Monitoring the cutting condition of structurally distinct aluminum oxide grinding wheels using acoustic emission signals and the Hinkley criterion

Abstract

Acoustic emission sensors (AE) have been extensively utilized as an indirect method for condition monitoring of grinding wheels, the essential tools in the grinding process. Statistical parameters like root mean square (RMS) and counts have been employed to process these signals, aiming to characterize the cutting state of the wheel and determine the optimal moment for interrupting the dressing operation. However, the Hinkley criterion statistic, despite being employed in scientific studies such as structural health monitoring, has not yet been explored for monitoring the dressing operation of aluminum oxide wheels. In light of this, the present study aims to assess the efficacy of the Hinkley criterion statistic in extracting features from AE signals collected during the dressing operation of structurally distinct aluminum oxide wheels. Dressing tests were conducted using two wheels, each subjected to different dressing conditions. The AE sensor-generated signals were subsequently collected and digitally processed, followed by the computation of the Hinkley criterion statistic. In the time domain, the Hinkley criterion statistic enables the precise extraction of detailed information regarding the behavior of AE signals during grinding wheel dressing procedures, eliminating the need for intricate frequency domain analysis. The outcomes unequivocally demonstrate the effectiveness of the Hinkley criterion statistic in classifying the wheel as either dressed or undressed, thereby facilitating the determination of the optimal moment to halt the dressing operation. Importantly, the method proves its efficiency in categorizing the dressing condition of structurally diverse wheels, even when subjected to varying dressing parameters. Consequently, its implementation promises enhanced efficiency and cost-effectiveness in dressing operations. Furthermore, it is noteworthy that this method exhibits potential for generalization, making it suitable for monitoring the dressing process across a wide array of wheel types. Ultimately, this methodology plays a pivotal role in optimizing the grinding process.