摘要
机械设备运行中得到的诊断信息往往存在信噪比低、信号混叠等问题,严重影响提取真实的故障信号特征,降低了诊断准确率。针对上述问题,提出一种新的基于快速独立分量分析与概率神经网络的设备故障诊断方法,FASTICA对振动信号降噪处理后提取特征,PNN实现故障识别。通过算法仿真以及LMS齿轮箱实验证明,该融合算法处理后的动态故障诊断能力和诊断精度都明显提高。
In the practical engineering,the operation equipment diagnosis information always has low SNR,aliasing source signal and so on,it increased the difficulty of the real feature extraction and reduced the fault diagnosis accuracy.This paper presents a technology which based on FASTICA and PNN of gearbox fault diagnosis,FASTICA used to vibration signal noise reduction and then Feature extraction,PNN used to fault identification.Through the signal simulation experiment and the LMS experimental data algorithm analysis,this fusion algorithm makes the dynamic ability of fault diagnosis and diagnosis accuracy are obviously improved.
出处
《煤矿机械》
北大核心
2013年第6期278-280,共3页
Coal Mine Machinery
