为识别轴承的工作状态及故障类型，针对非线性振动信号，基于分形盒维数与小波降噪方法，计算了轴承正常状态及不同故障状态的振动信号盒维数。结果表明：轴承的故障类型不同，其振动信号盒维数亦不同，正常状态盒维数最大，内环故障盒维数最小，其值分别为1.6和1.4。因此，根据盒维数能定量识别轴承故障状态与故障位置，本研究可为轴承状态监测和故障诊断提供理论依据。

With the rapid development of industry, the mechanical equipment becomes more and more complicated and precise. For complex mechanical systems, there are greatly deviates for fault diagnosis in theory and practice, based on traditional analysis method.Therefore, in order to identify the bearing working condition and different fault state, the boxdimension was quantitatively calculate by the vibration signal of the bearing in the normal state and different fault location in allusion to nonlinear vibration signals based on the fractal theory and the wavelet denoising method. The results showed that the boxdimension had obvious distinguish-ability when the bearing was in different fault state. The biggest boxdimension was in the normal state, followed by the outer ring, and the smallest was the inner ring. The fractal boxdimension could quantitatively identify the bearing in fault status and location to provide a reliable foundation for monitoring bearing condition and diagnosing fault state.

TH133

国家自然科学基金资助项目（51676131，51176129）