Performance Improvement of Two-Phase Steam Ejector based on CFD Techniques, Response Surface Methodology and Genetic Algorithm

The steam ejector is a crucial component in the waste heat recovery system. Its performance determines the amount of recovered heat and system efficiency. However, poor ejector performance has always been the main bottleneck for system applications. Therefore, this study proposes an optimization methodology to improve the steam ejector’s performance by utilizing computational fluid dynamics (CFD) techniques, response surface methodology (RSM), and genetic algorithm (GA). Firstly, a homogeneous equilibrium model (HEM) was established to simulate the two-phase flow in the steam ejector. Then, the orthogonal test was presented to the screening of the key decision variables that have a significant impact on the entrainment ratio (ER). Next, the RSM was used to fit a response surface regression model (RSRM). Meanwhile, the effect of the interaction of geometric parameters on the performance of the steam ejector was revealed. Finally, GA was employed to solve the RSRM’s global optimal ER value. The results show that the RSRM exhibits a good fit for ER ( R 2 =0.997). After RSM and GA optimization, the maximum ejector efficiency is 27.94%, which is 48.38% higher than the initial ejector of 18.83%. Furthermore, the optimized ejector efficiency is increased by 46.4% on average under off-design conditions. Overall, the results reveal that the combination of CFD, RSM, and GA presents excellent reliability and feasibility in the optimization design of a two-phase steam ejector.