Identification and Monitoring of Wear Stages in GCr15 with Acoustic Emission Signals via Variational Mode Decomposition
摘要
To address the challenge of real-time monitoring of sliding wear in the load-equalizing mechanism of nuclear main pump thrust bearings, friction and wear tests were conducted on GCr15 steel, with periodic collection of acoustic emission (AE) signals. Variational mode decomposition (VMD) was applied to analyze the signals. Furthermore, the classification of wear stages and underlying mechanisms were validated through SEM/EDS analysis, laser confocal microscopy, and coefficient of friction testing. Results indicate that compared to the wear curve, the AE signal showed more distinct stage-specific features, making it a superior means of monitoring the occurrence of severe wear, while avoiding the need for destructive testing to plot the wear curve. During the running-in wear stage, the first-order intrinsic mode function (IMF) exhibits noise characteristics; the second- and third-order IMFs remain stable (near 125 and 143 kHz); while the fourth and fifth-order IMFs fluctuate (160–166 kHz and 206–249 kHz). During the steady-state wear stage, the second and third-order IMFs exhibit noise; the first, fourth, and fifth-order IMFs remain stable (5, 124, and 146 kHz). During the severe wear stage, the first and second-order IMFs exhibit noise, while the third and fourth-order IMFs fluctuate within a narrow range (~123 and ~ 141 kHz). The fifth-order IMF exhibits wide-range fluctuations (198–267 kHz). The correlation coefficients of the fourth and fifth-order IMFs with noise consistently remain below the threshold, making them primary bases for determining sliding wear failure in GCr15.