Data-driven methods have shown promising performance in power converter fault diagnosis. However, the existing methods are bothered by the inadequacy of accuracy and robustness, due to the feature variation in the different operating points, especially in nonstationary fault signals. This article proposes a robust data-driven fault diagnosis method combining automatic feature extraction (AFE) and improved random forests (IRFs). In AFE, this method uses optimal discrete wavelet transform, optimized by the defined information component function, to extract multiresolution time-frequency features from nonstationary fault signals, and the improved ReliefF algorithm is used to select important fault features considering both the independent and interactive weights to overcome nonstationary influence. To enhance the robustness, the IRFs adjusts the weight of decision trees by robustness testing based on out-of-bag data, which extracts the mapping relationship between fault features and fault labels for fault location. The comparative experimental results show that the proposed method achieves 99.94% fault diagnosis accuracy, and has superior fault diagnosis robustness compared to existing methods.

Extensive research has demonstrated that memristors or trigonometric functions can enhance the complexity of discrete chaotic maps. This article introduces a four-dimensional trigonometric-based memristor hyperchaotic map (4D-TBMHM). By combining discrete memristors, sine, and cosine, the 4D-TBMHM exhibits complex dynamical behaviors. Bifurcation and multistability phenomena are demonstrated using numerical methods. The 4D-TBMHM demonstrates symmetric or attractor self-growth based on iteration length for various system control parameters and initial states, and its complicated complex fractal structure and exemplary performance metrics are also highlighted. Furthermore, an attractor hybrid control, capable of arbitrary positioning and shaping, is proposed. An field-programmable gate array (FPGA)-based hardware prototype is developed, and the attractors are experimentally captured. Moreover, by combining 4D-TBMHM with an arrayed linear feedback shift register, an ultra-fast pseudorandom number generator (UFPRNG) with a throughput of 195.2 Gbps is achieved on an FPGA, surpassing contemporary techniques. Last, the generated UFPRNG is employed for 2.8 Gbps signal generation with noise, and experimental results illustrate its remarkable real-time capability and the PRNG's inherent randomness, facilitating signal generation with any signal-to-noise ratios.

In this article, the asynchronous fault detection (FD) strategy is investigated in frequency domain for nonlinear Markov jump systems under fading channels. In order to estimate the system dynamics and meet the fact that not all the running modes can be observed exactly, a set of asynchronous FD filters is proposed. By using statistical methods and the Lynapunov stability theory, the augmented system is shown to be stochastic stable with a prescribed $l_{2}$ gain even under fading transmissions. Then, a novel lemma is developed to capture the finite frequency performance. Some solvable conditions with less conservatism are subsequently deduced by exploiting novel decoupling techniques and additional slack variables. Besides, the FD filter gains could be calculated with the aid of the derived conditions. Finally, the effectiveness of the proposed method is shown by an illustrative example.

Bearings are important part of aerospace equipments. Therefore, bearings fault diagnosis and fault detection is necessary for the safe operation. Generally, the bearing are core component and runs in a complex and heavy background noise environment. Thence, the bearing signals will be overwhelmed by noise. At present, the diagnosis methods of bearings are based on the experience of experts which costs a lot of labor and time-consuming. Consequently, to address these inadequacies, a novel method based on variational mode decomposition (VMD) algorithms and Alexnet neural network has been presented. Firstly, it decomposes the nonstationary bearing signals into intrinsic mode functions adaptively. However, background noise will seriously affect the number of signal decompositions and reduce accuracy even over-decomposition. Thence, this paper proposed the hierarchical fuzzy entropy method to adaptively extract weak fault characteristics of bearings. Then Alexnet neural network has been utilized to learn the relationship of fault features and bearings health conditions. The architecture of Alexnet diagnosis model is convenient and efficient. Finally, two tentative dataset are adopted to verify the effectiveness and feasibility of the method proposed in this article. The results show that the presented method has a higher fault diagnosis accuracy rate than traditional convolutional neural networks.

As a proactive maintenance approach, remaining useful life (RUL) prediction plays a key role in smart operation and maintenance of industrial systems. To enhance the interpretability of deep neural network, and to measure the uncertainty of complex systems in the degradation process, an RUL prediction approach based on interpretable serialized variational autoencoder with drift-diffusion stochastic equation (ISVAE-DDSE) is proposed. Specifically, considering a dynamic sequential modeling method, this article proposes a generative deep learning approach to ensure that the model effectively captures the distribution characteristics of degradation data. On this basis, from the perspective of probabilistic deep generative network, this article derives a new type of generative loss function with the aid of the Bayesian theory. Furthermore, this article proposes an interpretable latent variable construction pattern based on DDSE, which integrates the dynamic representation of states, and rate of state change. In this sense, the network model can understand, and predict the evolutionary behavior of complex systems over time. Moreover, a Gaussian distribution network is designed to evaluate the RUL prediction’s uncertainty. This article demonstrates the advantages of the ISVAE-DDSE using a NASA aircraft turbofan engine dataset.

Distribution alignment mechanism serves as a fundamental basis for a cross-domain fault diagnosis, directly affecting diagnostic accuracy. A typical joint distribution alignment which enables a fine-grained class-level subdomain alignment, has achieved notable success in some cases. However, this mechanism performs a synchronous class-wise alignment across all categories between the source and the target domains, without considering the confidence of pseudolabels. This approach may lead to the collapse of the entire knowledge transfer framework due to an unreliable alignment. By contrast, an asynchronous joint distribution alignment (AJDA) mechanism is developed to enhance the model stability. In AJDA, the asynchronous class-level alignment strategy is first proposed to adaptively select high-confidence class-conditional distributions from the training process. Furthermore, a new high-order significant discrepancy representation metric based on the mechanical vibration characteristics is constructed to improve the sensitivity in measuring cross-domain distribution shifts. Finally, the proposed AJDA mechanism is validated through two scenarios, including planetary gearbox cross-condition and public cross-bearing transfer diagnosis cases, achieving average diagnostic accuracies of 98.81% and 80.23%, respectively. AJDA yields an overall performance improvement of more than 20% compared with other well-known methods, confirming its effectiveness and advantage.