Machine Learning for Sheet Metal Bending Process
LightGBM Physics-Informed ML Feature Engineering Manufacturing Optimization

Overview
Traditional sheet metal bending relies on iterative trial-and-error, increasing production time and waste. This project introduced a physics-informed LGBM model to predict bending sequences with high precision.
Development Process



Final Result
By incorporating material properties and geometric parameters, prediction time was reduced from over an hour to under five minutes, improving accuracy from 84.7% to 93.0%.

Collaborators
This project was developed in collaboration with the manufacturing optimization team.