AN INTELLIGENT SENSOR SYSTEM APPROACH FOR RELIABLE TOOL FLANK WEAR RECOGNITION

An intelligent sensor system approach for reliable flank wear monitori ng in turning is described. Based on acoustic emission and force sensi ng, art intelligent sensor system integrates multiple sensing, advance d feature extraction and information fusion methodology. Spectral, sta tistical and dynamic analysis have been used to determine primary feat ures from the sensor signals. A secondary feature refinement is furthe r applied to the primary features in order to obtain a more correlated feature vector for the tool flank wear process. An unsupervised ART2 neural network is used for the fusion of AE and force information and decision-making of the tool flank wear state. The experimental results confirm that the developed intelligent sensor system can be reliably used to recognise the tool flank wear state over a range of cutting co nditions.