Error compensation in diamond turning via on-line metrology and robust control

Advances in technology increase the need for higher precision parts that can be best produced by the diamond turning process. Error compensation technology extends the performance of machine tool used in the diamond turning beyond the structure accuracy level to an intelligent level, thus improving the workpiece accuracy efficiently. Three issues have been addressed in this research for the error compensation in the diamond turning.; The first issue is the on-fine metrology, which is an approach used to measure the workpieces without removing them from the machine tool. To overcome the difficulty of on-line measurement of short workpieces, an Optimum Error Separation Technique (OEST) has been developed in the research. The new method enables the use of a much smaller sampling interval than is permitted by the existing technique. Consequently, it is suitable for the measurement of short workpieces.; The second issue is the modeling of hysteresis. In this research, the popular Preisach model has been used to model the hysteresis nonlinearity. To simplify the implementation of the Preisach model, a learning algorithm was developed to obtain the Preisach distribution function. This approach brings in a direct utilization of the Preisach model. In addition, it makes the dynamic and inverse model of hysteresis available for the control applications.; Micro positioning control is the third issue. The micro positioning system is used to accomplish the active error compensation. The control accuracy is the primary factor for the error compensation in the micron level. To improve the closed loop control accuracy and system stability, a multi-loop feedback control algorithm has been developed. This algorithm takes advantage of the hysteresis modeling technique to overcome the hysteresis nonlinearity, and further improve the robustness and stability of the micro positioning system. A simple PID controller tuned by the Powell's method has been introduced as a benchmark for the new control algorithm. In addition, a repetitive learning control algorithm, P-integrator control, has also been adopted to close the loop at the workpiece to further improve the workpiece accuracy.; In this research, a diamond turning lathe setup was retrofitted as a testing bed for the issues. The motion controls of the lathe are computerized, a magnetic coupling driving system has been designed and manufactured; the micro positioning system has been designed and fabricated; and a PC-DSP based implementation system has been set up. Real cutting experiments have been conducted to test the aforementioned techniques.