Initial position of Taylor vortices in supercritical carbon dioxide Taylor–Couette–Poiseuille flow

Supercritical carbon dioxide (sCO2) exhibits a characteristic Taylor–Couette–Poiseuille (T–C–P) flow within the narrow gap between the turbine rotor and stator, where the spatial distribution and variation of Taylor vortices significantly influence the performance of the turbine rotor. To predict the spatial variation and the initial position of Taylor vortices in this system, this study employs computational fluid dynamics (CFD) simulations to systematically investigate the effects of heat flux, pressure, Taylor number, axial Reynolds number, aspect ratio, and radius ratio on the initial Taylor vortex position. Based on the CFD results, a predictive model for the initial position of Taylor vortices in sCO2 T–C–P flow was developed. The results reveal that under high heat flux and elevated pressure conditions, the Taylor vortices form closer to the inlet of the axial flow. Increasing the mass flow rate, reducing the rotation speed, and reducing the radius ratio can suppress the formation of the Taylor vortex of sCO2 in the turbine rotor–stator gap. The predictive accuracy of the neural network model for determining the Taylor vortex initiation position is within a margin of ±10%, outperforming the nonlinear regression model. At the design point, the Reynolds number, Taylor number, radius ratio, and aspect ratio are 9000, 6×109, 0.1, and 30, respectively. The variation range is ±5%; the parameter priority in influencing the Taylor vortex is as follows: aspect ratio, radius ratio, Reynolds number, and Taylor number.