Feature extraction using frequency spectrum and time domain analysis of vibration signals to monitoring advanced ceramic in grinding process

Abstract

New alternatives for monitoring the ceramic grinding process have been studied. Monitoring vibration signals is one of the most successful methods because some characteristics that describe the behaviour and influence of the process on ground parts are only noticeable by studying such signals. This study aims to monitor the finishing of advanced ceramics during the surface grinding process via digital processing of the vibration signals. Experimental tests were performed using a surface tangential grinding machine with a diamond grinding wheel and alumina (Al2O3) test specimens. The vibration signal was measured by an accelerometer and recorded by an oscilloscope at a 2 MHz sampling rate. The tests were conducted at different depths of cut for two workpiece speeds (v(w)) under mild and severe machining conditions. Confocal microscopy and surface roughness R-a measurements were performed after grinding each workpiece to classify the samples. Digital signal processing was performed to achieve feature extraction. A frequency range of 800 Hz-2 kHz was most strongly related to the post-grinding ceramic condition. A correlation was found between vibration and integrity of the ceramic workpiece because the vibration signal was proportional to the surface roughness for each cutting depth used. To support the conclusion presented, a statistical analysis through variance by analysis of variance was performed.