基于多层训练干扰的CNN轴承故障诊断

Fault Diagnosis Based on CNN with Multi-layer Training Interference for Bearings

针对滚动轴承故障诊断方法在低信噪比情况下抗噪性能差以及在不同工况下自适应性不足的问题,提出了一种基于多层训练干扰的卷积神经网络滚动轴承故障诊断方法。首先,通过快速傅里叶变换将原始信号从时域变换到频域并作为卷积神经网络(CNN)的输入,再对多层卷积层进行实时训练干扰,以不同程度随机破坏网络训练,提高模型的抗噪性能和域自适应能力;同时引入批量归一化(BN)方法,改进卷积神经网络结构,改善模型性能。实验结果表明,所提方法不仅在噪声干扰下能达到较高的诊断准确率,同时也能有效地解决工况变化问题。

Aiming at the problem that the fault diagnosis method for rolling bearings has poor anti-noise performance under low signal-to-noise ratio(SNR) and insufficient self-adaptability under different working conditions, a fault diagnosis method based on convolutional neural network(CNN) with multi-layer training interference for rolling bearings is proposed. Firstly, the original signal is transformed from time domain to frequency domain by fast Fourier transform(FFT) and used as the input of the CNN. Then, the real-time training interference is performed on multiple convolutional layers to randomly destroy the network training to different degrees, so as to improve the anti-noise performance and domain adaptive ability of the model.Meanwhile, the batch normalization(BN) method is introduced to improve the structure of the CNN and the model performance. The experimental results show that the proposed method achieves high diagnostic accuracy under noise interference and solves the problem of working condition change.