Search

Fault diagnosis based on deep learning (DL) has been a research hotspot in recent years. However, the current neural networks are getting larger and larger, with more and more parameters and insufficient noise resistance, making it difficult to effectively apply these methods to real working conditions. To address these issues, we propose a novel deep learning method with fewer parameters and better noise resistance based on transfer adversarial subnetwork (TAS) and channel-wise thresholds (CWT), namely, anti-noise transfer adversarial convolutions (ANTAC). In the proposed method, the original data and feature vectors are mapped to reproducing kernel Hilbert space (RKHS) and processed by maximum mean discrepancy (MMD) and Wasserstein distance (WD), which makes the method more capable to distinguish the similar features without producing any additional training parameters. Furthermore, white Gaussian noise (WGN) and the soft thresholding method with CWT are used to reduce data noise and improve the robustness and noise resistance of the network. Finally, the superiority of the proposed method is validated through experiments on different datasets, network structures and the data with different SNRs. The results show that the proposed method has better feature discrimination ability, noise resistance, and fewer parameters compared with other methods. The highest accuracy of the proposed method is 99.90% on the test set. • In order to enhance the network ability in distinguishing similar types of samples, TAS is proposed. Experiments show that the proposed method can effectively improve the network stability and the feature discrimination ability with fewer parameters. • The anti-interference ability and robustness of the network is enhanced by adding a certain disturbance into it. In addition. The soft threshold function with channel-wise threshold is used to improve the denoising ability of the network. • To enhance the network generalization and sensitivity to similar features, AFV loss is calculated by WD and controlled by scaling factor λ , and finally sent into RMSprop optimizer together with classification loss for optimization. • The method can directly process the original signals and the procedures of denoising and feature extraction are finished automatically in the method, which improves the fault diagnosis efficiency and provides a further practical possibility for the deployment of end-to-end fault diagnosis platforms. [ABSTRACT FROM AUTHOR]

For the detection and identification problems of common faults in the use of gearbox,considering the nonlinear,non-stationary properties of the gearbox vibration response signal,a method for the gearbox fault diagnosis based on local mean decomposition（ LMD） approximate entropy and PSO-ELM is proposed.Firstly,the LMD decomposition method is used for gearbox vibration signal,with correlation coefficient method extracted the first four PF components which contain the main fault information. By using the approximate entropy to describe quantitatively,and the feature vector is formed. Finally,the input weights of ELM and the threshold value of the hidden layer neurons are optimized by the particle swarm optimization algorithm,the model of PSO-ELM is established,and the approximate entropy values are input into the ELM and PSO-ELM models to recognize and classify the fault types of the gearbox of different conditions. The results show that based on LMD approximate entropy and PSO-ELM has the higher classification accuracy,the feasibility of this method is verified.

Aiming at the non-stationary and nonlinear of gearbox vibration signals,a fault diagnosis method based on the multi-fractal and particle swarm optimization support vector machine（PSO-SVM）is put forward.Firstly,the fractal filter with short-time fractal dimension as fuzzy control parameters is used to filtering noise reduction the gearbox vibration signals with bigger background noises.Secondly,the multi-fractal spectrum algorithm is applied to analyze the signal after filtering,the results show that the characteristic parameters:Δa（q）、f（a（q））maxand box dimensions Dbcan give a good presentation for gearbox working condition.Finally,the particle swarm optimization（PSO）is applied to optimize the parameters of support vector machine（SVM）.Taking the multi-fractal characteristic vectors as input parameters of PSO-SVM and SVM to recognize the fault types of the gearbox.The results show that SVM based on particle swarm optimization can improve the classification accuracy.Meanwhile,the validity of gearbox fault diagnosis based on muti-fractal and PSO-SVM is proved.

Given the problems of low accuracy and complex process steps currently faced by the field of fault diagnosis, a fault diagnosis method based on multi-sensor data weighted fusion (MSDWF) combined with depth-wise separable convolutions (DWSC) is proposed. The method takes into account the temporal and spatial information contained in multi-sensor data and can realize end-to-end bearing fault diagnosis. MSDWF is committed to comprehensively characterizing the state information of bearings, and the weighted operation of the multi-sensor data is to establish the interactive information to tap into the inline relationship in the data; DWSC equipped with residual connection is used to realize the decoupling of the channel and spatial correlation of the data. In order to verify the proposed method, the data obtained by a different number of sensors with weighted and unweighted states are used as the input of DWSC, respectively, for comparison, and finally, the effectiveness of MSDWF is verified. Through the comparison between different fault diagnosis methods, the method based on MSDWF and DWSC shows better stability and higher accuracy. Finally, when facing different experimental datasets, the method has similar performance, which shows the stability of the method on different datasets. [ABSTRACT FROM AUTHOR]

The clearance between the main roller and the cylinder cam has an important impact on the operational status of the bolt. In order to achieve the clearance and recoil of the cylinder cam mechanism, a bolt model was built and studied. The leading dynamic roller group and the dynamic bolt were analyzed. Positive stress and friction force change quickly on the high-speed leading roller group, influencing multiple aspects. When the main roller suddenly rotates or reverses in several short intervals, the assembling clearance and the massive friction force between the main roller and the cam curve slot are the main wear factors in theoretical analysis and test experiments. The leading roller group, the pure rolling criterion of the cylindrical cam mechanism of the automaton, is derived in this paper. Moreover, a new fault diagnosis method is developed, based on the Variational Box Dimension Kernel Fuzzy Mean Clustering Algorithm (VK) algorithm, which combines the variational mode decomposition and fractal box-counting dimension (VMD-FBCD) and the fuzzy clustering algorithms with a kernelized Mahalanobis distance (KMD-FC). With the simulation validation from 2810 samples, diagnosis using the VK algorithm is found to have a higher accuracy rate and slightly lower fallacy rate and to be faster than other diagnostic methods of similar studies. The results in terms of the recoil cylinder cam mechanism's multi-body dynamic analysis and high-speed fault identification also have significance as a reference for solving similar problems. [ABSTRACT FROM AUTHOR]