Deep Transfer Learning-Based Fault Diagnosis Using Wavelet Transform for Limited Data

Bang, Junseong; Marco, Piergiuseppe Di; Shin, Hyejeon; Park, Pangun

doi:10.3390/app12157450

Cited by 10 publications

(4 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Then each output result is merged. In the literature, GoogLeNet confirms high efficiency for image classification[25][26][27][28].…”

mentioning

confidence: 86%

Fault diagnosis of electrical faults of three phase induction motors using acoustic analysis

Lucas,

Lui,

Brumercik

et al. 2024

Bulletin of the Polish Academy of Sciences Technical Sciences

View full text Add to dashboard Cite

Fault diagnosis techniques of electrical motors can prevent unplanned downtime and loss of money, production, and health. Various parts of the induction motor can be diagnosed: Rotor, stator, rolling bearings, fan, insulation damage, and shaft. Acoustic analysis is non-invasive. Acoustic sensors are low-cost. Changes in the acoustic signal are often observed for faults in induction motors. In this paper, the authors present a fault diagnosis technique for three-phase induction motors (TPIM) using acoustic analysis. The authors analyzed acoustic signals for three conditions of the TPIM: healthy TPIM, TPIM with twobroken bars, and TPIM with a faulty ring of the squirrel cage. Acoustic analysis was performed using Fast Fourier Transform(FFT), a new feature extraction method called MoD-7 (Maxima of differences between the conditions), and deep neural networks: GoogLeNet, and ResNet-50. The results of the analysis of acoustic signals were equal to 100% for the three analyzed conditions. The proposed technique is excellent for acoustic signals. The described technique can be used for electric motor fault diagnosis applications.

show abstract

“…Then each output result is merged. In the literature, GoogLeNet confirms high efficiency for image classification[25][26][27][28].…”

mentioning

confidence: 86%

Fault diagnosis of electrical faults of three phase induction motors using acoustic analysis

Lucas,

Lui,

Brumercik

et al. 2024

Bulletin of the Polish Academy of Sciences Technical Sciences

View full text Add to dashboard Cite

show abstract

“…Continuous wavelet transform is widely used to extract the time domain characteristics of the signal and the corresponding frequency domain spectrum content. It converts the original signal into a time-frequency distribution [28]. The representation of the original signal in the time domain and frequency domain is generated in the form of a time-frequency image [29].…”

Section: Data Pre-processingmentioning

confidence: 99%

An intelligent compound fault diagnosis method using generalized zero-shot model of bearing

Cen,

Zhao,

Liu

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Compound fault occurrence has been unpredictable, especially in industrial scenarios where it is difficult to collect a large number of labeled samples for compound fault. Based on this, this paper proposes a generative generalized zero-shot learning (GZSL) model aimed at synthesizing compound fault features through training with single fault samples. These synthesized features are then used for the recognition of compound fault. Firstly, in order to construct an accurate and effective semantic vector, the semantic generation module and discriminator are utilized to generate the semantics of compound fault. Secondly, a feature extraction module based on CNN is designed to extract various fault features from the two-dimensional time-frequency diagram. Finally, a fault semantic matching module is designed to match the feature vectors of compound faults with the generated fault semantic vectors. This enables the identification of unseen compound fault by computing their maximum similarity. The experimental results demonstrate that the proposed method achieved H scores of 75.83 and 69.24 on two real fault datasets, ensuring the correct classification of compound fault to the greatest extent possible.

show abstract

“…While the use of machine learning in different fields gained this reputation for categorization, many researchers raced to use different algorithms in the field of deep learning prior to preparing the input for categorization [32][33][34]. Even though they only used deep learning in different articles because of its ability to extract features internally by manipulating the system hyperparameters or even tuning them.…”

Section: Introductionmentioning

confidence: 99%

Transmission Line Fault Classification Based on the Combination of Scaled Wavelet Scalograms and CNNs Using a One-Side Sensor for Data Collection

Altaie,

Abderrahim,

Alkhazraji

2024

Sensors

View full text Add to dashboard Cite

This research focuses on leveraging wavelet transform for fault classification within electrical power transmission networks. This study meticulously examines the influence of various parameters, such as fault resistance, fault inception angle, fault location, and other essential components, on the accuracy of fault classification. We endeavor to explore the interplay between classification accuracy and the input data while assessing the efficacy of combining wavelet analysis with deep learning methodologies. The data, sourced from network recorders, including phase currents and voltages, undergo a scaled continuous wavelet transform (S-CWT) to generate scalogram images. These images are subsequently utilized as inputs for pretrained deep learning models. The experiments encompass various fault scenarios, spanning distinct fault types, locations, times, and resistance values. A remarkable feature of the proposed work is the attainment of 100% classification accuracy, obviating the need for additional algorithmic enhancements. The foundation of this achievement is the deliberate selection of the right input. The decision to employ an identical number of samples as the number of scales for the CWT emerges as a pivotal factor. This approach underpins the high accuracy and renders supplementary algorithms superfluous. Furthermore, this research underscores the versatility of this approach, showcasing its effectiveness across diverse networks and scenarios. Wavelet transform, after rigorous experimentation, emerges as a reliable tool for capturing transient fault characteristics with an optimal balance between time and frequency resolutions.

show abstract

Deep Transfer Learning-Based Fault Diagnosis Using Wavelet Transform for Limited Data

Cited by 10 publications

References 26 publications

Fault diagnosis of electrical faults of three phase induction motors using acoustic analysis

Fault diagnosis of electrical faults of three phase induction motors using acoustic analysis

An intelligent compound fault diagnosis method using generalized zero-shot model of bearing

Transmission Line Fault Classification Based on the Combination of Scaled Wavelet Scalograms and CNNs Using a One-Side Sensor for Data Collection

Contact Info

Product

Resources

About