Predicting lubrication performance between the slipper and swashplate in axial piston hydraulic machines

Andrew Schenk, Purdue University

Abstract

Engineering of the sliding interfaces within swashplate type axial piston machines represents the most complex and difficult part of the design process. The sliding interfaces are subject to significant normal loads which must be supported while simultaneously preventing component wear to ensure long lasting operation. Proper lubrication design is essential to separate the solid bodies from each other, but the complexity of the physics involved makes this a difficult problem. This work focuses on lubrication and the resulting energy losses at the sliding interface between the slipper and swashplate. To better understand the slipper lubrication performance, a numerical model has been developed to predict the behavior of a design. The numerical model considers the multi-physics, multi-scale, and transient nature of the lubrication problem by utilizing novel segmented physics solvers and numerical techniques. Partitioned solvers considering the fluid pressure and temperature distributions, structural deformation due to fluid pressure and viscous heating, as well as a solid body dynamics from transient loads have been originally developed and tightly coupled. Although the effort necessary to implement this was significant, by avoiding a more traditional co-simulation approach, high computational efficiency and model fidelity can be achieved. To validate the developed numerical model, a specialized test rig was designed and manufactured. Miniature high-speed inductive position sensors were mounted inside the swashplate of a commercially manufactured pump with only minimal modifications. These six sensors measured the distance between the sensor face and the slipper land as the slipper passed over the sensor, effectively measuring the direct film thickness in real time. The thickness of lubrication represents the greatest unknown predicted by the model and provides the most rigorous validation as well as experimental insight into actual slipper operation. New slippers were installed in the test rig, measured, and then following a period of operation, were measured again. A significant change in film thickness behavior was measured due to the presence of a worn slipper surface during the second period of testing, and this same behavioral change was captured with the simulation model. The developed numerical model was used to conduct case studies demonstrating the potential of virtual pump lubrication design. Slipper sensitivity to operating conditions and materials were explored. Operational changes such as slipper tipping and liftoff at high speeds were numerically observed and would serve to aid a designer in improving the robustness of a design. A multi-modeling approach using a surrogate model based upon a design of experiment study and the full numerical model explored the inter-dependence of variables in a multi-land slipper design. In particular, a decrease in total power loss while increasing the outer stabilizing land width at a constant hydrostatic balance factor was observed for low pressure operation.

Degree

Ph.D.

Advisors

Ivantysynova, Purdue University.

Subject Area

Mechanical engineering

COinS