针对风电机组传动系统时变转速及强噪声干扰等运行特点，首先采用改进小波阈值降噪方法，对风力机齿轮箱轴承振动信号进行降噪预处理，然后基于分形理论，计算变转速轴承振动信号盒维数及多重分形谱，定量描述轴承不同状态下振动信号的特征信息。结果表明：基于改进阈值与硬阈值函数相结合的小波阈值降噪方法对振动信号进行降噪预处理，其降噪效果优于传统的4种阈值选取原则；多重分形去趋势波动分析方法对于定速及变速轴承均能进行有效的故障识别；谱函数最大值所对应的奇异指数α(f max)当轴承处于内环故障时最小，可有效判断轴承运行状态及故障位置，能对轴承不同的故障状态做出准确的判断。

Aiming at the operating characteristics of the wind turbine transmission system time-varying speed and strong noise interference and other operational characteristics,the improved wavelet threshold noise reduction method is adopted to denoise the vibration signal of wind turbine gearbox bearing,and then the dimension and multifractal spectrum are determined based on fractal theory,to quantitatively describe the characteristic information of the vibration signal under different states of the bearing.The results show that the noise reduction effect of the proposed method is better than that of the four traditional thresholds.The MFDFA have a good performance on the fault identification of fixed speed and variable speed bearings.The singular index,which corresponds to the maximum value of the spectral function α(fmax),has very good identification; it can make accurate judgments under different fault conditions of bearings

TK83

国家自然科学基金（51676131，51176129，51875361）