Large-scale dwell time algorithm for computer controlled optical surfacing

This paper uses a non-negative least-squares method based on sparse matrix, to solve the dwell time in computer controlled optical surfacing (CCOS) of optical elements with large diameter and microstructure, and the regularization of the method is researched. The simulation results reveal that the large-scale non-negative least-squares method based on sparse matrix has higher accuracy and efficiency, in contrast with the traditional non-negative least-squares method. The RMS of surface residual error converges to 0.067 wavelength, when the method is used to simulate the manufacturing of an optical workpiece with large diameter, microstructures and average surface amplitude of 1.177 6 wavelengths.