Contouring Control of Biaxial Systems by Considering the Effects of Interpolation

• Main Authors: Hsien-Sheng Hsu, 許憲生
• Other Authors: Shyh-Leh Chen
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/86033600832257533120

碩士 === 國立中正大學 === 機械系 === 91 === This thesis is concerned with the contouring control problem of biaxial motion systems. It is well known that contour errors depend not only on the controller, but also on the reference command. This study is concentrated on the design of reference command. In general words, give a desired path, there are infinitely many possible reference commands that can form the desired path. Two problems are investigated. First, it is to find the reference command that will yield the minimum contour error. This is the problem of interpolation and acceleration/deceleration. Second, given a reference command, it is to find a modified reference command so that the system output will be as close to the reference command as possible. For a simple first order system, three popular acceleration/deceleration schemes (linear type, exponential type and bell-shaped type) are studied in the present work. They are compared under the condition of same average and maximum feedrate. It is found that the bell-shaped type is the best in terms of contour errors. The design of the modified reference command utilizes the inverse model of the closed-loop system. In other words, the modified reference command can be considered as the output of the inverse system with the desired reference command as the input. It can be obtained by the convolution technique. Both minimum phase and nom-minimum phase systems are discussed. It is found that the prep method works perfectly for minimum phase system. However, the method is sensitive to noise and numerical error for the non-minimum phase system. Also, the acceleration/ deceleration schemes are less significant if the modified reference command is employed. The analysis is verified by numerical simulation and experiments.