中文摘要：

针对风电机组齿轮箱局部微弱故障难于诊断的问题，提出全矢频带能量谱故障诊断方法。采用全矢理论对同源信号进行信息融合，获得相位不变、信息更完善的全矢信号，利用FIR滤波器对全矢信号进行分解，计算各频带能量作为识别工作状态的特征向量。分析风电机组齿轮箱的正常、齿根裂纹及均匀磨损信号的各频带能量，发现转频和啮合频率处的频带能量变化率能准确判别各类故障。通过对不同工况下50组信号的识别，证明该方法可有效区分风电机组齿轮箱的早期局部微弱故障。

英文摘要：

Because the local weak fault of wind turbine gearbox is difficult to diagnose, the fault diagnosis method based on full vector frequency band energy spectrum was proposed. In this method, the full vector theory was used to make the information fusion of homologous signals and obtain the full vector signal with constant phase and better information. Then the FIR filter was used to decompose the full vector signal, calculate the energy of each frequency band as the eigenvector to identify the working state. The energy of each frequency band of signals for the conditions of normal, tooth root crack and uniform wear was analyzed to find the energy change rate at frequency transfer and engage frequency for identifying all kinds of faults accurately. By identifying 50 groups of signals under different conditions, it shows that the method can distinguish early local weak failure of wind turbine gearbox effectively.