Thesis ANÁLISIS NO LINEAL DE OBSERVABILIDAD Y CONTROL SENSORLESS DE UNA MÁQUINA SINCRÓNICA DE RELUCTANCIA MEDIANTE FILTRO DE KALMAN EXTENDIDO
Loading...
Date
2018
Authors
Journal Title
Journal ISSN
Volume Title
Program
Campus
Universidad Técnica Federico Santa María UTFSM. Casa Central Valparaíso
Abstract
En esta tesis se abarca un m´etodo de control sensorless para un máquina sincrónica dereluctancia, o SynRM por sus siglas en inglés.Uno de los problemas recurrentes para el control sensorless de las máquinas AC radicaen la dificultad de la estimación de velocidad a frecuencias cercanas a cero. Esto dadoque la tensión denominada fuerza contra electromotriz, la cual aporta información sobre lavelocidad, se va reduciendo a medida que se reduce la velocidad. Esta reducción de información imposibilita, o al menos dificulta significativamente el uso de observadores basadosen modelos a frecuencia fundamental para la estimación de velocidad.Con el fin de realizar el control sensorless, se analiza la observabilidad no lineal deeste tipo de máquinas. Donde se analizan tres casos: dos modelos que estiman posición yvelocidad, de los cuales uno tiene el modelo completo de la dinámica de la velocidad angulary otro una versión simplificada; un tercer caso además de considerar el modelo completode la dinámica de la velocidad angular se agrega el par de carga a la estimación. De estaforma, se diseña un observador que satisfaga las condiciones de observabilidad, lograndoque el control funcione correctamente en todo el rango de velocidades admisibles inclusivea velocidad cero.En esta tesis, para estimar las variables no medidas, se implementa el control sensorlessusando un filtro extendido de Kalman como observador. Este observador se basa en un modeloque incorpora tanto el modelo a frecuencia fundamental como el efecto de la injecciónde señales de ricas espectralmente a velocidades bajas. De esta forma, se garantiza la observabilidada velocidades bajas y cero, se logra una composición de la información provistapor la inyección y el modelo fundamental a velocidades bajas, y se logra operación sin inyección a velocidades mayores. Este esquema garantiza mínimas pérdidas y ruido acústico avelocidades altas y una transición suave entre la operación con y sin inyección. Finalmentese valida el control propuesto mediante simulaciones, y resultados experimentales.
This thesis includes a sensorless control method for a synchronous reluctance machine(SynRM).One of the recurrent problems for the sensorless control of AC machines lies in thedifficulty of estimating the speed at frequencies close to zero. This problem is related to thereduction of the back electromotive force (Back EMF), which provides information aboutthe speed, as the speed decreases. This reduction of information makes it impossible orat least significantly hinders the use of observers based on fundamental models for speedestimation.In order to perform the sensorless control, the nonlinear observability of this type ofmachines is analyzed. Three cases are analyzed: two models that estimate position andspeed, of which one has the complete model of the angular velocity dynamics and anotherone a simplified version. A third case considering the complete model of the dynamics ofthe angular velocity, but with the load torque added to the state vector estimation is alsoconsidered. In this way, an observer that satisfies observability conditions is designed, makingthe control work correctly throughout the range of admissible speeds, even at zero speed.In this thesis, to estimate the unmeasured variables, the sensorless control will be implementedusing an extended Kalman filter as an observer. This observer is based on a modelthat incorporates both the fundamental model and the effect of injecting signals at lowspeeds. In this way, observability is guaranteed at low and at zero speed,achieving an optimalcomposition of the information provided by the injection and the fundamental modelat low speeds and operation without injection at higher speeds. This scheme guarantees minimumlosses and acoustic noise at high speeds and a smooth transition between operationwith and without injection. Finally, the proposed control is validated through simulationsand experimental results.
This thesis includes a sensorless control method for a synchronous reluctance machine(SynRM).One of the recurrent problems for the sensorless control of AC machines lies in thedifficulty of estimating the speed at frequencies close to zero. This problem is related to thereduction of the back electromotive force (Back EMF), which provides information aboutthe speed, as the speed decreases. This reduction of information makes it impossible orat least significantly hinders the use of observers based on fundamental models for speedestimation.In order to perform the sensorless control, the nonlinear observability of this type ofmachines is analyzed. Three cases are analyzed: two models that estimate position andspeed, of which one has the complete model of the angular velocity dynamics and anotherone a simplified version. A third case considering the complete model of the dynamics ofthe angular velocity, but with the load torque added to the state vector estimation is alsoconsidered. In this way, an observer that satisfies observability conditions is designed, makingthe control work correctly throughout the range of admissible speeds, even at zero speed.In this thesis, to estimate the unmeasured variables, the sensorless control will be implementedusing an extended Kalman filter as an observer. This observer is based on a modelthat incorporates both the fundamental model and the effect of injecting signals at lowspeeds. In this way, observability is guaranteed at low and at zero speed,achieving an optimalcomposition of the information provided by the injection and the fundamental modelat low speeds and operation without injection at higher speeds. This scheme guarantees minimumlosses and acoustic noise at high speeds and a smooth transition between operationwith and without injection. Finally, the proposed control is validated through simulationsand experimental results.
Description
Catalogado desde la version PDF de la tesis.
Keywords
CONTROL SENSORLESS, FILTRO DE KALMAN, MAQUINA SINCRONICA, OBSERVABILIDAD