航空启动发电机齿轮后轴承失效性故障分析

Failure Analysis of Rear Bearing Failure of Aviation Start-up Generator Gears

为了减少轴承故障引发的航空启动发电机运行故障,需要对轴承故障进行及时分析与排除,针对现有分析方法中存在的受噪声信号影响较大、准确率较低、所需时间长等问题,提出一种基于阈值去噪的航空启动发电机齿轮后轴承失效性故障分析方法。采用正交小波变换法对轴承故障振动信号进行处理,依据概率论的正态分布原则设置对应的阈值筛选小波变换系数,通过系数逆变换对信号进行重构,去除噪声信号。利用定子电流与故障振动信号间的关联性,将去噪后的信号进行分段并对各分段进行频谱分析,获取信号频率的信息传递函数,进一步获得故障振动信号的样本特征。根据故障信号特征构建故障分析模型,对故障特征进行对比分类,完成故障分析。仿真结果表明,上述方法能够降低噪声信号对故障分析结果的影响,增加分析结果的准确性且所用时间更短,具有一定的工程实用价值。

It is necessary to analyze and eliminate the bearing failure in time.At present,the analysis method is affected by the noise signal.Thus,the accuracy is low.This article presents a method to analyze the rear bearing failure .for gear wheel of aero-starter-generator based on threshold denoising.Firstly,the orthogonal wavelet transform method was used to process the vibration signal .of bearing fault.According to the normal distribution principle of probabilistic theory,the corresponding threshold was established to filter wavelet transform coeffcients.Then,inverse coefficient transformation was used to reconstruct signal and remove noise signal.Moreover,the correlation between stator current and fault vibration signal was used to segment the denoised signal and analyze .frequency spectrum of each segment.In addition,the information transfer function of signal frequency was obtained to further find the sample feature of fault vibration signed.Finally,the failure analysis model was built based on the fault signal features,and then.fault features were compared and classified to complete failure analysis.Simulation results show that the proposed method reduces the influence of noise signal on fault analysis result and increases the accuracy of analysis result.Meanwhile,this method costs lees time,which has certain engineering practical value.