High speed machining technology has been rapidly adopted in aerospace, die and mold manufacturing industry for its high productivity. High speed machine tools require a rigid structure, thermally and dynamically stable spindles with high power, and fast feed drives which are able to track complex tool paths accurately at feed speeds up to 40 [m/min] with high accelerations over 1 [g]. The design of trajectory generation and control algorithms play a crucial role in realizing the accuracy requirement for high speed feed motion. This thesis presents a systematic approach to designing a smooth trajectory generation algorithm and a high performance control system for machine tool feed drives.
A jerk limited trajectory generation algorithm employing trapezoidal acceleration profiles is developed to minimize discontinuity and harmonics in actuation force. The original position commands with varying interpolation period are re-sampled at control loop frequency via fifth order polynomials. The generated smooth trajectory commands for individual axes are delivered to a control system designed for accurate tracking and disturbance robustness. Axis dynamics are first stabilized via pole-placement control. Overall bandwidth is increased with a zero phase error tracking controller to minimize tracking errors. Disturbance rejection and parameter variation robustness is achieved using a Kalman filter based disturbance observer. Friction forces are compensated for in feedforward to improve the tracking accuracy at sharp corners and circular quadrants. On top of these, the contour error is also estimated in real-time and used in cross-coupling control via PID controllers, to achieve additional contouring accuracy in the presence of cutting forces.
The effectiveness of the proposed trajectory generation and control scheme is verified both in simulations and in experiments, where a high speed x-y table driven by linear motors is used.