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。 Diagnostic information obtained during the operation of mechanical equipment often has such problems as low signal-to-noise ratio and signal aliasing, which seriously affects the feature extraction of real fault signals and reduces the diagnostic accuracy. Aiming at the above problems, a new fault diagnosis method of equipment based on fast independent component analysis and probabilistic neural network is proposed. FASTICA extracts the features of the vibration signals after noise reduction and the PNN realizes the fault identification. Through the algorithm simulation and LMS gearbox experiment, it is proved that the dynamic fault diagnosis ability and diagnostic accuracy of the fusion algorithm are obviously improved.