Structural test rigs may incorporate between 1 and 30 hydraulic actuators each providing a load or displacement control. A closed loop PID controller is most commonly used to provide displacement or load control. Normally, industrial controllers provide a relatively high operational bandwidth of the actuator and a good performance within its operational range. The problem arises when the load is changing over a considerable range and sometimes in an unexpected manner. This causes undesirable interaction between actuators and results in degradation of the reference signal tracking.
The controller proposed in this thesis is designed to meet aforementioned challenges facing the control system for modern multi-axis rigs: to reach satisfactory reference signal tracking without the need for iteration and robust performance in the presence of interactions and disturbances. The controller is a novel extension of the Model Reference Adaptive Control method known as Minimal Control Synthesis (MCS). It is arranged to function seamlessly through integration with a conventional test rig controller. Furthermore the controller employs important features which allow model following behaviour of the hydraulic cylinder and, to a large extent, to compensate for the actuator dynamics by the inverse model extension. A theoretical tool has been proposed for the analysis of the adaptation dynamics of the controller. A multi-axis scheme for the controller utilises a decentralised approach which implies individual control of each cylinder of the multi-axis system with coupling effects treated as disturbances, which should be attenuated by the adaptive controller.
Experimental studies for single-axis and multi-axis variants of the control scheme were conducted on a six-axis MAST. Very good dynamic performance was demonstrated relative to a standard industrial controller. However, it was found that the bandwidth and disturbance attenuation property of the controller were compromised to some degree because it was not feasible to include acceleration feedback signal in the controller structure.
|Date of Award
|21 Jan 2009
|Kevin Edge (Supervisor)