Villasmil Urdaneta, Larry Alfonso (2006-08). Parameters defining flow resistance and the friction factor behavior in liquid annular seals with deliberately roughened surfaces. Doctoral Dissertation.
Thesis
Non-contacting annular seals are internal sealing devices used in rotating machinery, such as multistage centrifugal pumps and compressors. Their design affects both efficiency and rotor stability. Traditional plain and labyrinth seals are being replaced with stators containing different roughness patterns to reduce leakage and enhance rotor response. Several roughened seal experiments with liquid and air have produced leakage data indicating that the friction factor increases as the seal clearance is increased. Simplified models based on bulk flow theory and Moody??????s approach to characterize wall friction in pipes cannot explain this outcome. This research is an extension of a 2-D numerical analysis of flat plate experiments with water which found that friction factor of these surfaces is governed by the roughness?????? ability to develop high static pressures. An exhaustive 3-D numerical analysis of several experiments with liquid annular seals has been performed using a CFD code. Direct numerical simulations (DNS) of turbulent channel flow and smooth seals were replicated within 1% using Reynolds-averaged Navier-Stokes (RANS) equations and turbulence modeling. Similarly, measured groove seal leakage rates were reproduced within 2%. On the other hand, no turbulence model combination predicts the leakage in most 3-D pattern roughened seals with the same accuracy. Present results reproduce the friction factor ??????plateau?????? behavior predicted with the 2-D analysis and observed in the flat plate experiments. They also reproduce the friction-factor-to-clearance indifference behavior, the maximum friction factor observed in a specific roughness pattern size is independent of the actual clearance in a certain Reynolds number range, but clarify the role of the roughness length-to-clearance ratio and the actual roughness size in defining the friction-factor-toclearance proportionality. All simulations indicate that roughened surface area and roughness aspect ratios are the parameters defining the friction factor at a given seal clearance. The roughness pattern size, relevant in determining the friction-factor-to-clearance proportionality, plays a moderate role once the above cited ratios are defined. In any shape and size, shallow patterns are predicted and observed to provide larger friction factors than deep patterns. Predictions also confirm limited experimental data showing that friction factor is affected by the mean flow orientation relative to the roughness pattern. Solving RANS equations is sufficient to model simple seal geometries but might not be enough to replicate turbulent flow in liquid annular seals with roughened surfaces.
Non-contacting annular seals are internal sealing devices used in rotating machinery, such as multistage centrifugal pumps and compressors. Their design affects both efficiency and rotor stability. Traditional plain and labyrinth seals are being replaced with stators containing different roughness patterns to reduce leakage and enhance rotor response. Several roughened seal experiments with liquid and air have produced leakage data indicating that the friction factor increases as the seal clearance is increased. Simplified models based on bulk flow theory and Moody??????s approach to characterize wall friction in pipes cannot explain this outcome. This research is an extension of a 2-D numerical analysis of flat plate experiments with water which found that friction factor of these surfaces is governed by the roughness?????? ability to develop high static pressures. An exhaustive 3-D numerical analysis of several experiments with liquid annular seals has been performed using a CFD code. Direct numerical simulations (DNS) of turbulent channel flow and smooth seals were replicated within 1% using Reynolds-averaged Navier-Stokes (RANS) equations and turbulence modeling. Similarly, measured groove seal leakage rates were reproduced within 2%. On the other hand, no turbulence model combination predicts the leakage in most 3-D pattern roughened seals with the same accuracy. Present results reproduce the friction factor ??????plateau?????? behavior predicted with the 2-D analysis and observed in the flat plate experiments. They also reproduce the friction-factor-to-clearance indifference behavior, the maximum friction factor observed in a specific roughness pattern size is independent of the actual clearance in a certain Reynolds number range, but clarify the role of the roughness length-to-clearance ratio and the actual roughness size in defining the friction-factor-toclearance proportionality. All simulations indicate that roughened surface area and roughness aspect ratios are the parameters defining the friction factor at a given seal clearance. The roughness pattern size, relevant in determining the friction-factor-to-clearance proportionality, plays a moderate role once the above cited ratios are defined. In any shape and size, shallow patterns are predicted and observed to provide larger friction factors than deep patterns. Predictions also confirm limited experimental data showing that friction factor is affected by the mean flow orientation relative to the roughness pattern. Solving RANS equations is sufficient to model simple seal geometries but might not be enough to replicate turbulent flow in liquid annular seals with roughened surfaces.