论文部分内容阅读
由于传统基于傅立叶变换的利用频域对电机故障的信号分析中无法对奇异信号点的时域信息进行检测。针对上述问题,提出基于小波包神经网络的电机故障诊断的方法。结合电机振动的非平稳随机性的特点。利用小波包多分辨率分析方法对电机的采样信号进行分解,提取电机故障状态特征并作为BP神经网络输入样本的特征向量,利用神经网络的自学习和模式识别的特点最终输出电机故障类型。通过MATLAB仿真结果可以证实该方法可行性。
Due to the traditional Fourier transform based on the use of frequency domain motor fault signal analysis can not singular signal point of the time domain information is detected. In view of the above problems, a method of fault diagnosis of motor based on wavelet packet neural network is proposed. Combined with the non-stationary random vibration characteristics of the motor. The wavelet packet multiresolution analysis method is used to decompose the sampled signal of the motor to extract the fault characteristic of the motor and input the eigenvector of the sample as the BP neural network. Finally, the fault type of the motor is output based on the self-learning and pattern recognition of the neural network. The feasibility of this method can be confirmed by MATLAB simulation results.