Deep transfer learning provides a feasible fault diagnosis method for intelligent mechanical systems. However, this method usually assumes that the source domain and the target domain have the same label space, which greatly limits its application in the actual industry. Therefore, this article proposes a multidiscriminator deep weighted adversarial network (MDWAN) method, which is especially suitable for partial transfer learning where the number of target domain categories is less than the source domain. The proposed method is mainly composed of three parts: feature extractor, multidiscriminator, and classifier. In the feature extraction section, a deep separable convolutional neural network model (DSC) is proposed, which can greatly reduce parameter calculation amount under the premise of ensuring the extraction accuracy. In the multidiscriminator section, a multidiscriminator weighted learning strategy is proposed. This strategy comprehensively considers the domain information and source domain label information, and introduces a weight function to quantify the contribution of source domain samples to the domain discriminator and classifier, which can effectively identify and filter outlier source samples to promote the positive transfer of shared samples. In order to verify the effectiveness and feasibility of the proposed method, the method is applied to three types of bearing datasets of different machines. Comparing the classification results of different methods, the conclusion shows that this method is more beneficial for bearing fault classification.

Bearing is a key part of rotating machinery. Accurate prediction of bearing life can avoid serious failures. To address the current problem of low accuracy and poor predictability of bearing life prediction, a bearing life prediction method based on digital twins is proposed. Firstly, the vibration signals of rolling bearings are collected, and the time-domain and frequency-domain features of the actual data set are extracted to construct the feature matrix. Then unsupervised classification and feature selection are carried out by improving the self-organizing feature mapping method. Using sensitive features to construct a twin dataset framework and using the integrated learning CatBoost method to supplement the missing data sets, a complete digital twin dataset is formed. Secondly, important information is extracted through macro and micro attention mechanisms to achieve weight amplification. The life prediction of rolling bearing is realized by using fusion features. Finally, the proposed method is verified by experiments. The experimental results show that this method can predict the bearing life with a limited amount of measured data, which is superior to other prediction methods and can provide a new idea for the health prediction and management of mechanical components.

The fault vibration signal of rotating machinery system under strong background noise has the characteristics of non-stationary, non-Gaussian and complex components. In view of these characteristics, an improved method of signal separation based on Vold-Kalman filter (VKF) of adaptive instantaneous frequency estimation is proposed. First, a method for adaptive multiridge extraction of peaks detection based on synchro-squeezing wavelet transform (SWT) is proposed as the high-precision adaptive instantaneous frequency (IF) estimation method. The high precision IF estimation is used as the instantaneous frequency parameter of VKF, so that the complex multi-component non-stationary signal can be separated directly in the time domain and transformed into a signal combination composed of multiple stationary single-component signals and signal residues. Secondly, an improved method is proposed combining the adaptive IF estimation method with order tracking analysis and diagonal slice of bispectrum. In the improved method, the corresponding IF estimation of each component signal is taken as the reference frequency of its order tracking and the order spectrum analysis of each component signal is carried out respectively. Meanwhile, the signal residual is analyzed by diagonal slice of bispectrum, so as to suppress Gaussian noise and effectively separate and extract fault features in the vibration signal. Finally, the method is verified on simulation data and experimental data under different conditions. The results show that the improved method has higher extraction accuracy than other traditional methods. It has the superiority and the great potential for practical applications.

Nowadays, the methods of remaining useful life (RUL) prediction based on deep learning only use single model, or a simple superposition of two models, which makes it difficult for to maintain good generalization performance in various prediction scenarios, and ignores the dynamic sensitivity of features in the prediction, limiting the accuracy. This paper proposes a method of RUL prediction of bearing using fusion network through two-feature cross weighting (FNT-F). First, a fusion network with two subnets is proposed in this paper to adapt to the prediction problem in different scenarios. Meanwhile, a method of cross weighted joint analysis of the two features is proposed to make up for the shortcomings of feature analysis and achieve complementarity between time-domain and time–frequency features.

It is very important to accurately extract the instantaneous frequency (IF) of complex non-stationary signals, so the signal processing methods based on IF are widely used in engineering. However, in non-stationary complex vibration signals, it is difficult to extract the instantaneous frequency under strong noise because the frequency changes rapidly. This paper proposes an improved multi-ridge extraction method based on differential synchro-squeezing wavelet transform (DSWT). First, an improved method DSWT is proposed to reduce the computation time and image noise of synchro-squeezing wavelet transform (SWT). The running time of the algorithm can be further reduced by segmenting the signal and compressing the time-frequency matrix. The image noise can be significantly reduced by differentiating SWT time-frequency matrix. It lays a good foundation for the extraction of the time-frequency ridge line. Secondly, an improved multi-ridge extraction method is proposed. The ridge jump due to end effect can be eliminated adaptively in segmented extraction. It further improves the time-frequency aggregation and extraction effect of the wavelet ridgeline image. Finally, the improved multi-ridge extraction method based on DSWT (DSWT-IMRE) is verified by the simulation data and experimental data under different conditions. Compared with the method proposed in previous study, the results show that DSWT-IMRE has higher extraction accuracy.