A framework for computational dynamic characterisation of dry gas seals for assessing system response

Abstract

A computational framework to study the dynamic response of a dry gas seal (DGS) during multi-axis rotor vibrations is described in this work. The DGS comprises Rayleigh pads grooved in the primary ring and a pumping volume, and is used to explore the gas film behaviour for different rotor speeds and air supply pressures. Steady-state CFD simulations are conducted to determine the stiffness coefficients at various primary ring positions while the non-linear damping coefficients are computed from transient analyses. A novel methodology that allows the complete primary ring’s movement is developed by specifying a wall motion through a transfer function. A mesh deformation strategy is established in which the grid is controlled using a Laplacian equation along with smoothing iterations, and differential, gradual steps of wall motion enable high-quality “real-time” simulations. The gas film flow field is characterised by the generation of dynamic pressure from the Rayleigh pads and the flow entrapment in the pumping volume. The gas film thickness reduces circumferentially to reduce leakages thus improving the film stiffness. Higher gas film pressure is generated at higher air inlet pressures and rotor speeds, leading to an increase in stiffness coefficients. During the translational motion, the stiffness reduces as the film thickness increases while the damping response improves with the air inlet pressure. When the primary ring tilts, the stiffness and damping effects from the induced force in the radial direction are negligible. Negative stiffness and damping coefficients are also obtained, which have important implications for stable rotor operation. The computed opening forces depend on the primary ring’s position and the effects of rotational speed and air supply pressure are highlighted. Finally, the calculated dynamic coefficients are utilised in a rotordynamics model to determine its dynamic response characteristics. This study proposes a numerical framework to accurately study the dynamic stability of DGSs that are working under harsh dynamic conditions, leading to better performance calculations and initial designs.