基于扭振信号的行星齿轮传动故障诊断技术研究

Fault diagnosis of planetary gear transmission, which has been widely used in machines such as helicopter and aerogenerator, has great significance due to planetary gear tranmission's key role and high fault rate. Most of the current researches are based on the reciprocating vibration signals of the gear box. The gear fault signal feature in the reciprocating vibration signal is weak due to the complicated signal components and difficulties in separation of meshing vibration signals, however, the torsional vibration signals of gear shaft will avoid such problem. The research project aims to diagnose fault of planetary gears based on the reciprocating vibration signal. (1) Analysis of the torsional vibration signal frequency spectrum, and the influence of the tooth number, the planet gear number and its configuration mode. Research on the mechanism of signal amplitude modulation and frequency modulation both of fault types and fault levels of the gears including sun gear, planet gear and ring gear.(2)Research on the signal processing method(including Wavelet, spectral kurtosis,　Characteristic spectrum, cepstrum, time-domain index) on the time synchronous averaging torsional vibration signal and the original sampling torsional vibration signal. (3) Research on the torsional signal sampling method and its influence over the signal frequency spectrum. The development of the system of the torsional signal sampling and analysis of planetary gear box using virtual instrument technology. The fault diagnosis test using the proposed method. The project will simplify the fault diagnosis process of planetary gear box, and diagnose the early gear fault more effectively.

行星齿轮传动在航空器、风力发电机等领域应用广泛，作用关键，故障率高，故障诊断意义重大。现有的故障诊断方法基于齿轮箱往复振动信号。信号成分复杂，啮合振动信号分离困难，故障特征微弱。齿轮轴扭转振动信号能避免这些问题。将齿轮轴扭转振动信号应用于故障诊断。(1)研究扭转振动信号频谱特征，研究齿轮齿数、行星轮个数和配置方式对扭振信号频谱结构的影响，研究行星轮、太阳轮和齿圈的各类故障类型和故障程度对扭转振动信号的调频调幅机理及信号频谱特征。 (2)基于时域同步平均的扭振信号和原始采样扭振信号，对比采用现代信号处理方法（小波方法、谱峭度、细化谱/倒频谱、时域指标），提取故障信号特征。(3)研究扭振信号采集方式及对扭振信号频谱结构的影响，基于虚拟仪器技术，实现行星齿轮箱扭振信号采集和分析系统，齿轮故障啮合振动试验。本研究将简化行星齿轮传动故障诊断问题，更有效地分辨和诊断齿轮各类早期故障。

行星齿轮传动在航空器、风力发电机等领域应用广泛。利用齿轮轴扭转振动信号分析和诊断行星齿轮故障。1）建立了行星齿轮扭振动力学模型，计算了行星齿轮箱振动模态、不同程度故障对齿轮模态的影响，并对不同程度故障状态下啮合的齿轮做瞬态动力学分析。研究了扭转振动信号频谱特征，分析了扭振传感器安装位置（输入轴或输出轴）对扭振信号频谱结构的影响。2）研制了基于扭振信号的行星齿轮传动故障诊断试验台，提出一种基于增量式编码器和高频时钟脉冲计数的扭振测量方法。基于虚拟仪器技术，实现了行星齿轮箱扭转振动信号采集和分析系统。能够对扭转信号进行实时采集、存储、分析和显示。同时还能对齿轮箱横向振动信号进行分析处理。3）研究了基于扭振信号的行星齿轮故障诊断，提出了若干故障诊断方法，包括：最小熵反褶积、无时标时域同步平均、基于小波包分解的归一化能量谱分析、基于H-P滤波与变分模态分解的时域分析方法、基于小波包分析与距离评估技术的分析方法、基于最佳小波包分解和隐马尔可夫模型的故障识别方法、基于最小熵解卷积的谱峭度方法、时间序列分析法，等。比较了横向振动信号和扭振信号、输入轴扭振信号和输出轴扭振信号对故障的敏感程度。本研究将简化行星齿轮传动故障诊断问题，更有效地分辨和诊断齿轮各类故障。