This article presents a novel fault diagnosis method for open-circuit faults of insulated gate bipolar transistors (IGBTs) and current sensor faults in three-phase sinusoidal pulsewidth modulation inverters. The proposed method facilitates the simultaneous diagnosis of multiple IGBT and current sensor faults. A combination of Park transform and fast Fourier transform (FFT) is employed to extract fault features. The Park transform converts three-phase current into two-phase current, thereby reducing the dimensionality of the data. The FFT integrates time-domain and frequency-domain analyses to enhance the representation of fault characteristics. Additionally, fault features are extracted from the samples post-FFT to minimize the data volume of individual sample input by the model. For fault diagnosis, a temporal convolutional neural network model incorporating stacked residual blocks is developed to train and evaluate potential faults and demonstrate the model's validity and accuracy. The results of online simulations and hardware-in-the-loop experiments indicate that the proposed method can identify and locate the IGBT and sensor faults within 32 ms, and the accuracy is 98.75%. This method achieves higher training and testing accuracy and faster training and testing speed, which is superior to the existing diagnosis technologies.
Due to the high incidence and inconspicuous initial characteristics of rotor unbalance faults, the detection of incipient unbalance faults is becoming a very challenging problem. In this paper, a new method of small rotor unbalance fault diagnosis based on RIME-VMD and modified wavelet kernel network (modified-WKN) is proposed. Firstly, in order to extract the small unbalance fault information from the vibration signals with low signal-to-noise ratio (SNR) more efficiently, the RIME algorithm is used to search for the optimal location of the penalty factor and decomposition layer in the variable mode decomposition (VMD). Secondly, the most relevant decomposition components to the small unbalance fault information are selected by using Pearson Correlation Coefficients and utilized to reconstruct the signal. Finally, the modified-WKN diagnostic model that is used for multi-sensor data fusion is constructed. The model can acquire features of vibration signals from multiple position sensors, which enhances the ability of the modified WKN diagnostic model to deal with incipient fault modes. Based on the experimental analysis of rotor unbalance fault datasets with different SNRs, it is verified that the detection performance of the proposed method is better than the traditional WKN and VMD-WKN methods. Specifically, the proposed method is more sensitive to the initial unbalance faults.
Rotor unbalance faults are one of the high-frequency faults in rotating machinery. As such, their accurate and timely diagnosis is important. In contrast to traditional methods based on static features, the dynamics features and support vector machines (SVM) are combined for the accurate detection and classification of rotor unbalance faults. First, the dynamical trajectories of the rotor system associated with unbalance faults are accurately identified locally based on the deterministic learning theory, which is more sensitive to abnormal changes in the rotor system. Second, entropy dynamics features, including the sample entropy, fuzzy entropy, and permutation entropy, are extracted based on the obtained dynamical trajectory data. Finally, the dynamics features are used to train the fault classifier based on the SVM with a Gaussian kernel function. Experiments on a rotor unbalance fault test rig demonstrate the effectiveness of the proposed method. The accurate detection and classification of rotor unbalance faults were also achieved compared with the results of employing static time or frequency features.
As some of the high-frequency faults in rotating machinery, the early detection and accurate identification of rub-impact faults have attracted much attention. Due to the low vibration property and strong environment noise, it is still a challenge problem to deal with the early diagnosis of rub-impact faults. In this paper, a new rub-impact fault diagnosis method is proposed which consists four phases: the decomposition phase, the dynamic feature extraction phase, the diagnosis model construction phase, and the verification phase. In the first phase, the original vibration signals are decomposed based on empirical mode decomposition (EMD), and the most correlated components are selected using the Pearson correlation coefficient (PCC). Then, in the second phase, the fault dynamics of the selected components are identified based on dynamic learning, which means they are more sensitive to the small faults in the initial stage. In the last phase, a new fault feature deep fusion module is constructed based on the identified dynamics information and integrate into the convolution neural networks to derive the diagnosis model (dynamic CNN). Finally, the misjudgment self-correction mechanism is implemented based on the comparison of fault dynamics in the verification phase. Compared with traditional methods, the experiment results illustrate that the proposed method can detect the rub-impact faults more effectively.