基于同源和EFD方法的太阳轮故障特征提取研究

摘要:

行星齿轮箱具有复杂传动结构，在太阳轮出现故障时，其故障信息往往被无关或干扰成分所掩盖，导致故障特征难以辨识；为提取故障状态下振动信号的共性特征，采用同源响应的振源分离策略，依据旋转机械内部激励信号具有的周期性及低秩特性，挖掘与故障高度相关的同源响应片段，获取含有丰富故障信息的片段，并提取最能代表故障特征的主要成分，以突出故障特征信息并减少无关信息干扰；在传统经验傅里叶分解基础上，设置频带分割阈值，避免频谱局部分割；通过故障特征比指标自适应筛选最佳分解分量，用包络谱图验证故障特征提取效果；最终通过太阳轮裂纹故障振动仿真信号及实际齿轮箱运行数据进行验证，实现了太阳轮裂纹故障特征的清晰提取，验证了方法的有效性。

Abstract:

Planetary gearboxes possess intricate transmission mechanisms. Upon the failure of the sun gear, its malfunction messages are frequently obscured by unrelated or interfering components, complicating the process of pinpointing specific fault characteristics; To effectively extract common features from vibration signals under fault conditions, the strategy of employing a co-source response for vibration source separation is adopted.; Based on the periodic and low-rank nature of internal excitation signals in rotating machinery, this approach identifies homologous response segments highly related to sun gear faults, thereby obtaining segments rich in fault information; The main components that best represent fault characteristics are then extracted to highlight fault information and reduce interference from irrelevant data; Building on traditional empirical Fourier decomposition, a frequency band segmentation threshold is set to prevent local spectral segmentation. The optimal decomposition components are adaptively selected through a fault feature ratio index, and the envelope spectrum is used to verify the effectiveness of fault feature extraction; Finally, the method's effectiveness is validated using both simulated vibration signals of sun gear crack faults and actual gearbox operational data, achieving clear extraction of sun gear crack fault features;