Detecção e diagnóstico de falhas em conversores de potência back-to-back utilizando modelagem inversa
Carregando...
Data
2019-07-30
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Neste trabalho, é realizado o estudo da aplicação da inversão de sistemas dinâmicos para diagnóstico de falhas nos IGBTs dos conversores de potência de dois nı́veis CA-CC-CA bi-direcionais (back-to-back) aplicados na geração de energia elétrica com fonte eólica por meio de geradores de indução duplamente alimentados. Inicialmente à falha detectada, pode-se avançar para o diagnóstico do IGBT que apresenta a mesma; para tal propósito, utiliza-se o método do produto interno L 2 . Desta forma, com a detecção da falha e o di- agnóstico do componente, esta tese aborda um modo de tornar o conversor aqui estudado tolerante às falhas oriundas nos IGBTs. Com isso, adiciona a redundância fı́sica com a adição de um braço extra que permanece ocioso na ausência de falha e, caso contrário, assume o lugar do braço que apresenta a falha. O principal objetivo é demonstrar que a inversão dinâmica de sistemas pode ser utilizada como uma ferramenta de redundância analı́tica baseada em modelo; método que reduz custos por estimar sinais sem a neces- sidade de utilizar sensores, ou até mesmo utilizar o método para redundância do sensor. Para validar as ferramentas matemáticas aplicadas por esta tese, um conversor boost foi ensaiado em bancada de teste e sua inversa dinâmica foi discretizada, onde constatou sua funcionalidade em sistema embarcado utilizando um (DSP). Simulações no ambi- ente Matlab/Simulink/SymPowerSystems foram realizadas em pequena escala (gerador de 1,1kVA). Durante as simulações foram inseridas as falhas e verificou-se a robustez dos métodos estudados para detecção, diagnóstico e tolerância a falha, com o funcionamento do conversor CC-CA do lado da rede elétrica. Conclui-se que a inversão de sistemas dinâ- micos em conjunto com o produto interno L 2 podem ser utilizados como ferramentas de detecção e diagnóstico de falhas em conversores de potência; ainda ao adicionar a técnica de redundância fı́sica, o sistema apresentou tolerância a falha o que permitiu a operação do conversor sem danos ao sistema elétrico.
This thesis, the study of the application of the inversion of dynamic systems to diagnose failures in the IGBTs of two-level converter back-to-back applied in the generation of electric energy with a wind power source through double-fed induction generators. Ini- tially to the detected fault, one can proceed to the diagnosis of the IGBT that presents the same; for this purpose, the L 2 -inner product method is used. Thus, with the fault detection and component diagnosis, this thesis approaches a way to make the converter studied here fault tolerant from the IGBTs. This adds physical redundancy to the addi- tion of an extra arm that remains idle in the absence of failure and otherwise takes over the arm that has the fault. The main objective is to demonstrate that dynamic system inversion can be used as a model-based analytical redundancy tool; a method that reduces costs by estimating signals without the need to use sensors, or even use the method for sensor redundancy. In order to validate the mathematical tools applied by this thesis, a boost converter was tested on the test bench and its dynamic inverse was discretized, where it verified its functionality in an embedded system using a (DSP). Simulations in the Matlab/Simulink/SymPowerSystems environment were performed on a small scale (1.1kVA generator). During the simulations, the faults were inserted and the robustness of the methods studied for detection, diagnosis and fault tolerance was verified with the operation of the DC-AC converter on the side of the power grid. It is concluded that the inversion of dynamic systems together with the L2 -inner product can be used as tools for detecting and diagnosing failures in power converters; even when adding the technique of physical redundancy, the system presented fault tolerance which allowed the operation of the converter without damage to the electrical system.
This thesis, the study of the application of the inversion of dynamic systems to diagnose failures in the IGBTs of two-level converter back-to-back applied in the generation of electric energy with a wind power source through double-fed induction generators. Ini- tially to the detected fault, one can proceed to the diagnosis of the IGBT that presents the same; for this purpose, the L 2 -inner product method is used. Thus, with the fault detection and component diagnosis, this thesis approaches a way to make the converter studied here fault tolerant from the IGBTs. This adds physical redundancy to the addi- tion of an extra arm that remains idle in the absence of failure and otherwise takes over the arm that has the fault. The main objective is to demonstrate that dynamic system inversion can be used as a model-based analytical redundancy tool; a method that reduces costs by estimating signals without the need to use sensors, or even use the method for sensor redundancy. In order to validate the mathematical tools applied by this thesis, a boost converter was tested on the test bench and its dynamic inverse was discretized, where it verified its functionality in an embedded system using a (DSP). Simulations in the Matlab/Simulink/SymPowerSystems environment were performed on a small scale (1.1kVA generator). During the simulations, the faults were inserted and the robustness of the methods studied for detection, diagnosis and fault tolerance was verified with the operation of the DC-AC converter on the side of the power grid. It is concluded that the inversion of dynamic systems together with the L2 -inner product can be used as tools for detecting and diagnosing failures in power converters; even when adding the technique of physical redundancy, the system presented fault tolerance which allowed the operation of the converter without damage to the electrical system.
Descrição
Palavras-chave
Conversor CC-CA, Inversa dinâmica, Geração eólica, Back-to-Back, Diagnóstico e detecção de falhas, Sistemas tolerantes a falha, Converter DC-AC, Inversion of dynamic systems, DFIG, Fault-detection and diagnosis, Fault-tolerance