TITLE:
A Material Removal Prediction Framework for Ball EEM Polishing in Precision Lens Manufacturing
AUTHORS:
Tyler Young, Jacob Guymon, Mark Pankow, Gracious Ngaile
KEYWORDS:
Optics, Machine Learning, Metrology, Modeling, Polishing, Toolpathing
JOURNAL NAME:
World Journal of Engineering and Technology,
Vol.14 No.2,
April
7,
2026
ABSTRACT: Lens surface parameters are critical to optical system performance and require increasingly stringent precision due to rising performance demands and continued technological miniaturization. Although machining processes such as diamond turning can produce lenses with high form accuracy, they are not free from surface defects. These defects are typically addressed through post-processing techniques such as polishing; however, depending on defect size and the selected removal process, polishing can become a complex and iterative task. This complexity can be mitigated through the application of machine learning algorithms to predict material removal behavior. This paper presents the development of a material removal prediction framework for ball elastic emission machining (EEM) polishing of lens surfaces, incorporating machine learning tools to improve process predictability and efficiency. The resulting model accepts surface characteristics and process parameters as inputs and predicts the final surface parameters following polishing. The model’s root mean square error (RMSE) is approximately 0.1 µm. Surface parameters achieved using the removal strategy on which this model is based include a peak-to-valley (PV) value of 0.2841 µm and a root mean square (RMS) roughness of 0.032 µm.