Pulsation in Digital Displacement Hydraulic Systems

Abstract

Digital Displacement pumps (DDP) are a type of radial piston hydraulic pump in which the displacement of each piston can be selected on a stroke-by-stroke basis through control of a solenoid actuated valve at the inlet of each cylinder. These pumps are known to have an efficiency advantage compared to traditional swashplate pumps, however, a major challenge for digital hydraulics is the fluid-borne and air-borne noise associated with the transient switching events and cylinder selection. This is also exacerbated by the improved efficiency, since the swashplate pump is often a major source of damping in traditional hydraulic systems.
This thesis addresses the problem of fluid-borne noise with DDP through improvements to the control algorithm used to generate the cylinder enabling sequence necessary to satisfy the displacement demand. In the past, control algorithms developed for DDP seek to satisfy the demand by generating a specific enabling sequence, which may use the full or partial displacement of each cylinder. The most significant factor influencing the fluid-borne noise produced by the pump is the cylinder enabling sequence itself, since the frequency of commutation of the cylinders is directly reflected in the generated flow and pressure ripple.
In this work a novel concept is presented, whereby the pump command is separated into two terms: one controlling the cylinder enabling sequence and another for the cylinder stroke sizes used within that sequence. By separating the demand in this way, it is possible to determine and define the frequencies of fluid-borne noise which will be generated by the enabling sequence without strictly satisfying the demand. Then, the overall displacement can be adjusted by using the appropriate stroke sizes within that sequence. Separating the choice of cylinder enabling sequence from the stroke sizes within the sequence also allows the application of modulation to the stroke size command, which is shown to significantly alter the fluid-borne and air-borne noise characteristics of the pump, without any change to the overall delivered displacement. This gives a new way to control the tonal content in the fluid pulsation produced by the digital pump. Furthermore, a technique is demonstrated whereby the pump itself can act as a frequency source, applying a ‘chirp’ input to the downstream hydraulic system and allowing for the estimation of transfer functions of the system response, which can then be used to adjust the pump fluid-borne noise to avoid regions of frequency sensitivity in the wider system.
The enabling algorithm is developed first using a time-domain model of the pump and associated control software, and then this software is deployed and tested in a test rig with a 12-cylinder DDP. The test rig is equipped to measure the fluid-borne noise generated by the pump using an array of high frequency piezoelectric pressure sensors, which allows the accurate characterisation and comparison of the pump pressure and flow ripple when using a variety of control algorithms. The pump control software is also tested in a 20-tonne excavator, and the noise and vibration impact of some of the developed control algorithms assessed in the context of a realistic application.

Date of Award	10 Dec 2025
Original language	English
Awarding Institution	University of Bath
Supervisor	Nigel Johnston (Supervisor), Andrew Plummer (Supervisor) & Nathan Sell (Supervisor)