Identificação de padrões de escoamento horizontal bifásico gás-líquido através de distribuição tempo-freqüência e redes neurais
Data
2017-11-15
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Biblioteca Digital de Teses e Dissertações da USP
Universidade de São Paulo
Escola de Engenharia de São Carlos
Universidade de São Paulo
Escola de Engenharia de São Carlos
Resumo
Descrição
O presente trabalho tem como objetivo fundamental a construção de um sistema de identificação capaz de diagnosticar em tempo real as diferentes configurações de escoamentos bifásicos horizontais. É importante ressaltar que o desenvolvimento deste know-how é capital para a operação eficaz de instalações de manipulação e ou transporte de fluidos multifásicos, e representa, hoje, um dos grandes desafios nas indústrias do petróleo e termonuclear. O princípio de funcionamento do sistema proposto baseia-se nos sinais captados por um sensor de pressão flutuante de resposta rápida, e no seu pósprocessamento com auxílio da transformada de Gabor e de uma rede neural convenientemente treinada. A implementação é tal que a operação de diagnóstico pode ser feita on-line, desde a aquisição dos sinais até o pósprocessamento. Resultados experimentais foram obtidos no circuito experimental do NETeF - Núcleo de Engenharia Térmica e Fluidos da USP - Universidade de São Paulo, para uma secção de testes horizo ntal com 12 m de comprimento e diâmetro interno de 30 mm. Em específico foram ensaiados os seguintes padrões de escoamento ar-água: estratificado liso, ondulado, intermitente, anular e a bolhas. Os resultados mostram que, dependendo dos limites de detecção pré-estabelecidos, todos o principais padrões de escoamento bifásico horizontal são identificados corretamente.
The fundamental objective of this work is the construction of an identification system capable of diagnosing in real time different configurations of horizontal two-phase flow patterns. It is important to emphasize that the development of this know-how is capital to the efficient operation of facilities for manipulation and transportation of multiphase fluids, and represents, today, one of the most important challenges in the oil and thermonuclear industries. The working principle of the proposed system is based on the signals acquired by a rapid response fluctuating pressure sensor, and on its post processing through Gabor Transform and on a conveniently trained artificial neural network. The implementation is accomplished in way that the diagnosis operation is performed on-line, from signal acquisition to post-processing. Experimental results were obtained on the experimental circuit at NETeF - Núcleo de Engenharia Térmica e Fluidos of USP - Universidade de São Paulo, at São Carlos, using a horizontal test section, with 12 m length and 30 mm internal diameter. Experiments were done with the following air-water flow patterns: stratified smooth, wavy, intermittent, annular, and bubbly. Results show that, depending on the preset detection limits, all the main horizontal two phase flow patterns were correctly identified.
The fundamental objective of this work is the construction of an identification system capable of diagnosing in real time different configurations of horizontal two-phase flow patterns. It is important to emphasize that the development of this know-how is capital to the efficient operation of facilities for manipulation and transportation of multiphase fluids, and represents, today, one of the most important challenges in the oil and thermonuclear industries. The working principle of the proposed system is based on the signals acquired by a rapid response fluctuating pressure sensor, and on its post processing through Gabor Transform and on a conveniently trained artificial neural network. The implementation is accomplished in way that the diagnosis operation is performed on-line, from signal acquisition to post-processing. Experimental results were obtained on the experimental circuit at NETeF - Núcleo de Engenharia Térmica e Fluidos of USP - Universidade de São Paulo, at São Carlos, using a horizontal test section, with 12 m length and 30 mm internal diameter. Experiments were done with the following air-water flow patterns: stratified smooth, wavy, intermittent, annular, and bubbly. Results show that, depending on the preset detection limits, all the main horizontal two phase flow patterns were correctly identified.
Palavras-chave
Escoamento multifásico, Padrões de escoamento, Redes neurais, Transformada de Gabor, Flow patterns, Gabor transform, Multiphase flow, Neural networks