Forecasting Rectangular Pocket Deviations in Birch Plywood for Milling Process Optimization

Abstract

We explored the field of predicting and minimizing dimensional deviations when milling rectangular pockets in birch plywood, with the aim of improving the accuracy of manufacturing processes. We examined various machining conditions altering feed rate, depth of cut, and spindle speed and acquired high-resolution 3D scans to quantify pocket-shape deviations against the prescribed tolerances. To enhance the model’s robustness, the empirical dataset was augmented via synthetic data-generation techniques. We then compared several approaches using cross-validation. MLP proved the most accurate. These findings demonstrate the utility of readily available machine-learning algorithms for precise deviation prediction and lay the groundwork for future integration of automated hyperparameter tuning and feature-based process optimization.