No abstract

Precision requirements in ultra-precision machining are often given in the order of micrometers or sub-micrometers. Machining at these levels requires precise control of the position and speed of the machine tool axes. Furthermore, in machining of brittle materials, extremely low feed rates of the machine tool axes are required. At these low feed rates there is a large and erratic friction characteristic in the drive system which standard PID controllers are unable to deal with. In order to achieve the desired accuracies, friction must be accurately compensated in the real-time servo control algorithm. A learning controller based on the CMAC algorithm is studied for this task.

Diamond turning of brittle materials such as glass, ceramic, germanium, and zinc sulfide has been of considerable research interest in recent years due to applications in optics and precision engineering systems. When diamond turning brittle materials, material removal should be kept within the ductile regime to avoid subsurface damage (Evans, 1991; Nakasuji et al., 1990). It is generally accepted that ductile regime machining of brittle materials can be accomplished using extremely low depth of cut and feed rates. Furthermore, the tool positioning accuracy of the machine must be in the nanometer range to obtain optical quality machined parts with surface finish and profile accuracy on the order of 10 nm and 100 nm respectively (Nakasuji et al, 1990, Ueda et al., 1991). Nanometric level positioning accuracy of the machine tool axes is difficult particularly at low feed rates due to friction and backlash. Friction at extremely low feed rates is highly nonlinear due to the transition from stiction to Coulomb friction, and as such is very difficult to model. Standard proportional-integral-derivative (PID) type controllers are unable to deal with this large and erratic friction within the requirements of ultra precision machining. In order to compensate the effects of friction in the machine tool axes, a learning controller based on the Cerebellar Model Articulation Controller (CMAC) neural network is studied for servo-control. The learning controller was implemented using “C” language on a DSP based controller for a single point diamond turning machine. The CMAC servo control algorithm improved the positioning accuracy of the single point diamond turning machine by a factor of 10 compared to the standard PID algorithm run on the same machine and control system hardware.

The design and fabrication of a cerebellar model articulation controller (CMAC) controlled piezo-electric actuated fast tool servo has been reported [Pinsopon et al]. A capacitance type gap sensor with less than 1 nm resolution was used for position feedback. Cerebellar model articulation controller (CMAC) neural network control algorithm has been implemented parallel to the PID controller as the control scheme, and its performance was compared with the standard PID control. The CMAC control algorithm improves the tracking accuracy compared to the PID algorithm. However, the CMAC control system had a poor disturbance rejection capability as compared to the PID control system. In this research, state feedback with disturbance observer is implemented as the control scheme replacing PID controller to improve the disturbance rejection capability. CMAC control algorithm is also used in the feedforward loop to improve the tracking capability of the closed loop system.