Fuzzy logic controls for CNC machine tools.

Abstract

To improve the contouring accuracy of CNC systems in a cost-effective way, great efforts have been made to develop sophisticated servo-controllers. The objective of this research has been to develop advanced servo-controllers to improve the contour tracking accuracy of machining systems in the presence of large disturbances. To meet this objective, three rule-based fuzzy logic controllers for feed drive systems are presented: a fuzzy logic controller (FLC), a self-organizing fuzzy logic controller (SOFLC), and a fuzzy logic cross-coupling controller (FLCCC). These rule-based fuzzy logic controllers can overcome the drawbacks of conventional servo-controllers, which utilize model-based approaches requiring exact process models. In the proposed FLC, a proportional and derivative type of fuzzy logic controller is combined with a conventional integral controller. In addition, to provide a baseline for the controller design, stability analyses of fuzzy logic control systems are introduced. In the proposed SOFLC, the controller parameters are automatically tuned in real-time according to a continuous measurement of the performance of the controller itself and estimated disturbance values. This capability of self-tuning reduces the complex tuning procedure needed for each machine when installing fuzzy logic controllers, and also enables the controllers to cope with changing operating conditions during machining. In the proposed FLCCC, fuzzy logic control is incorporated into a cross-coupling architecture to overcome the disadvantages of the decoupled axial controllers and the conventional cross-coupling controller. The proposed fuzzy logic controllers, as well as a conventional PID controller, were simulated and implemented on a CNC milling machine. Evaluation and comparison were done in contour milling for various contour shapes and feedrates. Both the simulations and the experiments show that the proposed servo-controllers improve the contour tracking accuracy of machining systems with large disturbances. These fuzzy logic controllers do not have an advantage over conventional controllers for high-end machine tools which have a fine quality of feed drive systems. On the other hand, the fuzzy logic controllers can improve the performance of machine tools which have large and variable friction in the guideways, large variations in load, deflection of the leadscrews and backlash in the gears, and therefore cannot be controlled effectively by conventional control methods. In addition, the fuzzy logic controllers are not sensitive to feedrate changes and fit in high-feedrate machining for high productivity which usually causes large contour errors especially for nonlinear contours when conventional controllers, such as a PID controller, are utilized.