Neural network based surface shape modeling of stressed lap optical polishing

Area of Science:

• Optics and Optical Engineering
• Computational Modeling
• Materials Science

Background:

• Accurate modeling of actuator forces and lap surface changes is critical for polishing large, highly aspheric optical surfaces.
• Existing models may not fully capture the dynamic deformation of stressed laps used in advanced optical manufacturing.

Purpose of the Study:

• To develop a neural network-based model for representing the dynamic deformation of a stressed lap surface.
• To facilitate a computer-controlled optical polishing process for large and highly aspheric optical surfaces.

Main Methods:

• A neural network model was developed using original data from a microdisplacement sensor matrix.
• The model was trained to reflect the dynamic deformation of the stressed lap.

Main Results:

• The proposed neural network model accurately represents the stressed lap's surface shape.
• The model provides an analytical tool for controlling the stressed lap and active support systems.

Conclusions:

• The developed neural network model offers an accurate and dynamic representation of stressed lap surfaces.
• This model is valuable for computer-controlled polishing systems, particularly for large, aspheric optics, enhancing precision and control.