Parametric optimization to design a high-performance vaneless-diffuser for sCO2 centrifugal compressor

• Ano de defesa: 2022
• Autor(a) principal: Mattos, Vitor Cesar Nogueira
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: sCO2 centrifugal compressor; Optimization; Vaneless-diffuser; CFD; Compressor centrífugo de sCO2; Otimização
• Link de acesso: http://hdl.handle.net/11449/237249

Supercritical carbon-dioxide centrifugal compressors are machines with high potential for use in power generation plants and in the oil industry as it achieves high thermal efficiency in the Brayton cycle and assist in oil production through the capture and storage of CO2 by the EOR (Enhanced Oil Recovery) method. In this regard, to increase the performance of these machines, a three-dimensional numerical simulation (CFD) coupled with an optimization method is investigated. As well known, one-dimensional modeling for vaneless-diffuser designs is not able to predict all important flow phenomena due to several geometric parameters. Therefore, the present work aims to use the methodology optimization-surrogate coupled with CFD to optimize nine geometric input variables related to a vaneless-diffuser channel to identify the influence of each parameter on the turbomachinery performance and its sources of loss. Three different objective functions are submitted to single-optimization: Maximize total-to-total polytropic efficiency, minimize total pressure loss coefficient or maximize static pressure recovery coefficient. A reliminary geometry is created using Ansys Vista CCD for the impeller and the vaneless-diffuser is modeled as a channel of parallel plates. The sensitivity analysis is conducted using the Morris Elementary Effects method and SS-ANOVA through the response surface generated by Gaussian Process regression. The optimization procedure is fulfilled by the NSGA-II method. The main conclusions indicate that the optimized geometries increased by 2.9% the total-to-total polytropic efficiency, reduced by 24.0% the total pressure loss coefficient and increased by 11.4% the static pressure recovery coefficient at the design point operational condition. Moreover, the optimal configuration found by the optimization procedure remains with higher performance even operating at the off-design point. The strategy adopted in the present work through a combination of one-dimensional turbomachinery design with three-dimensional parametric sensitivity analysis and CFD-optimization of a vaneless-diffusers is a powerful tool for sizing high-performance equipment.