Global energy efficiency improvement of redundant hydraulic manipulator with dynamic programming

Abstract

Open-loop controlled hydraulic manipulators are still dominating in construction sites and forest fields thanks to little reliance on the expertise of human operators in manipulator control, which however exist large inlet/outlet and other energy losses during working cycles. And the situation further deteriorates with the increasing number of joints. This paper discusses the problem of global energy optimization of three-degrees-of-freedom (3-DOF) hydraulic manipulator in case of plane motion where exists an exceeding DOF for motion redundancy. Different from conventional energy optimization methods applied to electric driven manipulators, the proposed global energy-optimized dynamic programming (DP) algorithm models the whole system at the hydraulic level, which contains the dynamic characteristics of pressure-flow coupling between cylinders. The energy consumption models of constant pressure (CP), load-sensing (LS) and electrohydraulic load-sensing (ELS) systems are built in terms of cost functions formulated in the proposed algorithm together with penalty function containing physical actuator constraints at position, velocity and acceleration level. The DP algorithm is tested in various cases, and the convergence is proved consistent. To highlight the effectiveness of the proposed algorithm, the gradient projection (GP) method and the actuator velocity minimum norm (MA) method are introduced as a contrast. The results of the numerical examples of end-effector paths using the mechanical-hydraulic coupling AMESim model demonstrate that the DP algorithm saves around 10.49% hydraulic energy consumption in the CP system while saving around 76.55% in the LS system. Experiments performed on a typical-structured hydraulic manipulator further validate the effectiveness of the proposed algorithm, with approximately 4.61% energy saving in the CP system and 29.89% in the LS system. The energy-saving ratio is even larger in the ELS system, with around 80.03% comparing the max value. These contribute to the energy optimization resolution of redundant hydraulic manipulators to a degree.