Search Results

Generalized predictive control is a discrete time control strategy that was developed in the late 1980’s by Clark et al [1]. The controller tries to predict the future output of a system or plant and the take control action at present time based on future output error. It is the predictive nature of the controller that is utilized in having an effective noise attenuation in a 1-D noise propagation case (such as a cylindrical enclosure where noise propagates only in one direction). Simulation results are presented for the Generalized predictive controller (GPC) and then the results are compared against a more traditional controller.

A Generalized Predictive Kinetic Energy Controller (GPKEC), which has been previously developed, is implemented to control the machine tool vibration in a lathe turning process. The control variable, tool feed, is computed using acceleration feedback through the GPKEC algorithm. The feed is controlled by attaching a high torque/low inertia permanent magnet DC servomotor to the main feed rod through a high performance timing belt. Experiments are carried out for a number of cases. The system identification of the overall machining process is done on-line and precedes the control action. Accelerometers have been used to sense the vibration signal in the feed direction. The experimental results show that GPKEC can effectively suppress the chatter vibration in a single point turning process, even in presence of an appreciable change in the dynamics of the process. GPKEC has also been observed to be robust against step disturbances.