Condition Evaluation for Opening Damper of Spring Operated High-Voltage Circuit Breaker Using Vibration Time-Frequency Image

Method for Fault Diagnosis of Temperature-Related MEMS Inertial Sensors by Combining Hilbert–Huang Transform and Deep Learning

“…We compared the proposed two-layer BLSTM-based EMD algorithm, which is trained by the RI( ) dataset, to the other EMD methods, with respect to the EMD performance on our MEMS inertial datasets. The compared methods were as follows: traditional EMD [ 43 ], multiobjective optimization-based EMD [ 48 ], frequency-shifting-based EMD [ 49 ] and noise assistance analysis-based EMD [ 50 ]. We first evaluated EMD performance that includes the IMFs and Hilbert spectrum of the proposed method and the orthogonality of the IMFs.…”

“…We compared our proposed FD method with five state-of-the-art data-driven FD methods: Kordestani et al’s ANN-based method [ 21 ], Baskaya et al’s SVM-based method [ 31 ], Guo et al’s CNN-based method [ 20 ], Yang et al’s HHT-CNN-based method [ 43 ] and Wen et al’s CNN-based method [ 38 ]. We set training epochs to 300 and set the batch size to 16.…”

“…The results show that our method achieved the best fault classification accuracy, at 97.3007%. The CNN-based methods, which are proposed by Guo et al [ 20 ], Yang et al [ 43 ] and Wen et al [ 38 ] achieved mean FD accuracies of 94.6702%, 95.3258%, and 95.8644%, respectively. The machine learning-based methods, which are Kordestani et al’s ANN-based method [ 21 ] and Baskaya et al’s SVM-based method [ 31 ] achieved FD accuracies of 90.1147% and 91.1357%, respectively.…”

“…Researchers have studied FD methods that combine HHT-based feature extraction and deep learning-based FD methods. In [ 43 ], Yang et al proposed an opening damper evaluation method, which used HHT to convert a 1-D signal into 3-D time-frequency images and used these images for training CNNs in order to evaluate the vibration of power transmission systems. In [ 59 ], Liang et al used the intrinsic mode function (IMF) components of EMD to construct the fault features, which were then used to train the CNN-based FD network, which was called CRNN.…”

“…CNN is a powerful DL model for handling two-dimensional (2-D) images and has been used in FD research, such as mechanical systems FD [ 34 , 35 ], circuit systems FD [ 36 ], and avionics FD [ 37 ]. In FD applications, because raw data is often sampled in one-dimensional (1-D) format, researchers have turned to feature extraction operations that construct 2-D features for addressing FD problems using CNNs, such as sliding window [ 38 , 39 ], short time Fourier transform (STFT) [ 40 ], discrete wavelet transform (DWT) [ 41 , 42 ], and Hilbert–Huang transform (HHT) [ 43 , 44 ]. Structure health monitoring (SHM) is becoming a research hotspot in which CNN is applied and several methods have been proposed in the field of SHM combining CNN to solve mechanical system SHM problems [ 45 , 46 , 47 ].…”