Effective bearing fault diagnosis is vital for the safe and reliable operation of rotating machinery. In practical applications, bearings often work at various rotational speeds as well as load conditions. Yet, the bearing fault diagnosis under multiple conditions is a new subject, which needs to be further explored. Therefore, a multi-scale deep belief network (DBN) method integrated with attention mechanism is proposed for the purpose of extracting the multi-scale core features from vibration signals, containing four primary steps: preprocessing of multi-scale data, feature extraction, feature fusion, and fault classification. The key novelties include multi-scale feature extraction using multi-scale DBN algorithm, and feature fusion using attention mechanism. The benchmark dataset from University of Ottawa is applied to validate the effectiveness as well as advantages of this method. Furthermore, the aforementioned method is compared with four classical fault diagnosis methods reported in the literature, and the comparison results show that our proposed method has higher diagnostic accuracy and better robustness.
The dynamic wireless communication network is a complex network that needs to consider various influence factors including communication devices, radio propagation, network topology, and dynamic behaviors. Existing works focus on suggesting simplified reliability analysis methods for these dynamic networks. As one of the most popular modeling methodologies, the dynamic Bayesian network (DBN) is proposed. However, it is insufficient for the wireless communication network which contains temporal and non-temporal events. To this end, we present a modeling methodology for a generalized continuous time Bayesian network (CTBN) with a 2-state conditional probability table (CPT). Moreover, a comprehensive reliability analysis method for communication devices and radio propagation is suggested. The proposed methodology is verified by a reliability analysis of a real wireless communication network.
Degradation and overstress failures occur in many electronic systems in which the operation load and environmental conditions are complex. The dependency of them called dependent competing failure process (DCFP), has been widely studied. Electronic system may experience mutual effects of degradation and shocks, they are considered to be interdependent. Both the degradation and the shock processes will decrease the limit of system and cause cumulative effect. Finally, the competition of hard and soft failure will cause the system failure. Based on the failure mechanism accumulation theory, this paper constructs the shock-degradation acceleration and the threshold descent model, and a system reliability model established by using these two models. The mutually DCFP effect of electronic system interaction has been decomposed into physical correlation of failure, including acceleration, accumulation and competition. As a case, a reliability of electronic system in aeronautical system has been analyzed with the proposed method. The method proposed is based on failure physical evaluation, and could provide important reference for quantitative evaluation and design improvement of the newly designed system in case of data deficiency.
Convolutional neural networks (CNNs) are well suited to bearing fault classification due to their ability to learn discriminative spectro-temporal patterns. However, gathering sufficient cases of faulty conditions in real-world engineering scenarios to train an intelligent diagnosis system is challenging. This paper proposes a fault diagnosis method combining several augmentation schemes to alleviate the problem of limited fault data. We begin by identifying relevant parameters that influence the construction of a spectrogram. We leverage the uncertainty principle in processing time-frequency domain signals, making it impossible to simultaneously achieve good time and frequency resolutions. A key determinant of this phenomenon is the window function’s choice and length used in implementing the short-time Fourier transform. The Gaussian, Kaiser, and rectangular windows are selected in the experimentation due to their diverse characteristics. The overlap parameter ’s size also influences the outcome and resolution of the spectrogram. A 50% overlap is used in the original data transformation, and ±25% is used in implementing an effective augmentation policy to which two-stage regular CNN can be applied to achieve improved performance. The best model reaches an accuracy of 99.98% and a cross-domain accuracy of 92.54%. When combined with data augmentation, the proposed model yields cutting-edge results.
A fractional-order cumulative optimization GM(1,2) model based on grey theory is proposed to study the relationship between torpedo loading and working reliabilities. In this model, the average relative error function related to order and background value is established. Taking the average relative error function as the objective function, the optimal value of the two parameters is obtained through the optimization method, and the minimum value of the average relative error is determined. The calculation example shows that this method can greatly improve the accuracy of the model and more accurately reflect the relationship between torpedo loading and working reliabilities compared with the traditional GM(1,2) model.
The reliability-based selective maintenance (RSM) decision problem of systems with components that have multiple dependent performance characteristics (PCs) reflecting degradation states is addressed in this paper. A vine-Copula-based reliability evaluation method is proposed to estimate the reliability of system components with multiple PCs. Specifically, the marginal degradation reliability of each PC is built by using the Wiener stochastic process based on the PC’s degradation mechanism. The joint degradation reliability of the component with multiple PCs is established by connecting the marginal reliability of PCs using D-vine. In addition, two RSM decision models are developed to ensure the system accomplishes the next mission. The genetic algorithm (GA) is used to solve the constraint optimization problem of the models. A numerical example illustrates the application of the proposed RSM method.