On Bayesian Optimization-Based Residual CNN for Estimation of Inter-Turn Short Circuit Fault in PMSM

Song, Qiang; Wang, Mingsheng; Lai, Wuxuan; Zhao, Sifang

doi:10.1109/tpel.2022.3207181

Cited by 45 publications

(8 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…( 8), each mode uk(t) is calculated. Then the energy of each mode Ek and the energy matrix E of all modes can be structured as Suppose there are N samples, a data set can be constructed through different energy matrices and labels [31]. The energy matrix of the i-th sample is represented by xi and the label is yi, Then the data set {X, Y} can be described as…”

Section: Principle Of Variational Modal Decompositionmentioning

confidence: 99%

Variational Mode Decomposition for Deep Convolutional Neural Networks in Underwater Acoustic Communication

Jiang

Chen

Xing

2023

Preprint

View full text Add to dashboard Cite

Recently, the application of deep convolutional neural networks (DCNN) in underwater acoustic communication has become a growing research hotspot. However, it is hindered by the large amounts of training data required to achieve a complex communication system. In this paper, we propose a novel approach of combining variational mode decomposition (VMD) and DCNN networks to address this issue. Considering filter-bank multicarrier (FBMC) technologies offer promising performance in terms of spectral efficiency and robustness to interference, and the Bellhop model is also capable of providing effective and accurate communication data, this study adopts FBMC technologies for signal modulation and Bellhop models for communication data. The results show that the VMD-DCNN equalizer has a lower BER compared with the other seven contrast methods, thus further promoting the application and development of VMD-DCNN technologies in underwater acoustic communications.

show abstract

Section: Principle Of Variational Modal Decompositionmentioning

confidence: 99%

Variational Mode Decomposition for Deep Convolutional Neural Networks in Underwater Acoustic Communication

Jiang

Chen

Xing

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…At present, the method based on deep learning has made remarkable achievements in the field of fault diagnosis. The main methods include Convolutional neural network [3][4] generative adversarial network [5][6] , Skowron [7] uses FFT combined with CNN neural network to achieve fault diagnosis under 100 groups of samples. Although the accuracy reaches 95%, the network generalization ability of this method is poor due to the scarcity of data.…”

Section: Introductionmentioning

confidence: 99%

Research on fault diagnosis of motor interturn short circuit based on deep learning

ding,

Chang,

huang

2024

Ninth International Symposium on Sensors, Mechatronics, and Automation System (ISSMAS 2023)

View full text Add to dashboard Cite

Fault diagnosis is an effective means to improve the reliability of the electric drive system of new energy vehicles. At present, inter turn short circuit faults in new energy vehicle motors are mainly detected using single fault features and small sample fault datasets. This method has low fault diagnosis accuracy and poor robustness. In this paper, a combined sample expansion strategy of conditional generative adversarial networks for attention mechanism optimization is proposed, and a fault diagnosis method combining improved Convolutional neural network is proposed. First, the combined characteristic data set of third harmonic and negative sequence current is made, and input into Attention CGAN to realize the expansion of original data samples. Then the expanded data and the original data are combined into a new data set, which is input into the improved CNN for training, and finally the fault diagnosis results are obtained. The experimental results show that compared to traditional fault diagnosis methods, the proposed fault diagnosis method based on combined features and data expansion improves the fault diagnosis accuracy by about 6.92%.

show abstract

“…Song et al [10] proposed a residual expansion CNN based on residual expansion for ITSC fault diagnosis, using normalized threephase stator currents as inputs to the network, which is able to identify faults of different severity well. However, sometimes relying on only one signal can be difficult to extract fault information as studied in literature [11,12], which has shown that multiple signals have higher accuracy than single signals, and multi-scale CNNs have a better diagnosis performance than single-channel CNNs. In the end-to-end PMSM fault diagnosis model, the inputs to the system are raw currents or vibrations, and these signals are usually one-dimensional.…”

Section: Introductionmentioning

confidence: 99%

Inter-turn short circuit and demagnetization fault diagnosis of ship PMSM based on multiscale residual dilated CNN and BiLSTM

Yan,

2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Inter-turn short circuit (ITSC) and demagnetization of permanent magnet synchronous motors (PMSMs) can lead to serious ship accidents, timely and accurate fault diagnosis of these faults is very important. A multi-signal fusion fault diagnosis method (MD-CNN-BiLSTM) is proposed based on multi-scale residual dilated convolutional neural network (D-CNN) and bidirectional long and short-term memory (BiLSTM) for PMSM fault diagnosis. This method first takes three-phase current and vibration signals as input; uses a three-column parallel CNN structure with different scales to extract both global signal and local feature. A residual connection in the expanded CNN is then used to eliminate the problems of gradient disappearance or explosion; and finally, BiLSTM is used to further extract features and identify the fault. A 2.2 kW permanent magnet synchronous motor was used to build a fault simulation test rig. The motor stator was rewound to simulate the ITSC fault, and different sizes of permanent magnets were replaced to simulate demagnetization fault. ITSC, demagnetization and their coupled faults were simulated under 10 specific motor speeds and loads respectively. The test proved that the diagnostic accuracy of the proposed method was 4.2% higher than that of ordinary CNN and 29.06% higher than that of BiLSTM. It also had the best diagnostic effect under the noise interference of different intensities. It was verified that the proposed method has good noise interference and strong classification ability.

show abstract

On Bayesian Optimization-Based Residual CNN for Estimation of Inter-Turn Short Circuit Fault in PMSM

Cited by 45 publications

References 35 publications

Variational Mode Decomposition for Deep Convolutional Neural Networks in Underwater Acoustic Communication

Variational Mode Decomposition for Deep Convolutional Neural Networks in Underwater Acoustic Communication

Research on fault diagnosis of motor interturn short circuit based on deep learning

Inter-turn short circuit and demagnetization fault diagnosis of ship PMSM based on multiscale residual dilated CNN and BiLSTM

Contact Info

Product

Resources

About