Making Better Swirl Brakes Using Computational Fluid Dynamics: Performance Enhancement From Geometry Variation

Gas labyrinth seals (LSs) improve turbomachinery operational efficiency and mechanical reliability by reducing secondary leakage. As interlocking labyrinth seals (ILSs) restrict more leakage than conventional see-through LSs, attention is due to their performance. An earlier paper [1] details the performance of a particular ILS in an ad-hoc test rig via measurements of mass flow (leakage) and cavity pressures while supplied with pressurized air at ambient temperature and operating with a rotor speed to a maximum of 10 krpm (surface speed 79 m/s). The test seal comprises of two teeth on the rotor and three teeth on the stator to make a four cavity seal with radial clearance Cr = 0.2 mm. The experimental and numerical leakages for the ILS produce a modified flow factor (Φ¯) that is independent of the seal operating conditions, namely inlet pressure, discharge pressure and rotor speed. The finding leads to an orifice-like loss coefficient cd = 0.36 and an effective clearance (cd × Cr) for the test seal, thus evidencing its effectiveness in reducing leakage. To complement the former research, this paper reports measurements of the leakage and cavity pressures for the same geometry interlocking labyrinth seals configured with two other clearances Cr = 0.3 mm and 0.13 mm. For the ILS with Cr = 0.3 mm, a first configuration is without a swirl brake (baseline), the second is with a swirl brake with 0° teeth pitch (axial ribs), and the third configuration is with a swirl brake with teeth angled at 40° in the direction of shaft rotation. For tests conducted without shaft rotation and with rotor spinning at 7.5 krpm (surface speed = 59 m/s), the inlet air pressure (Pin) ranges from 0.29 MPa to 0.98 MPa, while the exit pressure (Pout) is set to pressure ratios PR = (Pout/Pin) = 0.3, 0.5, 0.8. As to the ILS with Cr = 0.13 mm, it operates with an upstream swirl brake with axial ribs (0° teeth pitch) and w/o rotor speed. The supply pressure (Pin) varies from 0.59 MPa to 1.4 MPa and PR = 0.3, 0.5. The measurements and bulk-flow model predictions show that the seal mass leakage is proportional to the inlet pressure (Pin), increases as PR decreases, and is not affected by either shaft speed or the swirl brake configuration. Seal cavity static pressures drop linearly for all inlet pressures (Pin) and PR = 0.5 and above; except under a choked flow condition at PR = 0.3. Processing of the test data to consolidate the numerous leakage measurements delivers a nearly invariant flow factor Φ¯ for each seal clearance, and from this follows a unique orifice-like loss coefficient cd = 0.36 for the ILS with Cr = 0.3 mm, and cd = 0.33 for the ILS with Cr = 0.13 mm. This finding is remarkable as the test results obtained for the ILS with Cr = 0.2 mm also deliver a similar loss coefficient (cd = 0.36). Finally, predictions of ILS leakage and cavity pressures are within 5% of the measurements for all test conditions. The test data and predictions are of significant value to better the selection and design of gas labyrinth seals in turbomachinery.

The modern compressor operation is challenged by the liquid presence in wet gas operating conditions. The liquid phase may affect the compressor stability by partially flooding the internal annular gas seals and inducing subsynchronous vibration (SSV). To improve the annular seal behavior and increase the rotor stability, high-precision results of leakage flow rates and rotordynamic force coefficients are needed for annular gas seals in wet gas conditions. In order to better understand the leakage and rotordynamic characteristics of the annular gas seal in wet gas conditions, a 3D transient CFD-based perturbation method was proposed for computations of leakage flow rates and rotordynamic force coefficients of annular gas seals with liquid phase in main gas phase, based on inhomogeneous Eulerian-Eulerian multiphase flow model, mesh deformation technique, and the multifrequency rotor whirling orbit model. Numerical results of frequency-dependent rotordynamic force coefficients and leakage flow rates were presented and compared for three types of noncontact annular gas seals, which include a smooth plain annular seal (SPAS), a labyrinth (LABY) seal, and a fully partitioned pocket damper seal (FPDS). These three seals were designed to have the identical rotor diameter, sealing clearance, and axial length. The accuracy and the availability of the present transient CFD numerical method were demonstrated with the experiment data of leakage flow rates and frequency-dependent rotordynamic force coefficients of the smooth plain seal with four inlet liquid volume fractions (LVFs) of 0%, 2%, 5%, and 8%. Steady and transient numerical simulations were conducted at inlet air pressure of 62.1 bar, pressure ratio of 0.5, rotational speed of 15,000 rpm, and inlet preswirl ratio of 0.3 for four inlet LVFs varying from 0% to 8% and 14 subsynchronous and synchronous whirling frequencies up to 280 Hz. The numerical results show that the inlet liquid phase has a significant influence on the leakage and rotordynamic coefficients for all three types of annular gas seals. The mixture leakage flow rate increases with the increasing inlet LVF, combining the decreasing gas-phase and linearly increasing liquid-phase leakage flow rates. The smooth plain seal leaks the most gas phase and liquid phase, followed by the pocket damper seal (PDS) and then the labyrinth seal. Increasing inlet LVF significantly decreases the direct stiffness and slightly increases the effective damping of the smooth plain seal. The labyrinth seal possesses evident negative direct stiffness and shows a noticeable decreasing effective damping with the increasing inlet LVF at the subsynchronous frequency range. Increasing inlet LVF obviously increases all the force coefficients of the pocket damper seal including the positive effective damping. From a rotordynamic viewpoint, the FPDS possesses a better liquid tolerant capability and so is a better sealing scheme for the balance piston seals and center seals of the centrifugal compressor in wet gas operating condition.

Secondary seal leakage in jet engine applications results in power losses to the engine cycle. Likewise, seal power loss in jet engines not only result in efficiency loss but also increase the heat input into the engine resulting in reduced component lives. Experimental work on labyrinth and annular seals was performed at NASA Glenn Research Center to quantify seal leakage and power loss at various temperatures, seal pressure differentials, and surface speeds. Data from annular and labyrinth seals are compared with previous brush and finger seal test results. Data are also compared to literature. Annular and labyrinth seal leakage rates are 2 to 3 times greater than brush and finger seal rates. Seal leakage decreases with increasing speed but increases with increasing test temperature due to thermal expansion mismatch. Also seal power loss increases with surface speed, seal pressure differential, mass flow rate, and radial clearance. Annular and labyrinth seal power losses were higher than those of brush or finger seal data. The brush seal power loss was 15 to 30 percent lower than annular and labyrinth seal power loss.

Gas labyrinth seals: On the effect of clearance and operating conditions on wall friction factors – A CFD investigation

Hole-pattern or honeycomb seals have been commonly used for many years in the Oil & Gas industry as damper seals for turbomachinery. The main motivation has been to introduce additional damping to improve the shaft rotordynamic stability operating under high-pressure conditions. Experience has shown that the dynamic and even static characteristics of those seals are very sensitive to the operating clearance profile as well as the installation tolerances. Rotordynamic stability is related not only to the seal effective damping but to the effective stiffness as well. In fact, for this kind of seal, the effective stiffness can be high enough to alter the rotor system's natural frequency. The seal stiffness is strictly related to the tapering contour: if the clearance profile changes from divergent to convergent, the effective stiffness may change from a strong negative to a strong positive magnitude, thus avoiding the rotor natural frequency drop as it is detrimental for the stability. Unfortunately, the effective damping is reduced at the same time but this effect can be mitigated using proper devices to keep the preswirl low or even negative (e.g., swirl brakes and shunt holes). This paper presents the results from an extended test campaign performed in a high-speed rotor test rig equipped with active magnetic bearings (AMBs) working under high pressure (14 krpm, 200 bar gas inlet pressure), with the aim to validate the rotordynamic characteristics of a negative preswirl, convergent honeycomb seal and demonstrate its ability to effectively act as a gas bearing as well as a seal. The test plan included variations of inlet pressure, differential pressure (given the same inlet pressure), as well as rotational speed in order to fully validate the seal behavior. This kind of test was performed in a “dynamic mode” that is to say exciting the spinning test rotor through a pair of AMBs along linear orbits. Additionally, the impact of the seal to rotor static eccentricity and the seal to rotor angular misalignment were both experimentally investigated and compared to relevant computational fluid dynamics (CFD) simulations. This kind of test was performed in a “static mode,” that is to say imposing through the AMBs the required eccentricity/angular misalignment and then measuring the forces needed to keep the rotor in the original position. Dynamic mode test was also performed in order to check the impact of the seal static eccentricity on its dynamic behavior. Finally, the test results were compared with predictions from a state of the art bulk-flow code in order to check the predictability level for future design applications.

The double suction pump (DSP) is a centrifugal pump variant widely utilized in applications requiring higher head (H) and volumetric flow rates of fluid, owing to its ability to deliver robust performance under such conditions. In this study, computational fluid dynamics (CFD) techniques were employed to evaluate the performance of a designed DSP using a steady-state approach. Specifically, the investigation focused on assessing the impact of impeller stagger angle (θ) on pump performance under designed conditions. A comparative analysis was conducted between pumps equipped with staggered and non-staggered impellers to gauge their suitability across a broad range of flow rates. The stagger angle ranged from θ=0° to θ=30°, with θ=0° denoting a non-staggered impeller configuration. Results from the CFD study revealed that utilizing a staggered impeller arrangement led to improved pump performance in terms of head and efficiency compared to non-staggered impellers. However, performance variations with stagger angle adjustments between θ=15° and θ=30° were marginal, suggesting that stagger angle alterations had minimal impact on hydraulic performance under designed conditions. This preliminary investigation laid the groundwork for future unsteady simulations aimed at comprehensively assessing and understanding the unsteady pressure pulsations inherent in the designed pump. Leveraging CFD has helped the efficient acquisition of results and provided insights into the fluid flow dynamics within the pump.

It is a common practice to use closed impeller in radial inflow turbine against the flow leakage from tip clearance of impellers, especially in small volume flow condition. It utilizes labyrinths between the shroud and the case to abate the higher pressure leakage. Experimental and computational investigations of shroud clearance flow in a radial inflow turbine with labyrinth seals are presented in this paper. Compared with the result without leakage, numerical computation result including the leakage of labyrinth seals agrees better with that of the experiment result, which indicates that the leakage of labyrinth seals cannot be neglected. Several geometrical arrangements with a series of different clearance of labyrinth seals are investigated experimentally and numerically, and the dimensionless shroud clearance is of 0%, 0.6%, 1.2%, 1.8%, 2.7%, 3.6%. Finally, the character of flow and loss is analyzed by computational fluid dynamics (CFD) tools. The results indicate that the labyrinth seal flow has no effect on the main flow passage and mainly causes different leakage mass flow.

Shunt injection is one of the methods used to improve rotordynamic stability of centrifugal compressors. It involves eliminating swirl by injecting fluid in an intermediate cavity of a seal. The fluid injection is usually directed from the discharge volute or diffuser section of the compressor. In this paper, shunt injection is modeled for a center seal in a back-to-back compressor arrangement. The test case is from an actual machine which was shown to be unstable during a factory loaded mechanical test (as per API standard 617) without shunt injection and stable with shunt injection. The center seal is a tooth-on-rotor type of labyrinth seal. Impeller back-face cavity passages at the entrance and exit of the seal are also modeled. To investigate the influence on shunt injection, models with and without shunt injection geometry are considered. Steady state Computational Fluid Dynamics (CFD) is used to study the influence on steady state characteristics such as leakage rate and swirl ratio. The CFD analysis shows a significant swirl reduction with shunt injection at the injection location. The estimated swirl ratio is 0.55 at the injection location without shunt injection and approximately 0 with shunt injection. Frequency dependent seal rotordynamic coefficients are evaluated using transient CFD. Both translational and moment coefficients are evaluated for the seal and impeller passages. For the case without shunt injection, CFD results show negative (destabilizing) effective damping from the seal at the rotor first damped natural frequency. Shunt injection case shows a positive (stabilizing) estimated effective damping at the frequency of interest. The CFD results are shown to be consistent with the factory test case observations.

Small oil particles generated from lubricant of rolling bearings in HDD sometimes cause fatal problems as head crash. A labyrinth seal is, therefore, often used in HDD in order to prevent oil particles from scattering on the disk surface. In this paper, the seal performance of labyrinth seals for small oil particles is investigated, numerically and experimentally. Two types of a labyrinth seal are prepared. One is an annular seal and the other is a bent one. It was consequently found that the seal capability of a labyrinth seal is generated by the centrifugal force exerted on oil particles when they pass through the seal gap and that the passing rate of particles largely depends on the air flow rate through the seal gap.

Experimental testing for the influences of rotation and tip clearance on the labyrinth seal in a compressor stator well

Neural network and CFD-based optimisation of square cavity and curved cavity static labyrinth seals

In the submersible pump used in an LNG cargo, the leakage flow (gap flow) occurs between the pump rotor and the casing, and the pump efficiency decreases. The labyrinth-shaped pattern is applied between these gaps to decrease the leakage flow, which is called Labyrinth Seal. This research studied the difference in flow characteristics when liquefied methane which accounts for most LNG composition passes through labyrinth seals of different shapes using computational fluid dynamics (CFD). Four shape variables were considered: clearance size(s), cavity width(b), cavity height(g), and number of cavity(z). Parametric study was performed by changing only the input value of one variable without changing other variables within a limited range. As the gap of the labyrinth seal increased, the leakage flow rate increased. But the effect of the cavity-related variables was different. As a result of analyzing the effects of three cavity variables using full factorial design, they were found to be almost independent of each other. As the cavity width and number increase, the leakage flow rate decreases linearly. And the cavity height has a maximum leakage flow rate at one point, but has the smallest influence among all variables.

To minimize fluid leakage and improve rotordynamic stability of traditional seals, a new kind of radial annular seal is presented in this article. The structure and the sealing principle are analyzed. Compared to the traditional labyrinth seal, the fluid leakage direction is altered from the axial to radial direction. The flow resistance is increased, and the flow-induced force is reduced greatly. There is also a tremendous improvement in the wear resistance of the seal. Three-dimensional numerical models of the traditional labyrinth seal and radial annular seal are set up. Calculated results show that the discharge coefficient of the radial annular seal is about 20% lower than that of the labyrinth seal. Though the rotating speed has a small influence on the radial seal force for both the radial annular seal and labyrinth seal, it has a large effect on the tangential seal force. The tangential seal force increases with the increasing speed, and the destabilizing force in the radial annular seal is reduced greatly when compared to that in the labyrinth seal. The radial seal force of the radial annular seal is about 1.5–2.5 times as strong as that of the labyrinth seal. The tangential seal force is about 30–50% lower than that of the labyrinth seal under the same eccentricity, and this ratio becomes larger and larger with the increasing speed. Finally, the dependency of the seal characteristics on the design parameters is studied. It is shown that the optimum ratios of the clearance s to the tooth thickness b ( s/ b) for the outward and inward flows are 4.5–5.5 and 11–13, respectively. The flow characteristics perform best with the ratio of the seal chamber width B to the tooth height H ( B/ H) on the same order of 5.5–7 for the two kinds of flows.

The ability to quantify leakage flow and windage heating for labyrinth seals with honeycomb lands is critical in understanding gas turbine engine system performance and predicting its component life. Variety of labyrinth seal configurations (number of teeth, stepped or straight, honeycomb cell size) are in use in gas turbines, and for each configuration, there are many geometric factors that can impact a seal's leakage and windage characteristics. This paper describes the development of a numerical methodology aimed at studying the effect of honeycomb lands on leakage and windage heating. Specifically, a three-dimensional computational fluid dynamics (CFD) model is developed utilizing commercial finite volume-based software incorporating the renormalization group (RNG) k-ε turbulence model with modified Schmidt number. The modified turbulence model is benchmarked and fine-tuned based on several experiments. Using this model, a broad parametric study is conducted by varying honeycomb cell size, pressure ratio (PR), and radial clearance for a four-tooth straight-through labyrinth seal. The results show good agreement with available experimental data. They further indicate that larger honeycomb cells predict higher seal leakage and windage heating at tighter clearances compared to smaller honeycomb cells and smooth lands. However, at open seal clearances larger honeycomb cells have lower leakage compared to smaller honeycomb cells.

In straight-through centrifugal pumps, a grooved seal acts as a balance piston to equilibrate the full pressure rise across the pump. As the groove pattern breaks the development of fluid swirl, this seal type offers lesser leakage and lower cross-coupled stiffnesses than a similar size and clearance annular seal. Bulk-flow models (BFMs) predict expediently the static and dynamic force characteristics of annular seals; however they lack accuracy for grooved seals. Computational fluid dynamics (CFD) methods give more accurate results, but are not computationally efficient. This paper presents a modified BFM to predict the rotordynamic force coefficients of shallow depth, circumferentially grooved liquid seals with an accuracy comparable to a CFD solution but with a simulation time of bulk-flow analyses. The procedure utilizes the results of CFD to evaluate the bulk flow velocity field and the friction factors for a 73 grooves annular seal (depth/clearance dg/Cr = 0.98 and length/diameter L/D = 0.9) operating under various sets of axial pressure drop and rotor speed. In a groove, the flow divides into a jet through the film land and a strong recirculation zone. The penetration angle (α), specifying the streamline separation in the groove cavity, is a function of the operating conditions; an increase in rotor speed or a lower pressure difference increases α. This angle plays a prominent role to evaluate the stator friction factor and has a marked influence on the seal direct stiffness. In the bulk-flow code, the friction factor model (f = nRem) is modified with the CFD extracted penetration angle (α) to account for the flow separation in the groove cavity. The flow rate predicted by the modified bulk-flow code shows good agreement with the measured result (6% difference). A perturbation of the flow field is performed on the bulk-flow equations to evaluate the reaction forces on the rotor surface. Compared to the rotordynamic force coefficients derived from the CFD results, the modified bulk-flow code predicts rotordynamic force coefficients within 10%, except that the cross-coupled damping coefficient is over-predicted up to 14%. An example test seal with a few grooves (L/D = 0.5, dg/Cr = 2.5) serves to further validate the predictions of the modified BFM. Compared to the original bulk-flow analysis, the current method shows a significant improvement in the predicted rotordynamic force coefficients, the direct stiffness and damping coefficients, in particular.