The Optimization of Surface Morphology Prediction Model for Aspheric Microlens Array and Machining Process

Aspheric microlens arrays (MLAs) are optical elements with exceptionally high performance and are widely used in various fields Slow tool servo machining (STSM) is preferred due to its characteristics, like high precision in one pass and machining, for machining aspheric MLAs. As the aspheric MLAs bite-sized optical elements, they have strict requirements for processing surface quality. However, their manufacturing process is elaborate, time-consuming and costly, which faces challenges to improve the surface quality of aspheric MLAs. Establishing a machining shape prediction model is an effective way to solve the above problem. Nevertheless, many studies are now focusing on predictive models of ideal conditions, and there is a gulf between the actual situation and the ideal result. In order to minimize the gap and reduce the error, we propose a method to consider the external disturbance error into account in the prediction model, which improves the accuracy of the prediction model for PV (Peak to valley) prediction by 22 \(\%\) $\%$ and the accuracy of surface roughness prediction by 52 \(\%\) $\%$ . To improve the surface quality of aspheric MLAs, we optimized a combination of surface extension and virtual substrate. The research content of our article points out a new path to improve the manufacturing quality of aspheric MLAs by STSM.