| dc.description.abstract | This thesis presents the test procedures and results of a study to create an alternative to plain annular seals as radial rotor supports in Electrical Submersible Pumps (ESPs). Currently, these pumps are assembled with annular seals which are susceptible to significant circumferential wear, resulting in increased radial clearance and degraded centering capabilities. Centering forces in annular seals arise due to hydrodynamic effects (fluid rotation) and hydrostatic effects (axial pressure drops that use the Lomakin Effect to get centering forces). The author’s attempt to circumvent this issue integrates pressure-dams to improve the hydrodynamic centering forces.
Three pairs of seals were manufactured for this study: (1) 1X-clearance smooth seals (baseline pair), (2) 1X-clearance pressure-dam seals, and (3) 2X-clearance pressure-dam seals. The objective was to compare the centering capabilities of the three seals to determine if the pressure-dams were effective. The test rotor diameter was 116.83 mm (4.5998 in.). The 1X-clearance smooth and pressure-dam seals were manufactured to have a minimum 0.127 mm (0.005 in.) radial clearance. The 2X-clearance pressure-dam seals had a radial clearance of 0.254 mm (0.01 in.). Both pairs of pressure-dam seals were machined to include three pressure-dams of equal arc angle and axial length, but different recess depths. The pressure-dams were designed via results from Nicholas.
The pressure-dam seals were tested in two load orientations; load on dam (LOD) and load on land (LOL). Test conditions for all seals includes four rotor speeds (1500, 3000, 4500, and 6000 RPM), four axial pressure drops (2.1, 4.1, 6.2, and 8.3 bar), and four eccentricity ratios (0.0, 0.4, 0.6, and 0.8). The lubricant is ISO VG46 at 115°Fahrenheit, which has a dynamic viscosity of 0.0306 Pa-s.
Static test results include seal loci and seal leakage. The 1X-clearance pressuredam seal leaks at least twice as much as the 1X-clearance smooth seal in every static position. The 2X-clearance pressure-dam seal leaks nearly 5 times more than the 1X-clearance pressure-dam seal.
Dynamic results include: (1) rotordynamic coefficients (direct stiffness, damping, and virtual mass as well as cross-coupled stiffness and virtual mass) and (2) whirl frequency ratios; the means by which these are arrived at is outlined in the proceeding sections. Results for the pressure-dam seals show a significant number of negative-direct-stiffness coefficients and whirl frequency ratios nearing 1. In general, the pressure-dam seals are out-performed by the smooth seals, which have whirl frequency ratios of approximately 0.5. In addition, their rotordynamic coefficients are sensitive to load orientation, making them poorly suited for vertical operation.
When the radial clearance was doubled, the pressure-dam seals did not have a stable equilibrium position under load; thus, the results are limited to centered (unloaded) conditions. Under these conditions, the pressure-dam seals do not retain their centering forces, and a large portion of the test data reflect negative direct stiffness coefficients (≥ 50%).
XLAnSeal of the XLTRC^2 software suite was used to predict the leakage and rotordynamic coefficients of the smooth seal. Results show good agreement with measurements. No code exists for predicting the static and dynamic characteristics of annular seals with pressure-dams.
All test-flow conditions remained in the laminar regime. The presented results include seal leakage, seal loci under varying load, direct and cross coupled stiffness, damping, and virtual mass coefficients, and whirl frequency ratio. | en |
