Supervisor:
Lasse Schmidt

Co-Supervisor:
Michael Møller Bech

Assessment Committee:
Anders Hedegaard Hansen(Chair)
Professor, Andrea Vacca, Purdue University, West Lafayette, Indiana, USA
Professor, Perry Y. Li, Department of Mechanical Engineering, University of Minnesota, Minneapolis, USA

Moderator:
Per Johansen

Abstract:

Excavators, injection moulding machines, and hydraulic presses are examples of equipment powered by oil hydraulics where very large forces and high impact resistance are required. In many applications, the overall energy efficiency is low, which converts a substantial share of input energy into heat rather than useful work. This has two main consequences: (i) higher total energy consumption and thus greater greenhouse-gas emissions regardless of the energy source, and (ii) a barrier to battery-powered mobile machinery, because low efficiency either demands larger and more expensive batteries or more frequent charging.

This thesis investigates hydraulic drive networks (HDN), in which hydraulics act as a mechanical gear between subsystems and hydraulic coupling mechanisms enable direct energy exchange between actuators (cylinders/motors. In an idealised sense, power transfer can occur without conceptual conversion losses (apart from friction, leakage, and similar effects) and regenerative energy can be utilized more directly. This provides potential for higher efficiency and smaller component sizes, reducing both capital costs and operating energy costs.

Two coupling mechanisms are considered. (1) A hydraulic short circuit between working chambers, which equalizes pressure levels and allows free oil exchange analogous to an electrical short circuit. (2) Cross-coupling via hydraulic machines, where machines are not only used to pump from tank but also to interconnect actuators. Because HDNs can be assembled in a very large number of network configurations, methods are presented for the systematic identification and enumeration of all physically meaningful configurations, enabling algorithmic multi-objective optimisation to compare architectures by, for example, component size and efficiency.

A two-cylinder hydraulic system is tested experimentally to quantify the effect of a hydraulic short circuit under different phase shifts between the cylinders’ sinusoidal motions. For parallel cylinder motion (no flow through the short circuit), the average efficiency is 37%. For offset cylinder motion (maximal flow through the short circuit), the efficiency reaches 40%, corresponding to a relative improvement of 8%. Assuming a lossless short circuit, the efficiency is estimated at 45%, i.e., a relative improvement of 21%. These results highlight the importance of low pressure drop and friction in the short-circuit path.

As a case study, the implement (three cylinders; boom, arm, bucket) of a 15 tonnes excavator is considered. A design study identifies a battery-powered HDN architecture that minimizes the sum of the peak torque demands on the electric machines as well as energy losses. The architecture comprises four drive units (electric motor + hydraulic motor/pump) and two hydraulic short circuits and is tested over digging cycles without rotation of the upper structure. Cylinder control is satisfactory and permits both individual and simultaneous motion. The average efficiency from the battery terminals to mechanical power at the cylinders is 38%. Significant losses are observed when transmitting hydraulic energy from the drives to the cylinders due to long, narrow hoses. If these hydraulic transmission losses are neglected, the efficiency from the battery to the HDN output is 59%.

The analyses support the hypothesis that an optimal HDN architecture can be identified through systematic comparison of all unique networks. A central caveat is that optimality depends on the machine’s operating tasks and requirements as well as on uncertainties in available component data (dimensions, efficiency, pressure/flow losses, and operational limits).