Modeling and simulation of an electro-hydraulic actuation system through subsystem characterization

An electro-hydraulic actuation system with flow and pressure transmitters and LVDT integrated to a data-acquisition system has been investigated here for identifying an appropriate model necessary for dynamic performance prediction through MATLAB/SIMULINK simulation. The actuation test system consists of an electrically driven hydraulic power pack feeding a linear actuator with a proportional valve controlling the flow to and from the actuator, while the power pack includes a reservoir, fixed-discharge axial piston pump, a pressure relief valve and a non-return valve. A non-linear state-space model for the system has been obtained with the command electrical signal to the proportional valve as the input and the actuator displacement as the output. Experimental characterization of actuator friction, the relief valve, the non-return valve and the transmission line along with a neural-network model for the proportional valve has been carried out and integrated to the model of the actuation dynamics. The model has then been used for performance prediction to assess the importance of each subsystem model through a comparison against experimental results. It has been found that the contributions of the valve modeling and actuator friction are significant. A need to characterize the end-cushioning effect has been identified.

I. Introduction

Electro-hydraulic systems have higher power-to-weight ratio and higher speed of response in comparison to electrical actuators. But, non-linear behavior of the system poses challenge to modeling and control of electro-hydraulic systems. These non-linearities arise from friction between the cylinder and piston of the actuator, unequal volume of liquid in the two chambers of the cylinder across the piston, characteristics of the different types of valves and the losses in the transmission lines connecting the power pack, valves, actuator and reservoir that comprise a typical electro-hydraulic system, as shown in Fig. 1. The major trend of control analysis [1] has been to arrive at a linearized model for designing the control and analyzing the system performance. Performance of such linear controllers is adequate only in the face of small excursions about an operating point. Advancement in the computational power and modern FPGA-based controller architecture makes the use of non-linear modeling and the corresponding controller design a viable alternative to linearized modeling. Though the development of non-linear control theories through Lyapunov approach, variable-structure system and soft computing is significant, the experimental analysis of the nonlinear behavior of the system has by and large remained confined to sub-system design and performance analysis.