Numerical study on the interaction of geometric parameters of a transcritical CO2 two-phase ejector using response surface methodology and genetic algorithm

Abstract

In this study, the effect of the interaction of geometric parameters on the performance of the rectangular transcritical CO2 two-phase ejector was investigated, and a geometric optimization of the rectangular CO2 ejector was conducted using computational fluid dynamics (CFD) technique, response surface methodology (RSM) and genetic algorithm (GA). The significance of geometric factors on entrainment ratio was analyzed firstly using CFD simulation, and then a second-order regression model was successfully developed using the RSM. Furthermore, a GA was used to solve the response surface model to obtain the optimized geometric parameters. The reliability of GA optimization was verified with CFD simulation under different operating conditions. The analysis results indicated that the entrainment ratio was highly sensitive to the nozzle exit position (NXP), suction chamber angle and mixing chamber width. Besides, the strong interactions between the mixing chamber width and mixing chamber length, the mixing chamber width and NXP, the mixing chamber length and diffuser angle, the mixing chamber length and NXP as well as the suction chamber angle and NXP had significant influence on the entrainment ratio. The results of CFD and GA optimization demonstrated that the optimized ejector increased the velocity of secondary flow and broke the vortex in the ejector, and optimized ejector improved the ejector efficiency more than 41% on average for different primary flow operating conditions and ejector outlet pressure. Additionally, the results confirmed the combination of CFD, RSM and GA had a considerable reliability and an excellent performance to optimize the CO2 two-phase ejector.