Improved preliminary geometry for SCO2 centrifugal compressors through full 1D modeling
Carregando...
Data
2023-03-08
Autores
Gonçalves, Gustavo Singulane
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Para prolongar a vida útil dos depósitos de petróleo e, portanto, o lucro deles, uma técnica chamada Enhanced Oil Recovery (EOR) pode ser usada. Um dos métodos para realizar essa técnica é a reinjeção de CO2 no reservatório de óleo, o que não apenas aumenta a vida útil do campo petrolífero, mas também fornece uma maneira de sequestrar permanentemente o dióxido de carbono. Para atingir as altas pressões necessárias para a reinjeção em tais profundidades, compressores supercríticos de CO2 devem ser desenvolvidos. A análise numérica do comportamento termo-hidráulico desses equipamentos geralmente leva a um alto custo computacional para simulações 3D. A fim de fornecer resultados mais rápidos e confiáveis, um código para o compressor completo foi escrito na linguagem de programação Python, sendo uma alternativa de código aberto para obter resultados preliminares que possam fornecer um ponto de partida para a geometria do compressor centrífugo para um projeto 3D. Para uma modelagem 1D completa, um compressor centrífugo de CO2 supercrítico com rotor, difusor sem palhetas e voluta é considerado. Os modelos 1D disponíveis na literatura relacionados à geometria e perdas foram estudados e implementados sob as hipóteses: estado estacionário, difusor sem palhetas, largura de passagem constante no difusor, ângulo constante entre as velocidades tangencial, meridional e absoluta ao longo do difusor e escoamento incompressível na voluta. Oito parâmetros geométricos são submetidos a um método de análise de sensibilidade de triagem usando o modelo 1D e seu impacto na eficiência isentrópica, potência necessária e recuperação de pressão são avaliados. Um modelo de compressor 3D consistindo de um impulsor e difusor é simulado, assumindo o fluxo de estado estacionário, modelo de turbulência SST k-ômega e equação de estado de Span e Wagner. Uma geometria 3D aprimorada é criada com base nos resultados obtidos por análise de sensibilidade de Morris, avaliações 1D e 3D. O modelo 1D é validado através de dados experimentais e o modelo 3D é usado para melhor compreensão das diferenças dos fenômenos físicos entre os designs originais e aprimorados. Esta metodologia pode aumentar significativamente o desempenho da máquina usando recursos computacionais relativamente baixos.
In order to extend the life of oil deposits, and therefore the profit from them, a technique called Enhanced Oil Recovery (EOR) can be used. One of the methods for performing this technique is the reinjection of CO2 in the oil reservoir, which not only increases the oilfield life, but also provides a way to permanently sequester carbon dioxide. To accomplish the high pressures required for the reinjection at such depths, supercritical CO2 compressors must be developed. The numerical analysis of the thermal-hydraulic behavior of this equipment usually requires complex and high computational cost for 3D simulations. In order to provide faster and reliable results, a code for the full compressor was written in Python programming language, being an open-source alternative to obtain preliminary results that can provide a starting point for the centrifugal compressor geometry for a 3D design. For a full 1D modeling, a supercritical CO2 centrifugal compressor with impeller, vaneless diffuser and volute is considered. The available 1D models in the literature related to geometry and losses were studied and implemented under hypothesis: steady state, vaneless diffuser, constant passage width in the diffuser, constant angle between tangential, meridional and absolute velocities throughout the diffuser and incompressible flow in the volute. Eight geometrical parameters are submitted to a screening sensitivity analysis method using the 1D model and their impact towards isentropic efficiency, power required and pressure recovery are assessed. A 3D compressor model consisting of an impeller and diffuser is simulated, assuming steady state flow, SST k-omega turbulence model and Span and Wagner equation of state. An improved 3D geometry is created based on insight obtained from Morris’ sensitivity analysis method, 1D and 3D evaluations. The 1D model is validated through experimental data and the 3D model is used for better understanding of physical phenomena differences between original and improved designs. This methodology can significantly increase the machine performance using relatively low computational resources.
In order to extend the life of oil deposits, and therefore the profit from them, a technique called Enhanced Oil Recovery (EOR) can be used. One of the methods for performing this technique is the reinjection of CO2 in the oil reservoir, which not only increases the oilfield life, but also provides a way to permanently sequester carbon dioxide. To accomplish the high pressures required for the reinjection at such depths, supercritical CO2 compressors must be developed. The numerical analysis of the thermal-hydraulic behavior of this equipment usually requires complex and high computational cost for 3D simulations. In order to provide faster and reliable results, a code for the full compressor was written in Python programming language, being an open-source alternative to obtain preliminary results that can provide a starting point for the centrifugal compressor geometry for a 3D design. For a full 1D modeling, a supercritical CO2 centrifugal compressor with impeller, vaneless diffuser and volute is considered. The available 1D models in the literature related to geometry and losses were studied and implemented under hypothesis: steady state, vaneless diffuser, constant passage width in the diffuser, constant angle between tangential, meridional and absolute velocities throughout the diffuser and incompressible flow in the volute. Eight geometrical parameters are submitted to a screening sensitivity analysis method using the 1D model and their impact towards isentropic efficiency, power required and pressure recovery are assessed. A 3D compressor model consisting of an impeller and diffuser is simulated, assuming steady state flow, SST k-omega turbulence model and Span and Wagner equation of state. An improved 3D geometry is created based on insight obtained from Morris’ sensitivity analysis method, 1D and 3D evaluations. The 1D model is validated through experimental data and the 3D model is used for better understanding of physical phenomena differences between original and improved designs. This methodology can significantly increase the machine performance using relatively low computational resources.
Descrição
Palavras-chave
Compressores, Fluidodinâmica computacional, Perdas, Centrifugal compressor, CFD, One-dimensional model, Loss models