The Online Monitoring for Milling Stability Boundary Considering Tool Wear

Abstract

Stability of the machining process is a prerequisite to ensure the dimensional accuracy and high-quality surface of parts in the milling process. Considering the milling process of titanium alloy with severe tool wear, the stability domain boundary changes significantly with changing machining dynamics. Therefore, the online stability prediction is especially important to improve machining efficiency and surface quality. Firstly, according to simulation and experimental results of the stability domain, the machine learning binary classification method Support Vector Machine is introduced to calculate the initial domain boundary. During the tool wear process, the stability identification method based on energy entropy is applied to determine whether the current machining state has changed, and an incremental learning model based on sequential minimal optimization algorithm to update the stability boundary. An online monitoring model for the position and shape of stability domain boundary is established. To verify the proposed method, the milling experiments with tool wear as the single variable is carried out to demonstrate the accuracy of method in practical application.