Machine learning-based optimization for chatter-free milling with maximum productivity

Official URL: https://www.utis.tc/gorseller/files/UTIS%202024_Proceedings.pdf

Optimization plays a critical role in modern manufacturing by balancing productivity, costs, and quality. In machining, this involves chatter-free operations, high material removal rates (MRR), extended tool life, and optimal form errors, while adhering to power/torque and tool breakage limits. In this regard, machine learning (ML)-based methods, especially Bayesian Optimization (BO) with Gaussian Process Regression (GPR), offer innovative solutions. This study optimized milling parameters to maximize productivity and suppress chatter, considering form error in the finishing process. According to the obtained results, the proposed approach improved the machining performance by approximately 31%. A sensitivity analysis using the Shapley value algorithm highlighted the input parameters importance level. A specialized software with a user-friendly GUI was developed, facilitating dataset loading, parameter/constraint specification, and algorithm execution, streamlining the optimization process for machining operations.