Automated modeling and energy saving adaptive robust control of electro-hydraulic systems with programmable valves

As the applications of electro-hydraulic systems become increasingly widespread, the demand for low cost, high-level control performance and significant energy saving schemes gets stronger and stronger. The thesis proposes an intelligent mechatronics approach---the integration of advanced control techniques with novel hardware reconfiguration and/or re-design---to meet these industrial needs. The Purdue programmable valves; a unique configuration of five independently controlled poppet type cartridge valves, enable not only the decoupled meter-in and meter-out flow controls but also the precise control of the cross-port regeneration flow. When properly controlled, the increased hardware flexibility and controllability lead to significant energy-saving, due to the reduced working pressures and the full use of free regeneration flows. However, the increased hardware flexibility also results in increased complexity in controlling such a multi-input-dual-objective system. A two-level coordinated control scheme is proposed in the thesis: the task-level configures the valve usage for maximal energy saving, and the valve-level utilizes Adaptive Robust Control (ARC) technique to guarantee the closed-loop system stability and performance under various model uncertainties and disturbances. Comparative experimental results were obtained to show the high-level control performance and significant energy saving achieved with the proposed low cost programmable valves. Relevant control issues, such as using the manufacturer supplied flow mappings and the deadband problem of electro-hydraulic systems, are discussed in detail in the thesis as well. Lack of suitable mathematical model has limited the cartridge valve to low cost applications where precision motion control was not of major concern. The thesis proposes an on-board modeling approach to decompose and approximate the unknown valve flow mapping with some localized orthogonal basis functions. The weighting parameters of the basis functions as well as the unknown system parameters were then estimated simultaneously based on the pressure dynamics of the cylinder for regions where sufficient on-board measurement data were available. Experimental studies have been obtained to demonstrate the feasibility of the proposed method and the improvement of control performance with the estimated flow mappings