Acoustic emission signal is more sensitive to faults and less disturbed. This paper proposes a deep compression learning method based on compression sensing(AEDCL), for diagnosing and recognizing faults in acoustic emission signals. The high signal frequency of acoustic emission signals makes them sensitive to faults but challenging to analyze and store due to the large amount of data. AEDCL mines fault features directly from compressed data, without reconstructing signals, using a combination of compression sensing, kurtosis selection criterion, and an improved VMD method. A CNN-based deep compression learning method is used to mine features from the optimal compressed IMF components and conduct training and learning. The proposed method shows improved accuracy in fault diagnosis, computational efficiency, and generalization performance compared to traditional shallow learning methods and other deep learning methods. The method is verified on self-made bearing and RV reducer test benches.
<abstract> <p>A bearing is an important and easily damaged component of mechanical equipment. For early fault diagnosis of ball bearings, acoustic emission signals are more sensitive and less affected by mechanical background noise. To cope with the large amount of data brought by the high sampling frequency and high sampling points of acoustic emission signals, a compressed sensing processing framework is introduced to research data compression and feature extraction, and a wavelet sparse convolutional network is proposed for resolved diagnosis and evaluation. The main research objective of this paper is to maximize the compression rate of the signal under the constraint of ensuring the reconstruction error of the acoustic emission signal, which can reduce the data volume of the acoustic emission signal and reduce the pressure of data analysis for subsequent fault diagnosis. At the same time, a wide convolution kernel based on a continuous wavelet is introduced when designing the neural network, and the energy information of different frequency bands of the signal is extracted by the wavelet convolution kernel to characterize the fault characteristics of the equipment. The energy pooling layer is designed to enhance the deep mining ability of compressed features, and the regularized loss function is introduced to improve the diagnostic accuracy and robustness through feature sparseness. The experimental results show that the method can effectively extract the fault characteristics of the bearing acoustic emission signal, improve the analysis efficiency and accurately classify the bearing faults.</p> </abstract>
The rotate vector (RV) reducer has a complex structure and highly coupled internal components. Acoustic emission (AE) signal, which is more sensitive to a weak fault, is selected for fault diagnosis of the RV reducer. The high sampling frequency and big data are the challenges for AE signal store and analysis. This study combines compressed sensing (CS) and convolutional neural networks. As a result, data redundancy is significantly reduced while retaining most of the information, and the analysis efficiency is improved. Firstly, the time-domain AE signal was projected into the compression domain to obtain the compression signal; then, the wavelet packet decomposition in the compressed domain was performed to obtain the information of each frequency band. Next, the frequency band information was sent into the input layer of the multi-channel convolutional layer, and the energy pooling layer mines the energy characteristics of each frequency band. Finally, the softmax classifier was used to classify and predict different fault types of RV reducers. The self-fabricated RV reducer experimental platform was used to verify the proposed method. The experimental results show that the proposed method can effectively extract the fault features in the AE signal of the RV reducer, improve the efficiency of signal processing and analysis, and achieve the accurate classification of RV reducer faults.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)