Automated onboard modeling of cartridge valve flow mapping

作者：Liu Song^*; Yao Bin

来源：IEEE/ASME Transactions on Mechatronics, 2006, 11(4): 381-388.

DOI：10.1109/TMECH.2006.878552

• electrohydraulics
• nonlinear system identification
• onboard modeling
• parameter estimation
• valves

Proportional poppet-type cartridge valves are the key elements of the energy-saving programmable valves, which have been shown to be able to achieve good motion control performance while significantly saving energy usage in our previous studies. Unlike costly conventional four-way valves, the cartridge valve has a simple structure and is easy to manufacture, but the complicated mathematical model of its flow mapping makes the controller design and implementation rather difficult. Although off-line individually calibrated or manufacturer supplied flow mappings of the cartridge valves can be used, neither method is ideal for wide industrial applications. The former method is time-consuming and needs additional flow sensors while the latter may lead to significantly degraded control performance due to the inaccuracy of the manufacturer supplied flow mappings. Furthermore, due to inevitable system worn out and/or changing working conditions, actual cartridge valve flow mapping may change significantly over the life span of the system and need to be updated periodically in order to maintain the same level of control performance. Sometime, it may be even impractical to do off-line calibrations once the valve leaves the manufacturing plant. To solve this practically significant problem, this paper focuses on the automated onboard modeling of the cartridge valve flow mappings without using any extra sensors and removing the valves from the system. The estimation of flow mappings is based on the pressure dynamics of the hydraulic cylinder with the consideration of effects of some unknown system parameters such as the effective bulk modulus of the working fluid. Localized orthogonal basis functions are proposed to bypass the lack of persistent exciting identification data over the entire domain of the flow mapping during onboard experiments. Experimental results are obtained to illustrate the effectiveness and practicality of the proposed novel automated modeling method.

• 出版日期2006-8
• 单位浙江大学