An error compensation scheme for polar coordinate glass edging machines using the contour registration method.

This paper proposes a comprehensive error compensation method for the glass edging machine using polar coordinates. An error model is established to evaluate the error source of the contour deviation from the ideal position after glass installation. According to the simulation results, it is determined that the misalignment of the linear axis and the rotary table and the inaccuracy of the fixture are the main reasons for the installation error of the work-piece. Then, a two-step contour registration method is introduced to accomplish the task of error identification. The glass contour is first measured by a laser displacement sensor after being installed on the rotary table. The coarse registration is presented to align the centroid positions of the measured and the ideal contours, and the fine registration is developed to match them precisely. After that, the installation error can be extracted from the transformation parameters by the least square method. Furthermore, a compensation model is proposed to map the original contour to the actual installation position using the transformation matrix constructed by the identified errors. Finally, a verification experiment is implemented that compares the grind edge without and with compensation to validate the efficiency and accuracy of the proposed method. After error compensation, the mean value and the variance value of the polar diameter deviation of the glass work-piece decreased by 87.5% and 83.0%, respectively. This paper provides a simple way to identify and compensate the installation errors online with high accuracy and efficiency. The proposed method can improve the machining accuracy of the polar coordinate glass edging machines and reduce glass waste. • An error compensation method for glass edging machines was presented. • A two-step contour registration method is developed to accomplish error identification. • The installation errors was identified and compensated online with high accuracy. • The proposed algorithm can save at least 6% of the glass material. [ABSTRACT FROM AUTHOR]