A Fast and Effective Method for the Optimization of the Valve Plate of Swashplate Axial Piston Pumps

Abstract

Abstract This research presents a lumped parameter numerical model aimed at designing and optimizing an axial piston pump. For the first time, it has been shown that a lumped parameter model can accurately model axial piston pump dynamics based on a comparison with computational fluid dynamic (CFD) models and experimental results. Since the method is much more efficient than CFD, it can optimize the design. Both steady-state and dynamic behaviors have been analyzed. The model results have been compared with experimental data, showing a good capacity in predicting the pump performance, including pressure ripple. The swashplate dynamics have been investigated experimentally, measuring the dynamic pressure which controls the pump displacement; a comparison with the numerical model results confirmed the high accuracy. An optimization process has been conducted on the valve-plate geometry to control fluid-born noise by flow ripple reduction. The nonlinear programming by quadratic Lagrangian (NLPQL) algorithm is used since it is suitable for this study. The objective function to minimize is the well-known function, the nonuniformity grade (NUG), a parameter directly correlated with flow ripple. A prototype of the best design has been realized and tested, confirming a reduction in the pressure ripple. An endurance test was also conducted. As predicted from the numerical model, a significant reduction of cavitation erosion was observed.