Design of controllers for nonlinear electrohydraulic systems with time delays

One particularly sensitive and difficult problem in current applications of the electrohydraulic control systems to agricultural machine and equipment is that the controlled systems possess unknown or varying time delays. The TUTSIM simulations and laboratorial tests have been conducted to evaluate the effects of the time delays on the performance of classical PD (proportional-derivative) controller and Smith dead-time compensator. The experimental results demonstrate that the PD controller can only handle a small constant delay problem at the sacrifice of high control gains or the response speed. The Smith dead-time compensator provides an effective way to control the system with known constant time delay. However, the system will become unstable under considerable delay modelling errors or varying time delay situations. Therefore, pole-placement self-tuning regulator was designed to resolve this special non-minimum phase problem occurred in the control practice in agricultural area. The control principle is based on a pole-placement method which places the system closed-loop poles to prescribed locations which define the required system transient response. A recursive least-squares approach is adopted for the process parameter identification which implicitly estimates the time delay by extending the numerator polynomial of the plant transfer function to include the time delay. Thus the delay approximation is simplified to the on-line estimation of the extended process parameters at each sampling time.; A unique graphical method generated from the test and simulation has been developed to evaluate and compare the general performance of the PD controller, Smith dead-time compensator, and self-tuning regulator, in terms of relative stability, absolute stability and response speed, in the presence of the process time delay. As compared to the PD controller and Smith dead-time compensator, the test results have proved that the pole-placement self-tuning regulator can produce more stable and accurate process responses in unknown or quickly varying time delay cases.