On the Feasibility of Catastrophic Cutting Tool Fracture Prediction Via Acoustic Emission Analysis

The prediction of catastrophic cutting tool fracture is explored through monitoring the acoustic emission (AE) from a cutting process. A prediction parameter is derived which combines the AE signal with a physical model of a cracked tool to form an estimate of the spatial energy release rate. Monitoring the energy release rate is found to be largely dependent on the detection of crack advancement. Experiments were performed with both new and artificially cracked inserts during interrupted cutting. Epoches denoting crack advancement were detected through high time homomorphic analysis of the acquired AE signals. AE bursts prior to and leading up to fracture were analyzed for crack advancement. The calculated energy release rate was found to exponentially increase as fracture was approached. Crack advancement could be feasibly detected approximately six cuts prior to fracture.