常规机械故障诊断方法需要信号预处理、特征提取、特征选择、模式识别等多个步骤,过程复杂,通用性差。卷积神经网络(convolutional neural network, CNN)是一种自学习性能好、抗干扰能力强的深度神经网络。为了简化步骤、提高效率,本文将CNN引入到机械故障诊断,直接使用传感器测得的原始数据进行故障识别。由于机械振动信号的特征具有典型的时移性,CNN需要大量数据才能自我学习到这种特性。结合故障信号的冲击特点和CNN的不足,提出了权值求和和大尺度最大值池化策略,有效解决了特征的平移不变性,增强了小样本时的泛化能力。通过对单点和多点故障的轴承进行诊断,验证了平移不变CNN的有效性。与常规故障诊断方法和其他机器学习算法对比显示,平移不变CNN不仅准确率高,而且使用简单,为故障诊断提供了一种新的途径。
Abstract
Traditional machinery fault diagnostic method usually requires multiple processing steps including signal preprocessing, feature extraction, feature selection and pattern recognition to lead to complicated process and poor universal application. Convolutional neural network (CNN) is a deep neural network with strong self-learning and anti-interference abilities. To simplify steps and improve efficiency, here, a novel machinery fault diagnosis method based on CNN was proposed directly utilizing raw data measured with sensors to identify faults. Due to features of mechanical vibration signals having a typical time shift property, CNN needed a lot of data to self-learn this property. Combining fault signals’ impact feature and shortcomings of CNN, the strategies of weight summing and large-scale maximum value pooling were proposed to effectively solve CNN’s shift invariance to enhance small samples’ generalization ability. Through diagnosing single point fault and multi-point faults of bearings, the effectiveness of the shift invariant CNN was verified. Compared with traditional fault diagnosis approaches and other machine learning algorithms, it was shown that the shift invariant CNN not only has a higher accuracy but also is simple to use, and it provides a new way for fault diagnosis.
关键词
故障诊断 /
卷积神经网络 /
深度学习
{{custom_keyword}} /
Key words
fault diagnosis /
CNN /
deep learning
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] Du W, Tao J, Li Y, et al. Wavelet leaders multifractal features based fault diagnosis of rotating mechanism[J]. Mechanical Systems and Signal Processing, 2014, 43(1): 57-75.
[2] Kankar P, Sharma S C, Harsha S. Fault diagnosis of ball bearings using machine learning methods[J]. Expert Systems with Applications, 2011, 38(3): 1876-1886.
[3] Cui L, Wang J, Lee S. Matching pursuit of an adaptive impulse dictionary for bearing fault diagnosis[J]. Journal of sound and vibration, 2014, 33(10): 2840-2862.
[4] Wang X, Zhu H, Wang D, et al. The diagnosis of rolling bearing based on the parameters of pulse atoms and degree of cyclostationarity[J]. Journal of Vibroengineering, 2013, 15(3): 1560-1575.
[5] Zhu H, Wang X, Zhao Y, et al. Sparse representation based on adaptive multiscale features for robust machinery fault diagnosis[J]. Proceedings of the Institution of Mechanical Engineers, Part C Journal of Mechanical Engineering Science, 2014, 229(12): 2303-2313.
[6] Chen X, Du Z, Li J, et al. Compressed sensing based on dictionary learning for extracting impulse components[J]. Signal Processing, 2014, 96: 94-109.
[7] Liu C, Wang G, Xie Q, et al. Vibration sensor-based bearing fault diagnosis using ellipsoid-artmap and differential evolution algorithms[J]. Sensors, 2014, 14(6): 10598-10618.
[8] Liu H, Liu C, Huang Y. Adaptive feature extraction using sparse coding for machinery fault diagnosis[J]. Mechanical Systems and Signal Processing, 2011, 25(2): 558-574.
[9] Tran V T, AlThobiani F, Ball A. An approach to fault diagnosis of reciprocating compressor valves using teager–kaiser energy operator and deep belief networks[J]. Expert Systems with Applications, 2014, 41(9): 4113-4122.
[10] Hinton G E, Salakhutdinov R R. Reducing the dimensionality of data with neural networks[J]. Science, 2006, 313(5786): 504-507.
[11] Abdel-Hamid O, Mohamed A-r, Jiang H, et al. (2012). Applying convolutional neural networks concepts to hybrid nn-hmm model for speech recognition. Acoustics, Speech and Signal Processing (ICASSP), 2012 IEEE International Conference on, IEEE.
[12] LeCun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.
[13] Boureau Y-L, Bach F, LeCun Y, et al. (2010). Learning mid-level features for recognition. Computer Vision and Pattern Recognition (CVPR), 2010 IEEE Conference on, IEEE.
[14] Krizhevsky A, Sutskever I, Hinton G E (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems.
[15] Sainath T N, Mohamed A-r, Kingsbury B, et al. (2013). Deep convolutional neural networks for lvcsr. Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on, IEEE.
[16] Sermanet P, Chintala S, LeCun Y (2012). Convolutional neural networks applied to house numbers digit classification. Pattern Recognition (ICPR), 2012 21st International Conference on, IEEE.
[17] Bin G, Gao J, Li X, et al. Early fault diagnosis of rotating machinery based on wavelet packets—empirical mode decomposition feature extraction and neural network[J]. Mechanical Systems and Signal Processing, 2012, 27: 696-711.
[18] Hu Q, He Z, Zhang Z, et al. Fault diagnosis of rotating machinery based on improved wavelet package transform and svms ensemble[J]. Mechanical Systems and Signal Processing, 2007, 21(2): 688-705.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}