摘要
针对得到的齿轮箱运行中振动信号信噪比低而引起的故障识别精度低的问题,提出了一种新的基于卡尔曼滤波与BP神经网络的设备故障诊断方法,卡尔曼滤波的作用是对振动信号降噪处理然后提取特征,而BP网络则可实现故障的识别。通过算法仿真以及齿轮箱实验可证明,结合此种算法便可解决信噪比低的问题,与此同时也提高了故障识别精度。
In the practical engineering ,the gearbox fault diagnosis information always has low SNR （Signal-to-noise ratio）,it increases the difficulty of the real feature extraction and reduces the fault diagnosis accuracy .This paper presents a technology of gearbox fault diagnosis ,which is based on Kalman and BP .Through analyzing the simulated data and the experimental data ,this fusion algorithm makes the diagnosis accuracy obviously improved .
出处
《机械工程与自动化》
2014年第1期134-135,共2页
Mechanical Engineering & Automation
