Machine learning applications in power system fault diagnosis: Research advancements and perspectives

Vaish, Rachna; Dwivedi, U.D.; Tewari, Saurabh; Tripathi, Saurabh

doi:10.1016/j.engappai.2021.104504

Cited by 80 publications

(32 citation statements)

References 112 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The diagnosis of power system faults using machine learning was thoroughly reviewed by [ 39 ]. The success of machine learning approaches was first attributed to attempts to include the problems with traditional fault diagnosis.…”

Section: Related Workmentioning

confidence: 99%

Optimized EWT-Seq2Seq-LSTM with Attention Mechanism to Insulators Fault Prediction

Klaar

Stefenon

Seman

et al. 2023

Sensors

View full text Add to dashboard Cite

Insulators installed outdoors are vulnerable to the accumulation of contaminants on their surface, which raise their conductivity and increase leakage current until a flashover occurs. To improve the reliability of the electrical power system, it is possible to evaluate the development of the fault in relation to the increase in leakage current and thus predict whether a shutdown may occur. This paper proposes the use of empirical wavelet transform (EWT) to reduce the influence of non-representative variations and combines the attention mechanism with a long short-term memory (LSTM) recurrent network for prediction. The Optuna framework has been applied for hyperparameter optimization, resulting in a method called optimized EWT-Seq2Seq-LSTM with attention. The proposed model had a 10.17% lower mean square error (MSE) than the standard LSTM and a 5.36% lower MSE than the model without optimization, showing that the attention mechanism and hyperparameter optimization is a promising strategy.

show abstract

Section: Related Workmentioning

confidence: 99%

Optimized EWT-Seq2Seq-LSTM with Attention Mechanism to Insulators Fault Prediction

Klaar

Stefenon

Seman

et al. 2023

Sensors

View full text Add to dashboard Cite

show abstract

“…The NBM received much attention because of its effectiveness and accessibility. However, training a NBM usually needs a large number of historical SCADA data [15]. But, there may not have enough historical data for condition monitoring of a newly installed wind turbine.…”

Section: Introductionmentioning

confidence: 99%

Condition Monitoring of Wind Turbine Generators Based on SCADA Data and Feature Transfer Learning

et al. 2023

View full text Add to dashboard Cite

In order to build an effective condition monitoring (CM) model for the target wind turbines (WTs) with few operational data, an approach based on the feature transfer learning and a modified generative adversarial network is proposed. First, a large amount of labelled data from WTs are analyzed to construct a CM model with the aid of an autoencoder. This forms the knowledge of CM for WTs in the source domain. Second, a generative adversarial network is trained to build a mapping relationship between the features of different WTs. Third, the health status of the target WT is determined by analyzing the data collected from it online based on the proposed approach. Two case studies are conducted to verify that the proposed method can transfer the CM knowledge from source WT to target WT and achieve good performance in the CM of target WT.INDEX TERMS Autoencoder, condition monitoring (CM), feature transfer learning, generative adversarial network (GAN), wind turbine (WT).

show abstract

“…Current signals achieved by the relay unit are passed through feature extraction, then to perform relay tripping, these features are evaluated in DT. It was seen from their results that a fault was detected in 24 ms. Vaish et al (2021) have indicated regarding fault diagnosis in real-time applications that the wide area management system (WAMS) and phasor measurement units have gained importance. Also, to integrate ML techniques in applications such as SCADA and WAMS, it was stated that real-time digital simulation programs should be used in terms of similarity to power system data.…”

Section: Introductionmentioning

confidence: 99%

Probabilistic weighted voting model using multiple machine learning methods for fault detection and classification

Ülker

Küçüker

2022

COMPEL

View full text Add to dashboard Cite

Purpose The individual machine learning methods used for fault detection and classification have accuracy performance at a certain level. A combined learning model composed of different base classifiers rather than an individual machine learning model is introduced to ensure diversity. In this way, this study aims to improve the generalization capability of fault detection and classification scheme. Design/methodology/approach This study presents a probabilistic weighted voting model (PWVM) with multiple learning models for fault detection and classification. The working principle of this study’s proposed model relies on weight selection and per-class possibilities corresponding to predictions of base classifiers. Moreover, it can improve the power of the prediction model and cope with imbalanced class distribution through validation metrics and F-score. Findings The performance of the proposed PWVM was better than the performance of the individual machine learning methods. Besides, the proposed voting model’s performance was compared with different voting mechanisms involving weighted and unweighted voting models. It can be seen from the results that the presented model is superior to voting mechanisms. The performance results revealed PWVM has a powerful predictive model even in noisy conditions. This study determines the optimal model from among voting models with the prioritization method on data sets partitioned different ratios. The obtained results with statistical analysis verified the validity of the proposed model. Besides, the comparative results from different benchmark data sets verified the effectiveness and robustness of this study’s proposed model. Originality/value The contribution of this study is that PWVM is an ensemble model with outstanding generalization capability. To the best of the authors’ knowledge, no study has been performed using a PWVM composed of multiple classifiers to detect no-faulted/faulted cases and classify faulted phases.

show abstract

Machine learning applications in power system fault diagnosis: Research advancements and perspectives

Cited by 80 publications

References 112 publications

Optimized EWT-Seq2Seq-LSTM with Attention Mechanism to Insulators Fault Prediction

Optimized EWT-Seq2Seq-LSTM with Attention Mechanism to Insulators Fault Prediction

Condition Monitoring of Wind Turbine Generators Based on SCADA Data and Feature Transfer Learning

Probabilistic weighted voting model using multiple machine learning methods for fault detection and classification

Contact Info

Product

Resources

About