Sensitivity analysis and optimization of centrifugal compressors train working with supercritical co2
Carregando...
Data
2023-06-16
Autores
Gasparin, Elóy Esteves
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
A crescente importância da indústria de captura, utilização e armazenamento de carbono (CCUS) devido ao aquecimento global tem demandado novas tecnologias de equipamentos. Entre os mais relevantes, nota-se o ciclo de potência de Brayton que reporta aumento na eficiência térmica de seu sistema ao utilizar o CO2 em estado supercrítico devido ao menor trabalho de compressão quando comparado ao ar. Além desse, sistemas de armazenamento de energia que utilizavam ar comprimido também demonstraram melhora em seu desempenho ao trabalhar com CO2. Sendo assim, a tecnologia de CCUS com maior viabilidade econômica devido ao seu produto final de alto valor agregado é o processo conhecido como Enhanced Oil Recovery (EOR), que utiliza um trem de compressores centrífugos para realizar o armazenamento de CO2 supercrítico em poços de extração de petróleo, e que foi objeto de estudo do presente trabalho. Normalmente, a abordagem utilizada em um projeto de sistema de compressão é simplesmente investigada por aspectos termodinâmicos, sem considerar fenômenos fluidodinâmicos que podem inviabilizar o sistema como um todo. Portanto, a metodologia desenvolvida nesta tese é composta por um modelo de compressor centrífugo de CO2 supercrítico e difusor sem palhetas unidimensional (1D), capaz de prever alguns aspectos do escoamento do fluido que não são modelados por meio de análise termodinâmica direta, seguido por uma Análise de Sensibilidade (SA) e estratégia de otimização. Posteriormente, a modelagem individual de dois compressores utilizando Computational Fluid Dynamics (CFD) fez-se necessária para corroboração dos resultados anteriores e para entregar melhorias adicionais aos equipamentos a partir da meta-modelagem por superfície de resposta (RS) destes modelos computacionalmente custosos e realização de otimização indireta. Associada a estratégia de otimização, uma restrição chamada Gas Behaviour Margin (GBM) é proposta para evitar grandes variações das propriedades termodinâmicas no processo de compressão próximo à linha de Widom, o que pode dificultar a convergência do modelo numérico (por CFD) ou ainda causar danos ao equipamento durante seu processo de operação. Os principais resultados da obtidos a partir da metodologia descrita neste trabalho mostram que o trem de compressores centrífugos com quatro estágios considerado no processo de otimização teve sua potência total reduzida em 14,09% quando comparado à configuração inicial, e garantindo uma configuração de compressores viável, já que as restrições: Margem de Aceleração para Condensação (AMC), GBM e o número de Mach na garganta foram atendidas. Além disso, a solução do modelo CFD para o quarto compressor demonstrou boa concordância com a modelagem 1D e dados experimentais, corroborando a validade e robustez da metodologia aqui desenvolvida. Além disso, esta abordagem é considerada como uma maneira rápida e de baixo custo para obter uma geometria preliminar para sistemas de trem de compressores centrífugos. Finalmente, os modelos CFD do primeiro e quarto estágios de compressão foram submetidos a estratégia de SA e otimização, aumentando ainda mais a eficiência politrópica destes dois estágios, essa modelagem permitiu a análise fenomenológica baseada em variação entrópica que foi conduzida para o processo de avaliação fluidodinâmica do escoamento dentro dos compressores centrífugos A estratégia aqui adotada pode ser aplicada a outros modelos de CFD considerados grandes e altamente dimensionais para reduzir o custo computacional do procedimento de otimização e acelerar a análise fenomenológica do escoamento.
The rising importance of carbon capture, utilization and storage (CCUS) industry due to global warming has increased the development of new technologies. Among the most relevant, note the Brayton power cycle, which reported an increase in the thermal efficiency of its system when using CO2 in a supercritical state due to the lower compression work when compared to air. Furthermore, energy storage systems that utilized compressed air also improved their performance when working with CO2. Therefore, the CCUS technology with greater economic viability due to its final product with high added value is the process known as Enhanced Oil Recovery (EOR), which uses a train of centrifugal compressors to store supercritical CO2 in oil wells, which was the object of study of the present work. Normally, the approach used in a compression system design is the simple thermodynamic investigation, without considering fluid dynamic phenomena that can deliver an unfeasible system. Therefore, the methodology developed in this thesis is composed by an optimization methodology based on an one-dimensional (1D) centrifugal compressor model, capable to predict a few fluid flow aspects that are not modeled through direct thermodynamic analysis, followed by a sensitivity analysis (SA) and optimization strategy. Subsequently, the individual modeling of two compressors using Computational Fluid Dynamics (CFD) was necessary to corroborate the previous results and to deliver additional improvements to the equipment from the metamodeling by response surface (RS) of these computationally expensive models and performing an indirect optimization. Associated to the optimization strategy, a constraint called Gas-like Behavior Margin (GBM) is presented to avoid high variations of thermodynamic properties on the compression process close to the Widom-line, which can difficult the numerical model convergence (By CFD) or also can cause damage to the equipment during its operation. The main results obtained throughout the described methodology shows that the train of centrifugal compressors with four stages considered in the optimization process had its total power reduced in 14.09% when compared to the initial configuration, and ensuring a feasible set of compressors, once the constraints: Acceleration Margin to Condensation (AMC), GBM and the Mach number at the throat were met. Moreover, a CFD model for the fourth stage demonstrated good agreement with 1D modeling and experimental data, proving the validity and robustness of the methodology developed herein. Besides, this approach is considered a fast and low-cost way to obtain a preliminary geometry for centrifugal compressors train systems. Finally, the first and fourth compression stages CFD models were submitted to the SA and optimization, further increasing the polytropic efficiency of both stages, this modeling allowed a phenomenological analysis based on entropic variation to perform the fluid dynamic behavior assessment inside centrifugal compressors. The strategy adopted herein could be applied to other CFD models considered large and high dimensional to reduce the computational cost of optimization procedure and speed-up the phenomenological analysis of fluid flow,
The rising importance of carbon capture, utilization and storage (CCUS) industry due to global warming has increased the development of new technologies. Among the most relevant, note the Brayton power cycle, which reported an increase in the thermal efficiency of its system when using CO2 in a supercritical state due to the lower compression work when compared to air. Furthermore, energy storage systems that utilized compressed air also improved their performance when working with CO2. Therefore, the CCUS technology with greater economic viability due to its final product with high added value is the process known as Enhanced Oil Recovery (EOR), which uses a train of centrifugal compressors to store supercritical CO2 in oil wells, which was the object of study of the present work. Normally, the approach used in a compression system design is the simple thermodynamic investigation, without considering fluid dynamic phenomena that can deliver an unfeasible system. Therefore, the methodology developed in this thesis is composed by an optimization methodology based on an one-dimensional (1D) centrifugal compressor model, capable to predict a few fluid flow aspects that are not modeled through direct thermodynamic analysis, followed by a sensitivity analysis (SA) and optimization strategy. Subsequently, the individual modeling of two compressors using Computational Fluid Dynamics (CFD) was necessary to corroborate the previous results and to deliver additional improvements to the equipment from the metamodeling by response surface (RS) of these computationally expensive models and performing an indirect optimization. Associated to the optimization strategy, a constraint called Gas-like Behavior Margin (GBM) is presented to avoid high variations of thermodynamic properties on the compression process close to the Widom-line, which can difficult the numerical model convergence (By CFD) or also can cause damage to the equipment during its operation. The main results obtained throughout the described methodology shows that the train of centrifugal compressors with four stages considered in the optimization process had its total power reduced in 14.09% when compared to the initial configuration, and ensuring a feasible set of compressors, once the constraints: Acceleration Margin to Condensation (AMC), GBM and the Mach number at the throat were met. Moreover, a CFD model for the fourth stage demonstrated good agreement with 1D modeling and experimental data, proving the validity and robustness of the methodology developed herein. Besides, this approach is considered a fast and low-cost way to obtain a preliminary geometry for centrifugal compressors train systems. Finally, the first and fourth compression stages CFD models were submitted to the SA and optimization, further increasing the polytropic efficiency of both stages, this modeling allowed a phenomenological analysis based on entropic variation to perform the fluid dynamic behavior assessment inside centrifugal compressors. The strategy adopted herein could be applied to other CFD models considered large and high dimensional to reduce the computational cost of optimization procedure and speed-up the phenomenological analysis of fluid flow,
Descrição
Palavras-chave
Compression system optimization, Gas-like behavior margin, Carbon capture and storage, S-CO2 centrifugal compressor, Sensitivity analysis