rectifiers,” Industrial Electronics, IEEE Transactions on, vol. 54, no. 6, pp. 3422–3425, 2007.
[16] M. Nussbaumer, T. Kazuaki and J. W. Kolar, “Design and comparative evaluationof three-phase buck+boost and boost+buck unity power factor pwm rectifier systemsfor supplying variable dc voltage link converters,” in 25th International Conference onPower Electronics (PCIM),, 2004.
[17] M. Baumann, U. Drofenik, and J. W. Kolar, “New wide input voltage range three-phaseunity power factor rectifier formed by integration of a three-switch buck-derived front-end and a dc/dc boost converter output stage,” in Proc. INTELEC TelecommunicationsEnergy Conf. Twenty-second Int, 2000, pp. 461–470.
[18] T. Nussbaumer, M. Baumann, and J. W. Kolar, “Comparative evaluation of modulationmethods of a three-phase buck + boost pwm rectifier. part ii: Experimental verification,”IET Power Electronics, no. 2, pp. 268–274, 2008.
[19] M. Baumann, T. Nussbaumer, and J. W. Kolar, “Comparative evaluation of modulationmethods of a three-phase buck + boost pwm rectifier. part i: Theoretical analysis,” IETPower Electronics, no. 2, pp. 255–267, 2008.
[20] M. Baumann and J. W. Kolar, “Minimization of the dc current ripple of a three-phasebuck-boost pwm unity power factor rectifier,” in Proc. Power Conversion Conf. PCCOsaka 2002, vol. 2, 2002, pp. 472–477.
[21] A. Stupar, T. Friedli, J. Miniboeck, and J. Kolar, “Towards a 99three-phase buck-typepfc rectifier for 400 v dcdistribution systems,” Power Electronics, IEEE Transactionson, vol. PP, no. 99, p. 1, 2011.
[22] A. Stupar, T. Friedli, J. Minibo¨ andck, M. Schweizer, and J. Kolar, “Towards a 99sys-tems,” in Applied Power Electronics Conference and Exposition (APEC), 2011 Twenty-Sixth Annual IEEE, march 2011, pp. 505 –512.
[23] T. Nussbaumer, M. L. Heldwein, and J. W. Kolar, “Differential mode input filter designfor a three-phase buck-type pwm rectifier based on modeling of the emc test receiver,”Industrial Electronics, IEEE Transactions on, vol. 53, no. 5, pp. 1649–1661, 2006.
[24] ——, “Common mode emc input filter design for a three-phase buck-type pwm rectifi-er system,” in Proc. Twenty-First Annual IEEE Applied Power Electronics Conf. andExposition APEC ’06, 2006.
[25] T. Nussbaumer, M. Baumann, and J. W. Kolar, “Comprehensive design of a three-phase three-switch buck-type pwm rectifier,” Power Electronics, IEEE Transactions on,vol. 22, no. 2, pp. 551–562, 2007.
[26] M. L. Heldwein and J. W. Kolar, “Design of minimum volume emc input filters foran ultra compact three-phase pwm rectifier,” in COBEP07. The 7th Brazilian PowerElectronics Conference, 2007.
[27] Z. John Shen, Y. Xiong, X. Cheng, Y. Fu, and P. Kumar, “Power mosfet switching lossanalysis: A new insight,” in Proc. 41st IAS Annual Meeting Industry Applications Conf.Conf. Record of the 2006 IEEE, vol. 3, 2006, pp. 1438–1442.
Referencias 121
[28] V. Kaura and V. Blasko, “Operation of a phase locked loop system under distortedutility conditions,” in Proc. 1996. Eleventh Annual Applied Power Electronics Conf andExposition APEC ’96, vol. 2, 1996, pp. 703–708.
[29] ——, “Operation of a phase locked loop system under distorted utility conditions,”Industry Applications, IEEE Transactions on, vol. 33, no. 1, pp. 58–63, 1997.
[30] M. Karimi-Ghartemani and M. R. Iravani, “A method for synchronization of powerelectronic converters in polluted and variable-frequency environments,” Power Systems,IEEE Transactions on, vol. 19, no. 3, pp. 1263–1270, 2004.
[31] B. Wu, High-Power Converters and AC Drives. Wiley-IEEE press, 2006.
[32] V. Blasko, L. Arnedo, P. Kshirsagar, and S. Dwari, “Control and elimination of sinu-soidal harmonics in power electronics equipment: a system approach,” in IEEE EnergyConversion Congress & Exposition ECCE 2011. Phoenix, Arizona, September 2011, pp.2827–2837.
top related