中文摘要：

为实现对风机轴承早期故障诊断,采用SPM42000传感器对轴承声发射信号进行实时采集。以SKF240球面滚子轴承为例,轴承内圈、滚动体、外圈和保持架故障频率分别为452.687 0、191.57、387.313、13.832 6 Hz。首先,采用小波阈值降噪方法去除轴承故障信号中掺杂的由各部件摩擦或机械振动引发的噪声信号,选择sym9小波函数,分解层数为5,阈值规则为启发式阈值的软阈值函数;之后,选择所需频带的小波函数进行重构;若初次分解不能达到要求,需要采用小波包再次分解;最后,做包络谱分析,结合声发射信号的峰值特征参数,对轴承故障诊断。仿真实验结果表明,该方法可以识别出轴承各部件故障。

英文摘要：

In order to diagnose the early fault of the wind turbine bearing,SPM42000 sensor was used to acquire the bearing real-time acoustic emission signal. Taking SKF240 spherical roller bearing as an example,the failure frequency of bearing inner ring,rolling body,outer ring and holder are 452.687 0,191.57,387.313,13.832 6 Hz,respectively. First of all,the wavelet threshold de-noising based on soft threshold function was used which wavelet function is sym9,wavelet decomposition layer is 5,and threshold rule is soft threshold function of heuristic threshold. Then,the wavelet function was selected in the required frequency band to reconstruct signal. If the initial decomposition can＇t meet the requirements,it needs to use wavelet packet to decompose again. Finally,combining with the peak value characteristic parameters of acoustic emission signal,the envelope spectrum was analysized to diagnose bearing faults. The result of simulation experiments indicate the method can identify the fault of bearing components.