Framework for prediction of two-phase R-744 ejector performance based on integration of thermodynamic models with multiphase mixture CFD simulations

Current computational models of two-phase R-744 ejectors require defined outlet pressure conditions for convergence, this brings a strong dependence on experimental data or pre-guessed outlet pressure values. This study presents a framework combining a thermodynamic model and a computational fluid dynamics (CFD) model to predict two-phase R-744 ejector performance, minimizing reliance on experimental data or guessed outlet pressure values. The proposed framework is developed in a broad operating range with 137 different experimental cases from the motive inlet pressure of 51.9 bar to 101.1 bar. Three different approaches are developed for the thermodynamic model in the MATLAB environment; i) a model with global coefficients which are unchanged for the entire operating range, ii) a model with local coefficients, which have unique values for each experimental case, and iii) a model with predicted-local coefficients, which uses a prediction algorithm to find the local values for each experimental case. The local and predicted-local coefficient approaches estimate the mass flow rate from the motive inlet with a relative error of 5 %. However, the mass flow rate predictions from the suction inlet showed high deviations above 30 % for the predicted-local coefficient approach in transcritical operating conditions, while the local coefficient approach keeps the relative error still lower than 5 %. All approaches estimated the ejector outlet pressure with less than 10 % error, which is an acceptable error; therefore, the thermodynamic model-based outlet pressure was defined as the outlet boundary condition to the CFD domain. All thermodynamic model embedded CFD simulations computed the temperature and Mach number values with less than 1 % deviation at the motive nozzle exit. The results show that the prediction algorithm can estimate the ejector outlet pressure within an acceptable deviation range, offering a promising direction for future research to reduce dependence on experimental data.