摘要
This study aims to monitor the occurrence of diffractions generated by periodic tool marks during single-point diamond turning (SPDT) of aspheric surfaces. The diffraction of lens surface caused by SPDT results in dizziness and discomfort to users. By monitoring the SPDT process, the diffraction of the lens can be effectively prevented. The acoustic emission (AE) signals were utilized to monitor the processing defects. Hence, this study proposes the method to apply AE signals analyzed by design of experiment (DOE), wavelet analysis, and support vector machine (SVM) in identifying the occurrence of diffraction in SPDT process. The original AE signals were decomposed by wavelet analysis, and the coefficients of energy were determined. The energy coefficient of the first detail layer (D1) and the process parameter, feed rate, were fed into the SVM to determine whether diffraction was formed. It is confirmed that the high-frequency energy of the AE signal is correlated with the diffraction occurrence. The prediction model referring to feed rate can predict the occurrence of diffraction in SPDT with 95.8% accuracy.