Machine Learning Based Modeling and Optimization of Drilling Parameters of Developed Bio-Based Composites

Abstract

Numerical and Artificial Neural Network (ANN) approaches are now frequently employed for modeling and optimizing the performance of industrial systems. Optimal machining parameters are of major relevance in production contexts, as machining operations efficiency is critical to market competitiveness. The optimal machining parameters (i.e., spindle speed, drill diameter, and feed rate) for drilling operations will be researched in this project in order to minimize the delamination factor. Regression modeling and Response Surface Methodology (RSM) was previously used to explore the effects of specified parameters on process variables (i.e., delamination factors). The data acquired during the machining operation is utilized to create machine learning (ML)-based surrogate models that test, assess, and optimize different input machining parameters. To predict different output reactions of bio-composites drilling, several ML approaches such as polynomial regression (PR), random forest (RF) regression, gradient boosted (GB) trees, and adaptive boosting (AB) based regression are utilized. The ML results will then be compared to the experimental and RSM results. Keywords: Artificial Neural Networks; Bio-Composites; Drilling; Machine Learning; Optimization;