Abstract
An integral squeeze film damper (ISFD) offers the advantages of a lower number of parts, a shorter axial span, a lighter weight, a split manufacturing, and high precision on its film clearance construction. An ISFD does not only add damping to reduce shaft vibration amplitudes and to enhance the stability of a rotor-bearing system but also can be used to tune a rotor-bearing system natural frequency, and thus increasing the operational safety margin between the running shaft speed and the system critical speed. In spite of the numerous commercial applications, the archival literature is scant as per the experimental quantification of force coefficients for ISFDs. This paper details the results of an experimental and analytical endeavor to quantify and to predict the dynamic force coefficients of an ISFD, hence bridging the gap between theory and practice. With an axial length of 76 mm, the test damper element has four arcuate film lands, 73 deg in arc extent at a diameter of 157 mm, and each with a clearance (c) equaling to 0.353 mm. As is customary, the damper has its axial ends sealed with end plates produced by a set of installed shims giving an axial gap (d) equal to 1.5c, 1.21c, and 0.8c. A baseline configuration, namely, open ends, is also tested without the end seals in place. In the test rig, the ISFD and its housing are flexibly mounted while the rotor is rigid and stationary (no spinning). The lubricant is an ISO VG46 oil supplied at a low pressure, 1 to 2 bar(g) and ∼47 °C inlet temperature, typical of compressor applications. The test procedure applies static loads on the ISFD and records the bearing static offset or eccentricity to verify the structure stiffness, and meanwhile, individual hydraulic shakers deliver dynamic loads along two orthogonal directions to produce motions over a set frequency range, 10 Hz to 160 Hz. The ISFD produces direct damping and inertia that increase with the journal static eccentricity albeit at a lower rate than predictions from a computational squeeze film flow model that includes lubricant compressibility. The end seals are effective in significantly raising the damping coefficient while reducing the oil through flow rate. The damper with the tightest sealed ends (d = 0.8c) shows nearly 20 times more damping that the open ends ISFD albeit also revealing a significant stiffness hardening (negative virtual mass) as the excitation frequency increases. On the contrary, the open ends ISFD and the sealed-ends configurations with gaps d = 1.21c and 1.5c produce a (positive) virtual mass that exceeds the test element physical mass and thus softens the test element direct dynamic stiffness. For the configurations with loose end seals (d = 1.21c or larger to open ends), the model predicts well the damping coefficients but under predicts the added masses by 50% or more. Note this virtual mass coefficient, largely ignored in practice, can make the test element either extremely stiff as with the sealed damper configuration with the smallest gap d = 0.8c, or very soft as with the ISFD with end seals gap = 1.21c or 1.5c. Hence, designers are cautioned not to pursue overly tight end sealed dampers as the mineral lubricant, nearly incompressible though always having a small amount of entrapped gas, may behave distinctly when confined to a squeezed film volume and having no adequate routes to escape or flow through.
Nonlinear vibration response characteristics of a dual-rotor-bearing system with squeeze film damper
