Steam-jet vacuum pump is widely used in a range of applications. This paper evaluated the performance of four well-known turbulence models for predicting and understanding the internal flow of a steam-jet vacuum pump first. With the help of a commercial computational fluid dynamics (CFD) code ANSYS-Fluent 6.3, the simulation results obtained from the concerned turbulence models were compared with experimental values, the k-omega-SST model was chosen as a tool model for carrying out numerical simulations. Then, based on the simulation results obtained from specific operating conditions, a method for locating the shock-mixing layer was put forward. The shape of the shock-mixing layer shows that the secondary steam does not mix with the primary steam immediately after being induced into the mixing chamber of the pump; actually, they maintain their independence till the shocking position instead. After the shock happens, the shock-mixing layer disappear, the two fluid in the pump begin to mix with each other and discharge to the next stage with almost the same state. Based on the shape of the shock-mixing layer and the supersonic region of the secondary steam, a detailed analysis for the flow duct of the secondary steam was carried out. It is found that the throat of the secondary steam flow duct plays a crucial role in maintaining a stable operating state and the length of the throat reflects the back pressure endurance for the pump.
