Wear state monitoring of ceramic-bonded diamond grinding wheels during ultra-precision grinding of sapphire A-plane

Due to the high hardness of the sapphire A-plane and the small grit size of ultra-precision grinding wheels, significant wheel wear occurs during grinding of its optical components. Acoustic emission (AE) and grinding force signals were monitored throughout the entire wheel life. Scanning electron microscopy was employed to analyze the surface topography of the ground surface and the morphology of the abrasive grains. By comparing the grain morphology, surface topography, and grinding force signals, three wear stages were identified for the grinding wheel: mild, medium, and severe. Furthermore, the study utilized Empirical Mode Decomposition (EMD) to decompose the AE signals into a composite of multiple Intrinsic Mode Functions (IMFs). IMFs with high correlation coefficients to the original AE signals were selected for the extraction of key features, which were then optimized. A Random Forest algorithm was utilized to develop a model that correlates grinding wheel states with extracted signal features. The determination of optimal model parameters was performed to ensure accurate recognition results. The grinding test results indicate that the identification accuracy of the grinding wheel wear states using AE frequency-domain feature values reaches 97.2%. This method serves as an effective means of monitoring the wear states of grinding wheels during sapphire ultra-precision grinding processes.