Development of Experiment-Based Mathematical Models of Acoustic Signals for Machine Condition Monitoring

In this work, acoustic signals were analyzed in order to develop mathematical models to evaluate machining operations. The preliminary experiments consisted of milling operations with a single carbide cutting edge. The process parameters included feed rate, spindle speed, depth of cut, cutting direction, and work material. The goal of these models is to determine the feasibility of using acoustic emission in real-time monitoring of machining operations and machine health. Multiple linear regression analyses were used in developing the mathematical models. It was found that the models accurately reflect the machining operation and thus can be used in machine monitoring.