Abstract:The roller bearing of the cutting part of shearer has a complex structure and is easy to be damaged, and it is difficult to locate and judge the fault based on human experience alone. An empirical mode decomposition method of adaptive time-varying filtering based on worker ant colony-simulated annealing algorithm is proposed to remove the low frequency interference in the original vibration signal. The band-pass filter parameters are set with the center frequency and bandwidth corresponding to the maximum kurtosis, so as to further highlight the high frequency characteristics of the signal. The envelope spectrum diagram is drawn after the square envelope of the processed signal, and the calculated theoretical fault frequency is compared with the frequency corresponding to the maximum value of the envelope spectrum, so as to distinguish the fault type. Shearer tests show that the proposed method can be applied to practical shearer engineering applications, effectively eliminate the low-frequency interference part of the signal, improve the kurtosis characteristic of vibration signal, and achieve the fault discrimination rate of 100% and 88.9% under different speed and different fault size conditions.