Identificação e classificação de curto entre espiras e desbalanço de tensão em motores de indução trifásicos utilizando sensores de corrente, emissão acústica e vibração com emprego de transformada Hilbert e transformada Wavelet
Carregando...
Arquivos
Data
2022-08-02
Autores
Rocha, Marco Aurélio
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
O motor de indução trifásico (MIT) é amplamente empregado em aplicações industriais para o acionamento de diversos tipos de cargas. Caso haja qualquer tipo de interrupção indesejada causada por falhas elétricas e mecânicas neste tipo de máquina, poderá haver prejuízos financeiros e produtivos diversos. Com isso, técnicas de monitoramento não invasivas vêm ganhando cada vez mais destaque no cenário industrial e acadêmico. Essas técnicas utilizam sensores que obtém informações físicas sobre a operação da máquina, tais como vibração, corrente e emissão acústica. A análise desses sinais se destacam como ferramentas tradicionais para a obtenção de características operacionais do motor de indução trifásico e diagnóstico de falhas. Dentre as falhas elétricas, o desequilíbrio de tensão (DT), bem como o curto entre espiras (ITSC, do inglês Inter-Turn Short-Circuit) de curta duração são falhas frequentes no MIT e que podem afetar o desempenho da máquina, gerando paradas inesperadas. Portanto, o objetivo deste trabalho é a proposição de técnicas de diagnóstico de falhas DT e ITSC utilizando sensores de corrente de efeito Hall, sensor MEMS e sensores piezoelétricos. Deste modo, o diagnóstico proposto permite a detecção da falha, a identificação da fase elétrica afetada e a classificação da magnitude da falta. Para isso, foram desenvolvidas duas técnicas de processamento baseadas no envelope do sinal analítico proveniente da Transformada Hilbert e na energia do sinal dos coeficientes de aproximação e detalhe da Transformada Wavelet Discreta (TWD). Por fim, verificou-se qual a influência que a variação do carregamento tem sobre as técnicas propostas para cada um dos três sensores utilizados. Os resultados demonstraram que tanto o sensor de corrente quanto o sensor piezoelétrico foram sensíveis para o diagnóstico das falhas elétricas ITSC e DT. No entanto, verificou-se que o sensor MEMS não apresentou resultados promissores para o diagnóstico completo de falhas elétricas levando em consideração todos os níveis de carregamento aplicados. Por fim, o principal resultado é o desempenho do sensores PZT juntamente com a TWD em comparação ao sensor de corrente, uma vez que o sensor acústico apresenta custo muito reduzido e mede indiretamente a variável afetada pela falha.
The three-phase induction motor (TIM) is widely used in industrial applications to drive various types of loads. If there is any kind of unexpected interruption caused by electrical and mechanical failures in this type of machine, there may be several financial and productive losses. Therefore, non-invasive monitoring techniques have been gaining more and more prominence in the industrial and academic scenario. These techniques use sensors that obtain physical information about machine operation, such as vibration, current, and acoustic emissions. The analysis of these signals stand out as traditional tools for obtaining operational characteristics of the three-phase induction motor and fault diagnosis. Among the electrical faults, voltage unbalance (VU), as well as the short duration inter-turn short circuit (ITSC) are frequent faults in the MIT and can affect the performance of the machine, generating unexpected stops. Thus, the objective of this work is the proposition of VU and ITSC fault diagnosis techniques using Hall effect current sensors, MEMS sensor and piezoelectric sensors. In this way, the proposed diagnostics allow the detection of the fault, the identification of the affected electrical phase and the classification of the fault magnitude. For this, two processing techniques were developed based on the envelope of the analytical signal originating from the Hilbert Transform and the signal energy of the approximation and detail coefficients of the Discrete Wavelet Transform (TWD). Finally, it was verified what influence the loading variation has on the proposed techniques for each of the three sensors used. The results showed that both the current sensor and the piezoelectric sensor were sensitive for the diagnosis of ITSC and DT electrical faults. However, it was found that the MEMS sensor did not show promising results for the complete diagnosis of electrical faults considering all applied loading levels. Finally, the main result is the performance of the PZT sensor combined with the TWD compared to the current sensor, since the acoustic sensor has very low cost and indirectly measures the variable affected by the fault.
The three-phase induction motor (TIM) is widely used in industrial applications to drive various types of loads. If there is any kind of unexpected interruption caused by electrical and mechanical failures in this type of machine, there may be several financial and productive losses. Therefore, non-invasive monitoring techniques have been gaining more and more prominence in the industrial and academic scenario. These techniques use sensors that obtain physical information about machine operation, such as vibration, current, and acoustic emissions. The analysis of these signals stand out as traditional tools for obtaining operational characteristics of the three-phase induction motor and fault diagnosis. Among the electrical faults, voltage unbalance (VU), as well as the short duration inter-turn short circuit (ITSC) are frequent faults in the MIT and can affect the performance of the machine, generating unexpected stops. Thus, the objective of this work is the proposition of VU and ITSC fault diagnosis techniques using Hall effect current sensors, MEMS sensor and piezoelectric sensors. In this way, the proposed diagnostics allow the detection of the fault, the identification of the affected electrical phase and the classification of the fault magnitude. For this, two processing techniques were developed based on the envelope of the analytical signal originating from the Hilbert Transform and the signal energy of the approximation and detail coefficients of the Discrete Wavelet Transform (TWD). Finally, it was verified what influence the loading variation has on the proposed techniques for each of the three sensors used. The results showed that both the current sensor and the piezoelectric sensor were sensitive for the diagnosis of ITSC and DT electrical faults. However, it was found that the MEMS sensor did not show promising results for the complete diagnosis of electrical faults considering all applied loading levels. Finally, the main result is the performance of the PZT sensor combined with the TWD compared to the current sensor, since the acoustic sensor has very low cost and indirectly measures the variable affected by the fault.
Descrição
Palavras-chave
Tensão desequilibrada, Curto entre espiras, Transformada Hilbert, Transformada Wavelet, Motor de idução, Unbalanced voltages, Inter-turn short-circuit fault, Hilbert transform, Wavelet transform, Induction motor